Methods for the Determination of Invasive Ability of Brain Tumour Cells and for the Diagnosis and Prognosis of Brain Tumour

Abstract

The present invention concerns a novel in vitro assay for determining the invasive ability of brain tumour cells, in particular brain tumour stem cells. This new assay, called 3D invasion assay, is based on the analysis and comparison of the area of neurospheres grown in vitro into a Matrigel extracellular-like matrix. The invention also concerns the use of said assay for diagnosing metastatic brain tumour in subjects, for determining its prognosis, as well as for the monitoring, for the stratification, for identifying subjects at risk of metastasis, and/or predicting the metastasis-associated risk in subjects diagnosed with brain tumour.

Claims

1.-16. (canceled)

17. An in vitro method for determining the invasive ability of glioblastoma stem cells, comprising the steps of: a) obtaining a neurosphere from a brain tumour primary cell culture, said neurosphere containing at least 75% of glioblastoma stem cells; b) embedding the neurosphere of step a) in a protein-based extracellular-like matrix, c) measuring the area of the embedded neurosphere of step b); d) incubating the embedded neurosphere of step c) at a temperature ranging from to 39 C., for 1 h to 72 h; e) measuring the area of the embedded neurosphere after step d); and f) determining the invasive ability of the brain tumour cells by comparing the area measured at step c) to the area measured at step e).

18. The method of claim 17, wherein the protein-based extracellular-like matrix comprises at least one basement membrane component, and wherein the protein-based extracellular-like matrix optionally further comprises at least one of the following components: at least one growth factor, Heparan sulfate proteoglycan.

19. The method of claim 17, wherein the solution of the protein-based extracellular-like matrix comprises at least 5 mg/mL of proteins.

20. The method of claim 17, wherein step b) of embedding comprises two steps b1) and b2), wherein step b1) comprises incorporating the neurosphere in a solution of the protein-based extracellular-like matrix; and step b2) comprises incubating said solution of step b1) at a temperature ranging from 35 C. to 39 C., until the protein-based extracellular-like matrix solution is polymerized to form the protein-based matrix.

21. The method of claim 17, wherein the solution of the protein-based extracellular-like matrix comprises a basement component extract.

22. The method of claim 17, further comprising the step of obtaining a brain tumour primary cell culture containing stem cells from a tumour sample from a subject prior to step a).

23. The method of claim 17, further comprising a step of obtaining a brain tumour primary cell culture containing stem cells from a tumour sample from a subject prior to step a) and a step of cultivating said primary cell culture in an appropriate stem cell culture medium, in appropriate conditions so that said culture eventually contains at least 75% of glioblastoma stem cells, prior to step a).

24. The method of claim 17, wherein the glioblastoma stem cells are diffuse midline glioma stem cells (DMG stem cells).

25. The method of claim 17, wherein step f) comprises calculating the invasion score of the brain tumour cells, wherein the invasion score is the ratio of the area of the embedded neurosphere measured at step e) to the area of the embedded neurosphere measured at step c).

26. The method of claim 17, wherein step f) comprises calculating the invasion score of the brain tumour cells, wherein the invasion score is the ratio of the area of the embedded neurosphere measured at step e) to the area of the embedded neurosphere measured at step c), and wherein the risk that the brain tumour cells have a high invasive ability is higher than 75%, when the calculated invasion score is equal to or higher than the invasion score of reference glioma metastatic cells.

27. The method of claim 17, wherein step f) comprises calculating the invasion score of the brain tumour cells, wherein the invasion score is the ratio of the area of the embedded neurosphere measured at step e) to the area of the embedded neurosphere measured at step c), and wherein the chances that the brain tumour cells have no invasive/metastatic ability are higher than 90%, when the calculated invasion score is lower than the invasion score of reference non-metastatic glioma cells.

28. An in vitro method for diagnosing a metastatic brain tumour in a subject and/or for the prognosis, for the monitoring, for the stratification, for identifying subjects at risk of metastasis, and/or for predicting the metastasis-associated risk in a subject diagnosed with a brain tumour, comprising the steps of: a) obtaining a neurosphere from a brain tumour primary cell culture, said neurosphere containing at least 75% of glioblastoma stem cells; b) embedding the neurosphere of step a) in a protein-based extracellular-like matrix, c) measuring the area of the embedded neurosphere of step b); d) incubating the embedded neurosphere of step c) at a temperature ranging from to 39 C., for 1 h to 72 h; e) measuring the area of the embedded neurosphere after step d); and f) determining the invasive ability of the brain tumour cells by comparing the area measured at step c) to the area measured at step e).

