Abstract
The presence of cytotoxic CD8 T cells correlates with improved survival and response to immunotherapies. However, patients with metastatic brain melanomas experience shorter-lasting survival than patients with extracranial metastases, for reasons which are not fully understood. To understand the underlying mechanisms of immune response to brain melanomas, we compared murine B16 melanomas implanted intracranially (IC) to well characterized tumors grown subcutaneously (SC). Compared to SC tumors, CD8 T cell activation in IC tumors was not altered but their representation was lower. Transferred CD8 T cells accumulated in similar numbers in tumors grown in both locations, suggesting that the vasculature does not differentially limit T cell representation in SC and IC tumors. Surprisingly, we did not see T cell activation in draining lymph nodes of either SC or IC tumor bearing mice, although this was consistent with the fact that there were relatively few dendritic cells (DC) that had acquired tumor antigen (Ag) in the draining lymph nodes, and these had an immature phenotype. In contrast, both SC and IC tumors contained substantial numbers of Ag+ myeloid cells. While the numbers of SC and IC intratumoral DC were comparable, DC in IC tumors captured less Ag, and were incompletely matured compared to DC in SC tumors. In addition, while microglia were the main Ag+ myeloid cell population in IC tumors, their presence did not influence Ag acquisition or the phenotype of other myeloid cells. Overall, this data suggest that the diminished representation of CD8 T cells in IC tumors is a consequence of alternatively matured DC and/or microglia that induce a distinct T cell activation state, which results in their failure to accumulate over time.
We also characterized the expression of homing receptor ligands (HRL), known to mediate T cell entry into tumors, on IC tumor endothelial cells (TEC) and compared it to their expression on well characterized SC TEC. We discovered that the fractions of cells expressing E-selectin, VCAM-1 and ICAM-1 on IC TEC were lower compared to SC TEC, but the fraction of cell expressing ALCAM was higher, and the fraction expressing P-selectin did not differ. Surprisingly, VCAM-1 and ICAM-1 expression on IC TEC was not upregulated by IFNγ or TNFα signaling, suggesting that IC TEC are less sensitive to those cytokines. Additionally, E-selectin and VCAM-1 expression on IC TEC was upregulated by the presence of vaccination induced effector CD8 T cells, but only when the tumor expressed an Ag that the mice had been previously vaccinated against. This suggests that activated T cells must persist and/or proliferate within the tumor to drive the upregulation of those two HRL. E-selectin was also transiently upregulated by focused ultrasound in conjunction with microbubble treatment, but it did not result in increased T cell infiltration in the investigated time-frame. Overall, this suggests that IC TEC are less sensitive to IFNγ or TNFα signaling than SC TEC, and expression of VCAM-1 and E-selectin on IC TEC is regulated by activated immune cells expressing other proinflammatory cytokines.
