PMC

A. Higher expression of PLK1 in EGFRvIII+ glioblastoma specimens relative to EGFRvIII- specimens. p = 0.01. B. Cyclin-normalized PLK1 expression did not correlate with PDGF-β mRNA expression. p = 0.67. C. Scatter plot showing a positive correlation between the PLK1 expression (x-axis) and DNA damage accumulation signature score (y-axis) in TCGA clinical glioblastoma dataset. D. Similar correlation in REMBRANDT clinical glioblastoma dataset. Linear regression was performed to generate the best-fitting line as indicated by the black line. Pearson Correlation Coefficient (R²) and p-value is as indicated.

A. (left) Enhanced p-PLK1 and γH2AX in Ink4a/Arf(−/−) EGFRvIII tumors. 3 representative tissue samples resected from murine Ink4a/Arf(−/−), Ink4a/Arf(−/−) EGFRvIII or Ink4a/Arf(−/−) PDGF-β, separately, were lysed and analyzed by immunoblotting using antibodies against p-PLK1, γH2AX, EGFRvIII and PDGF-β. Tubulin was loaded as loading control. (right) Quantitative densitometric assessment of pT210 PLK1 or γH2AX was normalized to tubulin and fold change represent the average of 3 samples, shown as mean±SD. *, p < 0.05; **, p < 0.01. B. BI2536 treatment exerted greater tumoricidal effect in Ink4a/Arf(−/−) EGFRvIII tumors compared to Ink4a/Arf(−/−) PDGF-β tumors. Ink4a/Arf(−/−) EGFRvIII or Ink4a/Arf(−/−) PDGF-β cells were implanted into nu/nu mice and the treatment started when the tumor size was > 500 mm³. Tumor growth was monitored twice per week and represented separately to represent the comparison between BI2536 or TMZ treatment.

A. (left) Representative immunoblots of U87MG parental (p) and U87MG EGFRvIII (vIII) cells for pS14 Rad51, Rad51, and Ku86. asyn, asynchronized cells; R-8 h, synchronized cells released from DTB for 8 h. (right) pS14Rad51 was normalized to the total Rad51 after correcting for protein loading using Ku86 level. **, p = 0.0026; *, p = 0.045. B. (left) Representative immunoblot of whole U87MG EGFRvIII cell lysate following 24 h treatment with control or 10 nM BI2536. (right) Quantitative densitometric assessment as above described. **, p = 0.0018.C. (left) Schematic summary of the DR-GFP assay. (right) Percentage of GFP-positive cells detected by FACS using U2OS DR-GFP and U87MG DR-GFP cells, separately. GFP-positive cells were scored after treatment with BI2536 (25 nM) or control for 24 h. **, p = 0.00056 and 0.00057 respectively. D. The effects of RAD51 and BRCA2 knockdown in U87MG parental and U87MG EGFRvIII cells. Cells were transfected with the various siRNAs for 24 h and re-plated overnight. BI2536 (25 nM) or control were then added. Clonogenic survivals were scored after additional 14 days. All results were shown as mean±SD.

A. U87MG EGFRvIII cells were treated with TMZ (100 μM) for 24 h followed by the addition of BI2536 at 12 nM. Clonogenic survival was measured after 14 days. B. Murine Ink4a/Arf(−/−) astrocytes and derived EGFRvIII cells were treated with TMZ (50 μM, 24h) followed by BI2536 at 5 nM. Clonogenic survivals were assessed as above described. *, p = 0.015; **, p = 0.0001. C. Growth curve of subcutaneous U87MG parental (left) and U87MG EGFRvIII (right) xenografts in nude mice. The xenograft harboring mice were treated with control, TMZ (for 3 days starting treatment at Day 7 after tumor implantation), BI2536 (starting at Day 13), or a combination of TMZ and BI2536 (T+B) (TMZ starting at Day 7 and BI2536 starting at Day 13). D. (left) Survival curve of murine Ink4a/Arf(−/−) EGFRvIII intracranial allografts bearing mice treated with control, TMZ (starting at Day 10), BI2536 (starting at Day 13), or combination (T+B) (starting with TMZ at Day 10, then with BI2536 at Day 13). (right) p values derived from survival comparisons. 5-6 mice per group for each in vivo experiment. All results were shown as mean±SD. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

A. (left) Representative immunoblots of U87MG and U87MG EGFRvIII cells treated with vehicle or 25 nM BI2536 for 24 h. (right) Quantitative densitometric assessment of γH2AX level normalized to β-actin. **, p = 4.92×10⁻⁵ and 0.0012 respectively. B. (left) Representative comet staining images in U87MG EGFRvIII cells and U87MG parental cells after BI2536 treatment (25 nM for 24 h). (right) Quantitative assessment of comet tail moment. **, p = 0.01; *, p = 0.037. C. U87MG EGFRvIII cells and U87MG parental cells were stained for γH2AX and pH3 after 24 h treatment with 5 nM BI2536 or control. (left) Representative immunofluorescence staining. (right) Quantitation of γH2AX foci in pH3+ (M) and pH3- (I) cells. Approximately 100 cells were scored. **, p = 0.0063; *, p = 0.035 and 0.036 for vehicle control respect to BI2536 treated samples correspondingly. D. (left) Representative images of aberrant mitotic progression in U87MG EGFRvIII H2B-GFP cells. Time elapsed was indicated in the right lower corner (h:min). Arrow marks a lagging chromosome (see arrow at hour 2:00). (right) Frequencies of aberrant mitotic events in U87MG parental and U87MG EGFRvIII cells. Length of mitosis was scored from the onset of prophase to the onset of anaphase. Approximately 100 cells were scored, respectively. E. Aberrant multipolar and monopolar mitotic spindles in the U87MG EGFRvIII cells. (left) U87MG parental and U87MG EGFRvIII cells were stained for α-tubulin and the centrosome protein Cep192. (right) Quantification of bipolar, multipolar, and monopolar mitotic figures in U87MG parental and U87MG EGFRvIII cells.