29. The method of claim 28, wherein step f) comprises calculating the invasion score of the brain tumour cells, wherein the invasion score is the ratio of the area of the embedded neurosphere measured at step e) to the area of the embedded neurosphere measured at step c).

30. The method of claim 28, wherein step f) comprises calculating the invasion score of the brain tumour cells, wherein the invasion score is the ratio of the area of the embedded neurosphere measured at step e) to the area of the embedded neurosphere measured at step c), and wherein the subject is diagnosed with a metastatic brain tumour, and/or the prognosis of the brain tumour is poor, the brain tumour has worsened, the subject is stratified and/or identified as a subject at risk of metastasis, and/or the metastasis-associated risk in the subject is high, when the invasion score of the glioma cells is equal to or higher than the invasion score of reference metastatic glioma cells.

31. The method of claim 28, wherein the subject has been diagnosed with diffuse midline glioma (DMG).

32. The method of claim 28, for selecting a therapy for treating said subject, or for assessing the efficacy of a therapy in a treated subject, or for adapting a therapy in a treated subject.

33. The method of claim 17, wherein the protein-based extracellular-like matrix comprises at least one basement membrane component selected from the group consisting of Laminin, Collagen, Nidogen, and any combination thereof.

34. The method of claim 17, wherein the protein-based extracellular-like matrix comprises at least one growth factor selected from the group consisting of: Transforming Growth Factor Beta (TGF-beta), Epidermal Growth Factor (EGF), Insulin-like Growth Factor 1 (IGF-1), Basic Fibroblast Growth Factor (bFGF), Tissue Plasminogen Activator, Platelet Derived Growth Factor (PDGF), and any combination thereof.

35. The method of claim 17, wherein the solution of the protein-based extracellular-like matrix comprises a basement component extract obtained from an Engelbreth-Holm-Swarm (EHS) mouse sarcoma.

Description

DESCRIPTION OF THE FIGURES

[0194] FIG. 1: Protein-based extracellular-like matrix (Matrigel) invasion assay and quantification. Example of imaging and quantification of the area of one neurosphere immediately after embedding in the extracellular-like matrix (A), 24 hours (B) and 40 hours (C) post-embedding. The considered area of the sphere is outlined in grey, post-acquisition. Only cells in physical contact with the structure are considered within the invasion area, whereas single cells or cluster are excluded. Images were acquired using an EVOS m5000 microscope (Thermo Fisher) equipped with a long working distance 4 objective and high-resolution camera. Scale bars correspond to 600 m.

[0195] FIG. 2: Protein-based extracellular-like matrix (Matrigel) invasion assay and prediction of metastasis risk. Panel A shows the difference between the invasiveness of biopsy-derived cells in 20 patients with and without metastasis during the course of the disease. Panel B shows the dispersion of the same results: the models with a high invasion coefficient are mostly those derived from patients who developed metastasis later. Panel C represents the results of a Receiver Operating Characteristic (ROC) analysis. This analysis is used to guide the observer in choosing a threshold that represents the best compromise between sensitivity and specificity of the analysis. Area under the curve (AUC) indicates a good (close to excellent) accuracy of the test in predicting metastatic relapse.

[0196] FIG. 3: Comparison of three methodologies for evaluating the metastatic risk with biopsy-derived models at diagnosis. Panel A shows the percentage of migrating cells 40 hours after seeding in biopsy-derived cell models from 13 patients with and without metastasis during the course of the disease. Panel B shows the dispersion of data obtained using a protein-based extracellular-like matrix (Matrigel) invasion assay in a cohort of 13 patient-derived models 40 hours post-embedding; threshold value defining the two groups (high and low invasion) was determined using a k-means analysis. Panel C shows the dispersion of migration rate (as a percentage of migrating cells 40 hours post seeding) obtained using a boyden-like migration assay for the same 13 patient-derived cell models; threshold value defining the two groups (high and low invasion) was determined using a k-means analysis. Panel D: Scratch-wound assay and metastasis risk. The graph shows dynamic of wound recovering (as a percentage of cell density) within 48 hours after seeding in biopsy-derived cell models from 6 patients with and without metastasis during the course of the disease. The results shown the lack of any correlation between cell migration as measured with this assay and clinical progression in those patients.

