Urine proteomic profiling of pediatric nephrotic syndrome

Abstract

The prognosis of pediatric nephrotic syndrome (NS) correlates with the responsiveness to glucocorticoid therapy. Steroid-resistant NS (SRNS) patients progress to end-stage renal disease, while steroid-sensitive NS (SSNS) and steroid-dependent (SDNS) patients do not. We have performed proteomic profiling of urine samples from a cross section of pediatric and adolescent subjects with SSNS, SRNS, and orthostatic proteinuria (OP) to identify urinary biomarkers of steroid resistance. We performed surface-enhanced laser desorption/ionization time of flight mass spectrometry (SELDI-TOF MS) on urine from 19 subjects with SSNS/SDNS in remission, 14 with SSNS/SDNS in relapse, 5 with SRNS in relapse, and 6 with OP. Genetic algorithm search of principal component space revealed a group of five peaks distinguishing SRNS subjects, with mass/charge (m/z) values of 3,917.07, 4,155.53, 6,329.68, 7,036.96, and 11,117.4. Our analyses identified the peak at m/z 11,117.4 with an accuracy of 95% for classifying SRNS. Multidimensional protein fractionation and mass spectrometric analysis of SRNS urine samples combined with immunodepletion identified the 11,117.4 protein as beta2-microglobulin (B2M). Using an unbiased protein profiling approach, we have validated previously reported findings of B2M as a biomarker associated with SRNS. Prospective studies are warranted to establish additional biomarkers that would be predictive of SRNS.

Fig. 1

Representative MS spectrum shown before (top) and after (bottom) noise reduction. For each spectrum, the algorithm determined the distribution of peak intensities, assumed a bimodal distribution, determined the intensity value at the nadir of the bimodal distribution, and removed all peaks from that spectrum that demonstrated intensity values below the nadir intensity value. The two spectra differ at the lower intensity levels

Fig. 2

Representative MS spectra demonstrating peak found in 11,117.4 in SR subjects (two spectra at the top) and not found in SD (two spectra in the middle) and SS subjects (two spectra at the bottom). Spectra are shown from three subjects in duplicate

Fig. 3

Scatterplot of SRNS (black) and SS/SD NS (gray) samples in principal component space, projected using five mass peaks selected by the genetic algorithm

Fig. 4

Box plots of the five mass peaks identified by the genetic algorithm within principal component space. These are the same five mass peaks that are used to project samples into principal component space in Fig. 3. IMAC30 refers to the anionic protein chip, m/z mass/charge ratio, SS/SD steroid sensitive/steroid dependent, SR steroid resistant

Fig. 5

A urine sample containing the protein of interest was incubated with antibody against β₂-microglobulin (a), water (b), normal goat immunoglobulin (c), or NF-κB p65 (d). The samples were then immunopre-cipitated and the supernatant was used to spot a gold chip for mass spectrometry analysis. The spectra shown are from the region of m/z 11,117.4. The major peak seen in this region is indicated by the arrow and is dramatically reduced with the B2M antibody and not with the controls used