A Proton Nuclear Magnetic Resonance-based Metabonomics Study In Serum Of Patients With Immunoglobulin A Nephropathy

Objective:Metabolites in serum of patients with differently pathological degrees of immunoglobulin A nephropathy (IgAN) and healthy volunteers were studied by proton nuclear magnetic resonance (1H-NMR)-based metabolomics technology in order to analysis biomarkers and stablish a human IgAN metabolic profile, as well as explore the pathogenesis of IgAN from metabolic pathways. We wanted to provide a new view to the early monitoring, diagnosis and pathogenesis for IgAN. Methods:The samples were recruited from Nephrology Department of181st Hospital Guilin, Guangxi, China from October2010to January2011.35cases primary IgAN patients were diagnosed by renal biopsy and biochemical indicators. Neither patients nor controls had not received hormones and immunosuppressive therapy. Patients who had systemic lupus erythematosus, Schonlein-Henoch purpura, chronic liver disease, malignancies, diabetes mellitus and other autoimmune disorder diseases, were excluded. According to Lee’s classification, IgAN were classed into grades Ⅰ-Ⅴ. Low-risk groups were consisted of twenty-three IgAN patients with grades Ⅰ-Ⅲin the name of IgAN-A, while high-risk groups were composed of twelve IgAN patients with grades Ⅳ-Ⅴin the name of IgAN-B. Twenty-three healthy volunteers who physical examinations were normal served as controls. The1H NMR spectra of serum samples from IgAN patients and controls were acquired on a Varian INOVA-600MHz NMR spectrometer. After data preprocessing of the acquisition of NMR spectra, principal component analysis (PCA) was performed to examine the intrinsic variation in the dataset, and then partial least squares-discriminant analysis (PLS-DA) were carried out for class discrimination. To further validate the quality of the PLS-DA model and the orthogonal partial least squares discriminant analysis (OPLS-DA) model,10fold cross-validation were carried out. Statistical analysis of correlation coefficients were able to identify the different metabolites in serum of IgAN patients and healthy controls. OPLS-DA model validated the sensitivity, specificity and resolution for diagnosing patients with IgAN. Finaly, the IgAN-related metabolic pathways network could provide insights into the molecular mechanisms involved in the pathogenesis of IgAN.Results:1.The biochemical results:The levels of total protein, albumin and globulin in serum of healthy groups were higher than IgAN patients(P<0.05), whereas the levels of urea nitrogen, creatinine and uricacid in serum of healthy controls were significantly lower than IgAN-A patients (P<0.01)and IgAN-B patients(P<0.001). Furthermore, the24h proteinuria were significantly higher in IgAN-B patients compared to IgAN-A patients(P<0.01).2. Histopathological immune typing with the IgAN groups:The simply type IgA was12cases (34.29%), type IgAG was5cases (14.28%), type IgAM was6cases (17.14%), type IgAGM was3cases (8.57%), type IgAMC3was4cases(11.42%), type IgAGMC4was2cases (5.71%), type IgAGC1q was3(8.57%). The deposition intensity of IgA, IgM and complement C3in the IgAN-A groups was significantly less than those in the IgAN-B groups(P<0.05). The deposition intensity of IgG, complement C4and C1q were no significant difference (P>0.05).3. The results of1H-NMR-based metabonomics study:The typical’H-NMR spectra of serum samples obtained from patients of IgAN-A and IgAN-B and healthy individuals. The PCA score plot showed the cluster of IgAN patients was located far away from that of healthy groups, while IgAN-A patients did not show obvious difference in the IgAN-B groups. The model parameters for the explained variation, R2, and the predictive capability, Q2, were significantly high(R2, Q2>0.5), indicating PLS-DA and OPLS-DA models had high predictive ability of differentiation between healthy individuals and IgAN patients. Additionally, R2=0.573and Q2=-0.160revealed lower predictive ability of differentiation between IgAN-A and IgAN-B patients. Model prediction showed the sensitivity of88.6%, the specificity of97.1%and the resolution of93.1%.4. Compared with the healthy individuals, the serum of IgAN patients were characterized by higher levels of pyruvate, phenylalanine,myo-Inositol, unsaturated lipids (such as L5:Lipid-CH2-C=O; L6:Lipid,=CH-CH2-CH=; L3:Lipid,-CH2-CH2-C=O), creatine and creatinine whereas lower levels of a-glucose, P-glucose, valine, tyrosine, phosphocholine, lysine, isoleucine, glycerolphosphocholine, glycine, glutamine, glutamate, alanine, acetate,3-hydroxybutyrate,1-methylhistidine. No significant difference between the low risk groups and high-risk groups(P>0.05).5. A total of24metabolites were identitied as potential biomarkers of IgAN, which involved in carbohydrate metabolism, lipid metabolism, protein metabolism, the citric acid cycle and creatine synthesis metabolism.Conclusion: 1.1H-NMR metabolomics technology combined with PCA, PLS-DA and OPLS-DA analysis was an effective research tool for high-throughput screening different metabolites in the serum of IgAN patients.2. Applying pattern recognition analysis of PCA, PLS-DA and OPLS-DA to analyze the NMR dates could clearly distinguished IgAN patients from healthy controls, but not well distinguished between differently pathological degrees of IgAN patients.3. The validitys of PLS-DA and OPLS-DA model were high, and metabolic profiling of IgAN could reflect renal pathological changes, which was better than the biochemical testing and histopathological analysis.4. The metabolic profiling of IgAN was significant changes.24kinds of characteristic metabolites might be new and effective biomarkers for early diagnosis and prediction.5. IgAN lead to the disorders of of carbohydrate metabolism, lipid metabolism, protein metabolism, the citric acid cycle and creatine synthesis metabolism, performing up-regulation of glycolysis, lipid and creatine synthesis metabolism and down-regulation of gluconeogenesis and protein metabolism. It would provide a new perspective for people to understand the pathogenesis of IgAN from the metabolite levels.