Glioblastoma (GBM) is an aggressive and incurable primary brain tumor that causes severe neurologic, cognitive, and psychologic symptoms. Symptoms are caused and exacerbated by the infiltrative properties of GBM cells, which enable them to pervade the healthy brain and disrupt normal function. Recent research has indicated that although radiotherapy (RT) remains the most effective component of multimodality therapy for patients with GBM, it can provoke a more infiltrative phenotype in GBM cells that survive treatment. Here, we demonstrate an essential role of the actin-myosin regulatory kinase myotonic dystrophy kinase-related CDC42-binding kinase (MRCK) in mediating the proinvasive effects of radiation. MRCK-mediated invasion occurred via downstream signaling to effector molecules MYPT1 and MLC2. MRCK was activated by clinically relevant doses per fraction of radiation, and this activation was concomitant with an increase in GBM cell motility and invasion. Furthermore, ablation of MRCK activity either by RNAi or by inhibition with the novel small-molecule inhibitor BDP-9066 prevented radiation-driven increases in motility both in vitro and in a clinically relevant orthotopic xenograft model of GBM. Crucially, treatment with BDP-9066 in combination with RT significantly increased survival in this model and markedly reduced infiltration of the contralateral cerebral hemisphere.

Significance: An effective new strategy for the treatment of glioblastoma uses a novel, anti-invasive chemotherapeutic to prevent infiltration of the normal brain by glioblastoma cells.Cancer Res; 78(22); 6509–22. ©2018 AACR.

Glioblastoma (GBM) is the most common and most aggressive primary brain tumor and is currently incurable. Poor outcomes are caused in part by the highly infiltrative nature of GBM tumor cells. This property enables them to disseminate through the brain via existing white matter tracts and perivascular spaces, making complete surgical resection unachievable and contributing to high recurrence rates. Therefore, despite aggressive treatment with surgery, radiotherapy (RT), and chemotherapy, overall survival remains extremely poor, with patients experiencing an average life expectancy of approximately 1 year (1). Targeting the molecular mechanisms that underlie the invasive nature of GBM has the potential to increase survival and alleviate some of the devastating neurologic disabilities associated with infiltrative disease.

The vast majority of patients with GBM receive RT as the central component of their first-line treatment. Studies in the late 1970s showed that radical RT doubles survival from approximately 6 to 12 months (2); in 2005, a further improvement from 12 to 14 months was achieved by adding concomitant and adjuvant temozolomide (TMZ) to RT (1). Every other phase III trial in this disease has been negative. It is universally accepted, therefore, that any novel therapies tested in the first-line setting must be added to a RT-based treatment schedule.

Despite its proven benefits, it is clear that radiation can also adversely affect tumor cell behavior through alterations in cell signaling and gene expression profiles. As early as 2001, there have been studies indicating that sublethal doses of radiation can promote GBM cell motility and infiltration in vitro and in vivo (3, 4). These early reports have recently been given renewed attention following the publication of a number of new studies (5–7). The concept that RT can promote a more aggressive, infiltrative phenotype in those GBM cells that survive treatment is consistent with the repeated finding that escalating RT dose does not improve outcomes for patients with GBM (8). It also has profound implications for the development of new therapeutic avenues and highlights the need for a greater understanding of the mechanisms that drive these changes so that new therapeutic targets can be identified and explored. For example, the failure of high radiation doses to achieve tumor control may be partly explained by tumor cells migrating outside of the irradiated volume during the treatment period.

Migration and invasion of tumor cells are driven by mechanical forces generated by changes in contractility of the actin-myosin cytoskeleton (9). The Rho family of small GTPases comprises key regulators of actin-myosin contractility and has been widely implicated in metastasis and invasion (10, 11). Indeed, the role of RhoGTPases in GBM has been interrogated by a number of researchers, and although their results support a pivotal role for small GTPases and their downstream effectors, teasing out their specific contributions has proved complex, with different studies returning contradictory results (12).

RhoGTPases act via downstream effector kinases such as Rho-associated protein kinase (ROCK) and myotonic dystrophy kinase-related CDC42-binding kinase (MRCK). These downstream kinases regulate contractility of the cytoskeleton by activating phosphorylation of myosin light chain proteins (MLC) and inactivating the myosin phosphatase subunit MYPT1 to facilitate the cytoskeletal changes that are responsible for different modes of cell motility. The “amoeboid” mode of invasion is characterized by rounded cell morphology and is highly dependent on ROCK activity (13). Considerable effort has been invested in the development of ROCK inhibitors that could oppose RhoA-driven cancer cell motility and metastasis. However, although ROCK inhibitors do indeed display strong efficacy in opposing cancer cell invasion and metastasis in vitro and in vivo (14), their clinical development has not progressed because of adverse effects on the cardiovascular system.

Significance: In contrast, mesenchymal invasion, which is classically adopted by infiltrating glioma cells, requires activity of the MRCK isoforms α and β that lie downstream of the RhoGTPase protein CDC42. This mode of invasion is characterized by an elongated cell body, actin-rich protrusions, and actin-myosin contractility at the rear of the migrating cells. Unbekandt and colleagues have recently demonstrated that MRCK can drive cancer cell migration, a process that can be inhibited by specific small-molecule inhibitors (15–17). Because CDC42 activity is upregulated in glioma (18, 19) and may drive the mesenchymal mode of migration employed by infiltrating glioma cells, we investigated a potential role for MRCK-driven GBM cell invasion in the pathogenesis of GBM. In this study, we demonstrate that MRCK activity is upregulated at the invasive edges of GBM tumors and is further enhanced by irradiation both in vitro and in vivo. We show that this response is essential to the phenomenon of radiation-induced migration of GBM cells in a relevant in vivo model of GBM. Furthermore, by demonstrating complete abrogation of radiation-driven invasion in this model using a novel small-molecule inhibitor of MRCK, we confirm the pivotal role of MRCK in driving radiation-induced infiltration and validate it as a novel and highly promising anti-invasive therapeutic target in GBM.

Cell culture and radiation treatments

E2 and G7 cell lines were obtained from Colin Watts (Cambridge) and are derived from anonymized patient resection specimens as previously described (20, 21). Cell lines were routinely cultured on Matrigel-coated plates (0.23 mg/L in AdDMEM, Life Technologies) in serum-free AdDMEM supplemented with 20 ng/mL EGF, 10 ng/mL FGF, 0.5% B27 supplement, and 0.5% N2 supplement (all Life Technologies), with the exception of E2 cells for ex vivo migration assays, which were cultured in MEM (Life Technologies) with 10% FBS. The commercially available U87MGluc2 line (Caliper Life Sciences) was cultured in MEM EBSS (Life Technologies), 10% FBS, 1% l-glutamine, 1 % NEAA (Life Technologies), and 1% NaPyruvate (Life Technologies).

Cells were irradiated using an Xstrahl RX225 radiation cabinet (195 kV X-rays, dose rate 1.39 Gy/minute).

Cell line testing

All cell lines were used in experiments between passage 2 and passage 12 from thawing and tested for mycoplasma every 3 months, most recent date August 2018. U87MGluc2 line is an authenticated cell line received from Caliper Life Sciences. E2 and G7 cell lines are primary patient-derived cell lines, and as such no separate cell authentication was under taken.

Subconfluent and ex vivo migration assays

For subconfluent migration assays, 2 × 105 cells per well were plated at in 6-well dishes and migration imaged by time-lapse microscopy capturing images every 15 minutes. Migration velocity was calculated using single-cell tracking via ImageJ analysis.

For ex vivo migration assays, cells were seeded onto fresh 1 mm coronal brain slices obtained from 6- to 8-week-old C57BL/6 mice in culture medium (described above) and allowed to establish overnight with incubation at 37°C with 5% CO2. Brain slices were then inverted onto Lumox 35 mm dishes (8 μm, Sarstedt) and secured with Nuclepore Track-Etch membrane (Whatman) sealed with Matrigel. Migration was captured via confocal time-lapse microscopy with images taken every 15 minutes. Migration velocity was calculated using single-cell tracking via ImageJ analysis.

