(H-282) Recapitulating Microglia Biophysical Dynamics in the Tissue-Engineered Glioblastoma Models to Study Tumor-Immune Crosstalk

Introduction:: Glioblastoma (GBM) is the most common and malignment form of primary brain cancer. While microglia are composed of one-third of total GBM tumor mass, the incorporation of microglia into cancer models are still relatively scarce. Microglia, resident macrophages in the brain, are responsible for primary immune defense for the foreign pathogens and tissue damage. The activation of microglia is a dynamic process regulated by the interplay of cellular composition, biochemical signaling and biomechanical cues of brain. In the context of brain cancers, infiltration of microglia into GBM tumor is one of the hallmarks of GBM progression. Nevertheless, the role of microglia during the GBM tumor development is still largely unknown despite several phenotypes of microglia have been recently identified. The challenges to decipher crosstalk between microglia and GBM cells are due in part to the lack of effective in vitro engineered models to recapitulate the microglia dynamics in GBM microenvironment because systematic analysis of this tumor-immune interaction is difficult to evaluate in vivo. Therefore, there is a critical need for developing experimental platforms to rigorously interrogate interactions between GBM cells and microglia, as well as to identify factors associated with this microglia-MG crosstalk that may alter drug response and therapeutic resistance. To this end, we apply integrative bioengineering approaches to evaluate microglia mechanobiological behaviors, such as 3D invasion and ECM remodeling in the microphysiological models of glioblastoma. Furthermore, we also explore the lymphangiogenic function of microglia in our tissue-engineered hydrogel models to correlate with microglia biophysical properties.


Materials and Methods:: To recapitulate GBM extracellular matrix, both Gelatin-based and type I collagen biomaterials were employed in this study. Gelatin was functionalized its backbone with methacrylamide to generate a library of Gelatin methacrylamide (GelMA) using the protocol previously reported by our group. The as-synthesized GelMA was lyophilized and the degree of functionalization (DoF) of GelMA was measured using 1H NMR. To prepare GelMA hydrogels, 5 wt% GelMA was dissolved in a PBS buffer solution containing 10% photoinitiator and polymerized under UV irradiation. To study 3D invasion of microglia, a three-parallel channels microfluidic device (AIM Bio) was applied. 3 mg/mL type I collagen was introduced to central channel while one of the lateral channels was seed with human microglial cell line (HMC3). To evaluate the microglia invasion into 3D matrix, LPS (pro-M1 cytokine), IL-4 (pro-M2 cytokine) and a conditioned medium from U87MG GBM cells were introduced to microfluidic device individually to stimulate microglia activations. For ECM contractility experiments, type I collagen (1 mg/mL) was mixed with microglia and cultured for 3 days in a 48 well-plate. The gel contracted area was analyzed using NIH Image J. Mechanical properties of GelMA hydrogels were characterized by a compression testing using Instron instrument. To generate lymphatic endothelial cells (LECs) vascular network, LECs (purchased from Promocell) were trypsinized and mixed with GelMA hydrogel solution prior to UV gelation. The hydrogels were cultured for 7 days to allow for LECs microvessel assembly. Two experimental conditions, LECs with and without Microglia, were utilized to examine the LECs network morphology.


Results, Conclusions, and Discussions:: We first tested the mechanical properties of GelMA hydrogels. The stiffness of 5 wt% GelMA hydrogel is ~ 3.2 kPa which value is in line with the brain tissue stiffness in vivo. We have previous reported that the stiffness of 3 mg/ml type I collagen is also similar with the brain tissue (~ 3.5 kPa). To investigate if the motility of microglia is the function of GBM-derived cytokines, we utilized a microfluidic device to study microglia invasion in 3D matrix. Our results showed that microglia did not significantly invade into matrix without any biochemical treatments (Figure A) while both LPS- and IL-4 treated microglia promoted microglia invasion in terms of invaded length and invasion area (Figure B and C). Interestingly, the U87MG-conditioned medium significantly increased microglia invasion compared to LPS and IL-4 treatments, suggesting GBM-secreted cytokines are potent to induce the microglia 3-D matrix invasion (Figure D). Since ECM remodeling is essential process prior to cell invasion, we applied ECM contractility assay to evaluate the capacity of matrix remodeling of microglia. Surprisingly, microglia demonstrate highly contractile phenotype even without any treatments on day 1 (~40% contraction). Moreover, both LPS and IL-4 further increased the gel contraction compared to untreated control (Figure E). Ongoing efforts are trying to measure the gel contraction by the U87MG-conditioned medium as well as to verify the immune phenotype of stimulated microglia by examining their secretome profiles. To explore the lymphangiogenic function of microglia, LECs were culture within GelMA hydrogel with and without microglia. Upon co-cultured with microglia in our 3-D GelMA hydrogels, LECs elongated and formed the microvessel morphology. On-going efforts are to study the effects of VEGF-C supplements on LECs network formation cocultured with microglia. In summary, our GBM microphysiological models have the great capability to mimic the important cues of glioblastoma to study GBM-microglia crosstalk. We are currently applying this developed model systems to further investigate the effects of other brain ECM components (e.g., hyaluronic acid) and patient-derived GBM cells (e.g., drug-resistance phenotypes) on biophysical dynamics of microglia to improve our understanding of tumor-immune regulatory circuit that orchestrates progression and immunosuppression of GBM.


Acknowledgements (Optional): : This work was supported by the NIH (R01CA256481).

References (Optional): :