# A six-gene expression signature in peripheral blood mononuclear cells effectively diagnoses osteoarthritis

## TL;DR

A six-gene model in blood cells can accurately diagnose osteoarthritis, offering a promising early detection method.

## Contribution

A novel six-gene expression signature in PBMCs for diagnosing osteoarthritis is developed and validated.

## Key findings

- A six-gene model (THBS1, USP36, GIMAP4, OSM, IL10, HDC) effectively distinguishes OA patients from healthy controls.
- The diagnostic model achieved AUCs of 0.928 in internal and 0.915 in external PBMC cohorts.
- The gene model also showed high accuracy (AUC = 0.910) when applied to synovial tissue data.

## Abstract

Osteoarthritis (OA) is a heterogeneous whole-joint disease that inconveniences more than 500 million people worldwide. Early diagnostic methods for OA remain lacking. Peripheral blood mononuclear cells (PBMCs) are ideal sample sources for the early diagnosis of different diseases. However, only a few studies have reported on the role of PBMCs in the early diagnosis of OA.

RNA sequencing was performed on PBMC samples from 27 patients with OA and 31 healthy controls. We integrated RNA sequencing data from our internal cohort and microarray data from external cohort to construct a diagnostic model of OA based on PBMC samples. The receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic model in PBMC samples and synovial tissue.

In this study, we screened and constructed a six-gene diagnostic model consisted of the genes THBS1, USP36, GIMAP4, OSM, IL10, and HDC, which could effectively distinguish patients with OA from healthy controls. The ROC curve analysis showed that the area under curve (AUC) of this diagnostic model was 0.928 for our internal cohort and 0.915 for the external cohort, respectively. Interestingly, the gene expression model also had high accuracy (AUC = 0.910) for diagnosing patients with OA based on expression data from synovial tissue.

Given that related studies on several signature genes in our diagnostic model for OA are lacking, our study provides novel potential biomarkers for the early diagnosis of OA based on PBMC samples.