The first evidence for expression of glutamate receptors in glioma cells came from an electrophysiological study performed in Xenopus
oocytes injected with mRNAs isolated from astrocytoma cell line R-111 
. Positive inward currents in response to glutamate and AMPA suggested that astrocytoma cells express AMPA-type glutamate receptors. Although no currents were elicited in oocytes injected with mRNAs from GBM cell lines in this study, a later study found that approximately two thirds of GBM cells express glutamate receptors, where currents in cultured GBM cells and acute slices made from tumor tissues in response to glutamate could be recorded 
. Other studies have also confirmed the expression of AMPA receptors, but also NMDA, kainate, and metabotropic glutamate receptors in primary brain tumors 
. Moreover, a study by de Groot and colleagues has shown that GBM cells express higher levels of GluR1 compared to anaplastic or low-grade astrocytomas, suggesting that level of GluR1 expression correlates with the grade of tumor 
. Our study is the first to show the up-regulated expression of calcium-permeable AMPA receptors in GBM BTICs, when directly compared to non-stem tumor cells from patient-matched tumor tissues.
We used a previously described selective culturing technique to enrich adherent GBM tumor cells expressing stem cell markers 
. Stem cell characteristics of GBM BTIC cultures were confirmed by the expression of stem cell markers, nestin and Sox-2, ability to form neurospheres, and ability to form differentiated tumors in xenotransplantation model, which are all hallmark features of BTICs 
. While this technique has limitations, the advantage of being able to stably culture an adherent BTICs is tremendous. A direct comparison of BTICs from non-BTICs from the same tumor tissue, for example, can give insights into key molecular differences that allow BTICs to act more aggressively, allowing tumor formation following xenotransplantation 
. Adherent GBM BTIC cultures also allows genetic and pharmacological manipulations to modulate and characterize key molecules involved in proliferation and invasion. These experiments can be followed by xenotransplantation experiments to see what role these molecules play in allowing BTICs to form tumors in animal models. One argument against the use of adherent GBM BTIC cultures is that protein expression may be falsely elevated due to EGF and FGF in the media. However, we have data suggesting that the expression levels of other key modulators of cellular proliferation (i.e. Akt and MAP kinase) are not different between GBM BTIC and differentiated GBM cultures (data not shown). Moreover, GluR2 subunit expression was undetectable in both cultures. Thus, we believe that up-regulated GluR1 and GluR4 expression in BTICs is a valid finding that warrants further studies to investigate the downstream effectors of these calcium-permeable AMPA receptors.
Although we did not test for expression of GluR3 subunit in BTICs, which has relatively low expression levels in glioma cells 
, GluR2 subunit could not be detected in either BTIC or differentiated GBM cultures. By contrast, GluR2 signal could be detected on immunoblots with normal mouse brain homogenates. This is consistent with previous findings where only calcium-permeable AMPA receptor subunits were found in GBM cells 
, although low expression of GluR2 subunit was detectable in low-grade gliomas in one study 
. In another study, GluR2 subunits were detected in malignant gliomas, but the post-transcriptional editing at the Q/R-site on the mRNA that makes GluR2 subunits calcium-impermeable was significantly under-edited, making these subunits calcium-permeable 
. By contrast, nearly 100% of GluR2 subunits in the adult brain are edited at the Q/R-site, making them calcium-impermeable. Interestingly, under-editing of GluR2 was less severe in the low-grade astrocytomas 
. Overall, these results together suggest that malignant gliomas tend to favor the expression of calcium-permeable AMPA receptors. Consistent with these findings, our results show up-regulated expression of calcium-permeable AMPA receptor subunits, GluR1 and GluR4, in GBM BTICs. The downstream effectors of these receptors and how their activity modulates GBM BTICs’ invasive potential is a subject of our future studies.
There is ample evidence currently to support the premise that glutamate receptors, and more specifically calcium-permeable AMPA receptors, do play a significant role in potentiating proliferation and invasion of GBM cells. Blocking glutamate receptor activity, for example, has been shown to block glioma proliferation 
, while direct pharmacological activation of glutamate receptors lead to increased invasion and proliferation 
. Although the exact mechanism for this is unclear, some downstream effectors have been identified. The growth and motility of GBM cells following activation of calcium-permeable AMPA receptors has been shown to involve Akt activation via phosphorylation of Akt at serine 473 
. Moreover, AMPA receptor activation causes increase in EGF receptor expression in U-87 MG cell line, which is also linked to downstream phospho-Akt pathway 
. Other key molecules downstream of phospho-Akt that mediate increased motility, proliferation, and invasion of GBM cells are yet to be determined.
