Abstract The recurrence of glioblastoma (GBM) remains unavoidable. Neural stem cells (NSCs) in the subventricular zone (SVZ) are considered a treatment target due to their potential role as the cellular origin of GBM. However, it is not yet confirmed whether GBM recurrence also originates from these cells. We first conducted deep sequencing on paired patient samples, including primary tumors (pT), recurrent tumors (rT), and SVZ tissues. The sequencing results revealed that in 5 out of 9 patients, the rT evolved independently from the SVZ rather than directly from the pT. Based on this, we developed a mouse model to simulate the GBM recurrence process by transforming SVZ NSCs into GFP-labeled cells carrying cancer mutations via electroporation. Following the resection of the tdTomato-labeled pT, we observed that 57.1% of the rT consists of a mixture of GFP- and tdTomato-positive tumor cells. To study the behavior of NSCs, we improved the existing mouse model by removing cortical tissue to simulate the complete resection of the pT. Four weeks post-surgical resection, 64.7% of tumors reconstructed in the resection cavity (RC), driven by the migration of OPC lineage cells. Time-course RNA-seq results indicated that the CXCL12-CXCR4 pathway might be responsible for driving the migration of SVZ NSCs to the RC. Further deconvolution and fluorescence staining analysis confirmed that CXCL12 secreted by endothelial cells around the RC could drive CXCR4-positive NSCs to migrate to the area. Additionally, the CXCR4 antagonist AMD3100 significantly reduced tumor recurrence in mice and extended survival. RNA-seq results from both mouse and human paired tumor samples confirmed that the CXCL12-CXCR4 pathway was enriched in locally recurrent tumor specimens and was correlated with shorter patient survival. In summary, our study provides evidence that GBM recurrence originates from SVZ NSCs. Moreover, we identified the CXCL12-CXCR4 pathway as a potential therapeutic target for GBM recurrence.
Read full abstract