Middle‐ and high‐molecular weight adiponectin levels in relation to nonalcoholic fatty liver disease

Abstract Objective Adiponectin (APN) circulates as high‐molecular weight (HMW), medium‐molecular weight (MMW), and low‐molecular weight (LMW) forms. Nonalcoholic fatty liver disease (NAFLD) is a common cause of chronic liver disease. Currently, the role of LMW, MMW, and HMW APN remains largely unclear in NAFLD. Methods We examined the variation of these forms and analyzed the related clinical characteristics in NAFLD. A total of 63 male NAFLD patients (mean age: 43.00 ± 6.10 years) and 70 healthy male subjects (mean age: 42.53 ± 7.98 years) were included in the study. Total APN and other clinical characteristics were measured. The changes in HMW, MMW, and LMW APN were determined in NAFLD patients and NAFLD patients on a high‐fat diet, and the association between the groups was further analyzed. Results Decreased levels of total APN and three APN isoforms were found in NAFLD. Significantly decreased levels of HMW (P < .01) and MMW (P < .001) were observed in NAFLD of high‐fat diet patients. In NAFLD patients, height (R = −.270, P = .032) and N‐epsilon‐(carboxymethyl) lysine (R = −.259, P = .040) significantly correlated with total APN. HMW APN was significantly associated with fasting plasma glucose (R = .350, P = .016), alanine aminotransferase (R = −.321, P = .029), and aspartate aminotransferase (R = −.295, P = .045). Additionally, MMW APN was significantly associated with total cholesterol (R = .357, P = .014) and high‐density lipoprotein (R = .556, P < .0001). Low‐density lipoprotein (R = −.283, P = .054) was also clearly associated with LMW APN in NAFLD patients. Conclusion These results suggest that HMW and MMW APN may be involved in the pathogenesis and progression of NAFLD.


| INTRODUC TI ON
Nonalcoholic fatty liver disease (NAFLD) is characterized by hepatic fat accumulation of equal to and greater than 5% and is a common cause of chronic liver disease. 1,2 Nonalcoholic fatty liver disease spans a clinicopathologic spectrum characterized by hepatic steatosis with or without other pathologic features in the absence of other specific causes of fatty liver. 3,4 Based on the epidemiological data, the global prevalence of NAFLD has been estimated as high as one billion cases. 3 The development of liver injury in NAFLD is considered a "multiple-hit process," which involves many stages, such as triglyceride (TG) and free fatty acid accumulation in hepatocytes, oxidative stress, lipid peroxidation, mitochondrial dysfunction, liver inflammation, insulin resistance, and perturbations of adipokine levels. 5,6 Extensive studies have reported an association between NAFLD and different indices of insulin resistance, illustrating that either insulin resistance may play a role in the pathogenesis and progression of NAFLD, or that insulin resistance shares a common pathogenic mechanism with liver metabolic disorder.
Adiponectin (APN) is a hormone produced by adipocytes that acts on specific receptors of several tissues through autocrine, paracrine, and endocrine signaling mechanisms. It exists as three distinct and basic oligomeric complexes in plasma as the homotrimer (low-molecular weight, LMW "mass," ∼70 kDa) APN, the hexamer (middle-molecular weight, MMW "mass," ∼140 kDa) APN, and 12-18 protomer (high-molecular weight, HMW "mass," >300 kDa) APN. 7,8 Adiponectin plays an important pathophysiological role in metabolic activities such as glucose, lipid, and branched-chain amino acid metabolism and also functions as an insulin sensitizer. The physiological effects or biological activity of multimer structures of these three forms has recently attracted sufficient attention. 9,10 Currently, HMW, total APN, and the ratio of HMW APN to total APN have been reported to be associated with diabetes mellitus, insulin resistance, and metabolic syndrome. 10,11 More importantly, serum APN levels are decreased in NAFLD patients, suggesting that low APN level is an independent risk factor for NAFLD. 12,13 Significantly lower serum HMW APN level was observed in Taiwanese NAFLD patients with type 2 diabetes. 14 Different percentages of distribution of the APN isoforms LMW, MMW, and HMW were observed in NAFLD patients. 15 However, the association of APN isoforms with NAFLD remains largely unclear.
Therefore, the aim of the present study was to investigate the variance in levels of the three oligomers of APN and assess their relation to other parameters in Chinese NAFLD patients.

