CD276 (B7-H3) as a Companion Diagnostic Biomarker for Glioblastoma: Multi-Platform Validation and Therapeutic Implications

Abstract Glioblastoma (GBM) remains the most lethal primary brain tumor, with median survival of 14-16 months despite aggressive multimodal therapy[1,2]. The failure of PD-1/PD-L1 checkpoint inhibitors in GBM (CheckMate-143)[3] has highlighted the need for alternative immunotherapeutic targets and companion diagnostics. CD276 (B7-H3) has emerged as a promising target, with multiple anti-B7-H3 therapies in clinical development including monoclonal antibodies, antibody-drug conjugates (ADCs), and CAR-T cells[4-6]. However, no validated companion diagnostic exists to stratify patients for these therapies. Here we present comprehensive validation of CD276 as a prognostic biomarker in GBM across multiple independent platforms. Using discovery analysis in TCGA (n=154) and independent validation in CPTAC proteomics (n=99)[7], we demonstrate that CD276-high expression is associated with significantly shorter survival (Δ=3.5-4.0 months, p=0.003-0.013). RNA expression correlates strongly with protein (r=0.75, p<0.0001), enabling flexible companion diagnostic development. Single-cell analysis of 338,564 cells from 110 patients[8] reveals CD276 is highest on tumor vasculature, supporting ADC targeting strategies that bypass the blood-brain barrier. Critically, we identify a novel therapeutic vulnerability: CD276-high tumors exhibit significantly reduced expression of ATP-binding cassette (ABC) drug efflux transporters ABCG2 (0.61-fold, p=0.0002) and ABCB1 (0.64-fold, p=0.005)[9]. Since these transporters actively efflux common ADC payloads including MMAE and DXd, their reduced expression suggests CD276-high tumors may be paradoxically more vulnerable to cytotoxic payloads despite their aggressive phenotype. This inverse relationship between target expression and drug efflux capacity provides mechanistic rationale for prioritizing CD276-high patients for ADC therapy. CD276 significantly outperforms PD-L1 across all metrics, consistent with PD-L1’s clinical failure in GBM. Competing Interest Statement The authors have declared no competing interest.