Clonogenic assays

Cells were plated onto 6-well plates in triplicate per biological repeat. Twenty-four hours after plating, cells were treated with DMSO or 250 μmol/L BDP-9066 for 2 hours, followed by radiation treatment and media replacement. Cells were fixed after 2 to 3 weeks, stained with crystal violet, and colonies counted manually. Data were fitted using a linear quadratic model.

siRNA transfection and Western blotting

G7 cells were transfected with 10 nmol/L siRNA targeting MRCKα and/or β or control-scrambled siRNA (Dharmacon; MRCKα AAGAAUAUCUGCUGUGUUU, MRCKβ GAAGAAUACUGAACGAAUU, MRCKα+β CGAGAAGACUUUGAAAUAAUU) using RNAiMAX reagent (Life Technologies) and incubated for 48 hours prior to imaging or protein extraction.

For Western blotting, protein was extracted using 1% SDS/50 mmol/L Tris, pH 6.8, supplemented with protease and phosphatase inhibitors (Roche) and subject to SDS-PAGE and protein transfer. Membranes were probed using antibodies listed in Supplementary Table S1.

Immunofluorescence

Note that 3 × 103 irradiated or untreated E2 cells were plated in each well of a 96-well glass-bottomed plate precoated with Matrigel and allowed to adhere for 2 to 3 hours before replacement of media with media containing DMSO or the indicated amount of BDP-9066 and incubated overnight. Cells were washed with PBS and fixed with 4% paraformaldehyde, followed by permeabilization with 0.1% TritonX/TBS. Cells were washed with TBS-Tween and blocked with 1% BSA/TBST, followed by incubation with 1:200 pMLC2 antibody (Abcam; cat#3675) for 1 hour at room temperature. Cells were washed and incubated with 20 μL/mL Alexa 488 anti-mouse secondary antibody, 1:40 TR phalloidin stain, 1:300 DAPI, and 1:10,000 whole-cell stain and incubated for 1 hour in the dark, followed by final wash steps. Images were captured using the Operetta HTP microscope (Perkin Elmer), and analysis was carried out using Columbus Image Analysis software (Perkin Elmer).

Animal studies

All animal experiments were performed under the relevant UK Home Office Project Licence and carried out with ethical approval from the University of Glasgow under the Animal (Scientific Procedures) Act 1986 and the EU directive 2010. Mice were maintained in individually ventilated cages with environmental enrichment and ARRIVE guidelines followed.

Intracranial tumor model

Female CD1 nude mice were orthotopically injected with 1 × 105 G7 cells into the subventricular zone as previously described (20, 22). Tumors were allowed to establish for 10 to 11 weeks before MRI to confirm presence of tumor. Brain irradiation was performed on an XStrahl Small Animal Radiation Research Platform (SARRP) using parallel opposed beams and a 10 × 10 mm collimator to ensure adequate tumor coverage in all animals. Mice were anaesthetized with isoflurane and immobilized on a cradle with a tooth bar attachment. A CT scan before irradiation allowed tissue segmentation and selection of an isocenter (approximately 5 mm below cell injection drill hole). Dose distributions were uniform across the brain area as shown in Supplementary Fig. S2C. Note that 5 mg/kg BDP-9066 or vehicle (20% propylene glycol/80% PBS) was given subcutaneously twice daily or at stated time before cull for pharmacokinetic (PK) analysis. Tumors were subdissected and fresh-frozen specimens sent for PK analysis (Vertex). Formalin-fixed, paraffin-embedded sections were stained for Ki67, HLA, or phosphoMYPT1 and then scanned using a Hamamatsu Nanozoomer Slide scanning machine with Leica SlidePath Slide imaging software. Algorithms were optimized for each stain individually and automated, quantitative analysis undertaken. The defining of contralateral regions was performed blinded.

MRI data acquisition and postprocessing

The animals were induced in an anesthetic chamber with 5% isoflurane and a 30:70 O2/N2O ratio before being transferred to the MRI instrument animal cradle allowing to monitor respiration and temperature. Imaging experiments were performed on a Bruker Biospec Avance 7 T imaging system with a 30 cm horizontal bore (Bruker). Homogeneous radiofrequency excitation was achieved using a birdcage volume resonator (diameter 72 mm, length 110 mm), and actively decoupled 4-channel phased array receive-only head surface coils were used for signal detection, with 22 mm length for mice and 35 mm length for rats (Rapid Biomedical). After standard spectrometer adjustments and geometry definition using a pilot sequence, T2-weighted imaging was performed using a RARE sequence (Rapid Acquisition with Relaxation Enhancement; echo time = 10 ms, repetition time = 4,300 ms, field of view = 176 × 176 mm, matrix = 176 × 176, slice thickness = 0.5 mm, 14 slices, RARE factor 8, 8 minutes). MRI data were exported in DICOM format for postprocessing using an in-house MATLAB code. Tumor-related abnormal regions were manually selected on each T2 slice. To evaluate tumor volumes, the number of voxels within abnormal regions was multiplied by the voxel volume.

MRCK activity is upregulated at the invasive edges of GBM tumors

Oncomine analysis of MRCK gene expression in GBM revealed modest but significant upregulation of MRCKα mRNA when compared with normal brain tissue (Fig. 1A; refs. 23, 24), suggesting a possible role in GBM pathogenesis. In light of the established role of MRCK in cancer cell motility, we explored whether its activity was spatially regulated within the tumor to support GBM cell infiltration. Indeed, using a phospho-sensitive antibody that detects MRCKalpha autophosphorylation on S1003 (pS1003) as a validated biomarker of kinase activity (pMRCK), immunohistochemical staining of matched patient tumor core and margin samples showed that positivity, and by implication, MRCKalpha activity was largely restricted to invading GBM cells at the tumor margins (Fig. 1B). In support of this observation, immunohistochemical analysis of primary human tumor cells in an intracranial GBM xenograft model revealed that MYPT1, the downstream target of MRCK, also showed increased phosphorylation levels in the cytoplasm of cells at the invasive edges of the tumors (Fig. 1C). These data are consistent with a role for MRCK in supporting GBM tumor cell motility and invasion.

Figure 1.

MRCK activity is upregulated at the invasive edge of GBM tumors. A, Oncomine analysis of data from two separate studies indicating increased MRCKα mRNA expression in clinical GBM samples compared with normal brain tissue. *, P < 0.05. B, Matched tumor margin and tumor core samples were obtained from patients with GBM and stained for an MRCK activity autophosphorylation site, S1003. Cytoplasmic pMRCK levels were quantified using automated analysis on SlidePath. Statistical analysis: two tailed, unpaired t test. **, P < 0.01. C, Example image from samples obtained from a G7 intracranial tumor mouse model stained for pMYPT1, a downstream target of MRCK.

Figure 1.

MRCK activity is upregulated at the invasive edge of GBM tumors. A, Oncomine analysis of data from two separate studies indicating increased MRCKα mRNA expression in clinical GBM samples compared with normal brain tissue. *, P < 0.05. B, Matched tumor margin and tumor core samples were obtained from patients with GBM and stained for an MRCK activity autophosphorylation site, S1003. Cytoplasmic pMRCK levels were quantified using automated analysis on SlidePath. Statistical analysis: two tailed, unpaired t test. **, P < 0.01. C, Example image from samples obtained from a G7 intracranial tumor mouse model stained for pMYPT1, a downstream target of MRCK.

Close modal

The downstream targets of MRCK are upregulated by radiation in vitro and in vivo, and this is associated with an increase in GBM invasion

Because the majority of patients with GBM receive RT, we explored whether MRCK activity was affected by radiation. As shown in Fig. 2A, Western blot analysis of two different primary human GBM cell lines, G7 and E2, showed that radiation induced an increase in phosphorylation levels of the MRCK biomarker, MYPT1. To confirm this observation, immunofluorescence analysis of pMLC2, another downstream biomarker of MRCK, was undertaken using a high-throughput imaging platform and automated analysis (Fig. 2B, i and ii). This unbiased technique clearly indicated a significant increase in pMLC2 levels upon irradiation of GBM cells in vitro. Furthermore, a significant increase in the average pseudopod length was observed in irradiated cells, suggesting that the activation of MLC2 by RT is driving actin-myosin cytoskeletal changes that may promote motility (Fig. 2B, iii).

Figure 2.

