Association of liver disease with postprandial large intestinal triglyceride-rich lipoprotein accumulation and pro/antioxidant imbalance in normolipidemic non-alcoholic steatohepatitis

Abstract

Background. Dietary fat excess and antioxidant deficiency, altered lipid metabolism, and increased lipoperoxidation have been associated with non-alcoholic steatohepatitis (NASH), but the relative importance of each of these factors is unclear.

Aims. To assess acute intestinal and hepatic very-low-density lipoprotein (VLDL) subfraction metabolism, lipid peroxidation, and pro/antioxidant imbalance after a fat load in NASH.

Methods. Dietary habits, circulating adipokines, fasting and postprandial lipids, intestinal and hepatic VLDL, oxidized low-density lipoproteins (oxLDL), and total antioxidant status (TAS) were correlated to postprandial liver enzymes and to liver histology in 28 non-obese non-diabetic normolipidemic patients with NASH and 28 healthy controls.

Results. Despite similar fasting profiles, NASH had more pronounced intestinal and hepatic VLDL1 accumulation, LDL lipid peroxidation and TAS fall postprandially. Postprandial intestinal VLDL1 independently predicted oxLDL and TAS responses in NASH. In NASH, hepatic steatosis was independently associated with postprandial intestinal VLDL1 and TAS; necroinflammation with postprandial serum gamma-glutamyltransferase, oxLDL and TAS responses; and fibrosis with adiponectin and postprandial TAS and oxLDL responses.

Conclusions. Postprandial intestinal VLDL1 accumulation is associated with a pro-oxidant imbalance in normolipidemic non-diabetic NASH, and both correlate with the severity of liver disease. Modulating postprandial lipoprotein metabolism may be beneficial in NASH, even if normolipidemic.

• Adiponectin
• GGT
• lipemia
• oxidized LDL
• TAS
• VLDL subfractions

Introduction

Non-alcoholic steatohepatitis (NASH) has been linked to insulin resistance and the metabolic syndrome in obese diabetic as well as in lean non-diabetic subjects 1, but mechanisms underlying this association are unclear. Increased lipid peroxidation and decreased antioxidant defense may have a role in the pathogenesis of NASH and of insulin resistance, alone or in association with altered adipokine signaling 1–4.

Experimentally, dietary fat induces fatty liver independently of primary changes in insulin sensitivity or body adiposity, but the exact mechanisms are unclear 5, 6. Similarly, postprandial lipemia induced an acute decrease in insulin sensitivity in healthy men 7. Taken together, these observations suggest both fatty liver and insulin resistance may be a consequence of impaired lipid metabolism.

Hepatic triglyceride-rich lipoproteins contain one apolipoprotein apoB100 molecule per particle; intestinally secreted triglyceride-rich lipoproteins contain one apoB48 molecule per particle throughout their life, are currently named VLDL, and contain those particles traditionally called chylomicrons. Thus hepatic and intestinal triglyceride-rich lipoproteins can be globally named VLDL, and their origin may be determined by the type of apoB content.

Abbreviations

Table

Apo	apolipoprotein
AUC	area under the postprandial curve
BMI	body mass index
Chol	cholesterol
FFA	free fatty acids
GGT	gamma-glutamyltransferase
IAUC	incremental area under the postprandial curve
IGT	impaired glucose tolerance
ISI	insulin sensitivity index
LPL	lipoprotein lipase
NAFLD	non-alcoholic fatty liver disease
NASH	non-alcoholic steatohepatitis
OGTT	oral glucose tolerance test
oxLDL	oxidized LDL
PUFA	polyunsaturated fatty acids
SFA	saturated fatty acids
TAS	total antioxidant status
Tg	triglycerides
TNF-α	tumor necrosis factor-α
VLDL	very-low-density lipoprotein

Postprandial lipemia is regarded as atherogenic and as a major determinant of oxidative stress in type 2 diabetes 8. During this phase oxidized LDL (oxLDL) and triglyceride-rich lipoprotein-derived remnants of both hepatic and intestinal origin impair vascular endothelial function and activate macrophages 8, 9.

The liver may also be a target organ of exaggerated postprandial lipemia, since it plays a key role in lipoprotein metabolism and takes up LDL and intestinal VLDL through the LDL receptor and the liver-related receptor protein (LRRP). Consistently, the severity of liver histology correlated with early atherosclerosis in NASH, with fibrosis stage predicting more accurately carotid intima-media thickness than the simple liver fat infiltration, independently of insulin resistance, metabolic syndrome, or other traditional cardiovascular risk factors. These findings suggest that inflammation and fibrosis share common biological mediators in the vessel wall and in the liver 10.

In NASH, most hepatic triglycerides (Tg) derive from circulating free fatty acids (FFAs), which are taken up by the liver in a dose-dependent fashion 11. Consistently, FFA concentration correlated with the severity of liver steatosis independently of clamp-measured insulin sensitivity 12, and postprandial FFA and Tg uptake are accelerated and contribute substantially to liver Tg pool in insulin resistance 11, 13.

