Association of serum and aqueous humor myonectin concentrations with diabetic retinopathy

Myonectin, a newly discovered myokine, enhances fatty acid uptake in cultured adipocytes and hepatocytes and suppresses circulating levels of free fatty acids in mice. This study is performed to evaluate the association between serum and aqueous humor myonectin concentrations with diabetic retinopathy (DR). This study was performed in a population of 228 patients with type 2 diabetes (T2DM) and 72 control subjects. Diabetic patients were then divided into T2DM patients without DR, non-proliferative diabetic retinopathy (NPDR) patients, and proliferative diabetic retinopathy (PDR) patients. Serum and aqueous humor myonectin concentrations were significantly lower in the case group than in the control group. PDR patients showed significantly decreased serum and aqueous humor myonectin concentrations than in the other two T2DM patients. In addition, NPDR patients showed significantly lower serum and aqueous humor myonectin concentrations than T2DM patients without DR. Logistic regression analysis demonstrated that serum and aqueous humor myonectin was correlated with a decreased risk of T2DM and DR. Simple linear regression analysis showed that serum myonectin was negatively correlated with duration of disease, body mass index (BMI), and HbA1c. Duration of disease and BMI were still correlated with the serum myonectin after a multiple linear regression analysis. Aqueous humor myonectin was negatively correlated with duration of disease, systolic blood pressure (SBP), and diastolic blood pressure. Duration of disease and SBP was still correlated with the aqueous humor myonectin after a multiple linear regression analysis. Our investigation indicates an inverse association of serum and aqueous humor myonectin with DR.

www.nature.com/scientificreports/ ing: more than 20 intraretinal hemorrhages in each of 4 quadrants; definite venous beading in 2+ quadrants; prominent intraretinal microvascular abnormalities in 1+ quadrant. And no signs of PDR); (3) PDR (one or more of the following: neovascularization, vitreous/preretinal hemorrhage). T2DM patients were divided as follows: T2DM patients without DR (n = 96), NPDR patients (n = 78), and PDR patients (n = 54). NPDR patients were then divided into mild NPDR (n = 34), moderate NPDR (n = 35), and severe NPDR (n = 27). Patients with intraocular surgery history, retina laser photocoagulation history, VEGF therapy history, glaucoma, other ocular disorders, and other systemic disorders were excluded. 72 healthy subjects who went on cataract surgery were considered as the control group. The subjects who had systemic disease were excluded from the control group. This study was approved by the ethics board of Qilu Hospital (Qingdao) and all patients provided written informed consent. All experimental protocols were approved by the licensing committee of Qilu Hospital (Qingdao). All methods were carried out in accordance with relevant guidelines and regulations. Laboratory methods. Serum was collected from all participants during a fasting status. A paracentesis was made in the peripheral cornea next to the limbus using a 1 mL syringe, and undiluted samples of aqueous humor (0.2 mL) were obtained just before cataract surgery. Serum and aqueous humor samples were deposited into Eppendorf tubes and stored at − 80 °C until sample analyses. Finally, serum and aqueous humor myonectin concentrations were investigated using enzyme-linked immunosorbent assay method (Aviscera Biosciences, Santa Clara, CA).

Statistical analysis.
The results were displayed as means ± standard errors or interquartile range. Chisquare tests, one-way ANOVA, or Kruskal-Wallis test were performed to compare the differences between the four groups. Serum and aqueous humor myonectin differences between mild, moderate, and severe NPDR subgroups were compared using Kruskal-Wallis test. The risk factors for the presence of T2DM and DR was determined by logistic regression analysis. The correlation between serum and aqueous humor myonectin and other variables was analyzed using simple and multiple linear regression. Pearson correlation analysis was utilized to determine the association between serum and aqueous humor myonectin. Levels of statistical significance were set at P < 0.05.

Results
Clinical characteristics. As presented in Table 1, higher blood pressure, HbA1c, triglycerides (TG), as well as lower high-density lipoprotein cholesterol (HDL-C) were observed in T2DM patients than in the controls. PDR patients presented higher blood pressure compared with the controls and the other two T2DM subgroups.
Myonectin concentrations. There were lower serum and aqueous humor myonectin concentrations in the three T2DM subgroups compared with healthy controls (Table 1). Decreased serum and aqueous humor myonectin concentrations were observed in PDR patients than in NPDR and T2DM without DR subgroups (Table 1). Moreover, NPDR patients presented lower serum and aqueous humor myonectin concentrations than T2DM without DR subgroup (Table 1).  Table 2, simple logistic regression analysis showed that systolic blood pressure (SBP), diastolic blood pressure (DBP), TG, HDL-C, serum and aqueous humor myonectin showed a trend toward an association with T2DM. Multivariate logistic regression analysis revealed that serum and aqueous humor myonectin remained a significant predictor of T2DM.

