# Urinary extracellular vesicle metabolomic profiling reveals a distinct molecular signature for the non-invasive diagnosis of lupus nephritis

## TL;DR

This study identifies a unique urine-based metabolic signature for diagnosing lupus nephritis without the need for invasive procedures.

## Contribution

The study introduces a non-invasive biomarker panel based on urinary extracellular vesicle metabolites for lupus nephritis diagnosis.

## Key findings

- 284 differential metabolites were identified in urinary extracellular vesicles from lupus nephritis patients.
- A panel of ten candidate metabolites was prioritized using machine learning for accurate diagnosis.
- Three metabolites showed strong discriminatory performance with AUCs above 0.89.

## Abstract

Lupus nephritis (LN) is a severe complication of systemic lupus erythematosus (SLE), underscoring an urgent need for non-invasive diagnostic biomarkers.

This study aimed to define the metabolomic signature of urinary extracellular vesicles (uEVs) in LN and to identify novel biomarkers for precision diagnosis.

uEVs were isolated from urine samples of 29 SLE patients with LN, 22 SLE patients without renal involvement, and 20 healthy controls (HCs) using a standardized precipitation-based protocol. uEVs were rigorously characterized in accordance with the Minimal Information for Studies of Extracellular Vesicles (MISEV) guidelines, including transmission electron microscopy, nanoparticle tracking analysis, and the assessment of canonical EV markers. Comprehensive untargeted metabolomic profiling of uEVs was subsequently performed using liquid chromatography–tandem mass spectrometry (LC–MS/MS).

A total of 284 differential metabolites were identified between LN patients and the SLE group, including 230 upregulated and 54 downregulated metabolites. Machine learning–based feature prioritization using a random forest algorithm identified a panel of ten candidate metabolites. Notably, three metabolites—Glucosylsphingosine (Lyso-Gb1), phosphatidylethanolamine N-methylated (PE-NMe), and PC(20:5/TXB2)—demonstrated excellent discriminatory performance for differentiating LN from non-renal SLE, with areas under the receiver operating characteristic curve (AUCs) of 0.912, 0.906, and 0.897, respectively.

We identified a distinct uEV metabolic signature in LN and developed a robust, non-invasive biomarker panel. This strategy holds significant promise for the early detection and personalized management of LN, offering a compelling alternative to invasive renal biopsy.