点击显示 收起
1 From the Emergency Medicine Department, Gradenigo Hospital, Turin, Italy (GM), and the Department of Internal Medicine, University of Turin, Turin, Italy (RG, FDM, GB, AP, GP, SB, MD, and MC)
2 Supported by the Piedmont Region Funds Comitato Interministeriale per la Programmazione Economica 2003. 3 Reprints not available. Address correspondence to G Musso, Emergency Medicine Department, Gradenigo Hospital, Corso Regina Margherita 8, 10124 Turin, Italy. E-mail: giovanni_musso{at}yahoo.it.
ABSTRACT
Background:Although nonalcoholic fatty liver disease (NAFLD) is associated with the metabolic syndrome, the mechanisms responsible for the development of NAFLD at different stages of the development of insulin resistance are unknown. Diet, adipokines, and nitrosative stress have been linked to both NAFLD and insulin resistance.
Objective:We aimed to identify the factors that are specifically associated with NAFLD at different stages in the development of insulin resistance and the metabolic syndrome.
Design:Circulating concentrations of adipokines (ie, tumor necrosis factor-, adiponectin, resistin, leptin, and interleukin-6), markers of nitrosative stress (nitrotyrosine), dietary habits, and MTP –493G/T polymorphism were cross-sectionally related to the presence and severity of insulin resistance (homeostasis model assessment index for insulin resistance: 2), the metabolic syndrome, and fatty liver in 64 nonobese nondiabetic patients with NAFLD (33 insulin-sensitive and 31 insulin-resistant subjects) and 74 control subjects without liver disease who were matched for sex, BMI, homeostasis model assessment index for insulin resistance status, and the various features of the metabolic syndrome.
Results:Persons with NAFLD had greater systemic nitrosative stress and a lower intake of vitamins A and E than did control subjects, but the 2 groups did not differ significantly in any other features. Nitrotyrosine and adiponectin concentrations and vitamin A intakes independently predicted alanine aminotransferase concentrations in NAFLD patients and liver histology in a subgroup of 29 subjects with biopsy-proven nonalcoholic steatohepatitis.
Conclusions:Oxidative stress is operating in NAFLD and nonalcoholic steatohepatitis, even in the absence of insulin resistance, the metabolic syndrome, and hypoadiponectinemia, which aggravate liver histology at more severe stages of metabolic disease. The possible pathogenetic role of reduced vitamin A intake in NAFLD warrants further investigation.
Key Words: Retinoid • microsomal triacylglycerol transfer protein • MTP polymorphism • nitrotyrosine • adipokines
INTRODUCTION
Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease, with a prevalence of 20% among Western adult population (1). Approximately 20% of patients with NAFLD will develop nonalcoholic steatohepatitis (NASH), which may progress to cirrhosis and hepatocellular carcinoma. NAFLD is associated with the metabolic syndrome, and it is present in lean glucose-tolerant subjects as well as in obese diabetic subjects (2). Nevertheless, this association is not universal: several studies show that only 30–55% of diabetic and 30–40% of obese subjects have NAFLD and that not every patient with fatty liver is insulin resistant (IR) (1, 3, 4).
Factors responsible for NAFLD at different stages of the development of insulin resistance and the metabolic syndrome are poorly understood. Most studies compared NAFLD patients with insulin-sensitive (IS) healthy control subjects. Because adipokines and oxidative balance are altered in NAFLD as well as in the metabolic syndrome (5), their specific role in the pathogenesis of liver disease would be better elucidated by comparing NAFLD patients with control subjects matched for the severity of insulin resistance, metabolic syndrome, and adipose tissue accumulation (1). Furthermore, the selection of nonobese nondiabetic subjects would be useful, because different mechanisms may intervene as the metabolic syndrome and hepatic disease progress, and the adipokines that are altered in overt obesity, diabetes, and cirrhosis may differ from those triggering the whole metabolic and inflammatory cascade.
Dietary factors may also predispose a person to develop fatty liver independently of the factors’ effects on body fat and insulin resistance. Retinol deficiency, in particular, has been experimentally linked to steatohepatitis and hepatocellular carcinoma (6, 7).
