Inhibition of Radiation and Temozolomide-Induced Glioblastoma Invadopodia Activity Using Ion Channel Drugs

Simple Summary Glioblastoma accounts for approximately 40–50% of all primary brain cancers and is a highly aggressive cancer that rapidly disseminates within the surrounding normal brain. Dynamic actin-rich protrusions known as invadopodia facilitate this invasive process. Ion channels have also been linked to a pro-invasive phenotype and may contribute to facilitating invadopodia activity in cancer cells. The aim of our study was to screen ion channel-targeting drugs for their cytotoxic efficacy and potential anti-invadopodia properties in glioblastoma cells. We demonstrated that the targeting of ion channels in glioblastoma cells can lead to a reduction in invadopodia activity and protease secretion. Importantly, the candidate drugs exhibited a significant reduction in radiation and temozolomide-induced glioblastoma cell invadopodia activity. These findings support the proposed pro-invasive role of ion channels via invadopodia in glioblastoma, which may be ideal therapeutic targets for the treatment of glioblastoma patients. Abstract Glioblastoma (GBM) is the most prevalent and malignant type of primary brain cancer. The rapid invasion and dissemination of tumor cells into the surrounding normal brain is a major driver of tumor recurrence, and long-term survival of GBM patients is extremely rare. Actin-rich cell membrane protrusions known as invadopodia can facilitate the highly invasive properties of GBM cells. Ion channels have been proposed to contribute to a pro-invasive phenotype in cancer cells and may also be involved in the invadopodia activity of GBM cells. GBM cell cytotoxicity screening of several ion channel drugs identified three drugs with potent cell killing efficacy: flunarizine dihydrochloride, econazole nitrate, and quinine hydrochloride dihydrate. These drugs demonstrated a reduction in GBM cell invadopodia activity and matrix metalloproteinase-2 (MMP-2) secretion. Importantly, the treatment of GBM cells with these drugs led to a significant reduction in radiation/temozolomide-induced invadopodia activity. The dual cytotoxic and anti-invasive efficacy of these agents merits further research into targeting ion channels to reduce GBM malignancy, with a potential for future clinical translation in combination with the standard therapy.

and 2 were generated via the analysis of gene expression datasets deposited within the Oncomine TM cancer-profiling database. The automated statistical analysis components that are inbuilt within the platform were used to perform logical differential expression analyses between samples. p-values are corrected for multiple hypothesis testing using the false discovery method by the Oncomine TM program. With regard to the definition of logical analyses, each gene assessed for differential expression using the Oncomine TM platform is conducted with a Student t-test. The Oncomine TM platform utilizes the R statistical computing package, and tests are conducted both as two-sided for differential expression analysis and as one-sided for overexpression analysis.
The Oncomine TM platform employs the following within its automated analyses.

t-Tests and p-values:
The most common t-Test is the two-sample t-Test, which is used to compare the means (averages) of two independent samples. The null hypothesis, which is presumed to be true, is that the two groups have the same average value. A t-Test generates a p-value, which indicates how likely the null hypothesis (no difference between the populations) is true. If the chance is less than 5% (pvalue of 0.05 or less), then, by convention, the null hypothesis is rejected, and we conclude instead that there is a real, statistically significant difference between the means of the two groups.

p-values and Effect Size:
p-values measure whether the difference in means between two groups is likely to occur solely by chance. Effect Size measures the amount of difference between the groups and is reflected in the t-Test statistic.

Fold Change and p-values:
Fold change is a valuable complement to p-values to assess large absolute differences between groups of samples (or classes) measured in an analysis. When relying on p-value to assess differences, there can be analyses where p-values are very significant due to a large number of samples and low sample variability within groups, but the actual difference in the magnitude, or fold change, between groups is low. Table 3 was generated via the analysis of gene expression datasets deposited within the online cancer-profiling database, SurvExpress. Patient survival analysis can be presented using a Kaplan-Meier plot, which is the graphical representation of the survival probability versus time. It is common to represent more than one Kaplan-Meier curve in the same plot, especially when comparing a lowrisk versus high-risk group. Instead of a visual inspection of these curves within the plot, a Log-rank test evaluates statistically the equality of the survival curves within the plot, and therefore, it can be defined as the difference between the observed and expected events within a group and is generated as part of the SurvExpress analyses of datasets.

SurvExpress Analyses
Concordance Index (CI) values as generated by the SurvExpress analyses are displayed for Figure 1 and Table 3 within the manuscript. The CI is a summary indicator that estimates the probability that subjects with higher-risk prediction will experience an event at a different time to the subjects of lower risk.
CI is a generalization of the AUROC (Area Under the Receiver Operating Characteristic Curve) used in classification problems. The CI is expressed in SurvExpress as follows: .
In the above formula, ri and rj represent the risk predictors given by the corresponding prognostic index for subjects i and j, respectively and Ω represents all subject pairs (i,j) where ti < tj and subject i is not censored. As in AUROC, higher CI values are associated to better prediction.
SurvExpress calculates the sensitivity and specificity using each data value in determining the cutoff values. This means that it calculates many pairs of sensitivity and specificity. The program employs a customized algorithm that decides how partitions change between risk groups or to evaluate the relation between risk groups and prognostic index. It changes the cutoff point one risk group at a time so that the p-value is at a minimum, and the process is repeated until no further changes are required.  Table 4 over a range of concentrations (0, 0.01, 0.1, 1, and 10 μM). Cell viability was determined using an MTT (3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) cell proliferation assay. Cell viability is represented as a percentage relative to the control cells. Mean of n = 3 experiments, error bars represent SEM, *p < 0.05 (relative to untreated (0 μM) control group).  Table S1. Invadopodia regulator genes and ion channel genes utilized in the analyses of online gene expression GBM datasets deposited within the Oncomine TM and SurvExpress databases. CTTN  CACNA1B  MMP2  CACNA1C  MMP9  CACNA1D  Nck1  CACNA1F  Nck2  CACNA1G  SH3PXD2A  CACNA1H  SH3PXD2B  CACNA1I  Src  CACNA1S  KCNA5  KCNB1  KCNH2  KCNJ10  KCNN4  SCN5A  SCN8A  Table S2. Candidate drug predicted blood-brain barrier penetrance properties. We examined information on the three candidate drugs, flunarazine dihydrochloride, quinine hydrochloride dihydrate, and econazole nitrate present in the online database, 'DrugBank'(go.drugbank.com). DrugBank is a knowledge base consisting of clinical information such as side effects, drug interactions, and molecular-level data including chemical structures, predicted properties, and protein interactions. Based on their chemical structures, the table lists the predicted water solubility (as determined by ALOGPS) and the predicted blood-brain barrier penetrance (as determined by ADMET -Absorption, Distribution, Metabolism, Elimination, Toxicity), as extracted from the DrugBank database. A '+' ADMET value indicates a 'yes' for the predicted property of blood-brain barrier penetrance. As defined by the predicted ADMET features in DrugBank, the blood-brain barrier penetrance values of 0.9789, 0.9382, and 0.9823 indicate that there is a 97.89%, 93.82%, and 98.23% probability that flunarizine dihydrochloride, quinine hydrochloride dihydrate, and econazole nitrate will cross the blood-brain barrier. logP-drug lipophilicity; logS-drug water solubility.