Acquired temozolomide resistance instructs patterns of glioblastoma cell motility, pro-angiogenic cytokine production, and response to repeated drug exposure in gelatin hydrogels

Introduction:: Glioblastoma (GBM) is the most common and lethal form of brain cancer. Despite an aggressive treatment strategy, including surgical resection followed by radio- and chemotherapy, GBM has one of the highest mortality rates among all human cancers. The leading causes of this devastating prognosis include rapid recurrence and acquired resistance to the standard-of-care chemotherapeutic, Temozolomide (TMZ). Despite extensive research efforts to better understand drug resistance in GBM and to develop new therapies, the prognosis remains abysmal, with a median overall survival of only 15 months. Commonly used GBM models, such as in vivo animal models and conventional in vitro 2D cell cultures, fail to recapitulate complex biological processes associated with acquired drug resistance (e.g., matrix remodeling). In vitro biomaterial models offer an opportunity to evaluate aspects of cell-cell and cell-matrix driven responses within the complex GBM tumor microenvironment (TME). In this project, we utilize gelatin methacrylamide (GelMA) biomaterials to model acquired TMZ resistance in GBM. Particularly, we investigate key features of cancer progression in isogenically matched pairs of TMZ-sensitive vs. TMZ-resistant GBM cell lines within 3D GelMA models. In this work, we show that acquired TMZ resistance leads to significant changes in 3D cell motility and cytokine production, including upregulation of pro-angiogenic factors. Additionally, we show that our model allows for robust evaluation of drug response metrics within 3D biomaterial models of GBM.


Materials and Methods:: GBM cell lines (8MGBA +/- TMZres, 42MGBA +/-TMZres) were provided by the Riggins lab where they were previously characterized. Cells were cultured in DMEM supplemented with 10% FBS (100 μM TMZ continuously applied to 8MGBA-TMZres) and encapsulated (D0) in 5wt%, 50%DoF gelatin methacrylamide (GelMA) hydrogels at 500,000 cells/mL. After 24 hours (D1), hydrogels were treated with varying TMZ concentrations (0 – 1000 μM) and further cultured until D7. Growth-rate inhibition and metabolic health were assessed 3, 5, and 7 days post treatment with alamarBlue HS™ assay. Cytokine production was evaluated at D7 of culture by first collecting conditioned media (CM) from the hydrogels and then analyzing CM using Proteome Profiler Human Angiogenesis Array Kit (R&D Systems). For the invasion experiments, Corning® spheroid microplates were used to form GBM spheroids (5,000 cells/spheroid). After 48 hours, formed spheroids were encapsulated (D0) in GelMA hydrogels and cultured until D3. Bright-field images were taken every 24 hours to monitor spheroid outgrowth.


Results, Conclusions, and Discussions:: To recapitulate acquired temozolomide (TMZ) resistance within three-dimensional models of glioblastoma (GBM), we cultured parental/wild-type (WT) lines, 8MGBA and 42MGBA, and their TMZ-resistant variants, 8MGBA-TMZres and 42MGBA-TMZres, in methacrylamide-functionalized gelatin (GelMA) extracellular matrix hydrogels. Our model allowed for robust detection of drug response and acquired resistance even at low, physiologically relevant TMZ doses. The 42MGBA model showed drastic differences in dose response between WT and TMZres lines with GR50 (50% inhibition of growth rate) value differences over 1000 μM (Fig. 1A). We further validated drug response at single low doses (5, 20, 25, 100 μM) and repeated low doses (5 daily doses of 5 and 20 μM) and observed that WT cells responded well to low doses (5 μM), while TMZres cells still exhibited robust resistance at repeated doses within our GelMA models (Fig. 1B).

Next, we sought to evaluate how acquired TMZ resistance altered key features of GBM progression associated with poor prognosis—cell motility (that leads to invasion) and angiogenesis. Our previously validated spheroid invasion model suggested that acquired TMZ resistance leads to reduced motility (Fig. 2). Furthermore, we probed the composition of conditioned media from 3D cultures of WT vs. TMZres cells for cytokines known to facilitate angiogenesis in cancer. This assay showed differential expression of angiogenic cytokines, namely we saw quite intriguing changes in the 42TMZres model (the more aggressive GBM cells); proteins often associated with poor patient survival, such as TIMP4, uPA, MCP1, etc., were significantly upregulated in the TMZres model (Fig. 3).

Overall, our model allowed us to determine the upper and lower bounds of detectable TMZ response in 3D GelMA hydrogels. These quantified shifts in TMZ-induced GR inhibition are a key benchmark for studies of the efficacy of novel therapeutics. Furthermore, we assess the effect of acquired drug resistance on local invasion and cytokine production. Ongoing work is focused on further evaluation of protein expression in WT vs. TMZres models and incorporation of co-cultures and TME (e.g., perivascular niche) to further probe cell-cell and cell-matrix interactions in GBM models in the context of standard-of-care drug treatment and acquired drug resistance.


Acknowledgements (Optional): :

References (Optional): :