Among genetic factors, a functional polymorphism of microsomal triacylglycerol transfer protein (MTP) gene has been linked to NAFLD, but the interaction of MTP polymorphism with acquired factors in the pathogenesis of liver disease is unknown (8). Identification of the factor or factors specifically responsible for the development of NAFLD would help us to understand its pathogenesis and ascertain the individual risk for liver disease and would allow individualized therapeutic interventions.
The aims of the study were to investigate MTP –493G/T polymorphism, dietary habits, adipokines, and markers of nitrosative stress in nonobese nondiabetic patients with NAFLD and in nonobese nondiabetic subjects without liver disease who had varied severities of insulin resistance and metabolic syndrome and to assess the combined effect of genetic and metabolic factors on the development and severity of liver disease in NAFLD patients.
SUBJECTS AND METHODS
Subjects
On the basis of available data (8-11), we considered 0.05 to be a type I error and 0.20 to be a type II error: 60 subjects with NAFLD and 30 subjects per arm were needed to detect a significant between-group difference in plasma nitrotyrosine, adipokines, metabolic syndrome, dietary intake of antioxidant vitamins, and –493G/T MTP polymorphism.
Insulin resistance was estimated by using the homeostasis model assessment index for insulin resistance (HOMA-IR), which was closely correlated with clamp measures in nondiabetic subjects in a large Italian cohort study (12). A HOMA-IR cutoff >2 was chosen [defining as IR those subjects with a HOMA-IR 2], because it predicted insulin resistance in an Italian population (12). This cutoff also identified subjects with a significantly increased risk of cardiovascular disease (CVD) in previous studies (13-15).
For the abovementioned reasons, we excluded obese subjects [obesity was defined by a body mass index (BMI; in kg/m2) 30 for males and 28 for females, based on the sex-related differences in total body fat (16)] and diabetic subjects [diabetes was defined as a fasting plasma glucose 126 or 200 mg/dL after 2 h on a standard oral-glucose-tolerance test (OGTT)]. NAFLD was defined as a persistent (for 6 mo) elevation in aminotransferases, ultrasonographic presence of bright liver without any other liver or biliary tract disease, and a daily alcohol consumption < 20 g/d in men and < 10g/d in women (1), as assessed by a detailed questioning of patients and relatives and daily completion of a validated questionnaire for 1 wk by the patients. Exclusion criteria for other causes of fatty liver were positive markers of viral, autoimmune, or celiac disease; abnormal copper metabolism, serum concentrations of 1-antitripsin, or thyroid function tests; or exposure to occupational hepatotoxins or to drugs known to be steatogenic or to affect glucose or lipid metabolism. Mutations in the hemochromatosis HFE and TRF2 genes were detected in patients and control subjects by using a single, multiplex amplification reaction and previously made, ready-to-use test strips (Nuclear Laser Medicine, Milan, Italy). Twenty-nine NAFLD patients (14 IS and 15 IR patients) underwent a biopsy of liver tissue the results of which were compatible with a diagnosis of NASH (steatosis involving 5% of hepatocytes, lobular inflammation, and zone 3 ballooning degeneration), as blindly assessed by a single pathologist (ED) using the score proposed by Brunt (17). Liver iron concentration was measured in 2 mg dry-weight tissue by using atomic absorption spectroscopy, and the hepatic iron index (HII) was calculated by dividing the liver iron concentration (µmol/g) by age (y).
The –493G/T MTP gene polymorphism was also assessed by using polymerase chain reaction–restriction fragment length polymorphism in a 2-step nested polymerase chain reaction. To further rule out subclinical liver disease in control group, an alcohol intake <20 g/d in men and <10 g/d in women and normal liver enzymes and abdomen ultrasound were inclusion criteria. To increase negative predictive value of a normal result, the healthy upper limit for ALT in controls was lowered to < 30 U/L for men and < 20 U/L for women (18, 19).
Of 1600 subjects taking part in a large population-based cohort study, 170 IS and 170 IR sex-matched nonobese nondiabetic subjects were randomly identified. In each group, 61 IS subjects with NAFLD and 73 IR subjects with NAFLD were extracted; in each of the 4 arms, subjects were subsequently split according to the presence or absence of abdominal obesity, of different features of the metabolic syndrome, and of the metabolic syndrome overall. At the end of the selection, 68 patients with NAFLD (35 IS subjects) and 79 control subjects (45 IS subjects) without liver disease remained.
