Back

Spatial vascular/BTB remodeling and malignant-state plasticity in glioblastoma

Zheng, L.; Gan, L.

2026-08-22 cancer biology
10.64898/2026.08.18.745357 bioRxiv
Show abstract

Background: Glioblastoma (GBM) contains spatially heterogeneous malignant and vascular states, but blood-tumor barrier (BTB) remodeling is often described as a binary functional phenotype. We asked whether anatomically distinct GBM compartments contain separable vascular programs that coexist with malignant-state plasticity. Methods: We performed donor-aware cross-sectional analyses of 38 histopathology-annotated spatial transcriptomic sections from 6 donors and a separately analyzed endothelial single-nucleus layer from the same GBM-Space atlas. Complementary external datasets tested patient-paired regional remodeling, anatomical replication, cross-technology source localization, and malignant-state architecture. Results: THSD1-FLT4 Recognition increased from leading edge to infiltrative tumor (median adjusted effect +0.02875; 4/4 donors positive). Priming increased across this boundary (+0.14814; 3/4) but decreased from infiltrative to cellular tumor (-0.16409; 0/4), whereas Gate remodeling increased from infiltrative to cellular tumor (+0.21296; 4/4). Remodeled endothelium showed higher PLVAP detection (+0.26409; 12/12 donors) and PLVAP pseudobulk expression (+1.61784 log1pCPM; 11/12), with lower MFSD2A pseudobulk expression (-0.71448; 10/12 negative). External cohorts supported regional vascular/BTB remodeling, while GSE131928 supported broad malignant-state architecture and an exploratory within-tumor pseudotemporal continuum. Conclusions: GBM contains spatially partitioned vascular/BTB-associated programs alongside malignant-state plasticity. Recognition-Priming-Gate is a cross-sectional discovery framework, not a validated temporal cascade, and the data do not establish BTB permeability, causal tumor-vascular signaling, or therapeutic-delivery benefit.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.