Key messages

• Postprandial intestinal large VLDL1 accumulation is associated with a dynamic pro-oxidant imbalance and an increase in serum gamma-glutamyltransferase levels; both phenomena correlate with liver injury.

• Modulating postprandial lipoprotein metabolism may beneficially affect oxidative stress and liver injury in non-alcoholic steatohepatitis, even in the absence of overt dyslipidemia.

Hypothesizing that lipid load acutely induces a disequilibrium in systemic pro/antioxidant balance and is related to hepatic fat, necroinflammation, and fibrosis, we assessed postprandial VLDL, oxLDL, and antioxidant responses following an oral fat challenge and correlated them to dietary habits, circulating adipokines, and to the severity of liver disease in non-obese non-diabetic normolipidemic patients with NASH.

Materials and methods

Patient selection

Twenty-eight patients were selected (Table I) according to the following criteria: persistently elevated liver enzymes; ultrasonographic presence of bright liver without any other liver or biliary tract disease. Exclusion criteria were: a history of alcohol consumption >40 g/week, as assessed by a detailed interview extended to family members and by a validated questionnaire filled in daily for one week by the patients; a body mass index (BMI) ≥30 kg/m²; positive markers of viral, autoimmune, or celiac disease; abnormal copper metabolism, serum alpha1-antitrypsin, or thyroid function tests; overt dyslipidemia (fasting serum cholesterol ≥200 mg/dL or plasma triglyceride ≥200 mg/dL); a diagnosis of diabetes mellitus (fasting plasma glucose ≥126 mg/dL or ≥200 mg/dL at +2h on a standard oral glucose load (OGTT)); exposure to occupational hepatotoxins or drugs known to be steatogenic, liver enzyme inducers, or to affect glucose/lipid metabolism. None of the participants in the study had clinically evident cardiovascular disease. Mutations in the hemochromatosis genes HFE and TRF2 were detected in patients and controls.

Table I.  Baseline characteristics of patients with NASH and controls. Data are presented as mean±SEM.

	Controls (n=28)	NASH (n=28)	P
Age (years)	39±3	38±2	0.741
Sex (M/F)	22/6	24/4	0.872
Smokers (n subjects)	8	10	0.916
BMI (kg/m^2)	25.8±0.8	25.6±0.5	0.869
Percent body fat	20±5	21±3	0.857
Waist (cm)	88±2	90±3	0.275
Waist-on-hip ratio	0.90±0.03	0.91±0.01	0.896
Systolic blood pressure (mmHg)	125±3	126±2	0.815
Diastolic blood pressure (mmHg)	78±1	88±1	0.0002
Triglycerides (mg/dL)^a	72±8	93±9	0.218
Total cholesterol (mg/dL)^b	189±9	182±10	0.687
HDL cholesterol (mg/dL)^b	59±3	46±3	0.009
LDL cholesterol (mg/dL)^b	100±5	107±9	0.386
Glucose (mg/dL)	88±3	95±3	0.032
Insulin (µU/mL)	7.8±1.2	16.7±4.6	0.002
Whole-body ISI	8.02±1.20	3.89±1.01	0.0002
Hepatic ISI	1.18±0.12	0.43±0.09	0.0003
AST (U/L)	21±4	46±4	0.021
ALT (U/L)	21±3	81±6	0.0002
GGT (U/L)	18±2	69±4	0.00002
Uric acid (mg/dL)^c	5.18±0.28	6.19±0.32	0.019
Adiponectin (ng/mL)	9860±982	5048±488	0.00002
TNF-α (pg/mL)	0.98±0.05	1.12±0.10	0.201
Leptin (pg/mL)	2579±648	2197±683	0.712
Resistin (ng/mL)	4.50±0.34	4.82±0.41	0.703
HFE mutation (H63D) heterozygotes (n subjects)	5	6	0.489
Serum iron (µg/dL)	84±9	93±6	0.238
Ferritin (µg/L)	142±28	183±31	0.286
Transferrin (% sat)	30±1	34±3	0.216

^aTo convert mg/dL to mmol/L, multiply by 0.01129.

^bTo convert mg/dL to mmol/L, multiply by 0.02586.

^cTo convert mg/dL to µmol/L, multiply by 59.48.

All patients had a histological diagnosis of NASH, as proposed by Brunt 14. Liver iron concentration (LIC) and hepatic iron index (HII) were determined from 2 mg dry weight tissue by atomic absorption spectroscopy.

The controls were 28 healthy subjects matched for age, sex, overall and central adiposity, with normal liver enzymes and abdomen ultrasound (Table I).

Patients and controls gave their consent to the study, which was conducted according to the Helsinki Declaration and was approved by the Ethical Committee of San Giovanni Battista Hospital.

