The Cytoskeleton Effectors Rho-Kinase (ROCK) and Mammalian Diaphanous-Related (mDia) Formin Have Dynamic Roles in Tumor Microtube Formation in Invasive Glioblastoma Cells

Abstract

Glioblastoma (GBM) is a progressive and lethal brain cancer. Malignant control of actin and microtubule cytoskeletal mechanics facilitates two major GBM therapeutic resistance strategies-diffuse invasion and tumor microtube network formation. Actin and microtubule reorganization is controlled by Rho-GTPases, which exert their effects through downstream effector protein activation, including Rho-associated kinases (ROCK) 1 and 2 and mammalian diaphanous-related (mDia) formins (mDia1, 2, and 3). Precise spatial and temporal balancing of the activity between these effectors dictates cell shape, adhesion turnover, and motility. Using small molecules targeting mDia, we demonstrated that global agonism (IMM02) was superior to antagonism (SMIFH2) as anti-invasion strategies in GBM spheroids. Here, we use IDH-wild-type GBM patient-derived cell models and a novel semi-adherent in vitro system to investigate the relationship between ROCK and mDia in invasion and tumor microtube networks. IMM02-mediated mDia agonism disrupts invasion in GBM patient-derived spheroid models, in part by inducing mDia expression loss and tumor microtube network collapse. Pharmacological disruption of ROCK prevented invasive cell-body movement away from GBM spheres, yet induced ultralong, phenotypically abnormal tumor microtube formation. Simultaneously targeting mDia and ROCK did not enhance the anti-invasive/-tumor microtube effects of IMM02. Our data reveal that targeting mDia is a viable GBM anti-invasion/-tumor microtube networking strategy, while ROCK inhibition is contraindicated.

Keywords: Rho-kinase; actin; cytoskeleton; glioblastoma; invasion; mDia formin; tumor microtube.

Figure 1

Pharmacological mDia agonism dynamically affects its expression and is accompanied by loss of GBM patient-derived sphere and tumor microtube integrity. (A) 10X phase-contrast images of free-floating Pat9 3D spheroids at indicated timepoints (in hours) maintained in DMSO (top) or 50 μM IMM-02 (bottom). Scale bars = 400 μm. (B) Western blots of cell lysates from free-floating Pat9 3D spheroids treated with DMSO or 50 μM IMM-02 at indicated time points (in h). Molecular weight markers (kDa) are listed on left. Blotting antibodies are listed on right of blot. (C) 4X phase-contrast images of Pat27 2.5D cultures at indicated time points maintained in DMSO (top) or 50 μM IMM-02 (bottom). Scale bars = 1000 μm. (D) Western blots of cell lysates from 2.5D Pat27 cultures treated with DMSO or 50 μM IMM-02 at indicated timepoints.

Figure 2

ROCK-directed contractility machinery regulates patient-derived GBM pro-invasive tumor microtube networks. (A) Confocal images of leading edge at T96 in fixed Y-27632-treated (90 $\mathsf{\mu }$M) Pat9 3D invasion assay stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m. (B) Confocal images of leading edge at T96 in fixed H₂O-treated Pat9 3D invasion assay stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m. (C) Tumor microtubule length in H₂O- or Y-276632-treated (90 $\mathsf{\mu }$M) Pat9 96 h 3D invasion assays. **** p $\le$ 0.0001; *** p $\le$ 0.001; ns = not significant. (D) Distance of cell body movement from the sphere core in H₂O- or Y-276632-treated (90 $\mathsf{\mu }$M) Pat9 96 h 3D invasion assays. **** p $\le$ 0.0001. (E) Increase in total area of invasion over T0 in H₂O- or Y-276632-treated (90 $\mathsf{\mu }$M) Pat9 96 h 3D invasion assays. (F) Confocal images of leading edge at T96 in fixed Pat9 3D invade-then-treat (ITT) assays treated with Y-276632 (90 $\mathsf{\mu }$M). Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m. (G) Tumor microtubule length in H₂O- or Y-276632-ITT (90 $\mathsf{\mu }$M) Pat9 96 h 3D invasion assays. Dotted line shows time of drug introduction. **** p $\le$ 0.0001. (H) Distance of cell body movement from the neurosphere core in H₂O- or Y-276632-ITT (90 $\mathsf{\mu }$M) Pat9 96 h 3D invasion assays. Dotted line shows time of drug introduction. **** p $\le$ 0.0001. (I) Increase in total area of invasion over T0 in H₂O- or Y-276632-ITT (90 $\mathsf{\mu }$M) Pat9 96 h 3D invasion assays. Dotted line shows time of drug introduction. Note: (G–I) experimental procedure was performed in the same experiment/time as (C–E), but results were split onto 2 graphs for clarity. The same controls are accordingly graphed in (G–I) as in (C–E).

