Abstract 613: Astrocytes and endothelial colony forming cells (ECFCs) influence the migration and drug response of glioblastoma cells in a 3D culture model

Glioblastoma, a brain cancer with low prognosis, is characterized by rapid infiltration and drug resistance that leads to tumor reappearance. Cell-extrinsic factors, such as the development of a supportive microenvironment and synergistic relationships between normal cells and GBM play an important role in the tumor successful progression. Despite the importance of the microenvironment as a regulator of tumor response, there is a lack of predictive in vitro models that represent the dimensionality and characteristics of the glioma microenvironment. To this end, we developed a 3D in vitro model that recapitulates the composition and physical features of the GBM microenvironment by generating a composite matrix of a hyaluronan supported by collagen. Human astrocytes and endothelial colony forming cells (ECFCs) were incorporated into the matrix and co-cultured with GBM cells to evaluate the effect of normal cells on the GBM 3D cell migration and drug response. Time-lapse microscopy was used to evaluate the 3D migration distance, velocity and persistence of GBM cell lines GBM10, GBM43 and GBAM1 (CD133+) during 18h after drug treatment with temozolomide and a STAT3-inhibitor. Survival of GBM cells was also evaluated 18h and 72 h after drug treatment. All results were compared to standard monolayer culture. Presence of astrocytes and ECFCs increased GBM net migration distance and velocity in the 3D model and in 2D monolayer culture. Additionally, viability of GBM after drug treatment was higher when astrocytes and ECFCs were present in the 3D model compared to only GBM 3D culture and standard monolayer culture. Our results support the importance of the 3D-microenvironment as modulator of tumor response and provide a tunable in vitro platform for primary brain tumor studies aimed to compliment in vivo model observations.

Citation Format: Marisol Herrera-Perez, Melissa Fishel, Sherry Harbin, Jenna Rickus. Astrocytes and endothelial colony forming cells (ECFCs) influence the migration and drug response of glioblastoma cells in a 3D culture model. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 613.

Abstract B37: Proteomic and metabolomic analyses of glioblastoma using mass spectrometry

Glioblastoma (GBM) is the most common and malignant form of primary brain tumor. Average survival rate of GBM is 12-15 months from diagnosis. Due to its highly invasive nature, current treatment options have not been able to improve the survival rate in past 20 years. In order to discover GBM therapeutic targets, omics technologies have been widely used in order to identify potential biomarkers. Identification of cancer biomarkers for diagnosis and prognosis continues to evolve as clinical specimen pipelines and technologies are optimized. This research study focused on investigating proteins, metabolites and lipids biomarkers and patterns of human GBM xenografts employing mass spectrometry. Human tumor cell lines GBM 10 and GBM 43 were injected in the right cerebral hemisphere and flank sites in NOD/SCID mice (n = 10 mice per group). After harvesting brain and flank tumors and control brain from mouse models, protein, metabolites, and lipids of tumor samples were collected through one simple extraction procedure, which may solve sample preparation difficulties in omics studies, developed by Dr. Jiri Adamec. These samples were analyzed by reverse phase high performance liquid chromatography - Fourier transform ion cyclotron resonance mass spectrometry (LC-FTMS). FTMS analyzers have the highest resolving power of all MS instruments, which is ideal for complex mixtures such as GBM tissue. Spectra obtained from the FTMS analysis were analyzed using Student t-tests to detect significant differences in tissue profiles at a level of p = 0.05. Compounds below this threshold were identified through a database using the m/z ratio. Lipidomic analysis indicated the possible up and down-regulation of several lipids classes in GBM tissue connected to metabolic pathways, tumor proliferation and immunodepression, such as glycerophospholipids, both up- and down-regulated (e.g. PC38:4 or PG 35:2, respectively), and down-regulated glycosphingolipids (Galβ1-3GlcNAcβ1-3Galβ1-4Glc-Cer, Galβ1-4GlcNAcβ1-3Galβ1-4Glc-Cer), and triglycerides (TG 55:9, 53:4,52:8). We expect that integrated studies of proteomics, metabolomics, and lipidomics increase the confidence of identification of potential biomarkers. Accompanying these studies is a collaborative effort to improve the effectiveness and efficiency of computational pipelines that are imperative to the analytics, visualization, identification, and interpretation of the omics data. Only by carefully integrating the computational pipelines can we successfully perform the types of integrative studies needed to advance the identification of cancer biomarkers for diagnosis and prognosis, and our integrative studies serve as a case study for our pipeline advancement efforts.