Alimentary record

Patients and controls filled in daily a dietary record: a list of foods was designed; for each item different portion sizes were specified according to the EPIC study 15. The recorded period included a complete week, and the record was collected within one week of the tolerance tests. The diet record was analyzed using the WinFood database (Medimatica, Teramo, Italy) according to the table of food consumption of the Italian National Institute of Nutrition and Food Composition Database for Epidemiological Study in Italy 16.

Anthropometry

Percent body fat was estimated by the bioelectric impedance method (TBF-202, Tanita, Tokyo, Japan), which closely correlates with measures obtained by dual X-ray absorption 17.

Glucose tolerance test (OGTT)

After completion of the alimentary record, patients and controls underwent a standard 75-g oral glucose tolerance test (OGTT). The whole-body and hepatic insulin sensitivity indices (ISI) were assessed from the OGTT, as previously described 18.

Oral fat load

Patients and controls underwent an oral fat load test, as previously described 19. Participants avoided strenuous physical efforts and followed their usual diet during the 24 h preceding the test. The fat load consisted of a mixture of 200 g dairy cream (35% fat) and 26 g egg yolk for a total energy content of 766 kcal: 78.3 g fat (55.6% saturated fatty acids, 29.6% monounsaturated fatty acids, 14.8% polyunsaturated fatty acids), 595 mg cholesterol, 8.8 g protein, 7 g carbohydrate. The fat load was consumed during a period of 5 min; subjects kept fasting on the test morning and avoided strenuous activity, since exercise can reduce postprandial lipemia.

Samples were drawn at 0 (baseline), 2, 4, 6, 8, and 10 hours. Plasma total cholesterol (Chol), triglyceride (Tg), and free fatty acids (FFA) were measured by automated enzymatic methods. ApoE genotype was determined by polymerase chain reaction (PCR) amplification of genomic DNA using specific oligonucleotide primers.

Laboratory analyses

Cytokines

Serum TNF-α, leptin, and adiponectin were measured by sandwich enzyme-linked immunosorbent assay (ELISA; R&D System Europe Ltd, Abingdon, UK). Resistin was measured by an enzyme immunoassay (Bio Vendor Laboratori Medicine, Inc., Brno, Czech Republic).

Liver enzymes

Aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyltransferase (GGT) were evaluated at each time of the fat load test with a kinetic determination according to the International Federation of Clinical Chemistry (Sentinel Ch., Milan, Italy).

Plasma glucose was measured by the glucose oxidase method (HITACHI 911 Analyzer, Sentinel Ch., Milan, Italy), and serum insulin by immunoradiometric assay (Radim S.p.A., Pomezia, Italy; intra-assay variation coefficients (CV): 1.6%–2.2%; inter-assay CV: 6.1%–6.5%). Plasma triglycerides and HDL-cholesterol were measured by enzymatic colorimetric assay (HITACHI 911 Analyzer, Sentinel Ch., Milan, Italy), the latter after precipitation of LDL and VLDL fractions using heparin-MnCl₂ solution and centrifugation at 4°C. Serum uric acid was evaluated with an enzymatic colorimetric method with uricase (HITACHI 911 Analyzer, Sentinel Ch., Milan, Italy).

Triglyceride-rich lipoprotein subfractionation

VLDL were defined by a Sf >20 on preparative ultracentrifugation. After preparative ultracentrifugation their Tg and total Chol contents were subsequently measured.

Briefly, plasma was brought to a density of 1.10 g/mL by adding solid KBr. The density gradient was prepared by adding to 4 mL of this plasma, 3 mL of a 1.065 g/mL solution containing 0.05% KBr/NaCl plus ethylenediaminetetraacetate (EDTA) (pH 7.4); 3 mL of a similar solution at 1.020 g/mL; 3 mL of physiological saline at 1.006 g/mL. The sample was ultracentrifuged in a Beckman L8-70M centrifuge at 20°C in stages allowing the separation of two VLDL subfractions with decreasing Sf values: VLDL1: Sf >100; VLDL2: Sf = 20–100. The automated methods mentioned above were used to determine Chol and Tg of the two fractions. The origin (hepatic or intestinal) of each subfraction was defined by the type of ApoB content (ApoB100 for hepatic VLDL and ApoB48 for intestinally derived VLDL). Large intestinal VLDL1 subfraction included the particles previously named chylomicrons 20. For the above-mentioned reasons, the concentration of hepatic/intestinal VLDL particles is equivalent to the concentration of apoB100/B48.

ApoB48 and ApoB100 contents of VLDL subfractions were quantified by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis as previously described 21. Non-delipidated samples were reduced in SDS buffer for 4 minutes at 96°C. Samples were applied to the gel and run at 40 mA in 0.025 M Tris, 0.192 M glycine, and 0.1% SDS. Gel was stained with Silver Stain (Bio-Rad). Since the chromogenicity of ApoB48 is similar to that of ApoB100, a protein standard was prepared from LDL isolated by sequential ultracentrifugation and used to quantify ApoB100 and ApoB48. The bands were quantified by densitometry using Gel Doc equipment (Bio-Rad). Density values were assigned to the ApoB100 bands of the standard LDL, and a standard curve was constructed. The values were recalculated by linear regression.

