Poster T7 - Overexpression of EGFR in glioblastoma cells promotes a methuosis phenotype

Introduction: Glioblastoma is the deadliest and most common form of brain cancer in adults. Even with standard of care resection, radiation, and chemotherapy, median survival is just 15 months. Our efforts focus on promoting a non-apoptotic form of cell death termed methuosis to combat glioblastoma chemo- and radio-resistance. Methuosis is characterized by dysregulated macropinocytosis leading to catastrophic cell vacuolization and rupture.[1] Macropinocytosis is used by RAS-driven cancer cells to engulf extracellular material containing nutrients in response to metabolic stress. Though RAS mutations are rare in glioblastoma, the ectopic expression of mutant HRAS drives a vacuolized phenotype consistent with the methuotic phenotype.[2] Epidermal growth factor receptor (EGFR) is a primary upstream effector of RAS and is the most commonly dysregulated protein in glioblastoma – EGFR is amplified in over half of glioblastoma tumors and mutated in over half of the EGFR-amplified tumors.[3,4] We hypothesize that EGFR can drive enough RAS activity when amplified in glioblastoma cells to beget the same catastrophic vacuolization (Fig. 1A-B) and that glioblastoma cells may have routes to adapt to this stress that can be therapeutically targeted.

Materials and

Methods: To study the effects of EGFR amplification in glioblastoma cells, human-derived U251 cells were engineered with doxycycline-inducible vectors encoding luciferase (control), HRAS G12V, wild type EGFR, and a glioblastoma-relevant EGFR point mutant (EGFR G598V). U251 cells are a common cell line in GBM research and harbor mutations to PTEN, p53, and NF1. To maximize wild type or point-mutant EGFR expression, transduced cells were treated with doxycycline and flow-sorted to retain the top 15% of EGFR-expressing cells. To test for a vacuolized phenotype consistent with methuosis, sorted cells were simultaneously induced with doxycycline and treated with super-saturating (16 nM) EGF for up to three days. HRAS G12V expression was induced with three days of doxycycline to produce vacuole accumulation consistent with prior findings.[2] For systematic quantification of cell phenotypes, a novel imaging and image analysis pipeline was developed to analyze the degree of vacuolization and correlate cell phenotype with protein expression changes on a cell-by-cell basis. The pipeline implements a machine learning algorithm in Ilastik to segment bright-field vacuole images and CellMask and NucBlue fluorescent stains to eliminate false positives, enabling quantification of sizes and numbers of vacuoles per cell. Using image registration techniques and object identification in Cell Profiler, immunofluorescence signals are overlaid with live-cell vacuole images to correlate protein expression and localization with vacuolization. Proteins analyzed in this pipeline include EGFR, c-Jun, and ERK.

Results, Conclusions, and Discussions: Results and

Discussion: In response to 16 nM EGF, obvious vacuolization occurred in U251 cells expressing wild type EGFR or EGFR G598V (Fig. 1C). Vacuoles were apparent as early as one day after treatment with EGF and grew in size and number by two days. Vacuoles were similar in appearance to those induced with the small-molecule PIKfyve inhibitor MOMIPP, a known driver of methuosis.[1] A wide range of EGF concentrations, from 0.016 to 160 nM, surprisingly produced similar degrees of vacuolization. Parallel immunoblotting revealed that ectopic EGFR expression resulted in a seven-fold increase in phosphorylated EGFR (Y1068) compared to luciferase control and that total EGFR expression was within the observable pathological range for glioblastoma cells. By three days, there was some evidence of cellular adaptation to macropinocytic stress and an absence of cell death as measured by cell permeability to propidium iodide. By comparison, U251 cells with ectopic RAS G12V expression continued to show vacuolization by three days and obvious signs of cell death. EGFR-driven vacuolization was inhibited by the EGFR kinase inhibitor gefitinib, confirming the EGFR-specific nature of the phenotype induced by EGF treatment. However, preliminary characterization of signals for phosphorylated c-Jun and ERK do not suggest that canonical EGFR downstream effector signaling is responsible for vacuole accumulation.

Conclusions: EGFR over-expression at levels encountered in glioblastoma tumor cells drives a surprising hyper-vacuolized methuotic phenotype in U251 cells reminiscent of phenotypes observed in response to RAS mutation or PIKfyve inhibition. This phenotype is not correlated with alterations in conventional EGFR/RAS effectors including ERK or c-Jun, motivating a systems-level analysis of the signaling regulatory pathways that connect EGFR to methuosis.

Acknowledgements (Optional): Sources: [1]Maltese and Overmeyer, Am J Pathol. 2014 [2]Overmeyer, et al. Mol Cancer Res. 2008 [3]Zhou, et al, Biol Pharm Bull. 2024 [4]Brennan, et al, Cell 2013

Funding: T32 CA 009109, NCI U54 CA274499