Prediction of Diabetic Nephropathy Using Urine Proteomic Profiling 10 Years Prior to Development of Nephropathy

Abstract

OBJECTIVE—We examined whether proteomic technologies identify novel urine proteins associated with subsequent development of diabetic nephropathy in subjects with type 2 diabetes before evidence of microalbuminuria. RESEACH DESIGN AND METHODS—In a nested case-control study of Pima Indians with type 2 diabetes, baseline (serum creatinine <1.2 mg/dl and urine albumin excretion <30 mg/g) and 10-year urine samples were examined. Case subjects (n = 31) developed diabetic nephropathy (urinary albumin–to–creatinine ratio >300 mg/g) over 10 years. Control subjects (n = 31) were matched to case subjects (1:1) according to diabetes duration, age, sex, and BMI but remained normoalbuminuric (albumin–to–creatinine ratio <30 mg/g) over the same 10 years. Surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) was performed on baseline urine samples, and training (14 cases:14 controls) and validation (17:17) sets were tested. RESULTS—At baseline, A1C levels differed between case and control subjects. SELDI-TOF MS detected 714 unique urine protein peaks. Of these, a 12-peak proteomic signature correctly predicted 89% of cases of diabetic nepropathy (93% sensitivity, 86% specificity) in the training set. Applying this same signature to the independent validation set yielded an accuracy rate of 74% (71% sensitivity, 76% specificity). In multivariate analyses, the 12-peak signature was independently associated with subsequent diabetic nephropathy when applied to the validation set (odds ratio [OR] 7.9 [95% CI 1.5–43.5], P = 0.017) and the entire dataset (14.5 [3.7–55.6], P = 0.001), and A1C levels were no longer significant. CONCLUSIONS—Urine proteomic profiling identifies normoalbuminuric subjects with type 2 diabetes who subsequently develop diabetic nephropathy. Further studies are needed to characterize the specific proteins involved in this early prediction.