Chromatin accessibility derived from cfDNA serves as a novel classification biomarker of glioma

Qiuyue Wu¹Jian Wu²Fan Wang³Qiang Wang⁴Wei Dai⁵Yang Yang¹Chuanyue Zhang³Ying Han³Min Chen³Hao Pan⁴Lan Luo^1*Xinyi Xia^6,7,8*

• ¹Institute of Laboratory Medicine, Jinling Hospital, Affiliated Hospital of Medical School, School of Life Sciences, Nanjing University, nanjing, China
• ²Department of Bioinformatics, Nanjing Medical University, nanjing, China
• ³Institute of Laboratory Medicine, Jinling Hospital, Affiliated Hospital of Medical School, Nanjing University, nanjing, China
• ⁴Department of Neurosurgery, Jinling Hospital, School of Medicine, Nanjing University, nanjing, China
• ⁵School of Animal Science and Food Engineering, Jinling Institute of Technology, nanjing, China
• ⁶Institute of Laboratory Medicine, Jinling Hospital, Affiliated Hospital of Medical School, nanjing, China
• ⁷State Key Laboratory of Analytical Chemistry for Life Science, Nanjing University, nanjing, China
• ⁸Jinling Hospital, The First School of Clinical Medicine, Southern Medical University, nanjing, China

Gliomas can be classified by their molecular characteristics, which are also closely associated with clinical outcomes. Cell-free DNA (cfDNA)-based liquid biopsy in gliomas is challenging because of the limited amount of tumor-derived cfDNA present in body fluids. In this study, we identified the open chromatin states of gliomas using cfDNA and demonstrated the potential of this technique for glioma detection. The chromatin accessibility of gliomas was investigated using tumor tissues donated by four donors. cfDNA derived from paired cerebrospinal fluid (CSF) and plasma samples was also sequenced. A total of 72 accessible chromatin regions in the tumor tissues were identified as open chromatin regions using CSF cfDNA. Furthermore, 16 open chromatin regions with significant differences in glioma grade were found using cfDNA extracted from plasma. A glioma grade classifier was constructed with 16 plasma cfDNA-derived accessible chromatin features, which could accurately differentiate low-grade from high-grade samples in the training dataset (AUC = 0.814). However, lower accuracy was obtained on the testing dataset (AUC = 0.736). The diversity of transcription factor binding sites among glioma patients was also illustrated using cfDNA. In conclusion, our study defines novel chromatin-accessibility-based biomarkers and illustrates their potential application in glioma liquid biopsy.