| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Medicine, Divisions of Cardiology (N.M., S.M., M.-R.T.) and Diabetes (J.W., S.V., H.Y.-J.) and Helsinki Medical Imaging Centre, Department of Radiology (N.L.), University of Helsinki, FIN-00029 Helsinki, Finland
Address all correspondence and requests for reprints to: Professor Marja-Riitta Taskinen, Department of Medicine, University of Helsinki, P.O. Box 700, FIN-00029 Helsinki, Finland. E-mail: Marja-Riitta.Taskinen{at}helsinki.fi.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Patients/Interventions: Twenty-nine men were allocated into subgroups with either low (
5%) or high (>5%) liver fat measured with magnetic resonance proton spectroscopy. Subjects underwent an oral fat tolerance test with measurements of postprandial TG, cholesterol, apolipoprotein B-48 (apoB-48), and apoB-100 in TRL fractions, a euglycemic hyperinsulinemic clamp, and determination of abdominal fat volumes by magnetic resonance imaging.
Results: Subjects with high liver fat displayed increased response of postprandial lipids in plasma, chylomicron, and very-low-density lipoprotein 1 (VLDL1) (Svedberg flotation rate 60–400) fractions. Liver fat correlated positively with postprandial responses (area under the curve) of TG (r = 0.597; P = 0.001), cholesterol (r = 0.546; P = 0.002), apoB-48 (r = 0.556; P = 0.002), and apoB-100 (r = 0.42; P = 0.023) in the VLDL1 fraction. Respective incremental areas under the curve correlated significantly with liver fat. Fasting adiponectin levels were inversely correlated with both postprandial lipids and liver fat content. Liver fat remained the only independent correlate in a multiple linear regression analysis for chylomicron and VLDL1 responses.
Conclusions: Liver fat content is a close correlate of postprandial lipids predicting the responses of TRL in chylomicrons and VLDL1 better than measures of glucose metabolism or body adiposity. Low adiponectin concentration is closely linked to high liver fat content and impaired TRL metabolism. High liver fat content associated with postprandial lipemia represents potential risk factors for cardiovascular disease.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Exaggerated response of postprandial TG-rich lipoproteins (TRL) and the accumulation of cholesterol-rich remnant particles stand as a risk marker for CHD and atherogenesis in subjects both with and without type 2 diabetes (3, 4, 5, 6, 7). The increase of TRL in the circulation has adverse effects on the metabolism of both LDL and HDL species as well as on the arterial wall (8). TRL particles may be directly trapped to vessel wall and cause endothelial dysfunction and oxidative stress (9, 10, 11). Therefore, excessive postprandial lipemia may be a more important contributor to CHD risk than can be assumed from TG values alone.
Nonalcoholic fatty liver disease is characterized by increased accumulation of TG in hepatocytes, i.e. hepatic steatosis reflecting imbalance in hepatic TG homeostasis. Hepatic steatosis has been recognized as a novel component of the metabolic syndrome (12). Coexistence of hepatic insulin resistance and hepatic steatosis is a common finding in people with type 2 diabetes (13, 14, 15). Interestingly, liver fat content correlates with the components of diabetic dyslipidemia including fasting plasma TG, HDL cholesterol, and LDL size (16, 17). The specific origin of intrahepatocyte TG is poorly understood in man. In insulin resistance, dysregulation of free fatty acid (FFA) metabolism resulting in overflow of FFA into the liver is one major cause for hepatic fat accumulation. As the amount of visceral fat increases, it directly contributes to enhanced hepatic FFA delivery (18). Consistently, visceral adipose tissue mass is a predictor of hepatic fat content (19).