| Analysis of adiponectin multimers
Adiponectin multimers were analyzed by Western blotting followed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis. Adiponectin in serum was diluted 10 times and combined with 5× Laemmli sample buffer without a reducing agent. Samples (15 μL) were loaded and run in 4%-15% criterion precast 10-well gels (Bio-Rad). Following electrophoresis, representative samples were transferred to nitrocellulose membranes (PALL) and subjected to immunoblotting analysis. Monoclonal anti-human/mouse Acrp30/APN (R&D Systems) and goat anti-rat IgG horseradish peroxidase (Santa Cruz Biotechnology) antibodies were used for Western blot analyses. The blots were developed using an ECL-Plus chemiluminescence reagent kit (Amersham Bioscience) and visualized using a UVP Bio-Imaging System. Blot densities were analyzed using Vision Works LS Acquisition and Analysis Software.

| Statistical analysis
Statistical analysis of the data was performed using SPSS Statistics V22.0 (SPSS Inc). The continuous variables were presented as the mean ± SD. The comparisons of the means between the two groups were tested by a Student's t test, and correlations among the parameters were tested by Pearson's correlation coefficient. A P value of <.05 was considered as statistically significant.

| RE SULTS
As shown in Table 1, the general clinical characteristics, including weight, body mass index (BMI), systolic blood pressure (SBP), diastole blood pressure (DBP), total APN, BCAA, CML, FPG, T.cho., TG (all P < .001), LDL-C (P < .01), and HDL-C (P < .05), were significantly higher in NAFLD patients when compared with age-matched healthy control subjects. Contrastingly, the levels of sRAGE (P < .05) and APN (P < .01) were significantly decreased in NAFLD patients. The two groups did not differ in the parameters of age, height, AST, or ALT.
In this study, 34 control subjects and 47 NAFLD patients were further analyzed for APN isoforms. A representative photograph of different APN isoform expression is shown in Figure 1A. We found that serum HMW, MMW, and LMW APN levels were lower in NAFLD patients ( Figure 1B). Subsequently, NAFLD patients following high-fat diets were separated from NAFLD patients according to the fat intake level in their daily diet. We further determined the levels of APN isoforms in NAFLD patients on high-fat diets and the control subjects, and found that serum HMW and MMW APN were significantly lower (P < .01 and P < .001, respectively) in NAFLD patients on a high-fat diet compared with those in the corresponding control subjects ( Figure 1C).
In the simple correlation analysis of the whole data set, serum total APN and the APN isoforms of HMW, MMW, and LMW were analyzed according to other clinical characteristics indicated in Tables 2-5 (Table 2). Furthermore, we analyzed the correlation between the APN isoforms of HMW, MMW, LMW, and other clinical characteristics to identify which factors affect the concentrations of these isoforms in control and NAFLD patients. As shown in Table 3, we found that BMI and DBP were positively associated with HMW APN in control subjects (R = .430, P = .011; R = .370, P = .031, respectively), similar to FPG (R = .350, P = .016) in NAFLD patients, while HMW APN was significantly and negatively correlated with ALT (R = −.321, P = .029) and AST (R = −0.295, P = .045) in NAFLD patients. As Table 4 demonstrates, positive correlations with MMW APN were observed in T.cho.

| D ISCUSS I ON
Metabolic syndrome is a cluster of metabolic abnormalities including diabetes, cardiovascular disease, and NAFLD. 16 In this study, we have reached the following findings: First, we found that the total APN and the distribution of APN isoforms were decreased in NAFLD patients. Second, in NAFLD patients following a high-fat diet, significantly decreased levels of HMW and MMW were observed, while the plasma BCAA levels were significantly increased.
Third, CML was significantly negatively correlated with total APN in NAFLD patients, which is in alignment with a study conducted  Thus, decreased APN may be the cause of elevated BCAA levels.
In this study, we observed significantly elevated serum BCAA levels and significantly reduced APN levels in the NAFLD patients, but the association between circulating BCAA and APN in the Chinese population was not significant. This result could be attributed to the specific population, the small sample size, and the specific detection   between plasma APN, which suppresses hepatic glucose output, and endogenous glucose production exists in NAFLD patients, 36  the causal associations. Thus, longitudinal studies should be adopted in further investigations as an approach to identify causal associations.
In conclusion, our findings suggest that decreased total, HMW, MMW, and LMW APN levels were observed in NAFLD patients.
Height and CML were significantly correlated with total APN. These results suggest that HMW and MMW APN may play an important role in the pathogenesis and progression of NAFLD. In addition, HMW APN and MMW APN may be closely associated with liver function and lipid metabolism, respectively, and can be considered potential novel therapeutic approaches for NAFLD.