Proteomics of Diabetic Nephropathy

Diabetes affects more than 6% of the British population, and the expenditure on patients with diabetes accounts for 15% of the UK health care budget. Almost 90% of patients have type 2 diabetes, and absolute numbers are expected to rise in parallel to the current obesity and metabolic syndrome epidemic. Recent advances in treatment have led to reduced mortality but, due to prolonging the duration of diabetes, have increased the likelihood of development of diabetic complications. Major complications are micro and macrovascular disease, which are highly interconnected. Diabetic nephropathy (DN) is a major microvascular complication of diabetes but is associated with substantial macrovascular disease and thereby cardiovascular morbidity and mortality. Already at early stages of DN, cardiovascular morbidity is increased, and it is further aggravated once end stage renal disease (ESRD) occurs.

Proteomic studies

We used CE-MS based proteomics to develop a urinary biomarker patternthat is specific for chronic kidney disease (CKD). As described in detail elsewhere [1] CE-MS data from healthy subjects (N=379, controls) and patients with various kidney diseases (N=230, cases) were compared for the discovery of CKD specific biomarkers. This resulted in the identification of 273 peptides that are significantly associated with disease. The CKD specific biomarkers include defined fragments of different collagens, alpha-1-antitrypsin, fibrinogen alpha chain, uromodulin, as well as various secreted protein fragments. A high dimensional, support-vector machine based classifier (CKD273) was established to distinguish between healthy subjects and individuals with CKD. The biomarker panel has been validated in a multicentre approach involving >1000 blinded samples. In these validation sets, ROC analysis demonstrated accurate prediction of CKD by the CKD273 classifier with an AUC under the ROC curve of >0.95. When only diabetic patients were evaluated, the accuracy was even higher (AUC of 0.99; 96% sensitivity and 98% specificity).

ROC analysis of the prediction of DN, based on urinary albumin excretion or the CKD273 pattern in 149 samples from patients who were normalbuminuric at the time of analysis. The AUC for albuminuria was 64%, AUC for the CKD273 pattern was 89%. The difference between both classifiers was highly significant; p<0.001.

Proposed study

The proteomic biomarkers of CKD can also be used to identify patients at very early stages of CKD in whom traditional biomarkers (albuminuria, creatinine, cystatin C) are still normal.

We propose a study using urinary proteomics to identify patients at high risk for development of diabetic nephropathy. These patients can then be subject to more intensive preventative therapy (“therapy +”) and compared with standard preventative therapy. Together with our collaborators we are currently developing appropriate study protocols.

Based on a conservative estimate from our previous studies, the expected relative proportions of patients with Type 2 diabetes who score positive in the urinary proteome analysis (UPA) is 20%. In the observation period of 3 years we estimate that in the absence of specific intervention, 60% of these patients at high risk will develop microalbuminuria. In contrast, we estimate that progression rate will be 45% in the “treatment +” group. Intervention will be targeted to patients at high risk, so that progression rate is estimated at only 6% in those who score negative in the proteome analysis. Sample sizes were calculated based on these assumptions.

Publications:

1. Good DM, Zürbig P, Argiles A, Bauer HW, Behrens G, Coon JJ, Dakna M, Decramer S, Delles C, Dominiczak AF, Ehrich JH, Eitner F, Fliser D, Frommberger M, Ganser A, Girolami MA, Golovko I, Gwinner W, Haubitz M, Herget-Rosenthal S, Jankowski J, Jahn H, Jerums G, Julian BA, Kellmann M, Kliem V, Kolch W, Krolewski AS, Luppi M, Massy Z, Melter M, Neususs C, Novak J, Peter K, Rossing K, Rupprecht H, Schanstra JP, Schiffer E, Stolzenburg JU, Tarnow L, Theodorescu D, Thongboonkerd V, Vanholder R, Weissinger EM, Mischak H, Schmitt-Kopplin P. Naturally occurring human urinary peptides for use in diagnosis of chronic kidney disease. Mol Cell Proteomics 2010;9:2424-2437.