Non-neoplastic cellular binding of PSMA-targeted probes limits radiation dose delivery and therapeutic efficacy in glioma

Abstract Background Prostate-specific membrane antigen (PSMA)-targeted theranostics has transformed prostate cancer management and is being explored in other PSMA-expressing malignancies, including glioma. However, a striking discordance exists between intense PSMA positron emission tomography (PET) signals and poor therapeutic response in glioma, challenging the “treat-what-you-see” paradigm. This study mechanistically resolves this discrepancy by defining the cellular origin of PSMA PET signals and quantifying its impact on radiation dose delivery for PSMA-targeted radioligand therapy (RLT). Methods Forty-nine patients with intracranial lesions underwent preoperative 68 Ga-PSMA-11 PET/magnetic resonance imaging (MRI). Neuronavigation-guided tissues from regions with varying uptake were analyzed by single-cell RNA sequencing, immunohistochemistry, and immunofluorescence. Spatial distances between PSMA-positive cells and tumor nuclei were quantified using artificial intelligence (AI)-driven image analysis. Pathology-informed dosimetry simulations estimated radiation dose delivered to tumor nuclei by 177 Lu- and 225 Ac-PSMA-617 using the dose-point-kernel method, with tumor cell survival modeled by the linear-quadratic approach. Results PSMA expression in glioma was predominantly confined to non-neoplastic oligodendrocytes and vascular endothelial cells, with minimal tumor cell expression, a pattern conserved in normal brain. PSMA-positive oligodendrocytes and endothelial cells were located at mean distances of 40.1 μm and 51.9 μm from tumor nuclei, respectively. This off-target binding reduced effective radiation dose to glioma cell nuclei to only 21.3% ( 177 Lu) and 21.9% ( 225 Ac) of that achievable with direct tumor targeting, corresponding to a modeled 5.2- to 6.9-fold increase in tumor cell survival probability. Conclusions In glioma, PSMA-targeted probes predominantly bind to non-neoplastic cells, creating a fundamental dosimetric barrier that explains therapeutic failure despite high PET uptake. These findings necessitate integrating cellular target validation and preemptive dosimetry into theranostic agent development. Trial registration This study was registered at the Chinese Clinical Trial Registry (ChiCTR) under the identifier No. ChiCTR2300076481 on October 10, 2023. The full record is publicly accessible at http://www.chictr.org.cn .