Increased basal endogenous (primarily hepatic) glucose

Transcription

1 GASTROENTEROLOGY 2007;133: Relationship Between Hepatic/Visceral Fat and Hepatic Insulin Resistance in Nondiabetic and Type 2 Diabetic Subjects AMALIA GASTALDELLI,*, KENNETH CUSI,*, MAURA ETTITI, JEAN HARDIES,* YOSHINORI MIYAZAKI,* RACHELE BERRIA,*, EMMA BUZZIGOLI, ANNA MARIA SIRONI, EUGENIO CERSOSIMO,* ELE FERRANNINI,*, and RALH A. DEFRONZO* *Division of Diabetes, the Department of Medicine, The University of Texas Health Science Center at San Antonio, San Antonio, Texas; Institute of Clinical hysiology, National Research Council, isa, Italy; Obstetrics-Gynecology Department, Case Western Reserve University, Cleveland, Ohio; and the Department of Internal Medicine, University of isa, isa, Italy Background & Aims: Abdominal fat accumulation (visceral/hepatic) has been associated with hepatic insulin resistance (IR) in obesity and type 2 diabetes (T2DM). We examined the relationship between visceral/hepatic fat accumulation and hepatic IR/accelerated gluconeogenesis (GNG). Methods: In 14 normal glucose tolerant (NGT) (body mass index [BMI] 25 1 kg/m 2 ) and 43 T2DM (24 nonobese, BMI 26 1; 19 obese, BMI 32 1 kg/m 2 ) subjects, we measured endogenous (hepatic) glucose production (3-3 H-glucose) and GNG ( 2 H 2 O) in the basal state and during 240 pmol/m 2 /min euglycemic-hyperinsulinemic clamp, and liver (LF) subcutaneous (SAT)/visceral (VAT) fat content by magnetic resonance spectroscopy/magnetic resonance imaging. Results: LF was increased in lean T2DM compared with lean NGT (18% 3% vs 9% 2%, <.03), but was similar in lean T2DM and obese T2DM (18% 3% vs 22% 3%; NS). Both VAT and SAT increased progressively from lean NGT to lean T2DM to obese T2DM. T2DM had increased basal endogenous glucose production (EG) (NGT, ; lean T2DM, ; obese T2DM, mol/min/kg ffm ;.02) and basal GNG flux (NGT, ; lean T2DM, ; obese T2DM, mol/min/kg ffm ;.02). Basal hepatic IR index (EG fasting plasma insulin) was increased in T2DM (NGT, ; lean T2DM, ; obese T2DM, ;.007). In T2DM, after accounting for age, sex, and BMI, both LF and VAT, but not SAT, were correlated significantly ( <.05) with basal hepatic IR and residual EG during insulin clamp. Basal percentage of GNG and GNG flux were correlated positively with VAT ( <.05), but not with LF. LF, but not VAT, was correlated with fasting insulin, insulin-stimulated glucose disposal, and impaired FFA suppression by insulin (all <.05). Conclusions: Abdominal adiposity significantly affects both lipid (FFA) and glucose metabolism. Excess VAT primarily increases GNG flux. Both VAT and LF are associated with hepatic IR. Increased basal endogenous (primarily hepatic) glucose production, despite fasting hyperinsulinemia, is a characteristic feature of type 2 diabetes mellitus (T2DM) 1,2 and indicates the presence of hepatic resistance to the action of insulin. 1 This is substantiated further by the inability of insulin to normally suppress the increased basal rate of hepatic glucose output. 3 Excess abdominal fat accumulation, both visceral and hepatic, has been associated with abnormalities in glucose and lipid metabolism. In particular, both increased visceral adipose tissue (VAT) and intrahepatic fat content have been associated with hepatic insulin resistance (IR). 4 6 VAT is highly lipolytic and, according to the portal hypothesis, increased delivery of FFA into the portal circulation, and hence to the liver, leads to enhanced gluconeogenesis (GNG) and hepatic IR. 7 However, the portal hypothesis has remained untested in human beings because the portal circulation is inaccessible. The past decade has witnessed an epidemic increase in the incidence of obesity 8 and nonalcoholic fatty liver disease (NAFLD), which includes both hepatic steatosis and steatohepatitis Both obesity and diabetes are important risk factors for the development of NAFLD. 9,10 Nondiabetic subjects with NAFLD manifest an impaired ability of insulin to suppress endogenous glucose production (EG), 11,12 and a recent study showed that the amount of insulin required to achieve normoglycemia in T2DM was related closely to liver fat content. 13 A strong association between NAFLD and hepatic IR also has been shown, 11,14 and this relationship cannot be explained by obesity, age, Abbreviations used in this paper: BMI, body mass index; C5, carbon 5; EG, endogenous glucose production; FI, fasting plasma insulin; GCRC, General Clinical Research Center; GNG, gluconeogenesis; IR, insulin resistance; NAFLD, nonalcoholic fatty liver disease; NGT, normal glucose tolerant; Ra, rate of appearance; Rd, rate of glucose disappearance; SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue; T2DM, type 2 diabetes mellitus by the AGA Institute /07/$32.00 doi: /j.gastro

