Dexamethasone-mediated inhibition of Glioblastoma neurosphere dispersal in an ex vivo organotypic neural assay

Abstract

Glioblastoma is highly aggressive. Early dispersal of the primary tumor renders localized therapy ineffective. Recurrence always occurs and leads to patient death. Prior studies have shown that dispersal of Glioblastoma can be significantly reduced by Dexamethasone (Dex), a drug currently used to control brain tumor related edema. However, due to high doses and significant side effects, treatment is tapered and discontinued as soon as edema has resolved. Prior analyses of the dispersal inhibitory effects of Dex were performed on tissue culture plastic, or polystyrene filters seeded with normal human astrocytes, conditions which inherently differ from the parenchymal architecture of neuronal tissue. The aim of this study was to utilize an ex-vivo model to examine Dex-mediated inhibition of tumor cell migration from low-passage, human Glioblastoma neurospheres on multiple substrates including mouse retina, and slices of mouse, pig, and human brain. We also determined the lowest possible Dex dose that can inhibit dispersal. Analysis by Two-Factor ANOVA shows that for GBM-2 and GBM-3, Dex treatment significantly reduces dispersal on all tissue types. However, the magnitude of the effect appears to be tissue-type specific. Moreover, there does not appear to be a difference in Dex-mediated inhibition of dispersal between mouse retina, mouse brain and human brain. To estimate the lowest possible dose at which Dex can inhibit dispersal, LogEC50 values were compared by Extra Sum-of-Squares F-test. We show that it is possible to achieve 50% reduction in dispersal with Dex doses ranging from 3.8 x10-8M to 8.0x10-9M for GBM-2, and 4.3x10-8M to 1.8x10-9M for GBM-3, on mouse retina and brain slices, respectively. These doses are 3-30-fold lower than those used to control edema. This study extends our previous in vitro data and identifies the mouse retina as a potential substrate for in vivo studies of GBM dispersal.

Fig 1. Dispersal of primary human GBM spheroids on extirpated mouse retina and brain slices.

(A). Spheroids composed of PKH-2-stained cells measuring approximately 400 micrometers in diameter were deposited onto mouse retinas (mRetina), or 100 micron-thick, formalin-fixed mouse brain (mBrain) slices and incubated either in the absence or presence of 1x10^-7 M Dexamethasone (Dex). After 24 hours in tissue culture, images were captured. Note significant dispersal of untreated (UT) spheroids relative to those treated with Dex. Scale bar = 300μm. Blocking the corticosteroid receptor reverses Dex-mediated inhibition of GBM dispersal on mouse brain (B). GBM-2 and GBM-3 spheroids were generated either in complete medium (UT), in the presence of 1x10^-7M Dex, or Dex and 1μm of the steroid receptor antagonist, RU-486. Mean dispersal area was analyzed by ANOVA and Tukey’s HSD. For each data set, n = 12. Bars are standard error of the means. Asterisks represent significant difference of p<0.0001 (****) and p<0.001 (***). Dex withdrawal profile of GBM-3 on mouse retina (C). Dispersal area was measured for cells pre-treated in 2D culture with Dex (Dex, D) prior to spheroid formation, for cells pre-treated with Dex and as spheroids in hanging drop cultures (Dex, D+HD), or with Dex throughout the course of the experiment, including placement on mouse retina (Dex, D+HD+R). Bars represent standard error of the means. For each data set, n = 12. Average dispersal area was compared by One-Factor ANOVA, α of 0.05. Asterisks represent significant difference by Tukey’s HSD test (p<0.01 **).

Fig 2. Dex-mediated inhibition of GBM dispersal on murine, porcine and human neural tissue.

Fluorescently-labeled spheroids of GBM-2 (A) and GBM-3 (B) were placed onto extirpated mouse retina (mRetina), or onto 100 micrometer-thick slices of mouse Brain (mBrain), pig brain (pBrain), or human brain (hBrain) and incubated for 24 hours. Mean dispersal area was compared by Two-Factor ANOVA and Tukey’s HSD. Data sets ranged from n = 9 to n = 12, depending on outlier removal by ROUT Test. Asterisks represent significant difference in dispersal area between untreated and Dex-treated spheroids. Note that the effects of Dex on dispersal of GBM-3 on human brain did not reach statistical significance by Tukey’s HSD after 24 hours in culture, but did so after 48-hours (Fig 2B).

Fig 3. Dex dose-response curves of GBM dispersal on extirpated mouse retina and brain slices.

Fluorescently-labeled spheroids of GBM-2 (A) and GBM-3 (B) were generated and plated onto mouse retina or brain slices in Dex concentrations ranging from 1x10^-10 M to 1x10^-6 M. Spheroids were incubated for 24 hours and dispersal area was plotted as a function of dose. For each data set, n = 10–12, depending on outlier removal by ROUT test. Curves were analyzed by non-linear regression. R-squared values for Goodness-of-fit ranged from 0.61 to 0.81. Log EC50 values were compared for mRetina and mBrain by Extra Sum-of-Squares F test (α of 0.05). p-values for differences in Log EC50 between mRetina and mBrain were 0.0033 and 0.0011 for GBM-2 and GBM-3, respectively.