Back

Combined blockade of VEGF, Angiopoietin-2, and PD1 reprograms glioblastoma endothelial cells into quasi-antigen-presenting cells

Amoozgar, Z.; Ren, J.; Wang, N.; Andersson, P.; Ferraro, G.; Rajan, S.; Lei, P.; Subudhi, S.; Kawaguchi, K.; Tay, R. E.; Gomes Dos Santos, I. L.; Huang, P.; Kim, H.-J.; Fukumura, D.; Jain, R.

2022-09-04 cancer biology
10.1101/2022.09.03.506476 bioRxiv
Show abstract

Glioblastoma (GBM) remains a highly aggressive and uniformly fatal primary tumor, which resists cytotoxic, targeted, antiangiogenic, and immune therapies, even when used in combination. Here we report that tumor endothelial cell dysfunction confers resistance to immunotherapy in preclinical GBM models. Anti-VEGF-therapy-induced vascular normalization is insufficient to fully restore the endothelial cell function. Strikingly, concomitant blockade of Ang2, VEGF, and PD1 reprograms dysfunctional endothelial cells to quasi-antigen presenting cells and upregulates receptors required for cytotoxic T lymphocyte entry into the tumor. Blocking VEGF, Ang2, and PD1 induces durable anti-tumor T cell responses. Upregulation of the transcription factor T-bet is both necessary and sufficient for generating resident memory T cells elicited by this combination therapy. In summary, our study reveals the role of Ang2 in resistance to PD1-blockade and provides a compelling rationale for clinical evaluation of blocking Ang2 along with VEGF and PD1 in GBM patients. Statement of SignificanceOur study is the first to demonstrate Ang2 as a resistance pathway for both VEGF and PD1 in GBM. Concomitant blockade of Ang2 reprograms endothelial cells to recruit, activate and retain CD8 T cells, overcomes resistance to VEGF and PD1, and imparts T cell memory formation via T-bet in GBM.

Matching journals

The top 7 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.