MRCK activity is stimulated by radiation in vitro and in vivo. A, Two primary cell lines, E2 and G7, were treated with 0, 2, or 5 Gy and protein lysates extracted after 24 hours. Western blot analysis was undertaken to assay levels of pMYPT1. Actin and tubulin were used as loading controls, and γH2AX as a marker of radiation-induced DNA damage. B, i, E2 cells were treated with 0 or 2 Gy and stained by immunofluorescence for pMLC2. Cells were imaged using an Operetta high-throughput imaging platform. Green, pMLC2; yellow, actin; red, whole cell dye; blue, DAPI; last panel, merge. Scale bar, 100 μm. ii, Automated image analysis was undertaken to compare pMLC2 levels in control and irradiated cells. iii, Pseudopod length was quantified using automated image analysis. Data are derived from two biological repeats, each analyzing >400 cells per condition. Statistical analysis: two tailed, unpaired t test. ****, P < 0.0001. C, Cohorts of mice bearing G7 intracranial tumors were subjected to 3 × 2 Gy fractions of whole brain irradiation or left untreated. Brain sections were stained by IHC for pMYPT1 levels. i–iii, Levels of nuclear and cytoplasmic pMYPT1 were quantified in mice sacrificed 5 days after RT. iv, Levels of cytoplasmic pMYPT1 were measured 12 days after RT. Quantification was done via automated analysis using SlidePath; n = 5 (no RT) and 6 (3 × 2 Gy), 5-day time point; n = 4 (no RT) and 4 (3 × 2 Gy), 12-day time point. Scale bar, 100 μm. Statistical analysis: two tailed, unpaired t test. NS, not significant; *, P < 0.05.

Figure 2.

MRCK activity is stimulated by radiation in vitro and in vivo. A, Two primary cell lines, E2 and G7, were treated with 0, 2, or 5 Gy and protein lysates extracted after 24 hours. Western blot analysis was undertaken to assay levels of pMYPT1. Actin and tubulin were used as loading controls, and γH2AX as a marker of radiation-induced DNA damage. B, i, E2 cells were treated with 0 or 2 Gy and stained by immunofluorescence for pMLC2. Cells were imaged using an Operetta high-throughput imaging platform. Green, pMLC2; yellow, actin; red, whole cell dye; blue, DAPI; last panel, merge. Scale bar, 100 μm. ii, Automated image analysis was undertaken to compare pMLC2 levels in control and irradiated cells. iii, Pseudopod length was quantified using automated image analysis. Data are derived from two biological repeats, each analyzing >400 cells per condition. Statistical analysis: two tailed, unpaired t test. ****, P < 0.0001. C, Cohorts of mice bearing G7 intracranial tumors were subjected to 3 × 2 Gy fractions of whole brain irradiation or left untreated. Brain sections were stained by IHC for pMYPT1 levels. i–iii, Levels of nuclear and cytoplasmic pMYPT1 were quantified in mice sacrificed 5 days after RT. iv, Levels of cytoplasmic pMYPT1 were measured 12 days after RT. Quantification was done via automated analysis using SlidePath; n = 5 (no RT) and 6 (3 × 2 Gy), 5-day time point; n = 4 (no RT) and 4 (3 × 2 Gy), 12-day time point. Scale bar, 100 μm. Statistical analysis: two tailed, unpaired t test. NS, not significant; *, P < 0.05.

Close modal

To confirm that this phenomenon also occurs in vivo, histologic staining and automated analysis of pMYPT1 levels were performed in samples from whole-brain irradiated (3 × 2 Gy fractions) and nonirradiated cohorts of mice bearing intracranial G7 xenograft tumors that were sacrificed 5 days after the final radiation dose. Although no significant change in nuclear pMYPT1 was detected, we observed significant upregulation of cytoplasmic pMYPT1 both at the invasive tumor edge (Fig. 2C, i and ii) and in the tumor core (Fig. 2C, iii). This suggests that MRCK activity is not only upregulated acutely, and throughout the tumor, by irradiation, but may also be maintained by a longer term “switch” in intracellular signaling. In support of this hypothesis, analysis of tumors from smaller cohorts of mice sacrificed 12 days after the final radiation dose also showed significant upregulation of pMYPT1 (Fig. 2C, iv).

Because radiation-induced migration of GBM cells has recently been reported by a number of groups, and MRCK activity is known to have promigratory effects, we questioned whether radiation would affect cell migration. To interrogate this important question, we used complementary experimental approaches in three different model systems. The invasive behavior of cancer cells can be affected by their intrinsic ability to produce the cytoskeletal rearrangements required for forward migration as well as their ability to modify their surroundings to produce a promigratory environment (e.g., by degradation of extracellular matrix). Having observed a radiation-induced increase in MRCK biomarkers, we first used a simple subconfluent migration assay coupled with single-cell tracking to assess whether radiation affects the intrinsic mechanisms of migration, without the complications of extracellular matrix and brain architecture (Fig. 3A). These experiments showed that irradiated cells from two different primary GBM cell lines (E2 and G7) migrated significantly faster than nonirradiated cells, indicating that radiation enhances the mechanical motility of GBM cells.

Figure 3.

Activation of MRCK by radiation is concomitant with increased motility of GBM cells in vitro and in vivo. A, E2 and G7 were treated with 0 or 2 Gy, and their motility analyzed in a subconfluent migration assay using time-lapse microscopy and single-cell tracking. i, Example track plots of individual control and irradiated E2 cells imaged over 16 hours, 15 minute intervals. ii, Comparison of control and irradiated cell speed. B, Fluorescently labeled E2 or G7 cells were irradiated with 2 Gy or left untreated and seeded onto fresh murine brain slices. Cell motility was analyzed using confocal time-lapse microscopy (i). ii, Cell speed was measured using single-cell tracking. Data from three biological replicates. Scale bar, 50 μm. Statistical analysis: Mann–Whitney test; **, P < 0.005; ****, P < 0.0001. C, i, Brain sections from control and irradiated mice bearing G7 intracranial tumors were stained via IHC for Ki67 to indicate presence of cycling GBM tumor cells. ii, Percentage of Ki67-positive cells in the tumor bulk was quantified using automated analysis in specimens from mice culled 10 days after initiation of treatment; n = 6 in both cohorts. The percentage of Ki67-positive cells (iii) or HLC (iv) in the contralateral hemisphere of mice culled 17 days after initiation of treatment was quantified using automated image analysis. Scale bar, 1 mm. Statistical analysis: two tailed, unpaired t test. N.S., not significant; **, P < 0.005.

Figure 3.

Activation of MRCK by radiation is concomitant with increased motility of GBM cells in vitro and in vivo. A, E2 and G7 were treated with 0 or 2 Gy, and their motility analyzed in a subconfluent migration assay using time-lapse microscopy and single-cell tracking. i, Example track plots of individual control and irradiated E2 cells imaged over 16 hours, 15 minute intervals. ii, Comparison of control and irradiated cell speed. B, Fluorescently labeled E2 or G7 cells were irradiated with 2 Gy or left untreated and seeded onto fresh murine brain slices. Cell motility was analyzed using confocal time-lapse microscopy (i). ii, Cell speed was measured using single-cell tracking. Data from three biological replicates. Scale bar, 50 μm. Statistical analysis: Mann–Whitney test; **, P < 0.005; ****, P < 0.0001. C, i, Brain sections from control and irradiated mice bearing G7 intracranial tumors were stained via IHC for Ki67 to indicate presence of cycling GBM tumor cells. ii, Percentage of Ki67-positive cells in the tumor bulk was quantified using automated analysis in specimens from mice culled 10 days after initiation of treatment; n = 6 in both cohorts. The percentage of Ki67-positive cells (iii) or HLC (iv) in the contralateral hemisphere of mice culled 17 days after initiation of treatment was quantified using automated image analysis. Scale bar, 1 mm. Statistical analysis: two tailed, unpaired t test. N.S., not significant; **, P < 0.005.