2 August 2007 HEATIC/VISCERAL FAT AND HEATIC IR 497 Table 1. Anthropometric and Clinical Characteristics NGT T2DM lean T2DM obese NGT vs all T2DM NGT vs lean T2DM lean vs obese T2DM Number Sex, F/M 4/10 8/16 8/11 NS NS NS Age, (y) NS HbA1c, % NS BMI, (kg/m 2 ) NS.0001 Waist, (cm) NS.0001 Total fat mass, % SAT, (cm 2 ) NS.0001 VAT, (cm 2 ) NS.0001 Liver fat, % NS lasma total cholesterol, (mmol/l) NS NS NS lasma high-density lipoprotein cholesterol, (mmol/l) NS NS NS lasma low-density lipoprotein cholesterol, (mmol/l) NS NS NS lasma triglyceride level, (mmol/l) NS NS AST level, (U/L) NS NS NS ALT level, (U/L) NS NS.03 Systolic blood pressure, (mm Hg) NS NS NS Diastolic blood pressure, (mm Hg) NS NS NS or visceral fat content, 15 even though the prevalence of liver steatosis is higher in obese subjects and liver/ visceral fat content are increased significantly in the elderly, even after accounting for total body fat content. 16 The goals of the present study were as follows: (1) to quantitate hepatic IR and to define the contribution of increased GNG/glycogenolysis to the hepatic IR, and (2) to examine the relationship between increased visceral/ hepatic fat content and hepatic IR/accelerated GNG in T2DM subjects during fasting and hyperinsulinemic conditions. Materials and Methods Subjects/Experimental Design Forty-three subjects with T2DM spanning a wide range of obesity (body mass index [BMI], kg/m 2 ; body fat, 21% 50%) and 14 lean normal glucose tolerant (NGT) subjects participated in the study. Obesity was defined as a BMI of greater than 30 kg/m 2 or a BMI of greater than 27 with a fat mass greater than 35%. Subjects were recruited from multiple advertisements in the newspapers and local community. All subjects who responded to the advertisement and who met entry criteria were recruited into the study. All studies were performed at the General Clinical Research Center (GCRC) of the University of Texas Health Science Center at San Antonio after an overnight fast. Subjects first received 75 grams of oral glucose tolerance test to establish the diagnosis of normal glucose tolerance or diabetes according to American Diabetes Association criteria. Within 5 10 days, subjects returned to the GCRC and euglycemic hyperinsulinemic clamp was performed with [3-3 H]- glucose to measure hepatic and total body (primarily represents muscle) insulin sensitivity. 17 The characteristics of the study population are shown in Table 1. None of the patients was treated with insulin, metformin, or thiazolidinediones. For subjects who were taking sulfonylureas, the medication was stopped 2 days before the study. Subjects were not taking any other drugs known to affect glucose tolerance. atients were in good general health without evidence of cardiac, hepatic, renal, or other chronic diseases as determined by history, examination, routine blood chemistries, urinalysis, and electrocardiography. Subjects were not taking any medications known to affect glucose metabolism. No subject was involved in strenuous physical activity and body weight was stable ( 3 lb) for at least 3 months before the study. The study protocol was approved by the Institutional Review Board of the University of Texas Health Science Center at San Antonio, and informed written consent was obtained from each patient before participation. Lean Body and Fat Mass Lean body mass was measured with an intravenous bolus (100 Ci) of 3 H 2 O at time 0 of the OGTT. lasma 3 H 2 O radioactivity was determined at 90, 105, and 120 minutes, and fat and lean body mass were calculated as previously described. 18 The quantitation of abdominal subcutaneous and visceral fat areas at L4 L5 was performed using magnetic resonance imaging (1.9 T; Elscint restige Ltd., Elscint, Haifa, Israel) 19 and liver fat

3 498 GASTALDELLI ET AL GASTROENTEROLOGY Vol. 133, No. 2 content was measured using magnetic resonance spectroscopy as previously described. 6 Euglycemic Hyperinsulinemic Clamp Subjects were admitted to the GCRC at 7:00 AM, after a 12- to 13-hour overnight fast, and a spontaneously voided urine sample was obtained. Subjects ate their last meal between 6:00 and 7:00 M and did not drink anything after the last meal. At 10:00 M on the evening before the study, the subjects drank 2 H 2 O (5 g/kg of fat-free mass; Isotech, Miamisburg, OH). A baseline blood sample was taken in the morning on the day before the study for determination of 2 H 2 O enrichment. On arrival at the GCRC, a polyethylene catheter was inserted into an antecubital vein for infusion of all test substances. A second catheter was inserted retrogradely into an ipsilateral wrist vein on the dorsum of the hand for blood sampling and the hand was kept in a heated box at 65 C. A primed (25 Ci [fasting glucose/5])-continuous (0.25 Ci/min) infusion of 3-[ 3 H]glucose (Duont- NEN, Boston, MA) was initiated and continued until study end. During the last 30 minutes of the basal equilibration period ( min), plasma samples were taken at 5- to 10-minute intervals for determination of plasma glucose and insulin concentrations and [ 3 H]glucose-specific activity. After the basal equilibration period, insulin was administered as a primed-continuous infusion at 240 pmol/min 1 /m 2 for 120 minutes. 17 The plasma glucose level was measured every 5 minutes after start of insulin, and a variable infusion of 20% glucose was adjusted based on the negative feedback principle to maintain the plasma glucose concentration at approximately 5 mmol/l with coefficient of variation less than 5%. lasma samples were collected every 15 minutes from 0 to 90 minutes, and every 5 10 minutes from 90 to 120 minutes, for determination of the plasma glucose and insulin concentrations and [ 3 H]glucose-specific activity. Four samples for determination of plasma FFA concentration were obtained during the basal state and during the last 5 minutes of the insulin clamp. lasma samples for the determination of GNG were obtained before the start of [ 3 H] glucose, at the end of the basal tracer equilibration period, and at the end of insulin clamp. Analytic methods. The plasma glucose concentration was determined by the glucose oxidase method (Beckman II Glucose Analyzer; Beckman Coulter, Fullerton, CA). The plasma insulin concentration was measured by radioimmunoassay (Diagnostic roducts Corporation, Los Angeles, CA). Hemoglobin A 1c (HbA 1c ) was measured by affinity chromatography (Biochemical Methodology, Drower 4350; Isolab, Akron, OH). The plasma FFA concentration was measured spectrophotometrically (Wako Chemicals Gmbh, Neuss, Germany). The 3-[ 3 H]glucose-specific activity was measured on barium hydroxide/zinc sulfate deproteinized plasma samples. The pattern of 2 H incorporation into plasma glucose after 2 H 2 O ingestion was determined by the modified method of Landau. 