Recently, a multicenter original manuscript on biomarkers of neurosyphilis (NS) by Professor Leng Ling’s lab from the Clinical Research Institute, and Professor Li Jun’s lab from the Department of Dermatology at PUMCH was published in “Advanced Science” (a tier 1 journal, or among the top 5%, as ranked by the Chinese Academy of Sciences, IF: 17.521), an influential journal in the field of basic research, and also featured on the cover. This study pioneered an AI-powered scoring approach based on proteomics, and revealed for the first time globally that specific extracellular matrix (ECM) proteins could serve as potential diagnostic biomarkers for NS. This study provides new methods and insights for addressing the diagnostic difficulties and high misdiagnosis rates of NS. It was supported by the National High Level Hospital Clinical Research Funding.

NS is a central nervous system infection caused by Treponema pallidum and manifests as a broad spectrum of clinical symptoms that damage the meninges, brain, blood vessels, and/or spinal cord, etc. It can occur at any stage of syphilis and can be life-threatening if left untreated. Currently, there is no gold standard for NS diagnosis, and existing diagnostic tools have limitations in sensitivity and specificity, leading to many cases of NS being misdiagnosed. Early identification and diagnosis as well as timely treatment are critical to NS patients.

In this study, proteome/matrisome technologies combined with an AI-powered machine learning model were used to characterize 223 cerebrospinal fluid (CSF) samples from 7 independent cohorts to identify and verify promising molecular biomarkers for diagnosing NS. The study also provides insights into the underlying ECM mechanisms triggered by the associated inflammatory responses.

The machine learning classifier constructed in this research achieved a 98.11% accuracy in distinguishing NS. The team simultaneously predicted and biologically verified the selected candidate ECM biomarkers (SEMA7A, SERPINA3, and ITIH4) in 115 suspected NS cases, with a prediction rate close to 100%. The protein expression patterns were consistent with the pathological features of NS patients’ brain tissues. During follow-up, the application of these biomarkers led to prompt detection of neurological damage in patients. Thus the study provides not only an effective tool for NS diagnosis and treatment monitoring but also important molecular evidence for understanding the pathological mechanisms of ECM and evaluating treatment efficacy.

This study increases the hope for improving the early diagnosis rate of NS and reducing the risk of complications. The results have been filed for a national invention patent and are expected to be applied clinically. (Nature Communications, 2022).

▲ Prediction and verification of ECM biomarkers in suspected NS populations

Written by Leng Ling

Edited by Chen Xiao and Gan Dingzhu

Translated by Liu Haiyan

Reviewed by Leng Ling and Wang Yao