Close modal

To explore whether this promigratory effect was maintained when GBM cells have to negotiate the complexities of the brain microenvironment, we employed an ex vivo migration assay (Fig. 3B; Supplementary Videos S1 and S2). Fluorescently labeled E2 and G7 GBM cell lines were seeded onto fresh murine brain slices, and their migration speed measured using confocal time-lapse microscopy and single-cell tracking. We tracked >50 cells per condition encompassing all routes of invasion (i.e., perivascular and along white matter tracts) to obtain accurate average migration speeds. We observed that the average speed of cells in this assay was slower that in 2D, suggesting the brain structure poses a barrier to the invading GBM cells and that their migration in this environment requires additional processes such as extracellular matrix remodeling. However, as with the 2D migration assay, we observed that cells irradiated prior to being seeded onto the brain slices moved significantly faster than control cells, indicating that the promigratory effect of radiation is maintained within the brain environment.

Finally, we demonstrated that radiation induces invasion in vivo by quantifying migration of GBM cells away from the primary tumor mass in control and irradiated cohorts of mice implanted with intracranial GBM xenografts (Fig. 3C). Primary human G7 GBM cells injected into the subventricular zone form a tumor mass in the ipsilateral hemisphere with invasive edges. Over time, cells leave the invasive tumor edge and migrate to the contralateral hemisphere. This allows for a simple analysis of the migratory potential of tumors cells by analyzing the number of tumor cells that have reached the contralateral hemisphere. To enable automated and unbiased quantification, brain sections from mice culled 17 days after initiation of radiation treatment were stained for the nuclear proliferation marker Ki67, which discriminates clearly between replicating tumor cells and the nonproliferating cells of the mouse brain (Fig. 3C, i–iii). Mice had undergone T2-weighted MRI to confirm the presence and equivalence of size of tumors 11 weeks after tumor cell injection (Supplementary Fig. S1A). No effect on Ki67 staining within the tumor bulk was detected, indicating that GBM cell proliferation was unaffected by these radiation doses and would therefore not affect quantification of invading cells. In contrast, analysis of the contralateral hemisphere showed significantly higher numbers of Ki67-positive tumor cells in irradiated mice, strongly supporting the concept that radiation promotes infiltration of tumor cells to distant sites within the brain in a relevant in vivo mouse model. Although Ki67 staining provides a clear nuclear stain that simplifies automated analysis of histology sections, a caveat to this approach it that it would fail to detect nonproliferating GBM cells. To validate the data, we also stained sections for the human Leukocyte Antigen (HLA Class I ABC). We found that Ki67 and HLA staining correlated well (Supplementary Fig. S1B) and observed increased numbers of HLA-positive cells in the contralateral hemispheres of the irradiated cohort, confirming the observations made with Ki67 (Fig. 3C, iii).

MRCK plays an essential role in driving radiation-induced migration

To probe whether increased cell migration requires enhanced MRCK activity, and to assess its relative contribution, we performed subconfluent migration assays in G7 cells in which MRCKα and β had been downregulated by siRNA targeting (Fig. 4A and B). Ablation of both MRCK isoforms concomitantly was found to inhibit pMYPT1 expression and to reduce migration speed to the levels observed in nonirradiated control cells. These findings strongly indicate that MRCK, and not ROCK, is primarily responsible for the downstream signaling to MYPT1 and MLC2 that drives radiation-induced migration in GBM cells.

Figure 4.

Inhibition of MRCK activity opposes radiation-driven motility but does not affect cell survival in vitro. A, G7 cells were transfected with siRNAs targeting MRCKα and MRCKβ, alone or in combination. i, Cell lysates were analyzed by Western blotting for MRCKα, MRCKβ, and pMYPT1 levels. ii, Treated cells were exposed to 0 or 2 Gy and imaged in a subconfluent migration assay. Cell speed was measured by single-cell tracking. Statistical analysis: Mann–Whitney test; ns, not significant; *, P < 0.05; **, P < 0.005; ***, P < 0.001. B, Novel MRCK-specific inhibitor, BDP-9066. C, i, Cell viability assays performed on G7 cells treated with increasing concentrations of BDP-9066. Data plotted relative to vehicle control. ii, Clonogenic survival assays performed on G7 and E2 cells irradiated in the presence of vehicle or 0.25 μmol/L BDP-9066. N.S., not significant. D, E2 cells treated with 2 Gy radiation (i) or untreated (ii) in the presence of increasing amounts of BDP-9066, and stained by immunofluorescence for pMLC2. Cells were imaged using an Operetta high-throughput imaging platform. Automated image analysis was used to compare pMLC2 levels. Data from two biological repeats with >400 cells per condition per biological replicate. Data plotted as a percentage of vehicle. E, E2 and G7 cells were exposed to 0 or 2 Gy radiation in the presence of DMSO or BDP-9066, and cell speed measured in a subconfluent migration assay using time-lapse microscopy and single-cell tracking. Data from three biological replicates. F, Fluorescently labeled E2 cells were exposed to 0 or 2 Gy radiation and seeded onto fresh murine brain slices. Cell motility was assayed in the presence of DMSO or BDP-9066 by confocal time-lapse microscopy and single-cell tracking using ImageJ. Data from three biological replicates. For all in vitro and ex vivo motility assays, statistical analysis was performed using Mann–Whitney test. N.S., not significant; *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0001. G, E2 and G7 cells were exposed to 2 Gy in the presence of increasing amounts of BDP-9066 (i) or Y-27632 (ii), followed by time-lapse microscopy and single-cell tracking. Data from three biological replicates. Baseline speed was calculated to be 34 from the measurement of >20 nonmigratory cells.

Figure 4.

Inhibition of MRCK activity opposes radiation-driven motility but does not affect cell survival in vitro. A, G7 cells were transfected with siRNAs targeting MRCKα and MRCKβ, alone or in combination. i, Cell lysates were analyzed by Western blotting for MRCKα, MRCKβ, and pMYPT1 levels. ii, Treated cells were exposed to 0 or 2 Gy and imaged in a subconfluent migration assay. Cell speed was measured by single-cell tracking. Statistical analysis: Mann–Whitney test; ns, not significant; *, P < 0.05; **, P < 0.005; ***, P < 0.001. B, Novel MRCK-specific inhibitor, BDP-9066. C, i, Cell viability assays performed on G7 cells treated with increasing concentrations of BDP-9066. Data plotted relative to vehicle control. ii, Clonogenic survival assays performed on G7 and E2 cells irradiated in the presence of vehicle or 0.25 μmol/L BDP-9066. N.S., not significant. D, E2 cells treated with 2 Gy radiation (i) or untreated (ii) in the presence of increasing amounts of BDP-9066, and stained by immunofluorescence for pMLC2. Cells were imaged using an Operetta high-throughput imaging platform. Automated image analysis was used to compare pMLC2 levels. Data from two biological repeats with >400 cells per condition per biological replicate. Data plotted as a percentage of vehicle. E, E2 and G7 cells were exposed to 0 or 2 Gy radiation in the presence of DMSO or BDP-9066, and cell speed measured in a subconfluent migration assay using time-lapse microscopy and single-cell tracking. Data from three biological replicates. F, Fluorescently labeled E2 cells were exposed to 0 or 2 Gy radiation and seeded onto fresh murine brain slices. Cell motility was assayed in the presence of DMSO or BDP-9066 by confocal time-lapse microscopy and single-cell tracking using ImageJ. Data from three biological replicates. For all in vitro and ex vivo motility assays, statistical analysis was performed using Mann–Whitney test. N.S., not significant; *, P < 0.05; **, P < 0.005; ***, P < 0.001; ****, P < 0.0001. G, E2 and G7 cells were exposed to 2 Gy in the presence of increasing amounts of BDP-9066 (i) or Y-27632 (ii), followed by time-lapse microscopy and single-cell tracking. Data from three biological replicates. Baseline speed was calculated to be 34 from the measurement of >20 nonmigratory cells.

Close modal

Specific targeting of MRCK, but not ROCK, using a novel small-molecule inhibitor, does not affect cell survival but elicits a robust biomarker and motility dose response at submicromolar concentrations

We have developed a novel small-molecule inhibitor, BDP-9066, that displays high selectivity toward MRCKα/β over related kinases such as ROCK (Fig. 4C; Supplementary Table S2; ref. 17). BDP-9066 demonstrates excellent selectivity for MRCKα and MRCKβ over the closely related kinases ROCK1 and ROCK2. Furthermore, BDP-9066 also shows excellent selectivity when screened in a larger kinase panel (1 μmol/L BDP-9066 against 113 kinases; 7/113 kinases show inhibition >80%; ref. 17). The biochemical activity and selectivity of BDP-9066 was maintained in cellular assays.