Sixty-four subjects with NAFLD (31 IS) and 74 control subjects (42 IS) (Table 1) gave written informed consent. The study was conducted according to the Helsinki Declaration and approved by the local Ethics Committee.
View this table:
TABLE 1. Main characteristics of nonalcoholic fatty liver disease (NAFLD) patients and control subjects grouped according to homeostasis model assessment index of insulin resistance (HOMA-IR) status1
Dietary intake
During an individual training session with a nutritionist, patients and controls were instructed to fill in a 7-d extensively validated and reproducible diet record, a list of foods was designed, and different portion sizes were specified for each item according to the European Prospective Investigation into Cancer and Nutrition (20, 21). The recorded period included a complete week. The diet record was analyzed by using the WINFOOD database (1997 version; Medimatica, Teramo, Italy) according to the table of food consumption of the Italian National Institute of Nutrition (22) and the Food Composition Database for Epidemiologic Study in Italy (23).
Oxidative stress
Nitrotyrosine was chosen as a marker of oxidative stress, because it correlated with the severity of liver oxidative injury in NAFLD and because nitrosative stress is a key feature of insulin resistance and diabetes (5, 24). Fasting plasma concentrations of nitrotyrosine were measured by using a commercial enzyme-linked immunosorbent assay kit (HyCult Biotechnology, Uden, Netherlands).
Cytokines
Serum concentrations of TNF-, leptin, IL-6, and adiponectin were measured by using a sandwich enzyme-linked immunosorbent assay (R&D System Europe Ltd, Abingdon, United Kingdom). Resistin was measured by using a biotin-labeled antibody-based sandwich enzyme immunoassay (Bio Vendor Laboratory Medicine Inc, Brno, Czech Republic).
Serum concentrations of TNF- and adiponectin were measured by using the sandwich enzyme-linked immunosorbent assay (R&D System Europe Ltd). For TNF-, the kit has a sensitivity of 0.12 pg/mL in a 200-µL sample size and a range of 0.5 to 32 pg/mL. The intraassay and interassay CVs were 5.9% and 12.6%, respectively. For adiponectin, the kit has a sensitivity of 0.25 pg/mL in a 50-µL sample size and a range of 3.9 to 250 ng/mL. The intraassay and interassay CVs were 3.4% and 5.8%, respectively. All samples were diluted 1/100.
Statistical analysis
Data were expressed as means ± SEMs. A 2-factor multivariate analysis of variance (ANOVA) with Wilks was used to measure the degree of correlation among variables. When a significant interaction was found between factors (A x B), those factors were identified with simple main-effects tests: differences across groups were analyzed by ANOVA and then by followed by Bonferroni correction, when variables were normally distributed; otherwise the Kruskal-Wallis test, followed by the post hoc Dunn test, was used to compare nonparametric variables. Normality was evaluated by using the Shapiro-Wilk test. The chi-square test or Fisher's exact test was used for categorical variables as appropriate. Univariate analysis (Spearman rank correlation coefficient, rs) was used to estimate the relation between different variables. Multiple linear regression analysis was applied when multiple associations between continuous variables were detected, after log transformation of skewed data.
Logistic regression analysis was applied to identify independent predictors for necroinflammation grade 2–3 or fibrosis stage 2–3. The covariates were age, waist circumference, HOMA-IR, serum concentrations of adiponectin, presence of metabolic syndrome, plasma concentrations of nitrotyrosine, dietary vitamin A or E intake, and HII. Differences were considered statistically significant if P < 0.05 (STATISTICA software, version 5.1; StatSoft Italia, Padua, Italy).
RESULTS
Subject characteristics
Main features of NAFLD patients and control subjects, along with the multivariate ANOVA interactions, are shown in Table 1. IR NAFLD patients had higher plasma nitrotyrosine concentrations than did IS NAFLD patients, who in turn had higher nitrotyrosine concentrations than did either IR or IS control subjects (Figure 1).