Measurement of LDL lipid peroxidation

LDL-conjugated dienes were determined during the fat load test, as follows. Capillary electrophoresis was performed as described by Stocks et al. 22. The cathode and anode electrolytes and the capillary run buffer were 40 mM methylglucamine-Tricine, pH 9.0. LDL samples were injected by low pressure for 4 seconds. Dimethylformamide was injected as an electroendosmotic flow (EOF) marker for 1 second. A voltage of 24 kV was applied ramping over 0.8 min. Migration of LDL particles was monitored at 200 and 234 nm. The amount of conjugated dienes was obtained from the percentage of the height of LDL peak at 234 nm related to the height of LDL peak at 200 nm.

Measurement of total antioxidant status

The evaluation of antioxidant levels in samples (TAS) is based on the reduction of Cu^+ + into Cu⁺ by the action of all present antioxidants. The amount of Cu⁺ is evaluated through measuring the complex formed by Cu⁺ and bathocuproine. This complex has a typical absorption at 490 nm (ANTOXT Kit by Fujirebio Diagnostics AB Göteborg, Sweden) 23, 24.

DNA analyses

DNA was isolated from anticoagulated blood using commercially available reagents (Qiagen, Genenco, M-Medical srl, Florence, Italy). HFE and TRF2 genes were amplified in a single, multiplex PCR reaction with primers of a commercial kit (Nuclear Laser Medicine, Milan, Italy).

ApoE genotypes were determined with polymerase chain reaction (PCR) amplification of genomic DNA with specific oligonucleotide primers 25.

Statistical analysis

Data were expressed as mean±SEM. Differences were considered statistically significant at P<0.05.

Differences between groups were analyzed by analysis of variance (ANOVA) when variables were normally distributed; otherwise the Mann-Whitney test was used. Normality was evaluated by Shapiro-Wilk test. Chi-square test or Fisher's exact test were used to compare categorical variables.

Data from the oral fat load were compared by ANOVA and Scheffè post hoc test after log normalization of skewed variables.

The area under the curve (AUC) and incremental AUC (IAUC) of plasma lipids, lipoproteins, and liver enzymes during the oral fat load were computed by the trapezoid method.

Simple and multiple regression analyses were used to estimate linear relationship between different variables in NASH and controls, after log transformation of skewed data.

Logistic regression analysis was used to identify independent predictors for necroinflammation grade 3 or fibrosis stage 3. The covariates were age, ISI index, adiponectin, presence/absence of metabolic syndrome, fasting and postprandial levels of GGT, FFA, Tg, oxLDL, TAS, and intestinal and hepatic VLDL1.

Results

Subject characteristics

The main features of NASH patients and controls are reported in Table I.

Adopting the Adult Treatment Panel (ATP) III criteria for definition of the metabolic syndrome 26, 72% patients had hypertension (systolic/diastolic blood pressure ≥130/85 mmHg), 11% were hypertriglyceridemic (fasting plasma triglycerides ≥150 mg/dL), 19% had low plasma HDL-cholesterol (HDL-C <40 mg/dL in men and <50 mg/dL in women), 31% had impaired glucose regulation (14% had impaired fasting glycemia, i.e. fasting plasma glucose ≥100 mg/dL but <126 mg/dL, and 17% had impaired glucose tolerance, i.e. plasma glucose ≥140 mg/dL at +2h on OGTT), and 11% had abdominal obesity (waist circumference >102 cm in men and >88 cm in women).

Twenty-nine percent of patients with NASH had the metabolic syndrome (three or more criteria met).

Whole-body and hepatic ISI were significantly lower in the NASH group than in controls: ISI: 3.89±1.01 versus 8.02±1.20; P=0.0002; hepatic ISI: 0.43±0.09 versus 1.18±0.12; P=0.0003.

Serum adiponectin levels were significantly lower in patients with NASH than in controls, while there was no difference in the other cytokines.

Histopathology

Fatty infiltration was mild (involving 5%–33% of hepatocytes) in 9 patients, moderate (33%–66% of hepatocytes) in 11, and severe (>66% of hepatocytes involved) in 8 patients. Necroinflammatory activity was grade 1 in 8 patients, grade 2 in 9, and grade 3 in 11. Fibrosis was stage 0 in 5 patients, 1 in 6 patient, 2 in 8, and 3 in 9 patients. Cirrhotic changes were absent.

Liver iron concentration was 14±4 µmol/g dry weight, and hepatic iron index was 0.52±0.05.