Figure 3

Combined targeting of ROCK and mDia halts GBM patient-derived sphere invasion, yet is not superior to mDia formin agonism alone. (A) Schematic of drug exposures in combination drug 3D invasion assays (Y-27632 = 90 $\mathsf{\mu }$M; IMM02 = 50 $\mathsf{\mu }$M). Bars at left indicate 48 h invasion prior to indicated drug treatment. (B) Tumor microtubule length in indicated Pat9 96 h 3D invasion assays. **** p $\le$ 0.0001; ns = not significant. (C) Distance of cell body movement from the neurosphere core in indicated Pat9 96 h 3D invasion assays. ** p $\le$ 0.01; *** p $\le$ 0.001, **** p $\le$ 0.0001. (D) Increase in total area of invasion over T0 in indicated Pat9 96 h 3D invasion assays. *** p $\le$ 0.001; ** p $\le$ 0.01; * p $\le$ 0.05. (E) Confocal images of leading edge at T96 in fixed Pat9 3D invasion assays treated with H₂O + DMSO. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 75 $\mathsf{\mu }$m. (F) Confocal images of leading edge at T96 in fixed Pat9 3D invasion assays treated with Y-27632 + DMSO. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 75 $\mathsf{\mu }$m. (G) Confocal images of leading edge at T96 in fixed Pat9 3D invasion assays treated with H₂O + IMM02. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 25 $\mathsf{\mu }$m. (H) Confocal images of leading edge at T96 in fixed Pat9 3D invasion assays treated with H₂O + IMM-02. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 25 $\mathsf{\mu }$m.

Figure 4

Altered the sequencing of combined ROCKi/mDia targeting does not modulate GBM invasion. (A) Schematic of drug exposures in drug-switch 3D invasion assays (Y-27632 = 90 $\mathsf{\mu }$M; IMM-02 = 50 $\mathsf{\mu }$M). (B) Tumor microtubule length in indicated Pat9 96 h drug-switch 3D invasion assays. Dotted line shows time of drug switch. **** p $\le$ 0.0001; * p $\le$ 0.05; ns = not significant. (C) Distance of cell body movement from the neurosphere core in indicated Pat9 96 h drug-switch 3D invasion assays. Dotted line shows time of drug switch. **** p $\le$ 0.0001; ** p $\le$ 0.01. (D) Increase in total area of invasion over T0 in indicated Pat9 96 h drug-switch 3D invasion assays. Dotted line shows time of drug switch. **** p $\le$ 0.0001; *** p $\le$ 0.001; ** p $\le$ 0.01; * p $\le$ 0.05. (E) Confocal images of leading edge at T96 in fixed Pat9 drug-switch 3D invasion assays treated with H₂O-then-DMSO. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m. (F) Confocal images of leading edge at T96 in fixed Pat9 drug-switch 3D invasion assays treated with Y-27632-then-DMSO. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m. (G) Confocal images of leading edge at T96 in fixed Pat9 drug-switch 3D invasion assays treated with H₂O-then-IMM02. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 50 $\mathsf{\mu }$m. (H) Confocal images of leading edge at T96 in fixed Pat9 switch-drug 3D invasion assays treated with Y27632-then-IMM02. Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m.

Figure 5

Sustained ROCKi postpones cellular responses to mDia agonists in invading GBM spheroids. (A) Schematic of drug exposures in add-drug 3D invasion assays (Y-27632 = 90 $\mathsf{\mu }$M; IMM-02 = 50 $\mathsf{\mu }$M). (B) Tumor microtubule length in indicated Pat9 96 h add-drug invasion assays. Dotted line shows time of drug addition. **** p $\le$ 0.0001; *** p $\le$ 0.001; * p $\le$ 0.05; ns = not significant. (C) Distance of cell body movement from the sphere core in indicated Pat9 96 h add-drug invasion assays. Dotted line shows time of drug addition. **** p $\le$ 0.0001; * p $\le$ 0.05. (D) Increase in total invasion area over T0 in indicated Pat9 96 h add-drug 3D invasion assays. Dotted line shows time of drug addition. ** p $\le$ 0.01; * p $\le$ 0.05. (E) Confocal images of leading edge at T72 in Pat9 add-drug 3D invasion assays treated with Y-27632-then-(Y-27632 + IMM-02). Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m. (F) Confocal images of leading edge at T96 in fixed Pat9 add-drug 3D invasion assays treated with Y-27632-then-(Y-27632 + IMM-02). Stained for $\beta$-tubulin, phalloidin, and DAPI. Scale bars = 100 $\mathsf{\mu }$m.