The hypothesis that activation of calcium-permeable AMPA receptors in GBM BTICs allows them to proliferate and move into the surrounding normal brain tissues also needs to be confirmed in in vivo models. After migrating into normal brain tissues, BTICs can proliferate and differentiate, resulting in increase tissue glutamate concentration. This process may lead to further BTIC activation via calcium-permeable AMPA receptors, resulting in the highly aggressive phenotype of GBM.
While increased glutamate concentration in the tumor milieu may be one key factor in increasing GBM proliferation and invasion, how GBM cells survive this elevated excitotoxic levels of glutamate remains unclear. The U-87 MG GBM cell line, for examples, has been found to have at least 45 fold in resistance to glutamate excitotoxicity compared to normal glial cells 
. A study by van Vuurden and colleagues found that GBM cells could tolerate high glutamate concentrations due to their down-regulated expression of AMPA receptor subunits compared to normal brain 
. Our results showing that differentiated GBM cultures express very little GluR1 and GluR4 compared to the normal mouse brain support these findings. How GBM BTICs survive and thrive in such high glutamate concentration environment is also unclear. One possibility is that the kinetics of calcium currents in glioma cells may be very different than that of normal neurons and glial cells. For example, electrophysiological recordings of glioma cells with stimulation with glutamate or kainate elicited action potentials in some cells, while others have demonstrated depolarization with slower kinetics than one would expect from a fast acting excitatory ion channels 
. In our study, we observed relatively slow increase in intracellular calcium in response to AMPA and CYTZ over time-scale of minutes, suggesting that kinetics of calcium currents are very different in GBM cells compared to normal neurons. Another possibility is that GBM cells may have significantly increased capacity to buffer intracellular calcium with different kinetics than normal neurons, allowing them to survive prolonged calcium influx. Overall, our results indicate that GBM BTICs do express high levels of calcium-permeable glutamate receptors and prolonged stimulation with AMPA and CYTZ, sometimes lasting up to 20 minute, can continue to elevate intracellular calcium concentrations. Factors involved in allowing GBM BTICs to survive the prolonged stimulation by increased glutamate concentration and calcium influxes could also lead to potential novel therapies.
Clinical trials are currently underway to determine whether glutamate receptor antagonist has any survival benefit for GBM patients. One particular drug called talampanel, an AMPA receptor antagonist, has undergone phase II clinical trial for newly diagnosed GBM patients. 
When compared to historical controls from the European Organisation for Research and Treatment of Cancer (EORTC), addition of talampanel to the standard radiation and temozolomide therapy prolonged the mean survival time from 14.6 months to 20.3 months and improved 24-month survival rate from 26.5% to 41.7% (p
Moreover, addition of talampanel as adjuvant therapy was well tolerated with minimal side-effects. However, these results should be taken with a caution, as talampanel as a single therapeutic agent for recurrent GBM patients had no significant effect on 6-month progression-free survival. 
Currently, phase III clinical trials are underway for talampanel in addition to the standard radiation and temozolomide therapy for newly diagnosed GBM patients. Our results showing consistent and highly up-regulated expression of calcium-permeable glutamate receptors in GBM BTICs warrant further studies exploring the clinical benefits of glutamate receptor antagonists, and specifically calcium-permeable glutamate receptor antagonists, in GBM patients. Therapies specifically targeting GBM BTICs by blocking the calcium-permeable glutamate receptors may allow improved control of recurrences and disease progression following the surgical resection.
Our results indicate that GBM BTICs overexpress calcium-permeable AMPA receptors, which conduct calcium influx in response to specific agonists. Presence of BTICs in GBM tissues with evidence supporting elevated glutamate concentrations in the tumor milieu warrant further studies exploring the role of increased intracellular calcium concentration mediated by calcium-permeable AMPA receptors in GBM BTICs.