N-cadherin dynamically regulates pediatric glioma cell migration in complex environments

Pediatric high-grade gliomas are highly invasive and essentially incurable. Glioma cells migrate between neurons and glia, along axon tracts, and through extracellular matrix surrounding blood vessels and underlying the pia. Mechanisms that allow adaptation to such complex environments are poorly understood. N-cadherin is highly expressed in pediatric gliomas and is associated with shorter survival. We found that inter-cellular homotypic N-cadherin interactions differentially regulate glioma migration according to the microenvironment, stimulating migration on cultured neurons or astrocytes but inhibiting invasion into reconstituted or astrocyte-deposited extracellular matrix. N-cadherin localizes to filamentous connections between migrating leader cells but to epithelial-like junctions between followers. Leader cells have more surface and recycling N-cadherin, increased YAP1/TAZ signaling, and increased proliferation relative to followers. YAP1/TAZ signaling is dynamically regulated as leaders and followers change position, leading to altered N-cadherin levels and organization. Together, the results suggest that pediatric glioma cells adapt to different microenvironments by regulating N-cadherin dynamics and cell-cell contacts.

Methods
To compare RNA transcriptomes from bulk cell populations, control or N-cad shRNA cells were dissociated using Accutase at room temperature for 10 min and resuspended in HBSS. 500 cells were collected in the center of a 1.5 ml centrifuge tube containing 4.75 μl SMART-seq reaction buffer (Takara) using a BD FACSymphony S6 (BD Bioscience), avoiding cell loss on the tube walls. To compare RNA transcriptomes from leader and follower cells, approximately 90 spheroids of PBT-05 cells expressing histone H2B-Dendra2 cells were allowed to migrate for 24 hrs on laminin and photoconverted as above. After photoconversion, cells were dissociated with Accutase at room temperature, resuspended in HBSS and transferred to ice, and approximately 200 leader and follower cells were sorted into SMART-seq reaction buffer as above. Each experiment was performed on four different occasions.

RNA was prepared and cDNA was synthesized with the SMART-Seq v4 Ultra Low Input RNA Kit for Sequencing (Takara) and ran the Agilent Tapestation to assess cDNA product. To construct RNA sequencing libraries, we used Illumina’s Nextera XT kit to fragment the cDNA and add barcoded sequencing adapters. Differential gene expression analysis was performed with the DEseq2 for paired sample R package (Love et al., 2014). Genes with a Benjamini-Hochberg adjusted p-value < 0.05 were defined as differentially expressed.