Glioma cell migration on three-dimensional nanofiber scaffolds is regulated by substrate topography and abolished by inhibition of STAT3 signaling

Abstract

A hallmark of malignant gliomas is their ability to disperse through neural tissue, leading to long-term failure of all known therapies. Identifying new antimigratory targets could reduce glioma recurrence and improve therapeutic efficacy, but screens based on conventional migration assays are hampered by the limited ability of these assays to reproduce native cell motility. Here, we have analyzed the motility, gene expression, and sensitivity to migration inhibitors of glioma cells cultured on scaffolds formed by submicron-sized fibers (nanofibers) mimicking the neural topography. Glioma cells cultured on aligned nanofiber scaffolds reproduced the elongated morphology of cells migrating in white matter tissue and were highly sensitive to myosin II inhibition but only moderately affected by stress fiber disruption. In contrast, the same cells displayed a flat morphology and opposite sensitivity to myosin II and actin inhibition when cultured on conventional tissue culture polystyrene. Gene expression analysis indicated a correlation between migration on aligned nanofibers and increased STAT3 signaling, a known driver of glioma progression. Accordingly, cell migration out of glioblastoma-derived neurospheres and tumor explants was reduced by STAT3 inhibitors at subtoxic concentrations. Remarkably, these inhibitors were ineffective when tested at the same concentrations in a conventional two-dimensional migration assay. We conclude that migration of glioma cells is regulated by topographical cues that affect cell adhesion and gene expression. Cell migration analysis using nanofiber scaffolds could be used to reproduce native mechanisms of migration and to identify antimigratory strategies not disclosed by other in vitro models.

Figures

Figure 1

Glioma cell morphology and migration are directed by substrate topography. (A) Representative photography of a multiwell culture plate with a nanofiber-coated backing film. The enlarged image shows the translucent coating of nanofibers, and the double-headed arrow indicates the direction of fiber alignment. (B) Scanning electron microscope images of randomly oriented and highly aligned nanofibers. Bars, 5 µm. (C–E) Images of calcein-stained U251 glioma cells cultured on highly aligned nanofibers (C), randomly oriented nanofibers (D), or conventional TCPS (E). Bars, 100 µm. (F) Quantification of cell adhesion (U251 cells/mm2) showed significantly lower adhesion of cells to nanofibers compared with TCPS (***P < .001), but no differences between randomly oriented and aligned nanofibers (analysis by one-way ANOVA and Bonferroni post hoc tests). (G) Radial dispersion of U251 cells from cell aggregates was measured using nanofiber scaffolds of different thickness. Cells gained motility much faster on aligned than on randomly oriented nanofibers as the fibers became sparser. Use of 70-µm-thick scaffolds (dashed line) resulted in the highest cell motility on aligned nanofibers but little motility of the same cells on randomly oriented nanofibers.

Figure 2

Glioma cell migration on aligned nanofibers is myosin II-dependent. (A) Representative images of U251 glioma cell aggregates after being cultured for 24 hours on aligned or randomly oriented nanofibers, in the presence of 10 µM blebbistatin (blebbistatin) or its vehicle (control). Notice the parallel and elongated migration profile of cells on aligned nanofibers. Dashed outlines indicate the size and shape of the same aggregates at t = 0 h. Bars, 200 µm. (B) Effect of blebbistatin on cell dispersion on nanofibers. Results indicate a significant inhibition of cell migration on aligned but not on randomly oriented nanofibers. ***P < .001 by two-way ANOVA. (C) Effect of blebbistatin on cell translocation through cell culture inserts. Results indicate a significant effect only at 25 µM blebbistatin. *P < .05 by one-way ANOVA and Bonferroni post hoc test. (D) Effect of blebbistatin on two-dimensional cell migration measured with a wound healing assay. Migration of U251 cells in this assay was not affected by the myosin II inhibitor.

Figure 3

Cell migration on nanofibers is less sensitive to stress fiber disruption than two-dimensional migration. (A) U251 glioma cell aggregates were cultured on aligned nanofibers, fixed after 24 hours, and stained with phalloidin. Alexa 594 to label actin F. Results show diffuse cortical staining in vehicle-treated cells (vehicle) versus punctuate staining (insets) along actin filaments in cells treated with the actin polymerization inhibitor cytochalasin D (cyt-D, 2 µM). Cell morphology was little or not affected in these conditions. Bars, 100 µm. (B) Quantification of radial dispersion of U251 cells shows that two-dimensional cell dispersion on TCPS was significantly reduced even at the lowest concentration of cytochalasin D tested (0.2 µM), whereas dispersion on nanofibers required 10 to 100x higher concentrations to be significantly inhibited. **P < .01, ***P < .001 by one-way ANOVA for each treatment.

Figure 4

Glioma cell migration on aligned nanofibers correlates with activation of STAT3 signaling. (A) Quantitative RT-PCR results showing differential gene expression levels in glioma cells cultured on aligned versus randomly oriented nanofibers (selected genes are a subset of those listed in Table W1). Many of the upregulated genes are known modulators or targets of JAK/STAT and are involved in positive regulation of cell migration (Table W2). (B) Dissociated U251 glioma cells (5 x 105 cells/ml) were plated on aligned (A) or randomly oriented (R) nanofibers and collected 6 or 24 hours after attachment. Cells recovered from aligned nanofibers showed a substantial increase in Y705-phosphorylated STAT3 (p-STAT3) compared with cells recovered from randomly oriented nanofibers. As a control, U251 cells were plated on conventional TCPS and processed in the same manner, showing considerable expression of total and active STAT3 at all times tested.

Figure 5

STAT3 inhibition reduces glioma cell migration on aligned nanofibers. (A and B) Cell dispersion was measured for two glioma cell lines (U251 and U87) and neurospheres from two glioblastoma-derived initiating cells (G8 and G9), in the presence of the STAT3 inhibitors stattic (A) and LLL12 (B). Both inhibitors significantly reduced cell migration at concentrations between 0.5 and 2 µM. *P < .05, **P < .01, ***P < .001 by one-way ANOVA for each cell type and inhibitor. (C and D) Western blot analysis of U251 and G9 cells collected from aligned nanofibers confirmed the inhibition of STAT3 phosphorylation (pSTAT3) using LLL12 (C) and stattic (D) at the same concentrations used to inhibit cell dispersion. STAT3 inhibition in G9 cells—which form large compact tumorspheres—was sometimes incomplete at low concentrations of LLL12 or stattic, but cell motility was, nevertheless, reduced. Complete STAT3 inhibition was achieved at higher concentrations of these inhibitors. (E) Migration of U251 cells using a wound healing assay was not affected by low concentrations of STAT3 inhibitors. NS indicates nonsignificant.

Figure 6

STAT3 inhibition reduces cell dispersion from tumor explants. (A) Representative image of a tissue explant from a GFP-expressing G9 glioma cultured on aligned nanofibers during 48 hours. Arrowheads indicate nonfluorescent areas of adjacent brain tissue. Bar, 200 µm. (B) Tumor explants were cultured in the presence of 1 µM stattic (stattic), 1 µM LLL12 (LLL12), or their vehicle (vehicle), and cell dispersion was measured as before. Both STAT3 inhibitors significantly reduced (LLL12) or abolished (stattic) outward cell migration. **P < .01, ***P < .001 by two-way ANOVA for repeated measures.