TG may be stored as lipid droplets in the liver, secreted as very-low-density lipoprotein (VLDL) particles, or metabolized via mitochondrial ß-oxidation. Importantly, defective insulin signaling interrupts the intricate network of nuclear receptors regulating the lipid homeostasis in the liver. The question of whether hepatic insulin resistance is a cause or a consequence of hepatic steatosis is unresolved (14, 20). Nevertheless, the lipid availability in the liver regulates the assembly and secretion of VLDL particles. Recently, we have shown that liver fat content is one of the best predictors for overproduction of large VLDL1 particles seen in type 2 diabetes (21). Because the components of diabetic dyslipidemia are closely linked to the elevation of VLDL1 particles, we hypothesize that liver fat content is also a predictor of postprandial lipemia.
The aim of the present study was to examine the relationship between liver fat content and the response of postprandial TRL species. Subjects with various degrees of insulin sensitivity were included in this study to obtain a wide range in liver fat content. We used the cutoff point of 5% liver fat quantified by magnetic resonance spectroscopy as the upper normal limit based on a recent population study (22) to allocate the subjects into two subgroups with either low (5%) or high (>5%) liver fat. We measured postprandial responses of intestinal and liver-derived TRL species to a standard oral fat tolerance test.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Twenty-nine male subjects were recruited for the study. The subjects were recruited by advertisement in local newspapers and from a Finnish database, Health 2000, which is a population-based cohort. Inclusion criteria included no known diagnoses other than type 2 diabetes or hepatic steatosis. Each subject underwent a physical examination and laboratory tests to exclude hepatic diseases (other than hepatic steatosis) and renal, thyroid, and hematological abnormalities. Study subjects with CHD, retinopathy, microalbuminuria, total cholesterol more than 6.2 mmol/liter, TG more than 5.0 mmol/liter, body mass index (BMI) more than 40 kg/m2, or regular daily alcohol consumption over two doses (i.e. 5 g pure alcohol) were excluded from the study. Subjects without known diabetes underwent an oral glucose tolerance test. The diagnosis of type 2 diabetes was based on the use of oral hypoglycemic agents or an oral glucose tolerance test. Twelve subjects fulfilled the criteria for type 2 diabetes. None of the subjects were taking lipid-lowering treatment or insulin. Three patients with type 2 diabetes were treated with diet alone, one with diet and sulfonylurea, three with diet and metformin, and five with a combination of diet, sulfonylurea, and metformin. The duration of diabetes was 6.4 ± 4.9 yr (range from 0–13 yr). The participants abstained from alcohol and physical exercise at leisure time for 2 d before each examination. The Helsinki University Central Hospital Ethics Committee approved the study design, and each subject gave written informed consent before participation in the study. All the samples were collected in accordance with the Helsinki declaration.
Oral fat tolerance test
After a 12-h fast, the subjects underwent a standardized fat-rich mixed test meal (oral fat tolerance test) consisting of bread, butter, cheese, sliced sausage, a boiled egg, fresh paprika, soured whole milk, orange juice, and coffee. The meal contained 72 g fat (polyunsaturated vs. saturated fatty acid ratio of 0.08, 490 mg cholesterol), 50 g carbohydrates, and 35 g protein. The total energy content was 1000 kcal. Blood samples were drawn before the meal and at 3, 4, 6, and 8 h after the meal.
Determination of liver fat content
Liver fat was measured with magnetic resonance proton spectroscopy as previously described in detail (19). Briefly, localized single voxel (2 x 2 x 2 cm3) proton spectra were recorded using a 1.5-T whole-body system device (Siemens Magnetom Vision, Erlangen, Germany) using a combination of whole-body and loop surface coils for radiofrequency transmitting and signal receiving. Chemical shifts were measured relative to water signal (Swater) intensity at 4.80 ppm. The methylene signal intensity, which represents intracellular TG, was measured at 1.4 ppm (Sfat). Signal intensities were quantified by using an analysis program VAPRO-MRUI (http://www.mrui.uab.es/mrui/). The percent liver fat was calculated by dividing 100 times Sfat by the serum of Sfat and Swater. The measurement of hepatic fat by proton spectroscopy has been validated against histologically determined lipid content of liver biopsies in humans (23) and against estimates of fatty infiltration by computed tomography (24).