To test whether BDP-9066 had any impact on tumor cell survival either as a single agent or in combination with radiation, we performed cell viability and clonogenic survival assays. No effect on viability was observed at submicromolar levels of BDP-9066 (Fig. 4C, i), and 250 nmol/L BDP-9066 had no impact on the radiation sensitivity of G7 or E2 cells as measured by clonogenic survival (Fig. 4C, ii) despite compound being in excess of the IC50 indicated in Fig. 4D, i. The loss of cell viability at higher concentrations is likely to represent off-target toxicity. Confirmation that BDP-9066 is inhibiting its target is provided by the robust dose response of the MRCK biomarker pMLC2 in irradiated GBM cells as measured by immunofluorescence (IC50 = 56 nmol/L; Fig. 3D, i). The response of nonirradiated GBM cells to the inhibitor was highly variable, suggesting a lack of synchronicity of MRCK activity in unirradiated cells. In contrast, when we tested BDP-9066 in our in vitro and ex vivo migration assays, we found that the compound completely blocked the radiation-induced increase in cell motility observed in subconfluent migration assays at a concentration of 100 nmol/L (Fig. 4E). This anti-invasive activity of BDP-9066 was also confirmed using the organotypic ex vivo brain slice assay (Fig. 4F). Further experiments showed a robust dose response of irradiated cells to BDP-9600 in subconfluent migration assays (Fig. 4G, i). Importantly, treatment with the ROCK-specific inhibitor Y27632 did not inhibit migration of irradiated cells (Fig. 4G, ii), providing further evidence that the proinvasive effect of RT is primarily driven through MRCK activation.

Pharmacologic inhibition of MRCK inhibits radiation-induced infiltration by GBM cells in vivo

Compound-level analysis of subdissected intracranial U87MG and G7 tumors indicated that BDP-9066 penetrated intracranial tumors at concentrations that would be expected to inhibit of MRCK in vivo, even when adjusted for free drug levels (based on plasma protein binding of 27.3%; Fig. 5A). Samples taken from the contralateral hemisphere showed that BDP-9066 has very low exposure in the brain in the absence of significant tumor burden, following subcutaneous dosing. To test the efficacy of BDP-9066 in inhibiting radiation-induced invasion in vivo, we established G7 intracranial tumors in four cohorts of mice: vehicle, BDP-9066, RT + vehicle, and RT + BDP-9066. Mice underwent T2-weighted MRI to confirm the presence and equivalence of size of tumors 11 weeks after tumor cell injection (Supplementary Fig. S2A); treatment was commenced the following week. Mice in the RT cohorts received 3 × 2 Gy whole brain RT over the course of 1 week (Supplementary Fig. S2B). Twice-daily dosing of vehicle or BDP-9066 was initiated at the same time as RT and continued for 10 days before sacrifice of the animals to ensure sustained inhibition of MRCK (Fig. 5B). Twice-daily dosing was chosen because, although the PK profile of BDP-9066 showed good bioavailability, rapid clearance was also observed (17). The compound was well tolerated, and PK analysis of blood taken at the time of culling confirmed its presence at micromolar concentrations (Supplementary Fig. S2C). A transient vasodilatory effect lasting 20 to 30 minutes was noted in some animals as evidenced by reddening of the ears.

Figure 5.

BDP-9066 penetrates intracranial GBM xenografts and inhibits radiation-induced infiltration of GBM cells in vivo. A, Mice bearing U87MG (i) or G7 (ii) intracranial tumors were injected subcutaneously with 5 mg/kg BDP-9066 30 minutes prior to cull. Tumors were subdissected from normal brain tissue and analyzed by mass spectrometry to determine total compound levels (“Total”). These levels were adjusted using a determined PPB-free value of 72.7% to estimate available BDP-9066 levels in the tumors (“Free”); n = 4 (U87MG) and 10 (G7). CL, contralateral. B, Outline of experiment to measure in vivo GBM cell response to BDP-9066. G7 intracranial tumors were allowed to establish for 12 weeks before initiation of treatment. C, Cohorts of mice from B were culled and excised brains subjected to IHC for Ki67, followed by automated analysis to determine extent of contralateral hemisphere invasion by GBM cells. *, P < 0.05.

Figure 5.

BDP-9066 penetrates intracranial GBM xenografts and inhibits radiation-induced infiltration of GBM cells in vivo. A, Mice bearing U87MG (i) or G7 (ii) intracranial tumors were injected subcutaneously with 5 mg/kg BDP-9066 30 minutes prior to cull. Tumors were subdissected from normal brain tissue and analyzed by mass spectrometry to determine total compound levels (“Total”). These levels were adjusted using a determined PPB-free value of 72.7% to estimate available BDP-9066 levels in the tumors (“Free”); n = 4 (U87MG) and 10 (G7). CL, contralateral. B, Outline of experiment to measure in vivo GBM cell response to BDP-9066. G7 intracranial tumors were allowed to establish for 12 weeks before initiation of treatment. C, Cohorts of mice from B were culled and excised brains subjected to IHC for Ki67, followed by automated analysis to determine extent of contralateral hemisphere invasion by GBM cells. *, P < 0.05.

Close modal

Mouse brains were excised and subjected to immunohisochemical analysis of Ki67 staining. Consistent with our previous experiment, increased numbers of GBM cells were observed in the contralateral hemispheres of mice in the “irradiated + vehicle” cohort (Fig. 5C). Importantly, irradiated mice that were treated with BDP-9066 showed no increase of tumor cell infiltration to the contralateral hemisphere. These data confirm that BDP-9066 prevents radiation-induced infiltration of GBM in vivo. Immunohistologic staining for the MRCK biomarker pMYPT1 (Fig. 5C, ii) showed reduced expression in the RT + BDP-9066 cohort compared with RT + vehicle, indicating that the anti-invasive effects of the inhibitor are likely “on target.”

BDP-9066 induces an aberrant morphology through the disruption of the actin-myosin cytoskeleton in GBM cells

In vitro and ex vivo time-lapse imaging of cells treated with BDP-9066 revealed the emergence of an aberrant morphology characterized by increased numbers of neurite-like structures emanating from the cell body that appeared to prevent migration (Supplementary Videos S3–S6). To characterize this further, high-throughput imaging was performed on irradiated E2 cells in which the actin cytoskeleton had been stained with fluorescently labeled phalloidin (Fig. 6A). Cells treated with BDP-9066 exhibited a highly disordered cytoskeletal structure, indicating that MRCK inhibition disrupts normal actin-myosin dynamics, leading to the aberrant morphology and reduced migratory potential observed. To explore this further, automated analysis was undertaken to measure the number of neurite roots in response to escalated doses of BDP-9066 in both irradiated and untreated cells (Fig. 6B). Neurite numbers increased in a dose-dependent manner under both conditions, but as with the pMLC2 biomarker response (Fig. 4D), data from nonirradiated cells were highly variable compared with irradiated cells. This provides further evidence that radiation synchronizes MRCK activity across the cell population, rendering more cells sensitive to inhibition by BDP-9066 and eliciting a more uniform and more striking response.

Figure 6.

BDP-9066 treatment disrupts the actin-myosin cytoskeleton, inducing an aberrant morphology that inhibits migration. A, i, E2 cells were exposed to 0 or 2 Gy radiation in the presence of DMSO or BDP-9066. Cells were imaged using an Operetta high-throughput imaging platform. Red, actin; blue, DAPI. Scale bar, 100 μm. ii, Automated image analysis was undertaken to quantify changes in neurite morphology. Data from >1,000 cells. B, Radiation stimulates MRCK activity, activating MLC2 and driving actin-myosin contractility and subsequent cell migration. Treatment with BDP-9066 inhibits MRCK activity, resulting in a disrupted actin-myosin cytoskeleton and loss of motility.

Figure 6.