Alimentary record

The daily caloric and macronutrient intakes of the two groups were similar: total calories: 2561±158 versus 2482±151 kcal, P=0. 901; carbohydrate: 50±3 versus 48±3% kcal, P=0.683; protein: 18±3 versus 19±3% kcal, P=0.642; fat: 32±3 versus 32±4% kcal, P=0.781. The NASH group consumed more saturated fat and less polyunsaturated fat than controls: saturated fatty acids (SFA): 13.6±0.9 versus 9.0±0.4% tot kcal, P=0.0007; polyunsaturated fatty acids (PUFA): 3.5±0.2 versus 5.1±0.3% tot kcal, P=0.001. NASH patients also had a significantly lower daily intake of antioxidant vitamins A (571±182 versus 1105±198 mg; P = 0.0006), C (118±16 versus 161±08 mg; P = 0.004), and E (5.7±0.3 versus 9.6±0.6 mg; P = 0.0002). Daily alcohol intake was similar in the two groups: 7±2 versus 9±4 g; P = 0.418. The dietary habits of controls were comparable to those of a large sample of healthy Piedmont population, as assessed by a recent alimentary survey 15, 16. Vitamin A and E intake of patients with NASH were lower than the recommended intake in the Italian adult population (recommended intake: vitamin E: 8 mg/day; vitamin A: 700 µg/day for males; 600 µg/day for non-pregnant females) 27.

Oral fat tolerance test

There was no significant difference in ApoE allelic frequency: 20 NASH patients and 18 controls were E3/E3, 6 NASH subjects and 7 controls were E3/E4, and 2 NASH subjects and 3 controls were E2/E3.

Postprandial plasma total Tg and FFA responses were significantly higher in the NASH group than in controls (Table II). Fasting LDL-conjugated diene levels did not differ between the two groups, but rose significantly in NASH postprandially, becoming significantly higher than controls at +2h and peaking at +6h. (Figure 1A). The incremental area under the curve (IAUC) of LDL-conjugated dienes was significantly higher in the NASH group than in controls (Table II). LDL-C levels did not change throughout the test.

Figure 1.  Oral fat load test. Postprandial plasma LDL-conjugated dienes (Figure 1A) and TAS (Figure 1B) responses in patients with NASH (n=28) and controls (n=28). Data are presented as mean±SEM. ★ P<0.05 versus controls P<0.01 versus controls ▴ P<0.05 versus fasting levels.

Table II.  Oral fat test parameters (mean±SEM) in NASH and controls.

	Controls (n=28)	NASH (n=28)	P
Fasting Tg (mg/dL)	72±8	93±9	0.218
IAUC Tg (mg/dL×h)	168±41	518±80	0.0001
Fasting FFA (mMol/L)	0.47±0.03	0.68±0.04	0.002
IAUC FFA (mMol/L×h)	1.89±0.38	5.61±2.89	0.003
Fasting LDL conjugated dienes (uA 234 nm/uA 200 nm×100)	6.30±0.68	7.25±0.89	0.129
IAUC LDL conjugated dienes (uA 234 nm/uA 200 nm×100×h)	0.3±0.2	11.7±2.1	0.00001
Fasting TAS (mMol/L)	0.378±0.091	0.356±0.151	0.445
IAUC TAS (mMol/L×h)	0.017±0.028	−0.179±0.032	0.0001
Fasting GGT (U/L)	18±2	69±4	0.00002
IAUC GGT (U/L×h)	1±1	54±3	0.00003
Fasting VLDL1-apoB48 (mg/dL)	2.29±0.42	2.27 ±0.23	0.967
IAUC VLDL1-apoB48 (mg/dL×h)	5.12±0.83	12.89±1.09	0.00009
Fasting VLDL2-apoB48 (mg/dL)	2.23±0.88	1.12±0.38	0.256
IAUC VLDL2-apoB48 (mg/dL×h)	6.27±2.01	4.19±1.63	0.362
Fasting VLDL1-apoB100 (mg/dL)	6.58±0.92	7.14±0.78	0.638
IAUC VLDL1-apoB100 (mg/dL×h)	10.73±1.89	23.60±3.22	0.002
Fasting VLDL2-apoB100 (mg/dL)	7.33±1.12	2.84±0.99	0.006
IAUC VLDL2-apoB100 (mg/dL×h)	11.28±2.69	8.59±1.46	0.384

Fasting plasma total antioxidant status (TAS) did not differ between patients and controls. It decreased postprandially in both groups, but the decrease was significantly more pronounced in NASH (Figure 1B; Table II)

Despite comparable fasting levels, postprandial large intestinal and hepatic VLDL1 responses were higher in NASH than in controls, while small VLDL2 responses did not differ between the two groups (Figure 2; Table II).

Figure 2.  Oral fat load test. Postprandial plasma VLDL subfraction responses in patients with NASH (n=28) and controls (n=28). Data are presented as mean±SEM. ★ P<0.05 versus controls P<0.01 versus controls ▴ P<0.05 versus fasting levels.

AST and ALT levels did not change in either group throughout the test. GGT levels increased slightly, but significantly at +8h and +10h in NASH, while they did not change in controls (Table II; Figure 3).

Figure 3.  Oral fat load test. Postprandial serum GGT in patients with NASH (n=28) and controls (n=28). ★ P<0.05 versus controls P<0.01 versus controls ▴ P<0.05 versus fasting levels.

