Abstract
Malignant brain tumors, including primary gliomas and secondary brain metastases (BrMs), present significant clinical challenges due to their high mortality and diagnostic complexity. While genomic studies have provided insights, the low correlation between transcriptomic and proteomic levels necessitates direct protein-level analysis. This study utilizes high-sensitivity mass spectrometry to map the proteomic landscape of 29 clinical samples, identifying over 8,000 protein groups. Our findings reveal distinct molecular programs—proliferation in BrMs versus invasion in gliomas—and establish a 5-protein biomarker panel capable of distinguishing these tumor types with 90% accuracy.
Introduction
Malignant brain tumors are among the most aggressive cancers. Distinguishing between high-grade gliomas and brain metastases (BrMs) preoperatively is critical for treatment planning but remains difficult using conventional imaging. Genomic and transcriptomic sequencing have advanced our understanding, yet a significant "disconnect" exists between RNA levels and functional protein expression, with correlations often as low as 0.54. As proteins are the primary functional units of the cell, proteomic profiling is essential to capture the true heterogeneity of these malignancies and identify actionable biomarkers.
Methodology
The study analyzed a cohort of 29 clinical tissue samples obtained from Ruijin Hospital, comprising 13 gliomas (WHO Grade II-IV) and 14 brain metastases originating from lung, breast, ovarian, and colorectal cancers. The researchers employed label-free quantitative (LFQ) mass spectrometry-based proteomics. Data acquisition was performed on a timsTOF mass spectrometer using Parallel Accumulation Serial Fragmentation (PASEF) technology, which ensures high-speed and high-sensitivity protein identification.
To ensure data reliability, commercial HeLa cell digests were interleaved every 10 samples for instrumental quality control. Bioinformatics analysis involved unsupervised hierarchical clustering and Principal Component Analysis (PCA) to identify proteomic patterns. A diagnostic model was developed using Partial Least Squares Discriminant Analysis (PLS-DA) and evaluated via Receiver Operating Characteristic (ROC) curves.
Results and Discussion
Deep Proteomic Profiling
The analysis yielded a comprehensive dataset of 8,370 protein groups and 116,526 peptides. Among these, 352 proteins were identified as significantly differentially expressed between BrMs and gliomas, highlighting the profound molecular differences between primary and secondary brain malignancies.
Colonization vs. Invasion Programs
A key finding of this study is the functional divergence between tumor types. Brain metastases were characterized by pathways related to cell proliferation and immunomodulation, suggesting a "colonization" program focused on establishing and expanding the tumor within the brain microenvironment. In contrast, gliomas were dominated by cytoskeleton remodeling and invasion processes, reflecting their characteristic diffuse infiltration into surrounding brain tissue. Notably, BrMs from different primary organs shared a conserved proteomic signature, indicating that the brain microenvironment exerts a strong selective pressure that converges metastatic cells toward a common survival program.
Ferroptosis and IDH1 Mutation
In high-grade gliomas, the researchers observed that IDH1 mutations were linked to the upregulation of ferroptosis—a form of iron-dependent programmed cell death. This upregulation correlates with improved patient prognosis, suggesting that ferroptosis pathways may represent a significant therapeutic vulnerability in IDH1-mutant tumors.
Diagnostic Biomarker Development
The study identified a 5-protein biomarker panel consisting of KRT8, KRT19, KRT7, TACSTD2, and CDH1. These proteins are primarily epithelial markers; since the brain lacks native epithelial tissue, their presence serves as a highly specific "red flag" for metastasis. This panel achieved an Area Under the Curve (AUC) of 0.973 and demonstrated 90% accuracy in distinguishing lung-origin metastases from Grade IV gliomas.
Conclusion
This research presents the first systematic proteomic comparison between primary gliomas and brain metastases. By mapping the proteomic landscape, the study identifies distinct molecular programs and a robust biomarker panel for clinical diagnosis. While the sample size is relatively small (n=29), these findings provide a critical framework for future large-scale validation and the development of proteome-informed therapeutic strategies to combat these deadly cancers.
Summary updated