BDP-9066 treatment disrupts the actin-myosin cytoskeleton, inducing an aberrant morphology that inhibits migration. A, i, E2 cells were exposed to 0 or 2 Gy radiation in the presence of DMSO or BDP-9066. Cells were imaged using an Operetta high-throughput imaging platform. Red, actin; blue, DAPI. Scale bar, 100 μm. ii, Automated image analysis was undertaken to quantify changes in neurite morphology. Data from >1,000 cells. B, Radiation stimulates MRCK activity, activating MLC2 and driving actin-myosin contractility and subsequent cell migration. Treatment with BDP-9066 inhibits MRCK activity, resulting in a disrupted actin-myosin cytoskeleton and loss of motility.

Close modal

Treatment with BDP-9066 significantly increases survival in a preclinical model of GBM

To assess whether inhibiting the invasive capacity of GBM cells in vivo confers a survival benefit, we performed an extended efficacy experiment using the G7 intracranial tumor model (Fig. 7A). Tumors were allowed to establish for 9 weeks before the mice underwent T2-weighted MRI to establish tumor establishment and enable randomization of mice into cohorts, stratified by tumor size. Treatment commenced at week 10 on mice randomized into 4 cohorts: no RT + vehicle, no RT + BDP-9066, RT + vehicle, and RT + BDP-9066. RT was fractionated into 6 × 2 Gy over 2 weeks. BDP-9066 or vehicle (blinded) was administered twice daily, Monday–Friday, for a total of 4 weeks (2 weeks concomitant with and 2 weeks adjuvant after RT). Treatment was limited to 4 weeks under veterinary guidance to reduce the stress of repeat dosing on the animals. Mice were culled upon presentation of clinical symptoms (weight loss, seizures, tilting, etc.). Only mice that were culled after the first week of treatment were included in the survival analysis to provide adequate time for treatment to take effect. The experiment was terminated at 104 days after T2-weighted MRI scans had been performed on the remaining mice.

Figure 7.

Combining BDP-9066 with RT confers a significant survival advantage on mice bearing intracranial GBM tumors. A, Outline of experiment to measure in vivo survival to BDP-9066 treatment. G7 intracranial tumors were allowed to establish for 10 weeks before initiation of treatment. Mice in the RT cohorts received 6 × 2 Gy whole brain RT over the course of 2 weeks. Only mice that were culled after the first three fractions of RT were included in the analysis to allow for any treatment benefit to take effect. The study was randomized, blinded, and the mice stratified across the cohorts based on starting tumor size as assessed by T2 MRI. B, Kaplan–Meier plot showing survival data. N.S., not significant; *, P < 0.05; **, P < 0.005. Statistical analysis: Log-rank (Mantel–Cox) test. C, Example histology from RT + vehicle and RT + BDP-9066 cohorts.

Figure 7.

Combining BDP-9066 with RT confers a significant survival advantage on mice bearing intracranial GBM tumors. A, Outline of experiment to measure in vivo survival to BDP-9066 treatment. G7 intracranial tumors were allowed to establish for 10 weeks before initiation of treatment. Mice in the RT cohorts received 6 × 2 Gy whole brain RT over the course of 2 weeks. Only mice that were culled after the first three fractions of RT were included in the analysis to allow for any treatment benefit to take effect. The study was randomized, blinded, and the mice stratified across the cohorts based on starting tumor size as assessed by T2 MRI. B, Kaplan–Meier plot showing survival data. N.S., not significant; *, P < 0.05; **, P < 0.005. Statistical analysis: Log-rank (Mantel–Cox) test. C, Example histology from RT + vehicle and RT + BDP-9066 cohorts.

Close modal

The results clearly indicate that, although RT alone provided a survival benefit, this was significantly enhanced when combined with BDP-9066, despite the inhibitor only being administered for a restricted period of time (Fig. 7B). In addition, although quantitation of contralateral invasion could not be applied in this experiment as the mice were culled at varying time points, histologic analysis indicates that tumors from the RT + BDP-9066 cohort are more contained with less invasive margins than RT + vehicle cohort (Fig. 7C). Indeed, qualitative assessment of invasion extent showed that 8 of 14 RT + vehicle mice showed extensive contralateral invasion compared with only 3 of 13 RT+ BDP-9066 mice. The variation in cull time also means that any potential effect on tumor size could not be assessed in this experiment as mice were culled at clinical endpoint and thereby all likely to be carrying a significant tumor burden. These data provide important evidence that adding a novel MRCK targeting, anti-invasive small-molecule treatment to RT has the potential to improve outcomes in this cancer of unmet need by negating the adverse, proinvasive effects of radiation.

GBM is the most aggressive primary brain tumor in adults and the least responsive to treatment. These adverse clinical features are partly attributable to the infiltrative nature of the disease. GBM cells undergo a number of biological and morphologic changes that allow them to migrate through the perivascular spaces and white matter tracts of the brain. The role that RhoGTPases and their effector kinases play in these processes has been interrogated by a number of researchers (12), and inhibition of CDC42, the upstream activator of MRCK, has been shown to inhibit migration of glioma cells (25–27). However, little is known about the effects of ionizing radiation on these pathways and the subsequent downstream effects on migration and invasion. In addition, by augmenting actin-myosin cytoskeleton plasticity through targeting the upstream regulator MRCK, we are hitting a regulatory node that is a convergence point for a number of signaling pathways that have previously shown to be important in GBM invasion such as integrin and FAK signaling (28–30).

Analysis of MRCK gene expression using Oncomine datasets indicated that, although expression is raised in tumor samples compared with the normal brain, this increase is modest. We postulated that overall gene expression levels might fail to reflect more pronounced upregulation at the tumor margins, where increased MRCK activity may be required to support tumor cell infiltration of the healthy brain tissue. Such spatial restriction of gene expression has been documented for other promigratory factors such as EGFR (31, 32). By measuring MRCK-specific autophosphorylation at Ser1003 in patient-derived tumor core and tumor margin samples, we observed that MRCK activity is indeed specifically upregulated in infiltrating GBM cells, indicating an important influence of this signaling pathway on invasion. Furthermore, we demonstrated that exposure to radiation further increases MRCK signaling to drive increased migration both in vitro and in a clinically relevant intracranial GBM tumor model. Interestingly, this radiation-induced activation of MRCK extended into the tumor core, whereas in the absence of radiation, MRCK activation biomarkers were observed only at the tumor margin. This indicates that MRCK activation by RT is not restricted to cells at the tumor margin and therefore may have potential to drive other biological processes within the tumor core.

The question of whether radiation promotes the invasive behavior of GBM cells has been controversial for many years. However, recent publications from a number of research groups (5–7) and the data presented here provide robust evidence in support of this phenomenon and highlight the potential value of incorporating anti-invasive strategies into standard first-line therapy for GBM. Although the observation that most GBMs recur within the maximally irradiated volume (33–35) has been used as an argument against radiation-induced invasion being of clinical relevance, an alternative explanation is that local tumor recurrence is caused by repopulation of the irradiated tumor bed by tumor cells from outside the irradiated volume. Evidence for repopulation of laser ablated regions of tumor by GBM cells was recently presented by Osswald and colleagues (36) and may represent a novel and important mechanism of treatment resistance. Such a model also is supported by the observation that a significant proportion of tumors recur at the periphery of the irradiated volume (33). This mechanism of tumor recurrence would require surviving GBM cells to retain their invasive capacity and would also be consistent with the well-documented failure of radiation dose escalation to either reduce tumor recurrence rates or reduce the incidence of recurrence within the irradiated volume (8, 37, 38).

Our data support the theory that novel anti-invasive chemotherapeutics for GBM should always be evaluated in the context of RT, not only because of the proinvasive effects of radiation, but also to ensure that candidate compounds are active against tumor cells whose signaling pathways have been fundamentally altered by RT. Our results show that although ablation of MRCK activity effectively inhibited the migration and infiltration of irradiated cells, there was a reduced effect on nonirradiated control cells in vitro and in vivo. Furthermore, our results indicate that radiation-induced effects on GBM cells in vivo are not transient but can persist for days after the conclusion of RT.

Our study utilized a novel, highly specific inhibitor of MRCK, BDP-9066, that was developed in-house by the CRUK Beatson Drug Discovery Programme. Treating primary GBM cells with this compound in vitro revealed that MRCK inhibition induces a highly aberrant morphology characterized by disordered, neurite-like protrusions. Not only do these morphologic changes appear to prevent GBM cells from migrating persistently, they may have additional implications for GBM biology and resistance to therapy. Winkler and colleagues recently published an elegant study that identified a network of communicating glioma cell protrusions, termed “tumour microtubes (TM)”, which contributed to chemo- and radioresistance in vivo (36, 39–41). Because TM formation is highly dependent on cytoskeletal dynamics and BDP-9066 specifically induces aberrant protrusions, it is possible that MRCK inhibition will exert therapeutic effects in vivo by disrupting the TM communication network.