Determination of abdominal fat volumes
Intraabdominal and sc fat volumes were determined using magnetic resonance imaging as previously described (19). A series of 16 T1-weighted transaxial scans were acquired from a region extending from 8 cm above to 8 cm below the fourth and fifth lumbar interspaces. Intraabdominal and sc fat areas were measured using an image analysis program (Alice 3.0; Parexel, Waltham, MA).
Hyperinsulinemic euglycemic clamp
On a separate day, whole-body insulin sensitivity was determined by using the euglycemic insulin clamp technique (25). The patients were admitted to the metabolic unit at 0730 h after an overnight fast. An indwelling cannula was inserted in an antecubital vein for infusions, and a second cannula was inserted retrogradely in a heated dorsal hand vein to obtain arterialized venous blood samples. Insulin (Actrapid Human; Novo Nordisk, Bagsvaerd, Denmark) was infused in a primed-continuous fashion (1 mU/kg·min) for 8 h. Plasma glucose was maintained at its fasting concentration by a variable rate of 20% glucose infusion. The infusion rate was determined empirically based on plasma glucose measurements performed every 5–10 min from arterialized venous blood. Whole-body insulin sensitivity (M-value) was calculated from the mean infusion rates of the second hour of glucose infusion after correcting for changes in the glucose pool size and was expressed per body weight (milligrams per minute per kilogram).
Separation of lipoproteins and biochemical analyses
Chylomicrons [Svedberg flotation rate (Sf) > 400], VLDL1 (Sf 60–400), VLDL2 (Sf 20–60), and intermediate-density lipoprotein (Sf 12–20) were separated by the density gradient ultracentrifugation as previously described (26). Aliquots of the isolated fractions were frozen immediately for subsequent determinations of TG, cholesterol, apolipoprotein B-48 (apoB-48), and apoB-100. Concentrations of apoB-48 and apoB-100 were measured using SDS-PAGE as described previously (26, 27). After photographing (electrophoresis documentation and analysis system 120; Kodak, Rochester, NY) the Coomassie blue-stained gels, the bands representing apoB-48 and apoB-100 were analyzed with Image-Master 1-D software (Amersham Pharmacia Biotech, Little Chalfont, UK). The detection limit for apoB-48 and apoB-100 ranged from 0.01–0.02 mg/liter.
TG and cholesterol concentrations were analyzed in total plasma and lipoprotein fractions by automated enzymatic methods using the Cobas Mira S analyzer (Hoffmann-La Roche, Basel, Switzerland). Concentrations of glucose, serum FFA, C-peptide, and insulin were measured in samples obtained during the fat tolerance test at each time point as previously described (26). Fasting plasma adiponectin levels were determined with the human adiponectin ELISA kit (B-Bridge International, San Jose, CA).
Statistical analyses
Statistical comparisons of data were performed using SPSS Statistical Package (version 11.0; SPSS Inc., Chicago, IL). Data are presented as mean ± SD or as median and range. The area under curve (AUC) and incremental AUC (IAUC) for postprandial variables were calculated according to the trapezoid rule (28). We used Shapiro-Wilk’s test to explore the distribution of each variable and applied log10 transformations with variables representing nonnormal distribution. Variables were compared between subjects with low liver fat (5%) and high liver fat (>5%) content using Student’s t test or Mann-Whitney U test when assumptions of normal distribution after log transformation were not met. We calculated also Spearman’s rank correlations between liver fat, visceral fat, sc fat, adiponectin, and postprandial lipid parameters in a pooled analysis. These results were confirmed with partial correlations controlling for plasma glucose, sc fat, and visceral fat. A forward stepwise multivariate linear regression model was used for searching for independent correlates of postprandial responses in chylomicron and VLDL1 fractions. Variables with univariate correlation with P < 0.05 were included in the multivariate regression. Also, we included sc and visceral fat volumes, BMI, and presence or absence of diabetes in the multivariate analysis, even though we did not find a univariate correlation. We considered a P value less than 0.05 as statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The characteristics of the subjects are summarized in Table 1
. Subjects with diabetes were well controlled with mean glycosylated hemoglobin of 6.7 ± 0.82%. Ten of 12 subjects with type 2 diabetes were in the high-liver-fat group and, respectively, five of 17 subjects without diabetes. As expected, the two groups showed significant differences in fasting insulin, C-peptide, and glucose levels and in M-value (Table 1). Likewise, total, visceral, and abdominal fat volumes were increased in subjects with high liver fat. The plasma adiponectin concentration was significantly lower in subjects with high liver fat. Fasting lipid and lipoproteins are represented in Table 1. Subjects with high liver fat had higher fasting TG and apoB and lower fasting lipoprotein A1 than subjects with low liver fat. Fasting or postprandial FFA concentrations did not differ between the groups.