Correlations between metabolic parameters in patients with NASH and controls

The main Pearson correlation coefficients between anthropometric, metabolic, and dietary parameters in patients with NASH (upper half of the Table) and in controls (lower half of the Table) are shown in Table III.

Table III.  Main Pearson correlation coefficients between different variables in patients with NASH (upper half of the Table) and in controls (lower half of the Table). Statistically significant correlations are written in bold characters.

NASH
	BMI	Waist	Tg	ISI	Fasting FFA	IAUC FFA	GGT	IAUC GGT	Fasting TAS	Adiponectin	IAUC TAS	IAUC Tg	Fasting LDL CD	IAUC LDL CD IAUC Tg	IAUC VLDL1 apoB48	IAUC VLDL1 apoB10
BMI		0.59^c	0.19	0.46^a	0.14	0.36	0.28	0.40	0.49^a	0.48^a	0.28	0.42	0.11	0.10	0.28	0.36
Waist	0.51^b		0.27	0.52^b	0.18	0.32	0.41	0.48^a	0.46^a	0.51^b	0.31	0.29	0.12	0.22	0.39	0.41
Tg	0.14	0.19		0.48^a	0.45	0.38	0.50^a	0.30	0.21	0.32	0.43	0.48^a	0.21	0.11	0.312	0.19
ISI	0.50^a	0.56^b	0.50^a		0.32	0.24	0.41	−0.47^a	0.42	0.61^c	0.41	0.36	0.50^a	0.10	0.48^a	0.50^a
Fasting FFA	0.26	0.14	0.43	0.27		0.22	0.52^b	0.37	0.39	0.28	−0.39	0.18	0.12	0.09	0.31	0.49^a
IAUC FFA	0.14	0.12	0.28	0.21	0.24		0.40	0.64^c	0.02	−0.54^b	−0.54^b	0.55^b	0.18	0.55^b	0.53^b	0.51^b
GGT	0.31	0.36	0.42	0.23	0.15	0.23		0.47^a	0.12	−0.54^b	0.28	0.41	0.48^a	0.49^a	0.46^a	0.45
IAUC GGT	0.02	0.13	0.12	0.12	0.09	0.12	0.49^a		0.22	0.46^a	−0.55^b	0.51^a	0.22	0.62^c	0.38^a	0.23
Fasting TAS	0.15	0.19	0.24	0.18	0.12	0.18	0.14	0.12		0.30	−0.51^b	0.12	0.13	0.10	0.14	0.08
Adiponectin	0.49^a	0.52^b	0.34	0.51^b	0.24	0.17	0.12	0.18	0.13		0.13	−0.63^c	0.28	−0.51^b	−0.68^c	−0.69^c
IAUC-TAS	0.12	0.28	0.15	0.12	0.27	0.13	0.05	0.32	0.40	0.15		−0.43	−0.51^b	−0.58^c	−0.63^c	−0.52^b
IAUC Tg	0.37	0.21	0.49^a	0.23	0.34	0.50^a	0.18	0.25	0.11	−0.51^b	−0.39		0.20	0.64^c	0.58^b	0.60^b
Fasting LDL CD	0.14	0.14	0.12	0.10	0.21	0.31	0.11	0.18	0.24	0.23	0.13	0.24		0.39	0.21	0.26
IAUC LDL CD	0.03	0.19	0.26	0.02	0.19	0.21	0.28	0.27	0.14	0.24	−0.49^a	0.20	0.50^a		0.63^c	0.61^c
IAUC VLDL1 ApoB48	0.11	0.25	0.13	0.34	0.14	0.49^a	0.13	0.12	0.26	−0.49^a	0.24	0.53^b	0.23	0.28		0.49^a
IAUC VLDL1 ApoB100	0.14	0.32	0.10	0.38	0.23	0.50^a	0.18	0.13	0.27	−0.51^b	0.27	0.49^a	0.03	0.31	0.50^a
	BMI	Waist	Tg	ISI	Fasting FFA	IAUC FFA	GGT	IAUC GGT	Fasting TAS	Adiponectin	IAUC TAS	IAUC Tg	Fasting LDL CD	IAUC LDL CD IAUC-Tg	IAUC VLDL1 apoB48	IAUC VLDL1 apoB10
Controls

^aP < 0.05.

^bP < 0.01.

^cP < 0.001.

Tg = triglyceride; Glu = glucose; C = cholesterol; CD = conjugated dienes.

On multiple regression analysis, IAUC Tg was independently predicted by basal adiponectin (β = − 0.59; P = 0.002) and IAUC VLDL1 apoB48 (β = 0.51; P = 0.009) in NASH, as well as in controls (for adiponectin β = − 0.48; P = 0.011; for intestinal VLDL1 β = 0.49; P = 0.010).

LDL-conjugated diene increase was independently predicted by IAUC-VLDL1 apoB48 (β = 0.51; P = 0.009) and IAUC Tg (β = 0.48; P = 0.011) in NASH but not in controls.