20,21 Briefly, the percentage of glucose produced via GNG from all precursors was quantified from the ratio of 2 H enrichment in carbon 5 (C5) to that of water. 2 H enrichment at C5 was obtained by converting glucose to xylose by removal of carbon in position 6, after purification by high-performance liquid chromatography. The C5 group was cleaved by oxidation with periodic acid and the formaldehyde was collected by distillation and incubated with ammonia to form a molecule of hexamethylenetetramine. Enrichment of hexamethylenetetramine obtained from C5 was determined by gas chromatography mass spectrometry by monitoring peaks of mass 140 and 141. The precision and accuracy of C5 have been reported previously. 22 Water enrichment in the total body water pool was measured by reacting a sample of urine with CaC 2 to obtain C 2 H 2, and the enrichment of C 2 H 2 then was determined by gas chromatography mass spectrometry by monitoring peaks of mass 26 and All samples were run through the gas chromatography mass spectrometry processing in duplicate or triplicate. Data analysis. Fat-free mass was measured using 3 H 2 O. 18 Subcutaneous adipose tissue (SAT) and VAT areas were quantitated by magnetic resonance imaging at the L4 L5 level, as previously described. 19 All glucose fluxes were expressed per fat-free mass because this normalization has been shown to correct for differences resulting from sex, obesity, and age. 24 During the last 30 minutes of the basal tracer equilibration period, both the plasma glucose concentration and the [ 3 H]glucose-specific activity were stable in all subjects. Following an overnight fast, steady-state conditions exist and basal EG equals total glucose rate of appearance (Ra) and was calculated as [3-3 H]glucose infusion rate (Inf_rate) divided by the plasma [3-3 H]-glucose-specific activity (SA) (mean, 5 determinations). During the insulin clamp, total glucose Ra was calculated using Steele s equation 25 Inf_rate C(t) V (d SA[t] dt) Ra(t) SA(t) where C is the plasma glucose concentration and V is the volume of distribution (162.5 ml/kg). Before applying Steele s equation, plasma tracer data were smoothed using a spline-fitting approach to stabilize the calculation of the derivative of specific activity. During the insulin clamp, the EG was calculated by subtracting the exogenous glucose infusion rate from the Ra. The tracer-determined rate of glucose disappearance (Rd) (time period, min) provided a measure of insulin-stimulated total body glucose disposal. The rate of glucose disappearance (Rd) was calculated as follows:

4 August 2007 HEATIC/VISCERAL FAT AND HEATIC IR 499 Rd(t) Ra(t) dc(t) V dt Because the fasting plasma insulin (FI) concentration is a strong inhibitory stimulus for EG, 26 an index of IR of basal EG (ie, hepatic IR) was calculated as the product of basal EG and FI concentration: hepatic IR index EG FI ( mol/min 1 /kg 1 ffm /pmol/l). Experimental validation of this index has been published previously. 3,27 eripheral insulin sensitivity was calculated as the Rd divided by the mean plasma glucose during the 90- to 120-minute time period (ie, the glucose metabolic clearance rate). Because the FI concentration is a strong inhibitory stimulus for lipolysis, 3 an index of adipocyte IR was calculated as the product of the fasting plasma FFA and insulin concentration: adipose IR index FFA FI (mmol/l/pmol/l). The hepatic insulin clearance was calculated by dividing the insulin infusion rate during the clamp by the mean peripheral plasma insulin concentration (mu/ min 1 ) during the 80- to 120-minute time period of the euglycemic clamp. The percentage contribution of GNG to plasma glucose was calculated as the ratio of enrichments in C5/ 2 H 2 O. 21 Gluconeogenic flux was calculated by multiplying the percentage of GNG by EG. Glycogenolytic flux was obtained as the difference between EG and GNG flux. The presence of NAFLD was established if the following 3 conditions were met: (1) liver fat of 5.5% or higher, (2) serum alanine aminotransferase (ALT) level greater than 30 U/L, and (3) serum aspartate aminotransferase (AST)/ALT ratio of less than 1.0. Statistical analysis. Data are given as the mean SE. Variables with skewed distribution are expressed as the median and interquartile range (in parantheses in tables). Group differences were analyzed by the Student t test, the Mann Whitney U test, and the 2 test for normally distributed, nonnormally distributed, and noncontinuous variables, respectively. Univariate and multivariate analyses were used to estimate associations among continuous variables in the whole data set and in the subgroup of diabetic patients. Figure 1. Relationship between age vs degree of obesity (BMI), liver fat content, VAT content, and SAT content in nondiabetic ( ) and diabetic (Œ) subjects.