FIGURE 1.. Mean (±SEM) plasma nitrotyrosine concentrations in patients with nonalcoholic fatty liver disease (NAFLD; n = 64) and control subjects (n = 74), grouped according to homeostasis model assessment index of insulin resistance status. IS, insulin-sensitive; IR, insulin-resistant. Significant effects were found for IR and NAFLD and an IR x N interaction. Between-group comparisons were made by using the Mann-Whitney test. , IR NAFLD patients compared with IS NAFLD patients (23.9 ± 3.7 and 17.8 ± 1.5 nmol/mL, respectively; P = 0.00003); , IR NAFLD patients compared with IR control subjects, P < 0.0001; , IR control subjects compared with IS control subjects (8.7 ± 2.0 and 4.8 ± 1.1 nmol/mL, respectively; P = 0.009).
NAFLD patients (n = 64) did not differ significantly from the subgroup of patients with biopsy-proven NASH (n = 29) for all of the main HFE and TRF2 gene mutations or in –493G/T MTP polymorphism, both when considered overall (NAFLD patients compared with control subjectss) and when grouped according to HOMA-IR status (Table 1). NAFLD patients did not differ significantly from the subgroup of patients with biopsy-proven NASH in any of the main anthropometric, dietary, and biochemical features (data not shown).
Dietary intake
IS and IR NAFLD patients had significantly lower daily vitamin A intakes than did IS and IR control subjects, whereas there was no difference in the dietary intake of macronutrients or other micronutrients between the 4 groups (Table 2). There was no significant difference in any anthropometric, dietary, or biochemical variables between alcohol-abstinent subjects (overall prevalence: 71% of males and 85% of females) and the remaining study population (range of daily ethanol consumption: 0.3–19.5 g/d in males and 0.2–9.8 g/d in females).
View this table:
TABLE 2. Daily intake of main dietary constituents in nonalcoholic fatty liver disease (NAFLD) patients and control subjects grouped according to homeostasis model assessment index of insulin resistance (HOMA-IR) status1
The dietary habits of the control subjects did not differ significantly from those of the general population of the Italian Piedmont, as assessed by a recent alimentary survey (21). Vitamin A and E intakes of the NAFLD patients were lower than the recommended intakes for the Italian adult population (ie, vitamin E: 8 mg/d; vitamin A: 700 µg/d for males and 600 µg/d for nonpregnant females) (25).
Metabolic syndrome
According to the Adult Treatment Panel III criteria, the prevalence of each feature and that of MS overall were higher in the IR control subjects and NAFLD patients than in the IS control subjects (3 criteria fulfilled) (Table 1). Adult Treatment Panel III criteria for the diagnosis of metabolic syndrome were hypertension (systolic and diastolic blood pressure 130/85 mm Hg, respectively, or taking antihypertensive therapy); hypertriglyceridemia (fasting plasma triacylglycerols 150 mg/dL or taking lipid-lowering therapy); low plasma HDL cholesterol (HDL cholesterol < 40 mg/dL in men and < 50 mg/dL in women); impaired glucose regulation (impaired fasting glycemia—ie, fasting plasma glucose 100 mg/dL but < 126 mg/dL; or impaired glucose tolerance—ie, plasma glucose 140 mg/dL at 2 h on OGTT); and abdominal obesity (waist circumference > 102 cm in men and > 88 cm in women) (26). When patients and control subjects were grouped according to the number of features of metabolic syndrome present in each subject, NAFLD patients had higher plasma nitrotyrosine concentrations and lower dietary vitamin A and E intakes than did their respective control subjects (Table 3).
View this table:
TABLE 3. Main characteristics of nonalcoholic fatty liver disease (NAFLD) patients and control subjects grouped according to the number of features of the metabolic syndrome1
Cytokines
Serum adiponectin concentrations were significantly lower in IR than in IS subjects. However, they did not differ significantly between NAFLD patients and their respective control subjects matched for HOMA-IR status (Table 1).
Biopsy-proven nonalcoholic steatohepatitis
In the 29 patients with biopsy-proven NASH, fatty infiltration was mild (involving 5–33% of hepatocytes) in 10 patients (6 IS), moderate (33–66% of hepatocytes) in 10 control subjects (6 IS), and severe (>66% of hepatocytes involved) in 9 patients (3 IS). Necroinflammatory activity was grade 1 in 9 patients (7 IS), grade 2 in 10 (5 IS), and grade 3 in 10 (5 IS). Fibrosis was stage 0 in 5 patients (4 IS), stage 1 in 7 patients (5 IS), stage 2 in 9 (5 IS), and stage 3 in 8 patients (4 IS). In the 29 patients with NASH, liver iron concentration was 15 ± 3 µmol/g dry weight and HII was 0.50 ± 0.04.