IAUC TAS was independently predicted by IAUC VLDL1 apoB48 (β = − 0.51; P = 0.010) and IAUC FFA (β = − 0.45; P = 0.031) in NASH but not in controls.

IAUC GGT correlated with IAUC LDL-conjugated dienes (β = 0.52; P = 0.008) and IAUC FFA (β = 0.46; P = 0.011) in NASH but not in controls.

Correlations between metabolic and histological parameters in patients with NASH

Liver fatty infiltration correlated with fasting adiponectin (r= − 0.62; P = 0.0001), IAUC TAS (r= − 0.58; P = 0.0003), IAUC-FFA (r=0.50; P = 0.010), waist circumference (r=0.46; P = 0.045), and postprandial large intestinal VLDL1 increase (defined by IAUC VLDL1 apoB48: r=0.55; P = 0.0009). On multiple regression analysis, adiponectin (β = − 0.52; P = 0.011), IAUC TAS (β = − 0.51; P = 0.019), and IAUC VLDL1 apoB48 (β = 0.48; P = 0.028) independently predicted hepatic steatosis.

Severe (grade 3) necroinflammation was predicted by IAUC TAS (OR 2.8; CI 1.5–7.1; P = 0.010), IAUC LDL conjugated dienes (OR 3.6; CI 1.5–6.9; P = 0.006), and IAUC GGT (OR 2.0; CI 1.4–6.1; P = 0.021).

Severe (stage 3) fibrosis was predicted by adiponectin (OR 3.2; CI 2.0–7.1; P = 0.003), IAUC TAS (OR 2.3; CI 1.4–6.1; P = 0.011), and IAUC LDL-conjugated dienes (OR 4.4; CI 1.6–18.1; P = 0.021).

Discussion

The main findings of our study are the following:

1) Despite comparable fasting values, normolipidemic non-diabetic subjects with NASH displayed an accumulation of large VLDL1 of both intestinal and hepatic origin, an increased LDL lipid peroxidation, and a reduced plasma total antioxidant capacity in the postprandial phase compared to controls.

2) We show here for the first time the magnitude of postprandial intestinal VLDL1 accumulation and of pro/antioxidant imbalance is associated with liver fat content and with the severity of hepatic necroinflammatory and fibrotic processes, respectively.

3) The slight, but significant postprandial increase in serum GGT correlates with liver necroinflammation and with intestinal VLDL1 and oxLDL responses in NASH.

Despite similar fasting plasma lipid, oxLDL, and TAS levels, subjects with NASH displayed a significantly higher postprandial lipemia, accounted for by an accumulation of large VLDL1, an increased plasma oxLDL concentration, and a more pronounced reduction in plasma TAS, all phenomena closely correlating with the severity of liver histology.

Postprandial lipemia is an established cardiovascular risk factor and an important source of oxidative stress in diabetes 8. The atherogenic effects of oxidized LDL and triglyceride-rich lipoprotein remnants on vascular endothelial function and macrophage activation have been extensively elucidated. However, the liver may be also involved in these phenomena, since it plays a crucial role in lipoprotein metabolism and is an important scavenger of LDL and intestinal VLDLs through the LDL receptor and the liver-related receptor protein (LRRP); furthermore, hepatic FFA uptake is driven exclusively by their plasma concentration and has been shown to contribute substantially to the liver Tg pool in non-alcoholic fatty liver disease (NAFLD) and diabetes 11, 13.

An increased fasting plasma oxidative stress, coupled with a reduced antioxidant activity, has been recently found in overweight–obese, hyperlipidemic subjects with NASH, with fasting plasma peroxide levels closely correlating with plasma Tg 4.

We are not aware of previous studies simultaneously assessing postprandial lipoprotein responses and dynamic pro/antioxidant balance and correlating them to different histological features of NASH. Our study links impaired postprandial VLDL metabolism and enhanced lipid peroxidation to liver injury in fatty liver even at early/mild stages of metabolic disease, i.e. in the absence of overt obesity, dyslipidemia, and diabetes. We found in fact the pro/antioxidant imbalance develops exclusively postprandially and is independently associated with hepatic steatosis, necroinflammation, and fibrosis in NASH; furthermore, the magnitude of postprandial FFA and intestinal large VLDL responses, rather than dietary antioxidant vitamin deficiency, was the main determinant of this pro-oxidant condition in patients with NASH, but not in controls.

Although our cross-sectional study cannot ascertain which is the cause and which is the consequence, experimental data suggest the liver may be a target organ of postprandial lipoprotein accumulation and pro-oxidant imbalance in NASH. Large VLDLs accumulation may in fact promote liver injury both directly, by stimulating liver hepatic stellate cells via a receptor-mediated mechanism 28, and indirectly, by FFA-induced liver injury and by enhancing LDL oxidation.