5 500 GASTALDELLI ET AL GASTROENTEROLOGY Vol. 133, No. 2 content, and slightly, although not significantly, higher liver fat content. Both VAT and liver fat content increased as a function of age and the increases were similar in NGT and T2DM subjects (Figure 1). VAT and SAT increased significantly both as a function of obesity (BMI) and diabetes, whereas liver fat content was associated with BMI only in NGT (Figure 2). Liver fat content and VAT were correlated strongly in both NGT and T2DM (Figure 3). There was no relationship between liver fat and SAT in either NGT or T2DM (Figure 3). The ALT level, but not the AST level, was increased as a function of obesity (Table 1). lasma high-density lipoprotein cholesterol tended to decline, whereas plasma triglyceride levels tended to increase as a function of obesity and diabetes, but these changes did not reach statistical significance. Both systolic and diastolic blood pressure tended to increase as a function of obesity, but these changes did not achieve statistical significance (Table 1). Glucose Metabolism During the fasting state, the plasma glucose concentration was similar in lean T2DM and obese T2DM Figure 2. Relationship between BMI vs liver, visceral, and subcutaneous fat content in nondiabetic ( ) and diabetic (Œ) subjects. The correlation between BMI and liver fat was found only in NGT subjects. Results Anthropometric, Clinical, and Laboratory Characteristics Lean T2DM and lean NGT subjects had similar BMI, waist circumference, and visceral and subcutaneous fat, but T2DM subjects had a higher percentage of body fat and liver fat content (Table 1). Compared with lean T2DM subjects, obese T2DM subjects had increased BMI, percentage fat mass, visceral and subcutaneous fat Figure 3. Relationship between liver fat and visceral and subcutaneous abdominal fat in nondiabetic ( ) and diabetic (Œ) subjects.

6 August 2007 HEATIC/VISCERAL FAT AND HEATIC IR 501 subjects, whereas the plasma insulin concentration tended to increase as a function of both obesity and diabetes (Table 2). Basal total EG was increased in T2DM and was not correlated with BMI, SAT, VAT, or liver fat content (Figure 4). GNG flux was increased in lean T2DM and obese T2DM subjects, whereas glycogenolytic flux was similar (Table 2). Increased GNG was correlated with VAT, but not with liver fat content (Figure 5). Glycogenolytic flux did not correlate with either VAT or liver fat content. Both hepatic and adipocyte IR indices increased as a function of obesity and diabetes (Table 2). During the euglycemic insulin clamp, T2DM subjects were insulin resistant compared with NGT subjects and showed impaired suppression of EG (Table 2). The impaired suppression of EG was correlated with both increased VAT and liver fat content (Figure 6). During the insulin clamp, diabetic subjects showed an impaired suppression of plasma FFA, and this was associated with increased liver fat content but not VAT. In this population, among the diabetic subjects, NAFLD was found in 10 of 43 (23%) T2DM (4 lean, 6 obese) subjects and in none of the lean NGT subjects (Table 3). Diabetic subjects with NAFLD had significantly more hepatic and adipocyte IR and tended to have more peripheral (muscle) IR than their diabetic counterparts without NAFLD. Consistent with this, T2DM NAFLD subjects had significantly increased FI concentrations compared with their diabetic counterparts without NAFLD. However, despite the more severe IR in T2DM NAFLD subjects, we did not observe an increased rate of GNG. Correlations In simple regression analysis VAT, but not SAT, was correlated with liver fat content (Figure 3). However, when age and BMI were included in the multivariate model, the correlation was lost. Correlations between metabolic variables and VAT, SAT, and liver fat are shown in Table 4. Variables first were analyzed using a univariate model. If significant associations were found, data were analyzed using a multivariate model accounting for age, sex, and BMI. In the overall population, in univariate analysis SAT was associated weakly with adipose tissue IR and rate of disappearance during clamp normalized to insulin (Rd/I), but these correlations were lost in multivariate analysis. Both increased hepatic and visceral fat were associated, in univariate analysis, with increased BMI, altered lipid profile, high blood pressure, increased fasting plasma glucose and insulin concentrations, reduced insulin clearance, impaired insulin suppression of lipolysis, impaired suppression of endogenous glucose production by insulin, and increased IR in liver, muscle, and adipose tissue. Neither VAT nor SAT nor percentage of liver fat were correlated with basal EG, and only VAT was correlated with increased GNG (expressed both as a percentage of EG and total flux) but not with glycogenolysis. Accelerated GNG was associated with increased basal plasma FFA (r 0.30,.03), and impaired suppression of GNG was associated with impaired suppression of plasma FFA by insulin (r 0.33,.01). In multivariate analysis, after accounting for age, sex, and BMI, VAT remained correlated with accelerated GNG, basal hepatic IR index, and residual EG during insulin clamp. In multiple regression analysis, increased liver fat content remained correlated with increased FI concentration, increased total/low-density lipoprotein cholesterol levels, increased hepatic (AST/ ALT) enzyme levels, basal hepatic IR index, impaired suppression of EG by insulin, and adipocyte IR index (Table 4). Increased liver fat also correlated with the adipose tissue IR index, decreased hepatic insulin clear- Table 2. Metabolic Characteristics NGT T2DM lean T2DM obese NGT vs all T2DM NGT vs lean T2DM lean vs obese T2DM Fasting lasma glucose-180, (mmol/l) NS lasma glucose 0, (mmol/l) NS lasma FFA, (mmol/l) NS NS NS lasma insulin, (pmol/l) 53 (42 70) 65 (55 87) 87 (64 161).008 NS.05 Hepatic IR index, (umol/min 1 /kg 1 ffm x/pmol/l) Adipose IR index, (mmol/l/pmol/l) NS.02 EG, (umol/min 1 /kg 1 ffm ) NS NS GNG, % NS NS NS Gluconeogenic flux, (umol/min 1 /kg 1 ffm ) NS.03 Glycogenolysis, (umol/min 1 /kg 1 ffm ) NS NS NS Insulin clamp lasma glucose level, (mmol/l) NS NS lasma insulin level, (pmol/l) 373 ( ) 381 ( ) 417 ( ) NS NS NS lasma FFA level, (mmol/l) NS EG, (umol/min 1 /kg 1 ffm ) NS Rd, (umol/min 1 /kg 1 ffm ) NS Metabolic clearance rate glu, (ml/min 1 /kg 1 ffm ) Hepatic insulin clearance, (ml/min 1 /kg 1 ) NS NS NS NOTE. 1 NGT vs T2DM, 2 NGT vs lean T2DM, 3 obese T2DM vs nonobese T2DM, by 2-way analysis of variance.