|
Subjects with high liver fat showed impeded disposal of postprandial lipids and lipoproteins in plasma, chylomicron, and VLDL1 fractions. Figure 1 highlights the findings in the VLDL1 fraction. Postprandial TG response (both AUC and IAUC) in total plasma and in the VLDL1 fraction were significantly higher in subjects with high liver fat than in those with low liver fat. Likewise, in the chylomicron fraction, the TG concentration tended to be higher among subjects with high liver fat. ApoB-48 and cholesterol responses in both chylomicron and VLDL1 fractions were elevated in subjects with high liver fat. ApoB-100 in the chylomicron fraction was also higher in subjects with high liver fat, and a similar trend was seen for the VLDL1 fraction. In contrast, AUC and IAUC responses of TG, cholesterol, apoB-48, and apoB-100 in VLDL2 and intermediate-density lipoprotein fractions were similar in the two groups (data not shown).
|
We calculated Spearman correlation coefficients to explore the relationship between liver fat content, fat compartments, and postprandial lipid and lipoprotein parameters. Postprandial responses (AUC) of TG, cholesterol, apoB-48, and apoB-100 in the VLDL1 fraction correlated positively with liver fat content as depicted in Fig. 2 and Table 2. Likewise, respective IAUC measures of the VLDL1 fraction displayed significant correlations with liver fat content (data not shown. Chylomicron TG, apoB-48, and apoB-100 AUC also correlated with liver fat content (TG AUC vs. liver fat, r = 0.36 and P = 0.055; apoB-48 AUC vs. liver fat, r = 0.504 and P = 0.005; apoB-100 AUC vs. liver fat, r = 0.597 and P = 0.001) as well as chylomicron apoB-48 and apoB-100 IAUC (data not shown). We also calculated partial correlations for respective postprandial parameters vs. liver fat, and the results prevailed even after controlling for plasma glucose, sc fat, and visceral fat (data not shown). M-value correlated with postprandial lipids and lipoproteins in the VLDL1 fraction (Table 2). Any measures of the postprandial TG, apoB-48, apoB-100, or cholesterol failed to correlate with visceral, sc, or total abdominal fat measurements.
|
|
Plasma adiponectin was negatively associated with liver fat content (Fig. 3). Adiponectin levels also correlated inversely with postprandial lipid responses in the VLDL1 fraction as shown in Table 2
. Significant correlations existed also between adiponectin and VLDL1 apoB-48 IAUC (r = –0.409; P = 0.028) chylomicron apoB-100 AUC (r = –0.401; P = 0.031), VLDL2 TG AUC (r = –0.446; P = 0.015), VLDL2 cholesterol AUC (r = –0.39; P = 0.037), and VLDL2 apoB-100 AUC (r = –0.483; P = 0.008). In contrast, no correlations were observed between adiponectin and sc fat volumes, visceral fat, or total abdominal fat. Postprandial FFA area responses failed to show any correlations with liver fat, abdominal fat compartments, or adiponectin levels. Postprandial FFA, glucose, insulin, or C-peptide responses (AUC) did not correlate with adiponectin levels.
