The paracaspase MALT1 controls cholesterol homeostasis in glioblastoma stem-like cells through lysosome proteome shaping.

Glioblastoma stem-like cells (GSCs) compose a tumor-initiating and -propagating population remarkably vulnerable to variation in the stability and integrity of the lysosomal compartment. Previous work has shown that the expression and activity of the paracaspase MALT1 control GSC viability via lysosome abundance. However, the underlying mechanisms remain elusive. By combining RNA sequencing (RNA-seq) with proteome-wide label-free quantification, we now report that MALT1 repression in patient-derived GSCs alters the homeostasis of cholesterol, which accumulates in late endosomes (LEs)-lysosomes. This failure in cholesterol supply culminates in cell death and autophagy defects, which can be partially reverted by providing exogenous membrane-permeable cholesterol to GSCs. From a molecular standpoint, a targeted lysosome proteome analysis unraveled that Niemann-Pick type C (NPC) lysosomal cholesterol transporters are diluted when MALT1 is impaired. Accordingly, we found that NPC1/2 inhibition and silencing partially mirror MALT1 loss-of-function phenotypes. This supports the notion that GSC fitness relies on lysosomal cholesterol homeostasis.

Neuropilin-1 modulates the 3D invasive properties of glioblastoma stem-like cells.

Glioblastoma multiforme (GBM) is a rare, yet devastating, primary brain tumor in adults. Current treatments remain generally ineffective and GBM almost invariably recurs, resulting in median survival of 15 months. This high malignancy sources notably from the resilience and invasive capabilities of tumor cells. Within GBM, exists a population of self-sustaining transformed cells with stem-like properties (GSCs), which are thought to be responsible for tumor initiation, growth, and invasion, as well as recurrence. In the tumor microenvironment, GSCs might be found in the vicinity of brain endothelial cells, which provide a protective habitat. Likewise, these resistant, quiescent GSCs may accumulate in hypoxic zones, away from the perivascular niche, or travel towards the healthy brain parenchyma, by eminently co-opting neuro-vascular tracks. Herein, we established an ex vivo model to explore GSC invasive behavior. We found that patient-derived cells massively invade the collagen matrix. In addition, we described that the glycoprotein Neuropilin-1 (NRP1) contributes to GSC spreading and invasion. Indeed, both RNA interference-mediated silencing and CRISPR-mediated gene editing deletion of NRP1 strongly impaired the 3D invasive properties of patient-derived GSCs and their close localization to the brain blood vessels. Of note, other typical features of GSCs, such as expansion and self-renewal were maintained. From a mechanistic standpoint, this biological effect might rely on the expression of the ß3 subunit integrin cell-extracellular matrix adhesive receptor. Our data, therefore, propose a reliable approach to explore invasive properties of patient glioma cells ex vivo and identify NRP1 as a mediator in this malignant process.