7 502 GASTALDELLI ET AL GASTROENTEROLOGY Vol. 133, No. 2 Figure 4. Relationship between basal EG and liver fat (%), VAT, SAT, and BMI in nondiabetic ( ) and diabetic (Œ) subjects ance (Figure 7), and impaired suppression of plasma FFA concentration during the insulin clamp. We also performed a stepwise regression analysis considering, at the same time, liver fat content and VAT as independent variables. The analysis confirmed that GNG was related only to VAT, whereas basal hepatic and adipose IR indices and residual hepatic EG during clamp were related to both VAT and liver fat content. On the other hand, peripheral IR (expressed as both Rd and Rd/I), as well as impaired suppression of plasma FFA during the insulin clamp, were related to liver fat content. Discussion referential abdominal accumulation of fat, both visceral and hepatic, has been associated with abnormalities in glucose and lipid metabolism. 28,29 Hepatic IR has been correlated with both increased visceral (VA) and hepatic fat content, whereas accelerated GNG has been shown in individuals with NAFLD, a disorder characterized by increased hepatic fat content. 32,33 However, those studies did not examine the relationships between GNG and VAT/liver fat content and the relationships between these variables and the degree of adiposity or state of glucose tolerance. We measured hepatic, visceral, subcutaneous, and total body fat content in 57 subjects with and without diabetes and with a wide range of obesity. In this population, advancing age was associated with both increased hepatic and visceral fat accumulation, whereas increasing BMI was associated with increases in both visceral and subcutaneous fat. Increased hepatic fat content was associated more closely with diabetes than with obesity. The fasting rate of EG was increased slightly in T2DM compared with NGT, but the hepatic IR index was increased in both diabetic groups, especially in obese T2DM subjects. GNG flux was increased in T2DM subjects, particularly in obese subjects. This is consistent with previous results from our laboratory that showed that obesity per se is a major determinant of the rate of GNG. 22,28 Subjects with increased VAT manifested increases in both percentage of GNG and GNG flux. Increased VAT also was associated with an increase in the basal hepatic IR index and impaired suppression of EG during the insulin clamp. These results are consistent with previous studies that showed that increased GNG is associated independently with visceral adiposity. 1,28 Al-

8 August 2007 HEATIC/VISCERAL FAT AND HEATIC IR 503 Figure 5. Relationship between basal gluconeogenic flux (GNG) vs liver fat (%) and VAT in nondiabetic ( ) and diabetic (Œ) subjects. though we did not measure portal FFA concentration, we found that accelerated GNG was associated with increased peripheral plasma FFA levels both during fasting and during insulin infusion. Subcutaneous fat content, on the other hand, was not correlated with the basal hepatic IR index, suppression of EG during the insulin clamp, or basal GNG flux. Collectively, these observations are in agreement with the portal hypothesis, which proposes that enhanced lipolytic activity in visceral fat results in increased delivery of FFA and gluconeogenic substrates to the liver, 7 leading to increased GNG. Consistent with this scenario, a recent study showed that the activities of hepatic phosphoenolpyruvate carboxykinase, the rate-limiting enzyme for GNG, and glucose-6-phosphatase, the rate-limiting enzyme for hepatic glucose production, increased in association with increased visceral fat content during high-fat feeding. 34 Of note, we also have shown that a decrease in VAT in thiazolidinedione-treated T2DM patients is associated with reductions in the basal rate of GNG and hepatic IR. 35,36 Increased liver fat was associated with hepatic IR, both under basal conditions and in response to insulin, but was not correlated with accelerated GNG. Liver fat content was associated with higher plasma insulin concentrations, decreased insulin clearance, altered hepatic enzyme profile, and dyslipidemia. We speculate that increased FFA flux to the liver from the increased VAT mass, together with high portal insulin concentrations, promotes hepatic fat accumulation, as has been shown in animal studies. 34,37 However, the increase in hepatic fat content did not correlate with accelerated GNG. Therefore, our results are most consistent with the postulate that it is the increased portal FFA flux, rather than the fat accumulation in the liver, that is responsible for the increase in GNG flux. A number of studies have shown an association between NAFLD and hepatic IR, 12,33,38,39 and a strong relationship between hepatic fat accumulation and hepatic IR has been shown in rats fed a high-fat diet. 37 The investigators also showed that increased hepatic content was associated with an increase in GNG. However, this study was not able to establish whether the Figure 6. Relationship between suppression of EG during the insulin clamp vs liver fat (%) and VAT in nondiabetic ( ) and diabetic (Œ) subjects.