|
Stepwise regression analysis
Table 3 shows results of the forward stepwise multivariate linear regression model. The independent variables in each model were liver fat, TG, HDL, glucose, insulin, M-value, sc and visceral fat volumes, presence or absence of diabetes, BMI, and adiponectin. Liver fat remained as a significant independent variable when IAUC for chylomicron apoB-48, chylomicron apoB-100, VLDL1 TG, and VLDL1 cholesterol were set as dependent variables.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Fatty liver, considered to be a key component of the metabolic syndrome (13, 32), is characterized by accumulation of TG in the hepatocytes. TG deposited in hepatocytes originates from increased FFA delivery from the plasma FFA pool reflecting the flux from the adipose tissue, from de novo lipogenesis, or from postprandial metabolism of chylomicrons. Regulation of fatty acid metabolism is highly different in fasting and fed states. Recent kinetic data (18) in obese hypertriglyceridemic and hyperinsulinemic subjects demonstrated that 59% of TG within hepatocytes arose from the plasma FFA pool in the fed state. Respectively, de novo lipogenesis and dietary TG accounted for 26 and 15% of the hepatocyte TG content. Because both intestinal and hepatic sources of TG contribute to development of fatty liver, in the postprandial period, both apoB-48 and apoB-100 in the chylomicron and VLDL1 fraction are expected to be elevated in subjects with high liver fat as observed in the present study. Our study design does not allow drawing conclusions whether increased TRL production, impeded lipolysis, decreased catabolism, or all of these account for postprandial lipemia. We demonstrated that the postprandial TRL response is strongly correlated with the liver fat content. Unexpectedly, postprandial FFA levels were not linked to liver fat. This may be due to the fact that only minor changes occurred in plasma FFA concentration during an oral fat load. Fatty liver, like type 2 diabetes, associates with impaired suppression of plasma FFA by insulin (21). We cannot exclude that FFA spillover from dietary fat and/or TRLs secreted from liver in the subjects with high liver fat in addition to excess FFA flow contributes to liver fat accumulation.
TG within the hepatocytes provide ample substrate for lipid synthesis. Our recent kinetic data demonstrate both that hepatic TG represent the driving force for VLDL assembly and that secretion of large VLDL1 increases with increasing liver fat content (21, 33). These mechanisms could partly explain our results. In the setting of VLDL1 overproduction, the competition of both intestinal and liver-derived particles may supersaturate the catabolic pathways. Elevation of apoB-48 in VLDL1 particles, including partly chylomicron remnants, suggests that remnant removal may be defective. Therefore, high liver fat content may impede both apoB-48 and apoB-100 TRL remnant clearance by hepatic receptors.
Excess abdominal obesity is associated with increased hepatic fat (19). In agreement, we observed a strong relationship between liver fat content and total abdominal, sc, and visceral fat volumes. Degree of insulin resistance and hyperinsulinemia, i.e. M-value, postprandial insulin, and C-peptide areas, also correlated strongly with hepatic fat content as well as with measures of abdominal fat compartments. However, measures of abdominal fat compartments were not related to postprandial TRL response. Taken together, our results are in concordance with the tenet that the liver fat is the key player in the metabolic syndrome. In agreement, Tiikkainen et al. (34) demonstrated that the hepatic fat content is more closely related to indexes of insulin resistance than degree of obesity. It is well documented that fatty liver fails to suppress glucose production in response to insulin in people both with and without diabetes (35, 36, 37). The question of whether hepatic insulin resistance is a consequence of increased liver fat or vice versa is a matter of a hectic debate (38). The fact that the subjects with high liver fat were predominantly people with diabetes is a potential weakness, and therefore our study can be considered as a pilot study. Recent data provide evidence that the prevention of liver fat accumulation in animal models is associated with reversal of hepatic insulin resistance (39, 40). Notably, the reduction of liver fat content achieved by weight loss improves hepatic sensitivity to insulin and restores the hepatic glucose production rate (41). Interestingly, in a study including subjects with normal and impaired glucose tolerance matched individually for similar volumes of visceral fat, the postprandial TRL-TG responses were comparable irrespective of hepatic glucose sensitivity. However, the caveat is that the liver fat content was not assessed (42).