As part of our in vivo studies, we demonstrated that BDP-9066 penetrates intracranial tumos in two different mouse models of GBM at levels that far exceeded the biomarker IC50 values calculated for irradiated GBM cell lines. Drug levels in the contralateral hemisphere were much lower, reflecting limited penetration of the intact blood–brain barrier (BBB), which may be of value in protecting normal brain cells and reducing toxicity. Conversely, delivery of BDP-9066 to regions of low tumor cell density may be compromised by limited BBB disruption, although evidence exists that suggests that focal disruption of the BBB can be elicited by single invading glioma cells (42). To maximize efficacy, it may be advantageous to develop the current series of compounds to include a more BBB-penetrant molecule that would penetrate areas of very low tumor burden. Most importantly, our results demonstrate conclusively that BDP-9066 completely blocks the proinvasive effects of radiation on GBM cells, preventing them from disseminating to intracranial sites distant from the primary tumor; this effect translated in to a survival benefit when BDP-9066 was given in combination with RT.

As mentioned previously, the development of effective ROCK inhibitors for clinical use has been hindered by severe adverse effects on the cardiovascular system. Although we observed a mild vasodilatory effect in some animals, to date, we have no evidence that MRCK inhibition would elicit a similarly intolerable response to ROCK inhibition, but such a possibility should be carefully considered during optimization of MRCK-specific compounds for clinical use.

This study did not extend to investigating the combination of BDP-9066 in combination with both RT and TMZ. Because TMZ is a component of standard of care for patients with GBM, it will be important to determine whether it modulates the interaction between RT and MRCK inhibition as this novel treatment strategy progresses toward the clinic.

In conclusion, our data provide novel and persuasive evidence that delivery of BDP-9066 or a related compound alongside RT has potential not only to extend survival of patients with GBM, but also to improve the devastating neurologic symptoms associated with these infiltrative tumors.

No potential conflicts of interest were disclosed.

Conception and design: J.L. Birch, L. Gilmour, D.R. Croft, D. McArthur, J. Bower, H.J. McKinnon, M. Drysdale, A.J. Chalmers

Development of methodology: J.L. Birch, K. Strathdee, D.R. Croft, C.H. Gray, M. Mezna, D. McArthur, M. Sime, H.J. McKinnon, A.J. Chalmers

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.L. Birch, K. Strathdee, A. Vallatos, L. McDonald, A. Kouzeli, R. Vasan, A.H. Qaisi, D. Crighton, K. Gill, C.H. Gray, J. Konczal, P. McConnell, W.M. Holmes, H.J. McKinnon, A.J. Chalmers

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.L. Birch, K. Strathdee, A. Vallatos, A. Kouzeli, A.H. Qaisi, P. McConnell, W.M. Holmes, J. Bower, H.J. McKinnon, A.J. Chalmers

Writing, review, and/or revision of the manuscript: J.L. Birch, K. Strathdee, A. Vallatos, A.H. Qaisi, W.M. Holmes, J. Bower, H.J. McKinnon, A.J. Chalmers

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Strathdee, L. Gilmour, L. McDonald, H.J. McKinnon

Study supervision: J.L. Birch, J. Bower, H.J. McKinnon, M. Drysdale, M.F. Olson, A.J. Chalmers

Other (synthesis of small-molecule BDP-9066): K. Gill

Other (contributed to the design and characterization of the small-molecule inhibitors used in this study): A.W. Schüttelkopf