9 504 GASTALDELLI ET AL GASTROENTEROLOGY Vol. 133, No. 2 Table 3. resence of NAFLD in Diabetic Subjects T2DM T2DM NAFLD n BMI NS % liver fat lasma glucose level, (mmol/l) NS lasma FFA level, (mmol/l) lasma insulin level, (pmol/l) 65 (58 82) 128 (97 216).0001 Hepatic IR index, (umol/min 1 /kg 1 ffm x/pmol/l) Adipose IR index, (mmol/l/pmol/l) EG, (umol/min 1 /kg 1 ffm ) NS GNG, (%) NS Gluconeogenic flux, (umol/min 1 /kg 1 ffm ) NS Glycogenolysis, (umol/min 1 /kg 1 ffm ) NS Rate of disappearance, (umol/min 1 /kg 1 ffm ) NOTE. Two-way analysis of variance is applied to T2DM vs T2DM NAFLD. increase in GNG was owing to increased hepatic fat content or to the increased plasma FFA levels that were associated with the high-fat diet. Both increased GNG and hepatic IR have been shown to be characteristic findings in patients with NAFLD. 32 However, NAFLD often is associated with obesity and diabetes, as well as Table 4. Diabetic Subjects: Matrix of Correlation VAT SAT % Liver fat Clinical variables Age 0.48 a a Sex a 0.09 BMI 0.70 a 0.80 a 0.29 a Serum triglyceride level 0.40 a Total cholesterol level a,b Low-density lipoprotein cholesterol level a,b High density lipoprotein cholesterol level 0.31 a Systolic blood pressure 0.39 a a Diastolic blood pressure 0.33 a AST level ALT level 0.29 a a,b Fasting Fasting glucose level 0.38 a a Fasting FFA level Fasting insulin level 0.43 a a,b EG GNG% 0.33 a,b GNG 0.37 a,b Glycogenolysis Metabolic clearance rate 0.44 a,b a,b Hepatic IR index (basal) 0.48 a,b a,b Adipose IR index 0.45 a 0.29 a 0.47 a,b Clamp Clamp FFA 0.36 a a,b Clamp residual EG 0.50 a,b a,b Rd/I 0.35 a 0.27 a 0.48 a,b Metabolic clearance rate 0.39 a a,b Insulin clearance 0.29 a,b a,b with increased visceral fat content. Therefore, it is not surprising that NAFLD is associated with increased GNG. In this population, among the diabetic subjects we found 10 subjects with NAFLD (6 obese, 4 nonobese) (Table 3). Diabetic subjects with NAFLD had significantly more hepatic and adipocyte IR and tended to have more peripheral (muscle) IR than their diabetic counterparts without NAFLD. Consistent with this, T2DM NAFLD subjects had significantly increased FI concentrations compared with their diabetic counterparts without NAFLD. However, despite the more severe IR in T2DM NAFLD subjects we did not observe an increased rate of GNG. This could be explained by the lower fasting FFA levels (substrate delivery is an important determinant of hepatic GNG) in T2DM NAFLD subjects. Because the total number of subjects with NAFLD in the present study was relatively small, resolution of this issue will require a larger number of diabetic subjects with NAFLD. Last, we observed a strong correlation between increased liver fat content and both peripheral (muscle) a Correlation coefficients:.05 or less for univariate analysis. b Significance after adjustment for age, sex, and BMI..05 or less in multivariate analysis. Figure 7. Relationship between hepatic insulin clearance during the insulin clamp vs liver fat (%) in nondiabetic ( ) and diabetic (Œ) subjects.

