Hypoxic Environment and Paired Hierarchical 3D and 2D Models of Pediatric H3.3-Mutated Gliomas Recreate the Patient Tumor Complexity

Abstract

Background: Pediatric high-grade gliomas (pHGGs) are facing a very dismal prognosis and representative pre-clinical models are needed for new treatment strategies. Here, we examined the relevance of collecting functional, genomic, and metabolomics data to validate patient-derived models in a hypoxic microenvironment. Methods: From our biobank of pediatric brain tumor-derived models, we selected 11 pHGGs driven by the histone H3.3K28M mutation. We compared the features of four patient tumors to their paired cell lines and mouse xenografts using NGS (next generation sequencing), aCGH (array comparative genomic hybridization), RNA sequencing, WES (whole exome sequencing), immunocytochemistry, and HRMAS (high resolution magic angle spinning) spectroscopy. We developed a multicellular in vitro model of cell migration to mimic the brain hypoxic microenvironment. The live cell technology Incucyte© was used to assess drug responsiveness in variable oxygen conditions. Results: The concurrent 2D and 3D cultures generated from the same tumor sample exhibited divergent but complementary features, recreating the patient intra-tumor complexity. Genomic and metabolomic data described the metabolic changes during pHGG progression and supported hypoxia as an important key to preserve the tumor metabolism in vitro and cell dissemination present in patients. The neurosphere features preserved tumor development and sensitivity to treatment. Conclusion: We proposed a novel multistep work for the development and validation of patient-derived models, considering the immature and differentiated content and the tumor microenvironment of pHGGs.

Hypoxia Inducible Factors’ Signaling in Pediatric High-Grade Gliomas: Role, Modelization and Innovative Targeted Approaches

“…This was demonstrated in patients' tumors, as well as in pHGG-derived cell lines [24,25]. Glutaminolysis and serinolysis were associated significantly to histone H3F3A mutation and phospholipid metabolism to a poorer outcome [25,59,60]. These metabolic alterations represent an onco-requisite factor of pHGG progression and probably tumorigenesis.…”

“…Glutamine addiction is usually associated to glycolysis impairment and Akt induction in those HGG cells [6]. This was demonstrated in patients' tumors, as well as in pHGG-derived cell lines [24,25]. Glutaminolysis and serinolysis were associated significantly to histone H3F3A mutation and phospholipid metabolism to a poorer outcome [25,59,60].…”

“…pHGGs, like most other solid tumors, are highly glycolytic cancers producing large amounts of lactate as a metabolic by-product, increasing resistance to radiotherapy. All these genes have a common objective: adapt the pHGG cells to the newly encountered hypoxia conditions and maintain their proliferation and homeostasis [24,25,56,57].…”

“…For cell homeostasis, in addition to the glycolytic switch, the mitochondrial oxidative phosphorylation system (OxPHOS) usually appears when Warburg effect decreases during tumor reoxygenation, and promotes a smaller pHGG cell viability [56] (Figure 3). However, HIF-1α and HIF-2α led the cells to a final complete metabolic reprogramming with the activation of glutaminolysis, serinolysis and/or phospholipid metabolism [24,25]. Glutamine addiction is usually associated to glycolysis impairment and Akt induction in those HGG cells [6].…”

“…Today, the most accurate in vitro model for pHGG is a patient-derived cell line initiated from tumor biopsy at diagnosis or at relapse time [13,24,25,69]. Therefore, a worldwide effort was dedicated in the last decade to develop such models to mimic accurately the intrinsic patient tumor behavior.…”