Urinary proteome profiling across normal glucose tolerance, impaired glucose tolerance, and type 2 diabetes mellitus
Year: 2026
Type: article
Abstract: Impaired glucose tolerance (IGT) and type 2 diabetes mellitus (T2DM) are distinct glycemic states, but their associated molecular profiles remain incompletely characterized. Urine is collected noninvasively and may capture systemic molecular changes relevant to biomarker discovery. We compared urinary proteome profiles across normal glucose tolerance (NGT), IGT, and T2DM in a cross-sectional study and developed candidate multiprotein classifiers. In this cross-sectional study, urine samples from 34 NGT individuals, 15 patients with IGT, and 17 patients with T2DM were analyzed by data-independent acquisition (DIA) liquid chromatography-tandem mass spectrometry (LC-MS/MS). Candidate proteins were screened using an exploratory dual threshold of fold change (FC) ≥ 1.5 or ≤ 0.67 and an unadjusted two-tailed t-test p < 0.05. Functional enrichment used DAVID, and individual-protein performance was assessed by receiver operating characteristic (ROC) curve analysis. Multiprotein logistic-regression classifiers were evaluated by fully nested cross-validation. Within each outer leave-one-out cross-validation (LOOCV) training set, the analysis repeated differential-protein selection, AUC ranking, combination selection, scaling, and model fitting. A total of 3,569 urinary proteins were identified. The exploratory screening criteria yielded candidate lists of 84 proteins for IGT versus NGT, 261 for T2DM versus NGT, and 128 for T2DM versus IGT. These proteins included C-reactive protein (CRP), catalase (CAT), and toll-like receptor 3 (TLR3), and were enriched in pathways related to advanced glycation end product-receptor for advanced glycation end product (AGE-RAGE) signaling, oxidative stress responses, and angiogenesis. The highest AUCs for individual proteins were observed for HLA-C in IGT versus NGT (AUC = 0.806), PRAP1 in T2DM versus NGT (AUC = 0.829), and APOC2 in T2DM versus IGT (AUC = 0.827). Fully nested cross-validation yielded outer LOOCV AUCs of 0.590 (95% CI 0.384–0.788) for IGT versus NGT, 0.713 (95% CI 0.541–0.872) for T2DM versus NGT, and 0.525 (95% CI 0.322–0.729) for T2DM versus IGT. Urinary proteome profiles differed among the three glycemic groups. Because the sample was small relative to the number of proteins tested, the candidate lists and nested models remain vulnerable to false discovery and model instability. The identified proteins should therefore be treated as hypotheses for multiplicity-controlled testing in larger independent samples. They may inform subsequent development of classifiers for conventionally defined glycemic phenotypes. Not applicable this was an observational study without an interventional component.
Source: Clinical Proteomics
Language: English
Subfield: Spectroscopy
Field: Chemistry
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,490
OpenAlex ID: W7225834749
FWCI:
Cites:
Cited by:
Related to: