Pretreatment of glioblastoma with bortezomib potentiates natural killer cell cytotoxicity through TRAIL/DR5 mediated apoptosis and prolongs animal survival

Abstract

Background: Natural killer (NK) cells are potential effectors in anti-cancer immunotherapy; however only a subset potently kills cancer cells. Here, we examined whether pretreatment of glioblastoma (GBM) with the proteasome inhibitor, bortezomib (BTZ), might sensitize tumour cells to NK cell lysis by inducing stress antigens recognized by NK-activating receptors. Methods: Combination immunotherapy of NK cells with BTZ was studied in vitro against GBM cells and in a GBM-bearing mouse model. Tumour cells were derived from primary GBMs and NK cells from donors or patients. Flow cytometry was used for viability/cytotoxicity evaluation as well as in vitro and ex vivo phenotyping. We performed a Seahorse assay to assess oxygen consumption rates and mitochondrial function, Luminex ELISA to determine NK cell secretion, protein chemistry and LC–MS/MS to detect BTZ in brain tissue. MRI was used to monitor therapeutic efficacy in mice orthotopically implanted with GBM spheroids. Results: NK cells released IFNγ, perforin and granzyme A cytolytic granules upon recognition of stress-ligand expressing GBM cells, disrupted mitochondrial function and killed 24–46% of cells by apoptosis. Pretreatment with BTZ further increased stress-ligands, induced TRAIL-R2 expression and enhanced GBM lysis to 33–76% through augmented IFNγ release (p < 0.05). Blocking NKG2D, TRAIL and TRAIL-R2 rescued GBM cells treated with BTZ from NK cells, p = 0.01. Adoptively transferred autologous NK-cells persisted in vivo (p < 0.05), diminished tumour proliferation and prolonged survival alone (Log Rank10.19, p = 0.0014, 95%CI 0.252–0.523) or when combined with BTZ (Log Rank5.25, p = 0.0219, 95%CI 0.295–0.408), or either compared to vehicle controls (median 98 vs. 68 days and 80 vs. 68 days, respectively). BTZ crossed the blood–brain barrier, attenuated proteasomal activity in vivo (p < 0.0001; p < 0.01 compared to vehicle control or NK cells only, respectively) and diminished tumour angiogenesis to promote survival compared to vehicle-treated controls (Log Rank6.57, p = 0.0104, 95%CI 0.284–0.424, median 83 vs. 68 days). However, NK ablation with anti-asialo-GM1 abrogated the therapeutic efficacy. Conclusions: NK cells alone or in combination with BTZ inhibit tumour growth, but the scheduling of BTZ in vivo requires further investigation to maximize its contribution to the efficacy of the combination regimen.