[0197] FIG. 4: Comparison of the test of the invention performed by two different operators, with two different batches of Matrigel lot.

[0198] A) invasion ability relative to the initial sphere size 40 hours after embedding (top panel); results represent the average+/s.e.m. of at least four technical replicates per cell model; bottom plot displays the radiological progression of patients from whom models are derived; metastatic displayed is represented in black, local- to locoregional progression are in grey. Matrigel lot used: 9322009

[0199] B) Comparison of invasion score at 40 hours after embedding for models derived from patients with local/locoregional (13) or distant metastatic progression (9); average+/s.e.m is shown for both groups; p-value as by two-tailed unpaired t-test.

[0200] C) ROC analysis performed on the results displayed in panel A; area under the curve (AUC) and statistical test are indicated.

[0201] D) Invasion ability relative to the initial sphere size 40 hours after embedding (top panel); results represent the average+/s.e.m. of at least five technical replicates per cell model; bottom plot displays the radiological progression of patients from whom models are derived; metastatic displayed is represented in black, local-to locoregional progression are in grey. Matrigel lot used: 0254001

[0202] E) Comparison of invasion score at 40 hours after embedding for models derived from patients with local/locoregional (13) or distant metastatic progression (9); average+/s.e.m is shown for both groups; p-value as by two-tailed unpaired t-test.

[0203] F) ROC analysis performed on the results displayed in panel D; area under the curve (AUC) and statistical test are indicated.

[0204] G) Pearson correlation of invasion score 40 h post-embedding in the 2 independent dataset displayed in panels A and D; Information relative to Matrigel lot used in the two experiments are reported on the axis, Pearson correlation and p-values are shown.

EXAMPLES

[0205] Although the present invention herein has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present invention. It is therefore to be understood that numerous modifications may be made to the illustrative embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims.

Example

1. Materials and Methods

1.1. General Presentation of the Method

[0206] The present Inventors have developed an original three-dimensional invasion assay that can be used to efficiently and accurately determine and measure the invasive ability of brain tumour cells, in particular glioma cells, more particularly glioma stem cells (GSCs). This innovative invasion assay has been successfully used on GSCs directly derived from stereotactic biopsies collected from patients at diagnosis. The results from this assay show a correlation with the risk of metastatic progression in DIPG patients for whom primary GSCs model were available.

[0207] Briefly, primary models of tumour stem cells were isolated from biopsies obtained by stereotactic biopsies performed at the Necker-Enfant Malades Hospital and cultured in a medium suitable for neural stem cell maintenance (Neurocult complete with bFGF, EGF and PDGF). The cultures were used at early passaging (p1-p4) for the formation of neurospheres for 72 hours. Those neurospheres were embedded in a protein extracellular-like matrix (e.g. Matrigel at an initial concentration>10 mg/ml), and the area of invasion was measured by microscopy at different times. The invasion score at 40 h post-inclusion, corresponding to the ratio between the surface area at 40 h and the surface area of the sphere after inclusion (at the initial time), was correlated with the clinical data obtained in the patients.

1.2. Establishment and Propagation of GSC Primary Models from Patients:

[0208] Tissue collection from newly-diagnosed patients was systematically performed by trained neuro-surgeons in Necker-Enfants Malades hospital (Paris, France) under informed consent. Tumour biopsy was preserved in DMEM with 100 U ml.sup.1 penicillin and 100 g ml.sup.1 of streptomycin (GIBCO).

[0209] Within 24 hours post-surgery, tumours were mechanically dissociated by vigorous pipetting into serum-free medium to obtain a single cell suspension.