IS NASH had milder histologic severity than did IR NASH (mean steatosis: 31 ± 3% and 55 ± 4% hepatocytes, respectively; P = 0.002; mean necroinflammatory grade: 1.8 ± 0.2 and 2.4 ± 0.1, respectively; P = 0.015; and mean fibrosis stage: 1.3 ± 0.3 and 2.3 ± 0.2, respectively; P = 0.028). The IS NASH and IR NASH groups did not differ significantly from the IS and IR NAFLD groups, respectively, in any of the main anthropometric, dietary, and biochemical features (data not shown). Adiponectin and nitrotyrosine concentrations in IS and IR NASH patients and in IS and IR control subjects are shown in Figure 2.
FIGURE 2.. Mean (±SEM) plasma nitrotyrosine and adiponectin concentrations in patients with nonalcoholic steatohepatitis (NASH; n = 29) and control subjects (n = 74), grouped according to homeostasis model assessment index of insulin resistance status. IR, insulin-resistant; IS, insulin-sensitive; N, NASH. Significant effects were found for IR and N and an IR x N interaction for both adiponectin and nitrotyrosine. Between-group comparisons were made by using the Mann-Whitney test. , IR NASH patients compared with IS NASH patients, P < 0.01; , IS NASH patients and IR NASH patients compared with IR control subjects, P < 0.0001; , IS NASH patients compared with IS control subjects, P < 0.01.
Correlative analysis
In the 64 NAFLD patients, ALT concentrations correlated significantly with adiponectin (rs = –0.42; P = 0.0008), HOMA-IR (rs = 0.34; P = 0.003) and HDL cholesterol (rs = –0.42; P = 0.0008) concentrations, the number of criteria for the metabolic syndrome (rs = 0.36; P = 0.001), dietary vitamin A (rs = –0.58 P = 0.0001; Figure 3) and vitamin E (rs = –0.48 P = 0.0004) intakes, and plasma concentrations of nitrotyrosine (rs = 0.59; P = 0.00009). On multiple regression analysis (adjusted R2 = 0.65; F = 15.7), ALT concentrations were independently predicted by nitrotyrosine concentrations (ß = 0.53; P = 0.001), vitamin A intake (ß = –0.51; P = 0.003), and adiponectin concentrations (ß = –0.22; P = 0.02).
FIGURE 3.. Correlation between concentrations of alanine aminotransferase (ALT) and daily intake of vitamin A in patients with nonalcoholic fatty liver disease (n = 64). rs: Spearman correlation coefficient.
Plasma concentrations of nitrotyrosine were correlated significantly with age (rs = 0.30; P = 0.02), HOMA-IR (rs = 0.27; P = 0.039), ALT (rs = 0.58; P = 0.0001) and HDL cholesterol (rs = –0.33; P = 0.011) concentrations, and dietary vitamin A intake (rs = –0.42; P = 0.0008). On multiple regression analysis (adjusted R2 = 0.51; F = 12.9), only dietary vitamin E (ß = –0.27; P = 0.01) independently predicted plasma nitrotyrosine concentrations. All of these correlations held in NAFLD patients who had and had not undergone liver biopsy (data not shown).
In the 29 patients with biopsy-proven NASH, fatty infiltration, expressed as the percentage of hepatocytes involved, correlated significantly with adiponectin (rs = –0.60; P = 0.005) and nitrotyrosine (rs = 0.59; P = 0.008) concentrations, vitamin A intake (rs = –0.59; P = 0.008), HOMA-IR status (rs = 0.45; P = 0.03), and waist circumference (rs = 0.50; P = 0.025) but not with other variables. On multiple regression analysis, only serum concentrations of adiponectin (ß = –0.40; P = 0.024), dietary vitamin A (ß = –0.32; P = 0.04), and plasma concentrations of nitrotyrosine (ß = 0.50; P = 0.01) predicted liver fatty infiltration. On logistic regression analysis, adiponectin and nitrotyrosine concentrations and vitamin A intake independently predicted both necroinflammatory grade 2–3 and fibrosis stage 2–3 (Table 4).
