Brain tumor detection based on transcription regulation features identified from public cerebrospinal fluid cell-free DNA sequencing data.

BACKGROUND

Characterization of the molecular features of a brain tumor is a critical step for patient treatment. Tissue-based detection methods are limited by the location of brain tumors and high intratumor heterogeneity, which also preclude repeat sampling to monitor tumor progression. Cerebrospinal fluid (CSF)-based noninvasive methods may provide an opportunity to solve these problems, but efficient markers are lacking. This study aims to develop and validate a CSF-based liquid biopsy approach to investigate the molecular characterization and transcriptional regulation features of brain tumors.

METHODS

In this study, we conducted genome wide analysis of CSF cell free DNA (cfDNA) data collected from Sonic hedgehog (SHH) pathway-activated medulloblastoma (MB) patients sourced from gene expression omnibus (GEO) database, to identify genome features that differed significantly between patients with MB and those with hydrocephalus (P<0.001) using the whole genome bisulfite sequencing (WGBS) dataset.

RESULTS

A total of 397 differential cfDNA genomic loci were identified and verified by assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-seq) as SHH-MB specific; 114 were located in promoter regions, and related to genes specifically expressed in SHH-MB and, combined with DNA methylation state in these regions, could be used to classify the SHH-MB subtype from 763 samples. Twelve of 283 non-promoter loci were identified as super-enhancers and binding sites for transcription factors related to brain tumors were also identified in associated genomic regions. Patients with SHH-MB were then classified using these CSF cfDNA derived transcription regulation features.

CONCLUSIONS

CSF cfDNA from patients with brain tumors was used to determine transcription regulation features, which could reflect the molecular characteristics of brain tumors. Further, these features represent biomarkers with potential to identify patients with tumors. Our study provides a new application for CSF cfDNA and extends its use for investigating tumor-specific gene transcription regulation.