Oxidized LDL were able to bind to scavenger receptors CD36 of cultured Kupffer cells and stellate cells and to trigger inflammatory cascade and extracellular matrix deposition 28. Finally, the liver of patients with NASH exhibited higher amounts of oxidized LDL than other chronic liver diseases, with the degree of oxLDL receptor expression on activated stellate cells paralleling the expression of profibrotic cytokines and the severity of fibrosis 29, 30.

FFA have been shown to trigger c-jun N-terminal kinase (JNK)-mediated hepatocyte apoptosis in cell cultures and to promote the development of NASH by activating the JNK1 pathway in mice 31, 32. This finding, coupled with the ability of a high-fat diet to induce NAFLD 5, 6, highlights the critical role of lipoperoxidative injury in the pathogenesis of NASH.

Taking these data together, oxidized LDL may be a key mediator connecting systemic lipid peroxidation to hepatic necroinflammatory and fibrogenic processes in NASH.

While further studies are required to elucidate which step of intestinal VLDL metabolism is impaired in NASH, from a therapeutical point of view our data suggest drugs like statins, peroxisome proliferator-activated receptor (PPAR)-α, or orlistat directly modulating postprandial lipoprotein metabolism, may be effective in NASH even in the absence of fasting hyperlipemia, alone or in adjunct to insulin-sensitizing measures.

A novel finding of our study is the dynamic postprandial serum GGT increase in NASH, closely correlating with hepatic necroinflammation and the magnitude of FFA and oxLDL responses. Although there are other potential sources of serum GGT, including circulating leukocytes and platelets, their overall contribution to total serum GGT activity is small in the absence of diabetes and clinical cardiovascular disease; so most of serum GGT of our subjects is likely to originate from the liver 33. Postprandial absolute GGT elevation was small. However, increases in fasting serum GGT as small as 5 U/L carried a 2.5-fold risk of incident diabetes, independently of obesity and insulin resistance 34. Whether postprandial serum GGT changes reflect a mere increase in hepatic lipid content, or a process connecting postprandial lipoprotein metabolism to vascular and hepatic oxidative injury, remains to be elucidated. More importantly, it remains unclear if GGT is a marker or an active player in oxidative injury. Circulating GGT can bind dose-dependently to β-lipoproteins and colocalize with oxLDL in the vessel wall, synergistically triggering oxidative stress and atherogenesis 35. The increased postprandial GGT and oxLDL correlated with hepatic necroinflammatory activity. Although hepatic oxLDL content and GGT activity were not directly measured, it can be hypothesized that postprandial lipemia triggers inflammation by promoting the influx of oxLDL and GGT in both the liver sinusoids and the vessel wall. Consistently, GGT has been linked to lipid metabolism in recent cross-sectional and interventional studies 36.

The mechanism underlying this association may be the involvement of GGT in the cellular redox balance 37–39. Whether GGT exerts an antioxidant or a pro-oxidant activity is controversial. GGT plays a pivotal role in the maintenance of cellular antioxidant defenses through its catabolism of extracellular reduced glutathione (GSH), allowing for reutilization of precursor amino acids for intracellular GSH synthesis (GSH ‘cycling’): GGT-deficient knockout mice exhibit a loss of GSH homeostasis, impaired mitochondrial respiration, and progressive liver injury, alterations that were reversed by N-acetylcysteine supplementation 40. Recently, however, GGT has been proposed to have a pro-oxidant function, being able to generate reactive oxygen species (ROS) during metabolism of GSH, especially in the presence of chelated iron 41. Interestingly, GGT activity expressed by macrophages promoted LDL oxidation and upregulated nuclear factor (NF)-kB, a key mediator of inflammatory gene transcription, in human atheromas 42, 43.

The association of adiponectin with postprandial lipid metabolism parameters is consistent with the ability of this adipokine to modulate LPL-mediated VLDL catabolism and FFA oxidation, independently of other adipokines or insulin resistance 44, 45.

Several limitations of this study need to be mentioned. First, its cross-sectional nature prevents any definitive causal inference, nor is it possible to ascertain whether pro-oxidant imbalance and VLDL accumulation are cause or consequence of NASH or insulin resistance. Second, since liver fat content was not directly measured, some controls might have fatty liver despite normal ultrasound and liver enzymes; even so, misclassification of NASH would attenuate the magnitude of the difference in oxidative stress and postprandial lipemia observed toward the null hypothesis, making our results a conservative estimate of the relationship between NASH, oxidative stress, and impaired lipoprotein metabolism. Third, we selected normolipidemic non-obese non-diabetic subjects to evaluate metabolic abnormalities occurring at mild/early stages of metabolic disease, but the generalizability of our findings to the large majority of obese diabetic dyslipidemic subjects with NASH needs to be demonstrated.

Acknowledgements

Giovanni Musso: study design, data analysis, writing of manuscript; Roberto Gambino: data analysis; Barbara Uberti: data analysis; Natalina Alemanno: data analysis; Franco De Michieli: data collection; Giampaolo Broli: data collection; Gianfranco Pagano: supervision; Maurizio Cassader: supervision.