Identification of Urinary CD44 and Prosaposin as Specific Biomarkers of Urinary Tract Infections in Children With Neurogenic Bladders

Forster, Catherine S; Haffey, Wendy D; Bennett, Michael; Greis, Kenneth D; Devarajan, Prasad

Five patients with UTI and 5 patients with UTC were included in the mass spectrometry analysis, while 10 patients with no growth, 15 with UTC, and 15 with UTI were included in the assay analysis. There were no differences in age, sex, race, etiology of neurogenic bladder, presence of Mitrofanoff, or bladder augmentation between those with UTC and UTI in the mass spectrometry analysis. In the assay analysis, there was a higher proportion of males in the no growth group compared with the UTC group. There were no other differences between groups (Table 1) . More than 200 proteins were identified using mass spectrometry. From this protein set, 100 proteins were consistently detected and quantified in at least 3 of the 5 patients from both groups and were subjected to t test analysis to establish the significance of the change (Supplemental Figure S1 ). There were a few examples of proteins that were detected in 3 or more of one group and less than 3 in the other group; however, in each case, the levels of the protein detected in the group of 3 were well outside of a significance P value of .05. After removing contaminating hemoglobin and several immunoglobulin proteins likely linked to minor blood contamination and/or inflammatory response, 8 proteins were identified with P <.05 as proteins of interest based on at least a 0.5-fold increase or decrease in the 2 test groups. These included haptoglobin, apolipoprotein D, alpha-amylase 2B, inter-a-trypsin inhibitor heavy chain H4, RNase2, CD44 antigen, liver fatty acid-binding protein, and prosaposin ( Figure 1 ). Of the 8 proteins of interest identified by mass spectrometry, 5 were measured by assay in an additional set of samples: apolipoprotein D, alpha-amylase 2B, RNase2, CD44, and prosaposin. Alpha-amylase 2B was not able to be analyzed as all samples had a value below the LOD. There were no significant differences between either normalized or non-normalized concentrations of either RNase2 or apolipoprotein D (Figure 2 ). Normalized prosaposin concentrations were significantly higher in UTI compared with no growth (P = .003) and UTC (P = .05). Normalized CD44 was significantly higher in UTI compared with UTC (P = .02), although there was no difference in normalized CD44 concentrations between UTI and no growth ( Figure 2 ). Non-normalized prosaposin was significantly higher in UTI compared with both no growth (P = .02) and UTC (P = .05). . At the identified cut-offs, both CD44 and prosaposin have poor sensitivity, but good specificity, for the outcome of UTI (Table 2 ).