We are not aware of previous studies reporting the interrelations between postprandial apoB-48 and apoB-100 responses in TRL fractions, adiponectin, and liver fat. In the present study, low plasma adiponectin levels were associated with high TRL response, especially in chylomicron and VLDL1 fractions. Recent data demonstrated a positive correlation between adiponectin and lipoprotein lipase activity and, on the other hand, a negative correlation between adiponectin and hepatic lipase activity both in nondiabetic and diabetic subjects (43, 44). These data suggest a role for adiponectin in the catabolic process of TRL (45). In agreement, Chan et al. (30) reported a correlation between adiponectin and fasting lipids including apoB-48, apoCII, VLDL TG, and remnant-like particle-cholesterol in nondiabetic males. Recent data from stable isotope kinetic studies have demonstrated a positive correlation between fractional catabolic rates of VLDL apoB and plasma adiponectin concentration (29). These associations were independent of degree of insulin sensitivity and size of adipose tissue compartments. We have reported that adiponectin correlated positively with VLDL1 apoB and VLDL1 TG fractional catabolic rates (21), further strengthening the role of adiponectin as a regulator of VLDL1 lipolysis.
However, in multivariate analysis, only liver fat and fasting TG, but not adiponectin, remained a significant predictor of postprandial VLDL1 response. Our results suggest that low adiponectin is closely associated with, but probably not directly related to, postprandial TRL metabolism. As expected, we observed a close correlation between adiponectin levels and liver fat content. On the contrary, adiponectin levels did not correlate with measures of abdominal obesity or postprandial insulin and C-peptide. Thus, low adiponectin may represent a surrogate marker of hepatic steatosis.
In the present study, we have linked postprandial elevations of TRL of both intestinal and hepatic origin to high hepatic fat content and low adiponectin levels. Elevations of postprandial TRL exposes liver to excess FFA and can influence the hepatocellular lipid metabolism leading to hepatic steatosis and disturbances of insulin signaling pathways. Postprandial lipemia is an important factor in the vicious circle of lipid dysmetabolism and consequently a potential target for therapies in patients with high liver fat content.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: N.M., S.M., J.W., S.V., N.L., H.Y.-J., and M.-R.T. have nothing to declare.
First Published Online May 8, 2007
Abbreviations: Apo, Apolipoprotein; AUC, area under curve; BMI, body mass index; CHD, coronary heart disease; FFA, free fatty acids; HDL, high-density lipoprotein; IAUC, incremental area under curve; LDL, low-density lipoprotein; S, signal; Sf, Svedberg flotation rate; TG, triglyceride; TRL, TG-rich lipoprotein; VLDL, very low density lipoprotein.
Received January 25, 2007.
Accepted May 1, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
prevents hepatic insulin resistance in nonalcoholic fatty liver disease. J Clin Invest 117:739–745[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  A. M. DeAngelis, G. Heinrich, T. Dai, T. A. Bowman, P. R. Patel, S. J. Lee, E.-G. Hong, D. Y. Jung, A. Assmann, R. N. Kulkarni, et al. Carcinoembryonic Antigen-Related Cell Adhesion Molecule 1: A Link Between Insulin and Lipid Metabolism Diabetes, September 1, 2008; 57(9): 2296 - 2303. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, Subjects with low liver fat; 
, subjects with high liver fat. The P value is given for AUC difference (see Subjects and Methods). Bars (white, subjects with low liver fat; black, subjects with high liver fat) in the insets represent respective IAUC values. Data are mean ± SE.