[0210] Culture flasks or dishes were laminin-coated (Invitrogen) at least 3 hours before cell plating. Cells were then cultured in these supports as an adherent monolayer using a NeuroCult NS-A medium with proliferation supplement (STEMCELL Technologies), also supplemented with human-basic FGF (20 ng m.sup.1, STEMCELL Technologies), human EGF (20 ng ml.sup.1, STEMCELL Technologies), PDGF-AA (10 ng ml.sup.1, Miltenyi Biotec), PDGF-BB (10 ng ml.sup.1, Miltenyi Biotec) and heparin (2 g ml.sup.1, STEMCELL Technologies).

[0211] Medium was completely renewed every 2-3 days and passaging performed when cells were at 70% confluence, using Accutase (GIBCO) to detach cells. Alternatively, cells can be frozen gradually in complete medium supplemented with 10% dimethyl sulfoxide (DMSO, Sigma-Aldrich).

1.3. Neurosphere Formation and Transferring to Matrigel Extracellular Matrix Droplets

[0212] Invasion assay was performed on patient-derived GSC models at passaging 1-4, using Matrigel basal matrix (Corning) at initial concentration superior to 10 mg/ml of protein.

[0213] Cells were detached with Accutase, and counted manually using a hemocytometer.

[0214] A volume of cells suspension equivalent to 20.000 cell per experimental replicate is centrifuged 400 g and resuspended in a volume equivalent to 100 l of complete medium per experimental replicates. A volume of 100 l of cells suspension (corresponding to 20.000 cells) is then dispatched into at least 24 wells in a U-bottom ultra-low attachment 96-wells plate (Corning). It is preferred that cells do not adhere to plastic.

[0215] Cells were let form neurospheres during 48-72 hours, until the spheres appear round-shaped with smooth edges.

[0216] The neurospheres are centrifuged at 200 g (unity of relative centrifugal force, RCF) to remove excess of medium, and resuspended in 5 l of fresh complete medium. A volume of 25 l of ice-cold Matrigel is added at a concentration of at least 10 mg/ml. The neurosphere-containing Matrigel droplets are incubated at 37 C. for 25 min, to let the Matrigel polymerize.

[0217] By using surgical forceps, each Matrigel droplet containing a neurosphere is carefully transferred into a well of a 24-well containing 1 ml of complete medium. Image acquiring is performed immediately after.

1.4. Imaging and Post-Acquisition Analysis.

[0218] Size measurement was performed using an inverted microscope equipped with a camera at 5 magnification. Initial images were acquired immediately after neurosphere embedding in Matrigel, and then 24 and 40 hours after. This short timing is intended to reduce the confounding effect of cell proliferation over invasion, since all GSC models analysed have a doubling-time longer than 48 hours (10).

[0219] Once collection of images is complete for a given patient-derived model, images are analysed using a measurement software (e.g. Fiji) to determine the area of the neurosphere, just after embedding in Matrigel and at later time points (FIG. 1 shows examples of such acquired images). Quantification is performed on multiple technical replicates to reduce experimental variation. Invasive ability at a given time point is determined as a ratio representing the average relative neurosphere area.

[0220] The invasion score is calculated with the following formula:

Invasion score=(Area at 24 or 40 hours)/(Area at 0 hours)

[0221] The in vitro results are then compared to the clinical data available for the patients.

1.6. Receiver Operating Characteristic (ROC)

[0222] Analysis was performed using Prism GraphPad statistics software, version 7a. This type of curve plays a central role in evaluating the ability of tests to discriminate the true state of subjects, finding the optimal cut off values. The area under the curve (AUC) represents a measure of the accuracy, frequently used for evaluating diagnostic test (16).

1.7. Chemotaxis (Boyden-Like) Assay (Technique of the Prior Art)

[0223] Both top and bottom parts of specific ClearView 96-well plates (4582, EssenBioscience) were coated with laminin (Thermo Fisher scientific). Primary gliomas stem cells (GSCs) prepared as in 1.2. were seeded at 1,000-2,000 cells per well in the top plate in presence of Neurocult complete medium deprived of growth factors. Cells were allowed to settle down for 30 minutes. Then the top plate was inserted onto the bottom reservoir plate containing complete Neurocult medium with growth factors (EGF, FGF, and PDGF). As an experimental negative control, the reservoir plate was filled with medium deprived of growth factors. Finally, cell migration was recorded by video-microscopy using a Zoom IncuCyte (EssenBioscience) for three days. Post-treatment analysis of acquired images was performed using the Incucyte analysis software (IncuCyte Zoom 2015A, EssenBioscience).