Citation Information: Mol Cancer Ther 2013;12(11 Suppl):B37.

Citation Format: Soo Jung Ha, Gordon Showalter, Jenna Rickus, Shanbao Cai, Haiyan Wang, Wei Michael Liu, Jann N. Sarkaria, John Springer, Jiri Adamec, Karen E. Pollok, Kari L. Clase. Proteomic and metabolomic analyses of glioblastoma using mass spectrometry. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2013 Oct 19-23; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2013;12(11 Suppl):Abstract nr B37.

Preparation of biomolecule gel matrices for electron microscopy

We report a new sample preparation method that allows the direct transmission electron microscopy evaluation of the architectural characteristics of biomolecules entrapped in gel matrices. We demonstrate that this sample preparation technique can be used for the identification and ultrastructural characterization of liposomes, collagen I and collagen III embedded in gel matrices, and has the potential to be useful for transmission electron microscopy (TEM) characterization of other biomolecule-gel matrix systems.

A multi-scale stochastic drug release model for polymer-coated targeted drug delivery systems

A multi-scale mathematical model for drug release of oral targeted drug delivery systems was developed and applied to a commercially available delayed release tablet (Asacol) that delivers 5-aminosalicyclic acid (5-ASA) to the colon. Underlying physical and biochemical principles governing the involved processes (diffusion and dissolution) were employed to develop the mathematical description. Finite element formulation was used to numerically solve the model equations. Molecular dynamics (MD) simulations were used to predict macro-scale transport properties of the drug and the biologic fluid. The effect of pH variability in the gastrointestinal tract environment on the dissolution of the polymeric enteric coating was investigated using the Monte Carlo method. The direct coupling method employed (MD) predicted a sufficiently accurate diffusion coefficient (5.7x10(-6) cm2 s-1) of the drug molecules in reasonable (3 h) computation times. The model was validated using experimental data from in vitro dissolution experiments and provided accurate prediction of the drug release from the delivery system (root mean square error of 5%). The amount of drug entering the systemic circulation, computed from the predicted drug release in varying pH environments in the small bowel, was 15-24%. This range was in good agreement with clinical in vivo data (13-36%) obtained from literature. This research shows that in silico experiments using mechanistic models and stochastic approaches can be used for drug design and optimization and as a decision making tool for physicians.

Impact of coenzyme regeneration on the performance of an enzyme-based optical biosensor: a computational study

A mathematical model of a reagent-less optical sensing scheme composed of an enzymatic reaction coupled to light-controlled photochemical coenzyme regeneration is described. The model is based on previous experimental work describing the regeneration of NADPH from NADP(+) by excited state thionine coupled to the oxidation of isocitrate by isocitrate dehydrogenase. The system is capable of repeated isocitrate measurements without the addition of exogenous coenzyme. The model is simulated using numerical integration to determine the effect of regeneration on the sensor sensitivity, response time and maximum sample throughput rate. Prediction of these effects without a model is difficult due to activation and inhibition of the dehydrogenase by both forms of the coenzyme. The regeneration parameters, including thionine concentration and thionine excitation pattern, are varied to determine optimal sensor conditions to maximize performance. A periodic regeneration approach is found to be superior to a continuous regeneration approach as the former maximizes sensitivity and minimizes response time in most cases. In addition periodic regeneration results in a maximum sample throughput frequency that is achieved at a single optimal thionine level and is independent of the analyte concentration. In contrast the optimal thionine concentration during continuous regeneration varies with the sample analyte concentration. These findings highlight the importance of designing controllable regeneration for dehydrogenase-based optical biosensors.