Understanding the mechanisms of diffuse midline glioma cell migration toward therapeutic targeting

Pediatric diffuse midline glioma (DMG) is a fatal disease without curative therapies.¹ Unlike other malignant brain tumors, DMGs are characterized by high cell infiltration within critical anatomic compartments that precludes surgery as a therapeutic option.¹ DMG cells integrate within a complex tumor microenvironment, hijacking interactions with normal neurons to promote tumor growth.² This complex biology underscores the relevance of studying the processes of DMG cell metastasis and invasion, which may correlate with patient outcomes and reveal new insights into brain tumor biology.

In this article, Bruschi and colleagues demonstrate that DMG prognosis can vary depending on invasion rate and metastatic relapse.³ In a large cohort of 72 patients with DMG, 3 major radiological classes were defined as local, locoregional, and distant metastasis at relapse. Distant metastases were associated with reduced survival in patients, which was independent of other well-known poor prognostic characteristics, such as H3.3-K27M and TP53 mutations. To further characterize the cell invasion phenotype, they compared the patients with local and locoregional relapse, which revealed infiltration patterns and parenchymal invasion rate that correlated with survival. Further, while invasion negatively correlates with survival, no recurrent genetic alteration is predictive of distant progression, and there are currently no known determinants associated with invasiveness.

To extend the biological findings, the authors used patient-derived DMG organoid models to predict metastatic relapse to elucidate the mechanisms behind different patterns of invasion. Models from patients with local to locoregional tumor progression were significantly less invasive than those from patients with metastatic relapse. This demonstrates that the 3D models generated recapitulate features of the primary disease and represent relevant models that can be used to investigate the mechanisms driving DMG invasion.

Using their DMG-derived organoids, the authors performed RNA-seq profiling, and found that gene expression data were able to differentiate the organoids into molecular groups that have been previously described. When using unsupervised hierarchical clustering of these organoids, their gene expression profiles separated them into 2 main groups, and interestingly, these clusters correlate with invasion capacity of the corresponding glioma-stem cells (GSCs). This is important since it suggests that transcriptional programs may contribute to the heterogeneity seen in the invasive capacity of these tumors. Additionally, the cluster of organoids that were highly invasive in GSCs had a distinctive oligodendrocyte progenitor cell (OPC) identity, whereas the moderately invasive group was much more mesenchymal in identity. Importantly, these findings were corroborated in patient DMG samples, which showed a similar pattern.

Characterization of the phenotypic differences in these GSCs showed distinct enrichment in the expression of genes related to focal adhesion and extracellular matrix interactions in the less invasive GSCs. These signatures were confirmed by RNA-seq of tumor biopsies. The highly invasive GSCs had an elongated morphology and an amoeboid-like migration. From this, the authors concluded that the differences in adhesion and migration underlie the differences in invasive capabilities of the 2 cohorts.

While looking at the differentially expressed genes between moderately invasive and highly invasive model cohorts, the authors found that bone morphogenetic pathway 7 (BMP7) was overexpressed by highly invasive GSCs. BMP7 has been previously identified as important for OPC migration during brain development, but its association with invasion is novel. Excitingly, this is recapitulated in the human data in pediatric DMGs, where BMP7 expression is associated with OPC-like tumor cells. It is suggested that BMP7 expression may promote DMG invasion, as genetic knockdown of BMP7 significantly delays invasion and migration in in vitro models. Additionally, their findings suggest that BMP7 is a major regulator of GSC motility and invasion, controlling a switch between 2 migration modalities. This finding importantly sheds light on the underlying mechanism as to how these tumors invade and migrate. From a therapeutic perspective, given that BMP7 kinase activation dictates migration, and MEK/ERK, Rho/ROCK kinases are activated downstream of BMP7, this might be an actionable target to prevent spread and metastasis. Supporting this concept, inhibition of MEK, ERK1/2, and ROCK in DMG models all impaired the invasion rate of highly invasive GSCs in vitro.

This research article provides a rich resource of 3D organoid DMG models that enable functional studies of the invasive and migratory properties of DMG cells. They demonstrate the utility of these models by characterization of heterogeneous patterns of cell migration in DMG tumors that correlate with the primary disease and may inform patient outcomes. Finally, this work sheds light on the cell-type functional diversity of DMG tumors and reveals new insights into therapies that target mechanisms of DMG migration and invasion.

Conflict of interest statement

None declared.