1.8. Scratch-Wound Assay (Technique of the Prior Art)

[0224] 96-well plates were coated with laminin. GSC were seeded at a density ranging from 50,000 to 66,000 cells/well. After seeding, cells were allowed to settle down and attach for 4 hours at 37 C. in an incubator at 20% oxygen. Then, plates were placed into the wound maker tool (EssenBioscience) and the scratch made using the 96-well comb. Wells were washed with cold D-PBS to remove debris and 100 l of medium containing growth factors (EGF, FGF, and PDGF) was added to each well. The wound healing closure was recorded by video-microscopy (IncuCyte Zoom, EssenBioscience) for 3 days and the post-treatment of acquired images was performed using the software provided by the manufacturer (IncuCyte Zoom 2015A, EssenBioscience). Percentage of migration was analyzed by the percentage of confluence of mask of cells.

2. Results

[0225] 2.1. A cohort of 20 patients was used for the Protein-based extracellular-like matrix (Matrigel) invasion assay. Results show that the invasion score at 40 h post-inclusion, corresponding to the ratio between the surface area at 40 h and the surface area of the neurosphere after inclusion (at the initial time), was correlated with the clinical data obtained in the patients (FIG. 2). The invasion area is significantly higher in models derived from patients with metastasis than in patients without metastasis (FIG. 2A). The models with a high invasion coefficient are mostly those derived from patients who developed metastasis later (FIG. 2B).

[0226] Table 1 and FIG. 2C represent the results of a Receiver Operating Characteristic (ROC) analysis. Accuracy of the test measured by the area under the ROC curve (AUC) was 0.889, where an AUC=1 represents a perfect test, whereas an AUC=0.5 represents a worthless test. This analysis is used to guide the observer in choosing a threshold that represents the best compromise between sensitivity and specificity of the analysis.

TABLE-US-00001 TABLE 1 sensitivity/specificity table generated using Prism-Graphpad 7. Invasion Likelihood score Sensitivity(%) 95% CI Specificity(%) 95% CI ratio >1.882 100 66.37% to 9.091 0.2299% to 1.1 100% 41.28% >2.326 100 66.37% to 18.18 2.283% to 1.222 100% 51.78% >2.726 100 66.37% to 27.27 6.022% to 1.375 100% 60.97% >3.279 100 66.37% to 36.36 10.93% to 1.571 100% 69.21% >3.783 100 66.37% to 45.45 16.75% to 1.833 100% 76.62% >4.241 100 66.37% to 54.55 23.38% to 2.2 100% 83.25% >4.541 100 66.37% to 63.64 30.79% to 2.75 100% 89.07% >5.007 100 66.37% to 72.73 39.03% to 3.667 100% 93.98% >5.39 100 66.37% to 81.82 48.22% to 5.5 100% 97.72% >5.838 88.89 51.75% to 81.82 48.22% to 4.889 99.72% 97.72% >6.721 77.78 39.99% to 81.82 48.22% to 4.278 97.19% 97.72% >7.257 66.67 29.93% to 81.82 48.22% to 3.667 92.51% 97.72% >7.315 55.56 21.2% to 81.82 48.22% to 3.056 86.3% 97.72% >7.694 55.56 21.2% to 90.91 58.72% to 6.111 86.3% 99.77% >8.107 44.44 13.7% to 90.91 58.72% to 4.889 78.8% 99.77% >8.355 33.33 7.485% to 90.91 58.72% to 3.667 70.07% 99.77% >9.201 22.22 2.814% to 90.91 58.72% to 2.444 60.01% 99.77% >10.41 22.22 2.814% to 100 71.51% to 60.01% 100% >12.07 11.11 0.2809% to 100 71.51% to 48.25% 100%

[0227] Quantification showed a strong heterogeneity in invasiveness (invasion score between 2.4 and 13.2) and a remarkable correlation between in vitro data and clinical profiles. In particular, models derived from patients who developed metastases during their progression showed a very high invasiveness and significantly different from models derived from patients without metastatic progression. For a representative cohort of 20 primary models, a ROC analysis showed a sensitivity (percentage of true positives) of 100% and a specificity (percentage of true negatives) of 81.8%.

