(411e) Graduate Student Award Session: The Combined Effect of Matrix Microenvironment and Hypoxia on the Activity of Glioblastoma Stem Cells

Introduction Glioblastoma (GBM) is the most common and lethal form of brain cancer. GBM is known to diffusely infiltrate throughout the brain tissue, which contributes to poor patient prognosis. The complex and heterogeneous cell population, structural and nutrient gradients along the tumor margin, complicates the study of GBM. Despite the standard-of-care treatment, which includes surgical debulking tumor mass as much as safely for the patient normal brain function followed by chemo- and/or radiotherapy, the short 14.6 months median survival rate, low long term (5+ year < 5%) survival rate and the fast and high recurrence rate (~6.9 months after standard treatment) suggesting the little-to-no effect of current therapy. Currently, there are limited tools approaches to rigorously assess the relationship between tissue microenvironment and the growth, invasion, and therapeutic outcomes. Several subsets of cells within the heterogenous GBM tumor niche has been considered to contribute to the poor prognosis. GBM stem cells (GSCs) also known as GBM-initiating cells are thought to be therapeutic resistant and hence are important for curing and avoiding recurrence of GBM. Tumor-associated-macrophages/microglias are immune cells recruited into tumor sites and the polarization (proinflammatory M1 and anti- inflammatory M2) is recently thought to be important to tumor growth and related to GBM invasive phenotype. This study focuses on developing an engineered hydrogel platform to mimic the brain tumor environment to investigate how matrix composition, metabolic environment and drug treatments alter the cellular heterogeneity of patient-derived GBM specimens.

Materials and Methods Hydrogels are generated from methacrylamide-functionalized gelatin (Gel-MA) and methacrylated hyaluronic acid (HAMA). Patient-derived xenograft (PDX) GBM cells (GBM6, GBM10, GBM12, GBM15, GBM26, GBM39) were obtained from our collaborator (Dr. Jann Sarkaria, Mayo Clinic). These specimens maintain patient-specific morphological and molecular characteristics and span a range of invasive and therapy resitant phenotypes. Hydrogels were maintained for 4 wt% and made for with or without 15 w/w% matrix-immobilized HA and photo-initiator (PI) was adjusted (-HAMA: 0.1%, +HAMA:0.02%) for maintaining elastic modulus. Pre-polymerized solutions were then mixed with cell suspension homogenously and polymerized using UV lamp (7.1 mW/mm², 30 sec). Hydrogels were then put into culture for either normoxia (20% O₂) or hypoxia (1% O₂) and culture media were made with different treatments: +Dimethyl sulfoxide (0.1 vol%, control), +Erlotinib (Erl, receptor tyrosine kinase inhibitor, 10 μM) or Temozolomide (TMZ, standard oral chemotherapy drug, 30 μM). Therapeutic outcomes were determined by overall metabolic reactivity MTT assay (ThermoFisher). The heterogeneity of the GBM cell populations were determined using flow cytometer before each experiment and after degradation of hydrogels at timepoints throughout culture. Hydrogels were obtained after culture and put into collagenase type IV (100 unit/ 500uL) and hyaluronidase (100 unit/ 500uL) containing PBS for 30 minutes for fully degraded hydrogel. Four cell populations were stained after degradation: GSCs (CD133), TAMs (CD68), astrocytes (GFAP) and endothelial cells (ECs, CD31).

Results and Discussion Hyrogel matrix microenvironment significantly influences the kinetics of glioblastoma stem cells within the PDX specimens. Interestingly, the subfraction of GSCs within the invasive GBM6 PDX popualtion showed a significant increase under hypoxia (1% O₂) culture for up to 7 days compare to groups cultured in normoxia (20% O₂) condition. Specimens cultured in the presence of matrix-bound HA showed a significant higher fraction of GSC⁺ cells, suggesting that GBM GSC% is influenced by both matrix composition as well as the metabolic environment. To study the complex heterogenous cell population between different PDX cells, a screening experiment of GSC%(CD133⁺) and TAM%(CD68⁺) was performed on 4 PDX lines: GBM6, GBM10, GBM12 and GBM39, each having different invasive potential (lo:0; 7:hi), Erlotinib response (0:lo, 100:hi) and molecular subtype (M: mesenchymal, C:classical). These speimens showed a wide range of initial and final cell compositions, suggesting profiling the heterogeneity of PDX specimens in biomimietic hydrogels may yield new, therapeutically actionable data. Ongoing work is looking at expansion of cell markers into four interesting cell groups (GSCs, TAMs, astrocytes and ECs) as well as in the presence of two therapeutic molecules (erlotinib, Temozolomide) to cell how the heterogeneous cell population changes as a response to microenvironmental cues and therapies.

Conclusion This study provides a rigorous way to study cell population, matrix influence and therapeutic outcome in a defined and controlled biomaterial in vitro platform. Data showed a huge heterogeneity existing in GBM and future work will examine more.