View this table:
TABLE 4. Multiple logistic regression analysis for factors associated with necroinflammatory grade 2–3 and fibrosis stage 2–3, respectively, in patients with biopsy-proven nonalcoholic steatohepatitis1
DISCUSSION
Insulin resistance and the metabolic syndrome have been associated with NAFLD, and the number and severity of features of the syndrome predict the severity of liver disease. Nevertheless, the sequence of events initiating and perpetuating metabolic and liver disease is unclear, beause only some of the IR subjects develop fatty liver, and not every patient with NAFLD is IR (3). Consistently, dietary manipulation induced hepatic fat accumulation and lipid peroxidation without primarily modifying insulin sensitivity in animal and human models (7, 27). Because steatosis and insulin resistance can cause or potentiate each other, the question remains as to whether steatosis or insulin resistance is the initiating step in NAFLD (1). Altered adipokine signaling and oxidative balance have been implicated in the pathogenesis of both NAFLD and insulin resistance, and so it is not clear which factor is specifically linked to the development of fatty liver rather than being associated with the metabolic syndrome. By comparing NAFLD patients with control subjects matched for the severity of insulin resistance, adiposity, and metabolic syndrome, we found 1) that an increased systemic nitrosative stress characterizes NAFLD even in the absence of insulin resistance, metabolic syndrome, and hypoadiponectinemia; 2) that the degree of nitrosative stress correlates with the severity of liver disease in both IS and IR NAFLD patients; 3) that hypoadiponectinemia is not a feature of IS NAFLD, but, rather, that it intervenes in IR subjects to aggravate both liver and metabolic disease; and 4) that a reduced intake of vitamin A independently correlates with the severity of liver disease.
An increased nitrosative stress is the common feature of NAFLD in our subjects: plasma nitrotyrosine concentrations in IS and IR NAFLD patients were, in fact, 2 and 3 times, respectively, those in IR control subjects without liver disease. Furthermore, the degree of nitrosative stress correlated with liver histology in biopsy-proven NASH, which suggested that this association may be a cause of liver disease. Because of its cross-sectional nature, our study cannot ascertain whether fatty liver or increased nitrosative stress comes first, or whether IS NAFLD patients will develop insulin resistance and metabolic syndrome. Nevertheless, it provides evidence that oxidative stress is operating in NAFLD even in the absence of insulin resistance, metabolic syndrome, and hypoadiponectinemia, which in turn, intervene at more advanced stages of metabolic and liver disease.
Increased oxidative stress can modulate both hepatic necroinflammation and fibrogenesis and the extent of triacylglycerol accumulation in NAFLD. An increased intracellular oxidative stress impaired VLDL secretion by reducing the expression of ApoB100 and acyl-coenzyme-A:cholesterol-acyltransferase2 mRNA, which led to triacylglycerol accumulation in cultured hepatocytes and mouse models, an effect totally reversed by antioxidant treatment (28, 29). Furthermore, oxidation products can activate cultured Kupffer and stellate cells and trigger inflammatory cascade and extracellular matrix deposition; their hepatic expression parallels the severity of liver fibrosis in NAFLD (30).
Hepatic steatosis and oxidative stress may eventually lead to hepatic insulin resistance, necroinflammation, and fibrogenesis via activation of protein kinase C and of the c-Jun N-terminal kinase pathway, thus perpetuating liver and metabolic disease (1).
Consistently, available evidence suggests that the direction of the association between NAFLD and insulin resistance is mutual, in which NAFLD predisposes a person—even normal-weight, glucose-tolerant persons—to develop metabolic syndrome or diabetes, and diabetes aggravates liver disease. Elevated liver enzymes prospectively predicted incident diabetes independentlof the metabolic syndrome and of classic risk factors, whereas diabetes substantially increased the risk of cirrhosis and increased liver-related mortality in NAFLD patients (1, 31).
The pathogenesis of oxidative stress in NAFLD patients most likely is multifactorial. An impaired mitochondrial ß-oxidation is likely to have a role, increasing hepatocyte production of reactive oxygen species (ROS) and storage of free fatty acids (FFA), both directly and indirectly, by peroxisomal and microsomal FFA oxidation (32). Poor dietary habits also may have contributed to oxidative stress in the patients in the present study, because they consumed lower amounts of antioxidant vitamins than did control subjects, and vitamin E intake predicted plasma concentrations of nitrotyrosine.