[0228] Indeed, table 1 shows that by imposing a minimum threshold of 5.39 it is possible to predict the totality (9/9) of patients who have developed metastasis, with a specificity>80% (in the present cohort, only two out of eleven patients who have not developed metastasis would be considered at risk).

[0229] To confirm that this assay is more sensitive and more specific in predicting metastasis risk than the tests classically used for determining cell migration or invasion abilities (Chemotaxis assay, also called Boyden-like assay, and scratch-wound assay), the results obtained with these 3 different assays have been compared for a cohort of 13 patients.

[0230] FIG. 3 shows that the invasion assay (FIG. 3B) provided a higher correlation between cell invasion/migration and clinical progression in those patients than the boyden-like migration assay (FIG. 3C) and scratch-wound assay (FIG. 3D). For both boyden-like migration assay and scratch-wound assay, sensitivity and specificity of the technique in detecting risk of metastasis were significantly reduced compared to extracellular matrix-like invasion test.

2.2. Invasion Assay is Predictive on the Metastasis-Associated Risk in DIPG Patients.

[0231] The present characterization has been extended to a larger number of models derived from patient biopsies (22 in total). Also, the experiments were reproduced twice by two different operators.

[0232] The results displayed on FIG. 4A to C describe the results obtained by Operator 1 using the Matrigel lot 9322009; the results displayed on panel D to F describe the results obtained by Operator 2 using the Matrigel lot 0254001.

[0233] The conclusions of this complementary study are as follows: [0234] 1) The models derived from tumors that have progressed to metastasis are generally the most invasive (panels A, B, D and E) [0235] 2) The invasion test is highly predictive of the risk of metastasis and represents an excellent compromise between sensitivity and specificity. [0236] 3) The results of the invasion test in terms of sensitivity and specificity (ROC analysis, panels 4C, 4F) to assay the individual risk of metastatic evolution are reproducible even if the operator and the batch of Matrigel vary.

[0237] FIG. 4G shows that all the results obtained by 2 different operators with 2 different batches of Matrigel products correlate very significantly. This means that the test can be performed by any scientist or technician if the experimental protocol detailed here is carefully respected.

3. Discussion

[0238] The data show that the 3D invasion assay developed by the Inventors is an efficient, reliable and rapid tool for the prediction of the risk of metastasis development in patients with midline infiltrating gliomas (MIG). The time to obtain an invasion score with this method can vary between 3 and 6 weeks from biopsy receipt due to the highly variable growth rate of the cell models after culture. This delay is compatible with the adaptation of the treatment regimen. The analyses suggest that an invasion score higher than the threshold determined by a Receiver Operating Characteristic (ROC) test in the in vitro assay performed under the conditions described in the protocol indicates a high risk (sensitivity close to 100%, specificity higher than 80%) for the patient to develop one or more metastases upon relapse.