10 August 2007 HEATIC/VISCERAL FAT AND HEATIC IR 505 and hepatic IR (Table 4). One potential explanation for these associations is that they all are correlated strongly with BMI. Thus, increased caloric intake with excessive fat deposition in liver and muscle (the latter was not measured in the present study) explains the link between percentage of liver fat, peripheral IR, and hepatic IR. We also have shown that in any given individual, there is a very strong correlation between hepatic and peripheral IR, 40 suggesting an inherited etiology of the liver and muscle IR. The hepatic IR, in turn, is associated with altered lipid metabolism, which results in increased liver fat content. Alternatively, increased fat deposition in the liver could lead to the release of inflammatory markers/ IR-provoking factors that contribute to the development of muscle IR. All of these explanations involve increased fat deposition in the liver (and muscle). Thus, interventions that mobilize fat out of the liver have the potential not only to ameliorate NAFLD, but also to improve insulin sensitivity in both muscle and liver. With respect to this, it is noteworthy that the thiazolidinediones have been shown to reduce liver fat content, improve nonalcoholic steatohepatitis, and to enhance peripheral (muscle) tissue sensitivity to insulin. 6,35,39 It is well known that FFA suppression in response to insulin is impaired in NAFLD patients. 6 In the present study, we found that basal adipose tissue IR and decreased FFA suppression during insulin infusion are associated with increased liver fat content, but not with increased VAT or SAT content. This interesting observation is deserving of further investigation to elucidate the mechanisms underlying this association. In summary, fat accumulation in the abdominal region significantly affects both lipid (FFA) and glucose metabolism. Excess VAT primarily increases GNG flux, whereas increased liver fat primarily is associated with hepatic IR. References 1. Gastaldelli A, Miyazaki Y, ettiti M, Buzzigoli E, Mahankali S, Ferrannini E, et al. Separate contribution of diabetes, total fat mass, and fat topography to glucose production, gluconeogenesis, and glycogenolysis. J Clin Endocrinol Metab 2004;89: Jeng C, Sheu W, Fuh M, Chen Y, Reaven G. Relationship between hepatic glucose production and fasting plasma glucose concentration in patients with NIDDM. Diabetes 1994;43: Groop LC, Bonadonna RC, Delrato S, Ratheiser K, Zyck K, Ferrannini E, et al. Glucose and free fatty acid metabolism in non-insulin-dependent diabetes mellitus. Evidence for multiple sites of insulin resistance. J Clin Invest 1989;84: Bugianesi E, McCullough AJ, Marchesini G. Insulin resistance: a metabolic pathway to chronic liver disease. Hepatology 2005;42: Barzilai N, She L, Liu BQ, Vuguin, Cohen, Wang J, et al. Surgical removal of visceral fat reverses hepatic insulin resistance. Diabetes 1999;48: Bajaj M, Suraamornkul S, Hardies LJ, ratipanawatr T, DeFronzo RA. lasma resistin concentration, hepatic fat content, and hepatic and peripheral insulin resistance in pioglitazone-treated type II diabetic patients. Int J Obes Relat Metab Disord 2004;28: Bjorntorp. Metabolic abnormalities in visceral obesity. Ann Med 1992;24: Wyatt SB, Winters K, Dubbert M. Overweight and obesity: prevalence, consequences, and causes of a growing public health problem. Am J Med Sci 2006;331: Marceau, Biron S, Hould FS, Marceau S, Simard S, Thung SN, et al. Liver pathology and the metabolic syndrome X in severe obesity. J Clin Endocrinol Metab 1999;84: Knobler H, Schattner A, Zhornicki T, Malnick SD, Keter D, Sokolovskaya N, et al. Fatty liver an additional and treatable feature of the insulin resistance syndrome. QJM 1999;92: Seppala-Lindroos A, Vehkavaara S, Hakkinen AM, Goto T, Westerbacka J, Sovijarvi A, et al. Fat accumulation in the liver is associated with defects in insulin suppression of glucose production and serum free fatty acids independent of obesity in normal men. J Clin Endocrinol Metab 2002;87: Bugianesi E, Gastaldelli A, Vanni E, Gambino R, Cassader M, Baldi S, et al. Insulin resistance in non-diabetic patients with non-alcoholic fatty liver disease: sites and mechanisms. Diabetologia 2005;48: Ryysy L, Hakkinen AM, Goto T, Vehkavaara S, Westerbacka J, Halavaara J, et al. Hepatic fat content and insulin action on free fatty acids and glucose metabolism rather than insulin absorption are associated with insulin requirements during insulin therapy in type 2 diabetic patients. Diabetes 2000;49: Kelley DE, McKolanis TM, Hegazi RA, Kuller LH, Kalhan SC. Fatty liver in type 2 diabetes mellitus: relation to regional adiposity, fatty acids, and insulin resistance. Am J hysiol 2003;285: E906 E Garg A, Misra A. Hepatic steatosis, insulin resistance, and adipose tissue disorders. J Clin Endocrinol Metab 2002;87: Cree MG, Newcomer BR, Katsanos CS, Sheffield-Moore M, Chinkes D, Aarsland A, et al. Intramuscular and liver triglycerides are increased in the elderly. J Clin Endocrinol Metab 2004;89: DeFronzo R, Tobin D, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J