<title>Abstract</title>
<sec>
<title>Background</title>
Gliomas typically escape surgical resection and recur due to their “diffuse invasion” phenotype, enabling them to infiltrate diffusely into the normal brain parenchyma. Over the past 80 years, studies have revealed 2 key features of the “diffuse invasion” phenotype, designated the Scherer’s secondary structure, and include perineuronal satellitosis (PS) and perivascular satellitosis (PVS). However, the mechanisms are still unknown.


</sec>
<sec>
<title>Methods</title>
We established a mouse glioma cell line (IG27) by manipulating the histone H3K27M mutation, frequently harboring in diffuse intrinsic pontine gliomas, that reproduced the diffuse invasion phenotype, PS and PVS, following intracranial transplantation in the mouse brain. Further, to broadly apply the results in this mouse model to human gliomas, we analyzed data from 66 glioma patients.


</sec>
<sec>
<title>Results</title>
Increased H3K27 acetylation in IG27 cells activated glucose transporter 1 (Glut1) expression and induced aerobic glycolysis and TCA cycle activation, leading to lactate, acetyl-CoA, and oncometabolite production irrespective of oxygen and glucose levels. Gain- and loss-of-function in vivo experiments demonstrated that Glut1 controls the PS of glioma cells, that is, attachment to and contact with neurons. GLUT1 is also associated with early progression in glioma patients.


</sec>
<sec>
<title>Conclusions</title>
Targeting the transporter Glut1 suppresses the unique phenotype, “diffuse invasion” in the diffuse glioma mouse model. This work leads to promising therapeutic and potential useful imaging targets for anti-invasion in human gliomas widely.


</sec>