This work was funded by a grant awarded by The Brain Tumour Charity (grant ref. number 26/160; J.L. Birch, A.J. Chalmers, K. Strathdee, L. Gilmour, and A. Vallatos) and in part by Cancer Research UK (M. Drysdale, J. Bower, H.J. Mckinnon, L. McDonald, D.R. Croft, K. Gill, C.H. Gray, J. Konczal, M. Mezna, D. McArthur, A.W. Schüttelkopf, P. McConnell, and M. Sime) and the Beatson Endowment Fund (J.L. Birch). Primary GBM cell lines and patient GBM tumor samples were obtained from Colin Watts, Cambridge.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Stupp
R
,
Mason
WP
,
van den Bent
MJ
,
Weller
M
,
Fisher
B
,
Taphoorn
MJ
, et al
Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma
.
N Engl J Med
2005
;
352
:
987
96
.
2.
Walker
MD
,
Alexander
E
 Jr
,
Hunt
WE
,
MacCarty
CS
,
Mahaley
MS
 Jr
,
Mealey
J
 Jr
, et al
Evaluation of BCNU and/or radiotherapy in the treatment of anaplastic gliomas. A cooperative clinical trial
.
J Neurosurg
1978
;
49
:
333
43
.
3.
Wild-Bode
C
,
Weller
M
,
Rimner
A
,
Dichgans
J
,
Wick
W
. 
Sublethal irradiation promotes migration and invasiveness of glioma cells: implications for radiotherapy of human glioblastoma
.
Cancer Res
2001
;
61
:
2744
50
.
4.
Zhai
GG
,
Malhotra
R
,
Delaney
M
,
Latham
D
,
Nestler
U
,
Zhang
M
, et al
Radiation enhances the invasive potential of primary glioblastoma cells via activation of the Rho signaling pathway
.
J Neurooncol
2006
;
76
:
227
37
.
5.
Edalat
L
,
Stegen
B
,
Klumpp
L
,
Haehl
E
,
Schilbach
K
,
Lukowski
R
, et al
BK K+ channel blockade inhibits radiation-induced migration/brain infiltration of glioblastoma cells
.
Oncotarget
2016
;
7
:
14259
78
.
6.
Kegelman
TP
,
Wu
B
,
Das
SK
,
Talukdar
S
,
Beckta
JM
,
Hu
B
, et al
Inhibition of radiation-induced glioblastoma invasion by genetic and pharmacological targeting of MDA-9/Syntenin
.
Proc Natl Acad Sci U S A
2017
;
114
:
370
5
.
7.
Xiong
Y
,
Ji
W
,
Fei
Y
,
Zhao
Y
,
Wang
L
,
Wang
W
, et al
Cathepsin L is involved in X-ray-induced invasion and migration of human glioma U251 cells
.
Cell Signal
2017
;
29
:
181
91
.
8.
Lee
SW
,
Fraass
BA
,
Marsh
LH
,
Herbort
K
,
Gebarski
SS
,
Martel
MK
, et al
Patterns of failure following high-dose 3-D conformal radiotherapy for high-grade astrocytomas: a quantitative dosimetric study
.
Int J Radiat Oncol Biol Phys
1999
;
43
:
79
88
.
9.
Hall
A
. 
The cytoskeleton and cancer
.
Cancer Metastasis Rev
2009
;
28
:
5
14
.
10.
Bishop
AL
,
Hall
A
. 
Rho GTPases and their effector proteins
.
Biochem J
2000
;
348
Pt 2
:
241
55
.
11.
Pandya
P
,
Orgaz
JL
,
Sanz-Moreno
V
. 
Modes of invasion during tumour dissemination
.
Mol Oncol
2017
;
11
:
5
27
.
12.
Khalil
BD
,
El-Sibai
M
. 
Rho GTPases in primary brain tumor malignancy and invasion
.
J Neurooncol
2012
;
108
:
333
9
.
13.
Croft
DR
,
Olson
MF
. 
Regulating the conversion between rounded and elongated modes of cancer cell movement
.
Cancer Cell
2008
;
14
:
349
51
.
14.
Rath
N
,
Olson
MF
. 
Rho-associated kinases in tumorigenesis: re-considering ROCK inhibition for cancer therapy
.
EMBO Rep
2012
;
13
:
900
8
.
15.
Unbekandt
M
,
Croft
DR
,
Crighton
D
,
Mezna
M
,
McArthur
D
,
McConnell
P
, et al
A novel small-molecule MRCK inhibitor blocks cancer cell invasion
.
Cell Commun Signal
2014
;
12
:
54
.
16.
Unbekandt
M
,
Olson
MF
. 
The actin-myosin regulatory MRCK kinases: regulation, biological functions and associations with human cancer
.
J Mol Med
2014
;
92
:
217
25
.
17.
Unbekandt
M
,
Belshaw
S
,
Bower
J
,
Clarke
M
,
Cordes
J
,
Crighton
D
, et al
Discovery of potent and selective MRCK inhibitors with therapeutic effect on skin cancer
.
Cancer Res
2018
;
78
:
2096
114
.
18.
Tysnes
BB
,
Mahesparan
R
. 
Biological mechanisms of glioma invasion and potential therapeutic targets
.
J Neurooncol
2001
;
53
:
129
47
.
19.
Alonso
MM
,
Diez-Valle
R
,
Manterola
L
,
Rubio
A
,
Liu
D
,
Cortes-Santiago
N
, et al
Genetic and epigenetic modifications of Sox2 contribute to the invasive phenotype of malignant gliomas
.
PLoS One
2011
;
6
:
e26740
.
20.
Ahmed
SU
,
Carruthers
R
,
Gilmour
L
,
Yildirim
S
,
Watts
C
,
Chalmers
AJ
. 
Selective inhibition of parallel DNA damage response pathways optimizes radiosensitization of glioblastoma stem-like cells
.
Cancer Res
2015
;
75
:
4416
28
.
21.
Fael Al-Mayhani
TM
,
Ball
SL
,
Zhao
JW
,
Fawcett
J
,
Ichimura
K
,
Collins
PV
, et al
An efficient method for derivation and propagation of glioblastoma cell lines that conserves the molecular profile of their original tumours
.
J Neurosci Methods
2009
;
176
:
192
9
.
22.
Gomez-Roman
N
,
Stevenson
K
,
Gilmour
L
,
Hamilton
G
,
Chalmers
AJ
. 
A novel 3D human glioblastoma cell culture system for modeling drug and radiation responses
.
Neuro-oncol
2017
;
19
:
229
41
.
23.
Bredel
M
,
Bredel
C
,
Juric
D
,
Harsh
GR
,
Vogel
H
,
Recht
LD
, et al
Functional network analysis reveals extended gliomagenesis pathway maps and three novel MYC-interacting genes in human gliomas
.
Cancer Res
2005
;
65
:
8679
89
.
24.
Bredel
M
,
Bredel
C
,
Juric
D
,
Harsh
GR
,
Vogel
H
,
Recht
LD
, et al
High-resolution genome-wide mapping of genetic alterations in human glial brain tumors
.
Cancer Res
2005
;
65
:
4088
96
.
25.
Yiin
JJ
,
Hu
B
,
Jarzynka
MJ
,
Feng
H
,
Liu
KW
,
Wu
JY
, et al
Slit2 inhibits glioma cell invasion in the brain by suppression of Cdc42 activity
.
Neuro Oncol
2009
;
11
:
779
89
.
26.
Furukawa
K
,
Kumon
Y
,
Harada
H
,
Kohno
S
,
Nagato
S
,
Teraoka
M
, et al
PTEN gene transfer suppresses the invasive potential of human malignant gliomas by regulating cell invasion-related molecules
.
Int J Oncol
2006
;
29
:
73
81
.
27.
Malchinkhuu
E
,
Sato
K
,
Horiuchi
Y
,
Mogi
C
,
Ohwada
S
,
Ishiuchi
S
, et al
Role of p38 mitogen-activated kinase and c-Jun terminal kinase in migration response to lysophosphatidic acid and sphingosine-1-phosphate in glioma cells
.
Oncogene
2005
;
24
:
6676
88
.
28.
Reyes
SB
,
Narayanan
AS
,
Lee
HS
,
Tchaicha
JH
,
Aldape
KD
,
Lang
FF
, et al
αvβ8 integrin interacts with RhoGDI1 to regulate Rac1 and Cdc42 activation and drive glioblastoma cell invasion
.
Mol Biol Cell
2013
;
24
:
474
82
.
29.
Webb
DJ
,
Donais
K
,
Whitmore
LA
,
Thomas
SM
,
Turner
CE
,
Parsons
JT
, et al
FAK-Src signalling through paxillin, ERK and MLCK regulates adhesion disassembly
.
Nat Cell Biol
2004
;
6
:
154
61
.
30.
Kwiatkowska
A
,
Kijewska
M
,
Lipko
M
,
Hibner
U
,
Kaminska
B
. 
Downregulation of Akt and FAK phosphorylation reduces invasion of glioblastoma cells by impairment of MT1-MMP shuttling to lamellipodia and downregulates MMPs expression
.
Biochim Biophys Acta
2011
;
1813
:
655
67
.
31.
Snuderl
M
,
Fazlollahi
L
,
Le
LP
,
Nitta
M
,
Zhelyazkova
BH
,
Davidson
CJ
, et al
Mosaic amplification of multiple receptor tyrosine kinase genes in glioblastoma
.
Cancer Cell
2011
;
20
:
810
7
.
32.
Okada
Y
,
Hurwitz
EE
,
Esposito
JM
,
Brower
MA
,
Nutt
CL
,
Louis
DN
. 
Selection pressures of TP53 mutation and microenvironmental location influence epidermal growth factor receptor gene amplification in human glioblastomas
.
Cancer Res
2003
;
63
:
413
6
.
33.
Weber
DC
,
Casanova
N
,
Zilli
T
,
Buchegger
F
,
Rouzaud
M
,
Nouet
P
, et al
Recurrence pattern after [(18)F]fluoroethyltyrosine-positron emission tomography-guided radiotherapy for high-grade glioma: a prospective study
.
Radiother Oncol
2009
;
93
:
586
92
.
34.
Chen
L
,
Chaichana
KL
,
Kleinberg
L
,
Ye
X
,
Quinones-Hinojosa
A
,
Redmond
K
. 
Glioblastoma recurrence patterns near neural stem cell regions
.
Radiother Oncol
2015
;
116
:
294
300
.
35.
Minniti
G
,
Amelio
D
,
Amichetti
M
,
Salvati
M
,
Muni
R
,
Bozzao
A
, et al
Patterns of failure and comparison of different target volume delineations in patients with glioblastoma treated with conformal radiotherapy plus concomitant and adjuvant temozolomide
.
Radiother Oncol
2010
;
97
:
377
81
.
36.
Osswald
M
,
Jung
E
,
Sahm
F
,
Solecki
G
,
Venkataramani
V
,
Blaes
J
, et al
Brain tumour cells interconnect to a functional and resistant network
.
Nature
2015
;
528
:
93
8
.
37.
Fitzek
MM
,
Thornton
AF
,
Rabinov
JD
,
Lev
MH
,
Pardo
FS
,
Munzenrider
JE
, et al
Accelerated fractionated proton/photon irradiation to 90 cobalt gray equivalent for glioblastoma multiforme: results of a phase II prospective trial
.
J Neurosurg
1999
;
91
:
251
60
.
38.
Chan
JL
,
Lee
SW
,
Fraass
BA
,
Normolle
DP
,
Greenberg
HS
,
Junck
LR
, et al
Survival and failure patterns of high-grade gliomas after three-dimensional conformal radiotherapy
.
J Clin Oncol
2002
;
20
:
1635
42
.
39.
Osswald
M
,
Solecki
G
,
Wick
W
,
Winkler
F
. 
A malignant cellular network in gliomas: potential clinical implications
.
Neuro-oncol
2016
;
18
:
479
85
.
40.
Weil
S
,
Osswald
M
,
Solecki
G
,
Grosch
J
,
Jung
E
,
Lemke
D
, et al
Tumor microtubes convey resistance to surgical lesions and chemotherapy in gliomas
.
Neuro-oncol
2017
;
19
:
1316
26
.
41.
Winkler
F
. 
Tumour network in glioma
.
ESMO Open
2016
;
1
:
e000133
.
42.
Watkins
S
,
Robel
S
,
Kimbrough
IF
,
Robert
SM
,
Ellis-Davies
G
,
Sontheimer
H
. 
Disruption of astrocyte-vascular coupling and the blood-brain barrier by invading glioma cells
.
Nat Commun
2014
;
5
:
4196
.