Pediatric and adult high-grade gliomas are the most common primary malignant brain tumors, with poor prognosis due to recurrence and tumor infiltration after therapy. Quiescent cells have been implicated in tumor recurrence and treatment resistance, but their direct visualization and targeting remain challenging, precluding their mechanistic study. Here, we identify a population of malignant cells expressing Prominin-1 in a non-proliferating state in pediatric high-grade glioma patients. Using a genetic tool to visualize and ablate quiescent cells in mouse brain cancer and human cancer organoids, we reveal their localization at both the core and the edge of the tumors, and we demonstrate that quiescent cells are involved in infiltration of brain cancer cells. Finally, we find that Harmine, a DYRK1A/B inhibitor, partially decreases the number of quiescent and infiltrating cancer cells. Our data point to a subpopulation of quiescent cells as partially responsible of tumor invasiveness, one of the major causes of brain cancer morbidity.

Lifespan is determined by complex and tangled mechanisms that are largely unknown. The early postnatal stage has been proposed to play a role in lifespan, but its contribution is still controversial. Here, we show that a short rapamycin treatment during early life can prolong lifespan in Mus musculus and Drosophila melanogaster. Notably, the same treatment at later time points has no effect on lifespan, suggesting that a specific time window is involved in lifespan regulation. We also find that sulfotransferases are upregulated during early rapamycin treatment both in newborn mice and in Drosophila larvae, and transient dST1 overexpression in Drosophila larvae extends lifespan. Our findings unveil a novel link between early-life treatments and long-term effects on lifespan.

Pediatric and adult high-grade gliomas are the most common malignant brain tumors, with poor prognosis due to recurrence and tumor infiltration after surgical removal and chemotherapy. Quiescent/slow cycling stem cells have been proposed to be one of the main players of tumor relapse but their involvement in in the infiltration remain unclear. Despite they have been described in mouse models or after transcriptional profiling of human tumor samples, their direct visualization, targeting and ablation remains a challenge. Here, we identified a population of malignant cells expressing Prominin-1 in a non-proliferating state in pediatric high-grade glioma patients. We next used a fluorescent cell cycle sensor to visualize quiescence tumor cells in mouse brain cancer and human cancer organoids. In particular, we characterized them within the tumor revealing the invasiveness capacity of slow cycling tumor cells. Furthermore, we generated a new system to specifically trace and ablate such cells. Indeed, lineage tracing experiments allowed to trace quiescent Prom1 progeny in the tumors after temozolomide treatment. In addition, the selective ablation of quiescent Prom1+ cells in mouse brain cancer reduced tumor infiltration and improved survival. Furthermore, time-lapse experiments showed that slow cycling cells are also able to infiltrate co-cultured human brain cancer organoids. Finally, using our new cancer organoid- and GBM spheroids-based models we identified a drug acting on quiescent cells leading to a reduction in cell invasion. Overall, our data show that quiescence/slow cycling cells are key driver of tumor invasiveness, the major malignant feature of high-grade brain cancer.

BACKGROUND High-grade gliomas are the most common malignant brain tumors, with poor prognosis due to recurrence and tumor infiltration after surgical removal and chemotherapy. Quiescent/slow cycling stem cells have been proposed to be one of the main players of tumor relapse but their involvement in in the infiltration remain unclear. Despite they have been described in mouse models or after transcriptional profiling of human tumor samples, their direct visualization, targeting and ablation remains a challenge. MATERIAL AND METHODS Tumors were induced over-expressing oncogenic forms of MET and p53 in the subventricular zone (SVZ) of P2 mouse brain as well as human forebrain organoids. The co-expression with specific cell cycle sensors as well as lineage specific CreERT2 under control of stem cells promoters allowed to visualize and target glioma stem cells. RESULTS Here, we used a fluorescent cell cycle sensor to visualize quiescent tumor cells in mouse brain cancer and human cancer organoids. In particular, we characterized them within the tumor revealing the invasiveness capacity of slow cycling tumor cells. Furthermore, we generated a new system to specifically trace and ablate such cells. Indeed, lineage tracing experiments allowed to trace quiescent Prom1 progeny in the tumors after temozolomide treatment. In addition, the selective ablation of qProm1 in mouse brain cancer reduced tumor infiltration. Finally, time-lapse experiments showed that slow cycling cells are also able to infiltrate co-cultured human brain cancer organoids. CONCLUSION Overall, our data show that quiescent/slow cycling cells are key driver of tumor invasiveness, the major malignant feature of high-grade brain cancer.