Multi-omic Analysis Reveals Distinct Molecular Signatures of Brain Radiation Necrosis

Introduction: Cranial radiotherapy for neoplasms and cerebrovascular lesions often leads to radiation-induced complications included memory loss, impaired cognition, and seizures. Imaging and histopathological studies show that these long-term effects result from structural changes and tissue necrosis in the brain post-radiation, yet the exact pathophysiological mechanisms remain poorly understood. This study employs multi-omic approaches to identify distinct molecular signatures of radiation necrosis (RN) and elucidate the underlying biological processes.

Methods: Samples were collected from patients with glioma, brain metastasis, and AVM undergoing surgery post-radiotherapy. RN tissues were resected, along with paired normal brain parenchyma. Using next-generation RNA-sequencing and mass spectrometry, tissues were analyzed for differentially expressed genes (DEGs) and differentially abundant proteins (DAPs), respectively. Network analysis was then conducted, and the rich-club coefficient was calculated to assess network topology and identify core nodes. Finally, patient MRIs were processed to quantify edema volume (normalized-to-lesion volume).

Results: Principal-component-analysis revealed clear clustering between normal parenchyma and RN tissues, using both transcriptomic and proteomic data, confirming the presence of distinct disease signatures. Further analysis identified 393 upregulated and 116 downregulated DEGs along with 75 upregulated and 63 downregulated DAPs. Significant overlapping upregulated DEGs and DAPs included C3, ITGB3, and IGKC, among others. Functional annotation of these findings revealed positive regulation of leukocyte-mediated immunity, regulation of lipid metabolism, and 'de novo' protein folding. We then constructed an interactive network of the DEGs and DAPs, using rich-club coefficients to identify network cores. Core nodes for the DEG network included PTEN, MMP9, and TGFB1, while the DAP network included APP, GSN, and ICAM1 among others. The expression of these network cores demonstrated a strong correlation with normalized brain edema, with an R-squared value of 0.8.

Conclusion : Multi-omic analysis provides novel insights into the molecular characteristics of brain RN and highlight potential biomarkers and therapeutic targets for improved diagnosis and treatment.