The correlation of serum and aqueous humor myonectin with other variables. As presented
in Table 4, serum myonectin was negatively correlated with duration of disease, body mass index (BMI), and HbA1c. Duration of disease and BMI were still correlated with the serum myonectin after a multiple linear regression analysis. Simple linear regression analysis showed that aqueous humor myonectin was negatively correlated with duration of disease, SBP and DBP ( Table 5). Duration of disease and SBP were still correlated with the aqueous humor myonectin after a multiple linear regression analysis (Table 5).

Discussion
The present study indicated that serum myonectin concentrations were decreased in T2DM patients. Zhang and Li also reported that circulating myonectin levels were significantly decreased in T2DM patients compared with the controls 8,9 . However, another study performed in Chongqing of China showed that newly diagnosed Table 3. Logistic regression analysis for determining the risk factors of developing DR. Abbreviation as Table 1. Univariate logistic regression analysis was performed and the variables with a P < 0.05 were then entered into a multivariate logistic regression model to assess the significant independent factors associated with DR.  Table 4. The correlation between serum myonectin concentrations and various parameters. Abbreviation as Table 1. The correlation between serum myonectin and other parameters were analyzed using simple linear regression analysis. Then the variables with a P < 0.05 were entered into a multiple linear regression model to determine the contribution of various factors to serum myonectin. www.nature.com/scientificreports/ T2DM and impaired glucose tolerance (IGT) subjects had higher circulating myonectin concentrations than normal subjects 10 . And circulating myonectin levels were higher in newly diagnosed T2DM patients than in IGT subjects 10 . These four studies are all performed in China. Therefore, the conflicting results caused by ethnic difference can be excluded. We speculate that the contradictory results may be due to different ELISA kits or population enrolled. Myonectin is associated with macrovascular disease. Serum myonectin was higher in coronary artery disease (CAD) patients compared with controls 11 . Serum myonectin was correlated with disease severity of CAD. Zhang reported that serum myonectin levels were decreased in CAD patients compared to the non-CAD group 12 . Serum myonectin alteration may serve as a marker for CAD 12 . Myonectin-knockout mice with ischemia-reperfusion showed the enhancement of myocardial infarct size, cardiac dysfunction, apoptosis, and proinflammatory gene expression compared with wild-type mice 6 . Myonectin treatment inhibited hypoxia/reoxygenation-induced apoptosis in cultured cardiomyocytes 6 . All these results point to the close association of myonectin with macrovascular disease. It is hypothesized that serum myonectin may be correlated with macrovascular complications of diabetes. Our results demonstrated that decreased serum and aqueous humor myonectin concentrations were correlated with DR. This is the first report about the correlation between serum and aqueous humor myonectin and DR.  www.nature.com/scientificreports/ The potential role of inflammatory mediators in DR has been demonstrated by a variety of clinical evidence 13 . High inflammatory molecules levels have been observed in serum and vitreous sample from DR patients and animal model of diabetes 14,15 . Recent studies demonstrated the role of myonectin in regulating inflammation. Proinflammatory gene expression were increased in myonectin-knockout mice 6 . In addition, myonectin inhibited inflammatory response stimulated by lipopolysaccharide in macrophages 6 . There was a significant relation between serum myonectin and interleukin-6, tumor necrosis factor-α in CAD patients 11 . Myonectin treatment could alleviate the anemia of inflammation induced by in mice 16 . These results indicates that myonectin is closely correlated with inflammation. Therefore, myonectin may be involved in the pathogenesis of DR by inhibiting inflammation. However, as the angiogenesis plays a key role in the mechanism of DR, further investigation may focus on the possible role of myonectin protecting from DR through the anti-angiogenic effects.
Where does the aqueous humor myonectin come from? Myonectin has been found in the eye tissue of mouse 4 . The mRNA expression of myonectin in the mouse eye tissue is about one fifth of those in the skeletal muscle tissue 4 . However, we did not know exactly which cells in the eye tissues contribute to the production of myonectin. Further study is needed to investigate the original cell in the eye to secrete the myonectin.
This study has several potential limitations. First, the conclusion is limited by relatively small sample size. Secondly, the cross-sectional nature of the data limited the strength of conclusion. The causative relation must be confirmed by future longitudinal studies. Thirdly, we did not explain the exact mechanism of myonectin protecting from DR. Further basic science study is needed to clarify the mechanism of myonectin involved in DR development or progression.
In short, serum and aqueous humor myonectin concentrations are negatively correlated with DR.