Abstract A47: Interstitial flow increases patient-derived glioma stem cell invasion via CXCR4, CXCL12, and CD44-mediated mechanisms in distinctive cell populations

Abstract

Abstract Background: Glioblastoma (GBM) is the most common and aggressive brain tumor with an average survival of 15 months. Standard of care treatment strategies, which include surgical resection, radiotherapy, and chemotherapy, do not result in long-term survival(1). Much of this poor response is due to the invasiveness of GBM cells. We have shown that interstitial fluid flow (IFF) can increase invasion of glioma cell lines in vitro through both activation of the CXCR4 receptor and formation of autologous pericellular gradients of CXCL12(2). In other studies, mechanotransduction via the CD44 receptor has been implicated in IFF-mediated invasion(3). We hypothesize that these mechanisms are interrelated and vary depending on populations of receptor positive glioma cells. Therefore, we examined both CD44 and CXCR4-CXCL12 mechanisms in multiple primary patient-derived glioma stem cells (GSCs), which show increased heterogeneity compared to traditional cell lines, to identify both the clinical relevance and potential mechanisms of IFF-induced invasion. Methods: Four patient-derived GSC lines were used for our experiments (G2, G34, G62, and G528) from four individual patients and were derived/maintained as described (4). GSCs were seeded at 1E6 cells/mL into hyaluronan-collagen gels atop 8μm-pore tissue culture inserts(2). A pressure head was applied to yield a superficial velocity of 1 μm/s through the cell/gel compartment for flow, or media held level for static conditions. After 18h, gels were removed and inserts analyzed for transmigrated cells. Blocking antibodies for CD44, CXCL12, or CXCR4 were applied in media. For generation of irradiated cells, gamma-irradiation (0.3, 1 and 2 Gy) was applied 7 days before flow experiments. Flow cytometry for CXCR4, CD44, and CXCL12 was conducted to identify positive populations. We inoculated 15,000 GSCs (n=6 per GSC) into the cortices of male NOD-SCID mice. Prior to euthanasia, intravenous Evans blue was delivered to identify high/low flow regions. Tissue was processed and GSCs identified by human nuclear antigen to quantify invasion. Results: Characterization of our four patient-derived GSC lines for baseline migratory potential and expression of CXCR4, CXCL12, and CD44 showed separate populations of CXCR4+, CXCL12+, CD44+, CXCR4/CXCL12+ and CXCR4/CD44+ cells which varied by GSC, though baseline migration was relatively similar across GSCs. Using our 3D in vitro assay we demonstrated that IFF increased invasion in three of four cell lines. This flow-mediated invasion was reduced by blockade of CXCR4, CXCL12, CD44, or all, revealing that GSCs are migrating via both mechanisms, and that this varies depending on the GSC. Radiation increased flow-mediated invasion in G62, while eliminated it in G34. In vivo we see increased invasion in high flow regions compared to low flow that correspond to our in vitro results. Conclusion: Together, these data indicate heterogeneous flow response across patient-derived GSCs and contribute to mechanisms of invasion in vitro and in vivo. These results indicate, for the first time, that within a single patient, there are subpopulations of GSCs that respond to flow via both CD44-CXCR4 and CXCR4-CXCL12 mechanisms and these responses are affected by standard of care radiotherapy.