A. Schematic depiction of siRNA library screen. B. The top siRNA targets that, when silenced, were preferentially toxic to the U87MG EGFRvIII cells relative to the U87MG parental cells. Red: DDR genes involved in HR. The “% increase in cytotoxicity” was calculated based on the mean of two independent experiments. C. (left) Immunoblot of PLK1 following knockdown with two independent siRNAs, siPLK1-1 (si-1) and siPLK1-2 (si-2). Negative control siRNA is indicated as n. Whole cell lysates were collected 48 h after siRNA transfection. (right) Clonogenic survival following PLK1 siRNA transfection in U87MG parental and U87MG EGFRvIII cells. **, p = 0.0092 and 0.0019 respectively. D. Effect of BI2536 on murine Ink4a/Arf(−/−) EGFRvIII cells and parental Ink4a/Arf(−/−) astrocytes. Clonogenic survival was determined after 14 days treatment. **, p = 0.0027 and 0.0036 respectively. E. (left upper) Schematic depiction of cell synchronization by double thymidine blocking (DTB). (left lower) Cell cycle distribution of U87MG parental and U87MG EGFRvIII cells after DTB. (middle) Representative immunoblots of whole cell lysates derived from synchronized and asynchronous U87MG parental (p) and U87MG EGFRvIII (vIII) cells. (right) Quantitative densitometric assessment of pT210 PLK1 was normalized to the total PLK1 after correcting for protein loading using β-actin level respectively. asyn, asynchronized cells; R-8 h, synchronized cells released from DTB for 8 h. *, p = 0.044; **, p = 0.0014. F. (left upper) Cell cycle distribution of U178MG tet-EGFRvIII cells with or without doxycycline (Dox) treatment at 1 μg/mL for 96 h; (left lower) Representative immunoblots of U178MG tet-EGFRvIII cell lysates with or without Dox treatment; (right) Quantitative densitometric assessment of pT210 PLK1 as above described. **, p = 0.0058. The densitometric results represent the average of three experiments, shown as mean±SD.

A. (upper) Clonogenic survival of murine Ink4a/Arf(−/−) EGFRvIII cells (parental) and established EGFR inhibitor resistant cells (erlotinib-resistant E4, E5 and gefitinib-resistant G1, G5, G12, GR-1, GR-7, GR-11) with BI2536 treatment. (lower) Representative colony formation images. B. (upper) Tumor growth curve of the subcutaneous GR-7 allografts. Nude mice bearing established GR-7 tumors in the flank were treated with control, TMZ (for 3 days starting treatment at Day 10 after implantation), BI2536 (starting at Day 13), or combined (T+B) (starting with TMZ at Day 10, then with BI2536 at Day 13). Mean tumor volume±SD are shown in 5-6 mice per group. (lower) Typical tumors isolated from each group. C. (left) Survival curve of intracranial G12 allografts bearing mice. The mice were treated with control, TMZ (starting at Day 10), BI2536 (starting at Day 13), or combined (T+B) (starting with TMZ at Day 10, then with BI2536 at Day 13) in 5-6 mice per group. (right) p values derived from survival comparisons. *, p < 0.05; **, p < 0.01; ***, p < 0.001. D. Tumor growth of the subcutaneous Ink4a/Arf(−/−) EGFRvIII allografts. Nude mice bearing established Ink4a/Arf(−/−) EGFRvIII tumors in the flank were treated as indicated in Methods. T, TMZ; B, BI2536; G, Gefitinib. Mean tumor volume±SD are shown in 5-6 mice per group. E. Schematic representation of “multi-orthogonal” approach. Upper panel: EGFRvIII expressing glioblastomas adapt to EGFR inhibition (EGFRi) by activation of alternative oncogenic signaling cascade, such as ones mediated by the urokinase receptor (uPAR). Other resistance mechanisms involving activation of cytoplasmic proteins, such as Src, have also been reported []. Despite the change in oncogenic signaling, the intrinsic physiological “architecture” of the transformed cells and the cellular dependence on DDR remain largely unaltered. Red arrow: EGFRvIII signaling. Yellow arrow: EGFR inhibition induced up-regulation of uPAR signaling. Bottom panel: As such, simultaneous inhibition of DDR and EGFR inhibition impose independent and parallel selection against glioblastoma cells. The therapeutic efficacy of the regimen is further magnified induction of additional DNA damage by temozolomide (TMZ), a DNA alkylating agent and the standard-of-care chemotherapy for glioblastomas.