BIBLIOGRAPHIC REFERENCES

[0239] 1. Sturm D, Witt H, Hovestadt V, Khuong-Quang D-A, Jones D T W, Konermann C, et al. Hotspot Mutations in H3F3A and IDH1 Define Distinct Epigenetic and Biological Subgroups of Glioblastoma. Cancer Cell. 2012; 22:425-37. [0240] 2. Louis D N, Perry A, Reifenberger G, von Deimling A, Figarella-Branger D, Cavenee W K, et al. The 2016 World Health Organization Classification of Tumors of the Central Nervous System: a summary. Acta Neuropathol. 2016; 131:803-20. [0241] 3. Chen R, Smith-Cohn M, Cohen A L, Colman H. Glioma Subclassifications and Their Clinical Significance. Neurotherapeutics. 2017; 14:284-97. [0242] 4. Caretti V, Bugiani M, Freret M, Schellen P, Jansen M, van Vuurden D, et al. Subventricular spread of diffuse intrinsic pontine glioma. Acta Neuropathol. 2014; 128:605-7. [0243] 5. Castel D, Philippe C, Calmon R, Le Dret L, Truffaux N, Boddaert N, et al. Histone H3F3A and HIST1H3B K27M mutations define two subgroups of diffuse intrinsic pontine gliomas with different prognosis and phenotypes. Acta Neuropathol. 2015; 130:815-27. [0244] 6. Werbrouck C, Evangelista C C S, Lobn-Iglesias M-J, Barret E, Le Teuff G, Merlevede J, et al. TP53 pathway alterations drive radioresistance in Diffuse Intrinsic Pontine Gliomas (DIPG). Clin Cancer Res. 2019; [0245] 7. Kluiver T A, Alieva M, van Vuurden D G, Wehrens E J, Rios A C. Invaders Exposed: Understanding and Targeting Tumor Cell Invasion in Diffuse Intrinsic Pontine Glioma. Front Oncol. 2020; 10:92. [0246] 8. Liang C-C, Park A Y, Guan J-L. In vitro scratch assay: a convenient and inexpensive method for analysis of cell migration in vitro. Nat Protoc. 2007; 2:329-33. [0247] 9. Guy J-B, Espenel S, Vallard A, Battiston-Montagne P, Wozny A-S, Ardail D, et al. Evaluation of the Cell Invasion and Migration Process: A Comparison of the Video Microscope-based Scratch Wound Assay and the Boyden Chamber Assay. J Vis Exp [Internet]. 2017 [cited 2020 Jul. 29]; Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5755419/ [0248] 10. Plessier A, Le Dret L, Varlet P, Beccaria K, Lacombe J, Mriaux S, et al. New in vivo avatars of diffuse intrinsic pontine gliomas (DIPG) from stereotactic biopsies performed at diagnosis. Oncotarget. 2017; 8:52543-59. [0249] 11. Reynolds B A, Weiss S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science. American Association for the Advancement of Science; 1992; 255:1707-10. [0250] 12. Vinci M, Molinari V, Taylor K, Carvalho D, Burford A, Mackay A, et al. HG-99A PATIENT-DERIVED PAEDIATRIC HIGH GRADE GLIOMA AND DIPG CELL CULTURE PANEL RECAPITULATING THE GENOTYPIC AND PHENOTYPIC DIVERSITY OF THE DISEASE. Neuro Oncol. 2016; 18:iii71. [0251] 13. Vinci M, Burford A, Molinari V, Kessler K, Popov S, Clarke M, et al. Functional diversity and cooperativity between subclonal populations of pediatric glioblastoma and diffuse intrinsic pontine glioma cells. Nat Med. 2018; 24:1204-15. [0252] 14. Pollard S M. In vitro expansion of fetal neural progenitors as adherent cell lines. Methods Mol Biol. 2013; 1059:13-24. [0253] 15. Pollard S M, Yoshikawa K, Clarke I D, Danovi D, Stricker S, Russell R, et al. Glioma Stem Cell Lines Expanded in Adherent Culture Have Tumor-Specific Phenotypes and Are Suitable for Chemical and Genetic Screens. Cell Stem Cell. 2009; 4:568-80. [0254] 16. Hajian-Tilaki K. Receiver Operating Characteristic (ROC) Curve Analysis for Medical Diagnostic Test Evaluation. Caspian J Intern Med. 2013; 4:627-35. [0255] 17. Baskaran J P, Weldy A, Guarin J, Munoz G, Shpilker P H, Kotlik M, Subbiah N, Wishart A, Peng Y, Miller M A, Cowen L, Oudin M J. Cell shape, and not 2D migration, predicts extracellular matrix-driven 3D cell invasion in breast cancer. APL Bioeng. 2020 May 5; 4(2):026105 [0256] 18. Prager, Briana C., Shruti Bhargava, Vaidehi Mahadev, Christopher G. Hubert, and Jeremy N. Rich. 2020. Glioblastoma Stem Cells: Driving Resilience through Chaos. Trends in Cancer 6 (3): 223-35. https://doi.org/10.1016/j.trecan.2020.01.009 [0257] 19. Werbrouck et al, TP53 Pathway Alterations Drive Radioresistance in Diffuse Intrinsic Pontine Gliomas (DIPG), Clinical Cancer Research 2019 Nov. 15; 25(22):6788-6800