hysiol 1979;237:E214 E Bonora E, Del rato S, Bonadonna R, Gulli G, Solini A, Shank M, et al. Total body fat content and fat topography are associated differently with in vivo glucose metabolism in nonobese and obese nondiabetic women. Diabetes 1992;41: Lancaster J, Ghiatas A, Alyassin A, Kilcoyne R, Bonora E, De- Fronzo R. Measurement of abdominal fat with T1-weighted MR images. J Magn Reson Imaging 1991;1: Schumann WC, Gastaldelli A, Chandramouli V, revis SF, ettiti M, Ferrannini E, et al. Determination of the enrichment of the hydrogen bound to carbon 5 of glucose on 2H2O administration. Anal Biochem 2001;297: Landau BR, Wahren J, Chandramouli V, Schumann WC, Ekberg K, Kalhan SC. Contributions of gluconeogenesis to glucose production in the fasted state. J Clin Invest 1996;98: Gastaldelli A, Baldi S, ettiti M, Toschi E, Camastra S, Natali A, et al. Influence of obesity and type 2 diabetes on gluconeogenesis and glucose output in humans: a quantitative study. Diabetes 2000;49: revis S, Hazey J, Diraison F, Beylot M, David F, Brunengraber H. Assay of the deuterium enrichment of water via acetylene. J Mass Spectr 1996;31: Natali A, Toschi E, Camastra S, Gastaldelli A, Groop L, Ferrannini E. Determinants of postabsorptive endogenous glucose output in non-diabetic subjects. European Group for the Study of Insulin Resistance (EGIR). Diabetologia 2000;43:

11 506 GASTALDELLI ET AL GASTROENTEROLOGY Vol. 133, No Altszuler N, De Bodo RC, Steele R, Wall JS. Measurement of size and turnover rate of body glucose pool by the isotope dilution method. Am J hysiol 1956;187: Sindelar DK, Chu CA, Venson, Donahue E, Neal DW, Cherrington AD. Basal hepatic glucose production is regulated by the portal vein insulin concentration. Diabetes 1998;47: DeFronzo RA, Ferrannini E, Simonson DC. Fasting hyperglycemia in non-insulin-dependent diabetes mellitus: contributions of excessive hepatic glucose production and impaired tissue glucose uptake. Metabolism 1989;38: Gastaldelli A, Miyazaki Y, ettiti M, Matsuda M, Mahankali S, Santini E, et al. Metabolic effects of visceral fat accumulation in type 2 diabetes. J Clin Endocrinol Metab 2002;87: Miyazaki Y, Glass L, Triplitt C, Wajcberg E, Mandarino LJ, De- Fronzo RA. Abdominal fat distribution and peripheral and hepatic insulin resistance in type 2 diabetes mellitus. Am J hysiol 2002;283:E1135 E Eguchi Y, Eguchi T, Mizuta T, Ide Y, Yasutake T, Iwakiri R, et al. Visceral fat accumulation and insulin resistance are important factors in nonalcoholic fatty liver disease. J Gastroenterol 2006; 41: Marchesini G, Brizi M, Morselli-Labate AM, Bianchi G, Bugianesi E, McCullough AJ, et al. Association of nonalcoholic fatty liver disease with insulin resistance. Am J Med 1999;107: Haque M, Sanyal AJ. The metabolic abnormalities associated with non-alcoholic fatty liver disease. Best ract Res Clin Gastroenterol 2002;16: Sanyal A, Campbell-Sargent C, Mirshahi F, Rizzo W, Contos M, Sterling R, et al. Nonalcoholic steatohepatitis: association of insulin resistance and mitochondrial abnormalities. Gastroenterology 2001;120: Kabir M, Catalano KJ, Ananthnarayan S, Kim S, Van Citters GW, Dea MK, et al. Molecular evidence supporting the portal theory: a causative link between visceral adiposity and hepatic insulin resistance. Am J hysiol 2005;288:E454 E Bajaj M, Suraamornkul S, ratipanawatr T, Hardies LJ, ratipanawatr W, Glass L, et al. ioglitazone reduces hepatic fat content and augments splanchnic glucose uptake in patients with type 2 diabetes. Diabetes 2003;52: Gastaldelli A, Miyazaki Y, ettiti M, Santini E, Ciociaro D, Defronzo RA, et al. The effect of rosiglitazone on the liver: decreased gluconeogenesis in patients with type 2 diabetes. J Clin Endocrinol Metab 2006;91: Samuel VT, Liu ZX, Qu X, Elder BD, Bilz S, Befroy D, et al. Mechanism of hepatic insulin resistance in non-alcoholic fatty liver disease. J Biol Chem 2004;279: Marchesini G, Brizi M, Bianchi G, Tomassetti S, Bugianesi E, Lenzi M, et al. Nonalcoholic fatty liver disease: a feature of the metabolic syndrome. Diabetes 2001;50: Belfort R, Harrison SA, Brown K, Darland C, Finch J, Hardies J, et al. A placebo-controlled trial of pioglitazone in subjects with nonalcoholic steatohepatitis. N Engl J Med 2006;355: Matsuda M, DeFronzo RA. Relationship between insulin sensitivity in adipose tissue, liver, muscle and components of the insulin resistance syndrome. Diabetes 1997;46:68A. Received August 7, Accepted April 19, Address requests for reprints to: Amalia Gastaldelli, hd, Institute of Clinical hysiology, National Research Council, via Moruzzi 1, isa, Italy. amalia@ifc.cnr.it; fax (39) Supported by a National Institutes of Health grant (DK-24092), a General Clinical Research Center grant (M01-RR-01346), a Veterans Administration Merit Award, and funds from the Veterans Administration Medical Research Service. The authors thank their nurses for assistance in performing the insulin clamp studies, S. Frascerra, hd, S. Baldi, hd, and E. Santini, BS, for technical assistance, and L. Albarado and E. Chapa for secretarial assistance.