A novel and intriguing finding is the association of reduced dietary vitamin A intake with the presence and the severity of liver disease. Although data on dietary intake of vitamin A in NAFLD patients are sparse, transgenic mice lacking retinoid acid receptor function developed severe steatohepatitis and hepatic carcinoma that were accompanied by a decreased hepatic mitochondrial FFA ß-oxidation and an increased peroxisomal and microsomal FFA oxidation and ROS production. Such effects were totally reversed by dietary all-trans retinoic acid supplementation (6, 33). Furthermore, retinoid acid supplementation reversed hepatic lipid peroxidation and carbon tetrachloride–induced liver fibrosis in mouse models (34-36).
The molecular basis for these antioxidant, antisteatotic, and antifibrotic activities of retinoids may be provided by the direct antioxidant scavenger activity of ß-carotene (37) and by the function of retinoid as nuclear receptor ligands: liver retinoid X receptor- plays a key role in the regulation of hepatic peroxisomal FFA ß-oxidation and glutathione homeostasis (38, 39); furthermore, retinoids were able to suppress transforming growth factor-ß1–induced proliferation and the migration of cultured hepatic stellate cells (40). Consistently, retinoid X receptor agonists reduced lipid peroxidation and enhanced glucose and lipid oxidation in human skeletal muscle (41, 42).
Taking available data together, we propose that IR NAFLD may be the final step of both IS NAFLD and metabolic syndrome that is initially free of liver disease. If these findings are confirmed by larger prospective studies, the correction of oxidative stress at early stages of fatty liver may help stop the progression of liver disease in these subjects.
As is consistent with previously reported findings (11), serum adiponectin concentrations predicted the severity of liver disease in NAFLD patients but were comparable to those in control subjects matched for the degree of insulin resistance. This finding suggests either that hypoadiponectinemia is absent at mild or early stages of liver disease, whereas it characterizes IR NAFLD patients, or that hepatic adiponectin or adiponectin receptor expression are more tightly related to liver injury than are plasma concentrations in NAFLD patients (43).
Some limitations of this study should be mentioned. Its cross-sectional nature prevents any causal inference. Although the diagnostic accuracy of modern ultrasound techniques for fatty liver is good (44), several control subjects may have NAFLD despite normal ultrasound and liver enzymes. Even in such a case, however, misclassification would attenuate the magnitude of the difference observed toward the null hypothesis, and these results can therefore be considered a conservative estimate of the relation between NAFLD and oxidative stress.
Although plasma concentrations of different antioxidant vitamins were not measured, a reproducible dietary record may reflect body stores and long-term consumption of these vitamins more reliably than do plasma concentrations (45-47).
In conclusion, nitrosative stress may initiate hepatic injury in NAFLD—it is detectable in IS subjects with a normal plasma adipokine profile—and also may contribute to its progression, behaving as both first and second hit. The association of insulin resistance and hypoadiponectinemia with more advanced liver disease confirms that those conditions may intervene and aggravate liver damage in NAFLD patients once the process has begun. If these data are prospectively confirmed, antioxidants would be a useful therapy in IS NAFLD patients, who fail to or do not need to lose weight. They also may be synergistic to insulin sensitizers in IR subjects.
The use of vitamin A supplementation warrants further evaluation, given the controversial liver toxicity of an excessive intake of its synthetic analogues, especially in alcoholic liver disease (48). With respect to this issue, the possibility of a lower threshold for ethanol-induced liver toxicity in cannot be discarded in NAFLD. According to our data, a diet rich in antioxidant vitamins may be more effective in protecting against the development of NAFLD than is supplementation with a single vitamin at different stages of the development of insulin resistance and metabolic syndrome.
ACKNOWLEDGMENTS
We thank Natalina Alemanno and Barbara Uberti for their help in laboratory analyses.
The authors’ responsibilities were as follows—GM: study design, data elaboration, statistical analysis, and manuscript writing; RG: laboratory analyses and data elaboration; FDM: dietary data collection and elaboration; GB, AP, and SB: data collection; and GP, MD, and MC: critical review of the manuscript. None of the authors had a personal or financial conflict of interest.
REFERENCES