ATM kinase inhibition preferentially sensitizes p53-mutant glioma to ionizing radiation

Abstract

Purpose: Glioblastoma multiforme (GBM) is the most lethal form of brain cancer with a median survival of only 12 to 15 months. Current standard treatment consists of surgery followed by chemoradiation. The poor survival of patients with GBM is due to aggressive tumor invasiveness, an inability to remove all tumor tissue, and an innate tumor chemo- and radioresistance. Ataxia-telangiectasia mutated (ATM) is an excellent target for radiosensitizing GBM because of its critical role in regulating the DNA damage response and p53, among other cellular processes. As a first step toward this goal, we recently showed that the novel ATM kinase inhibitor KU-60019 reduced migration, invasion, and growth, and potently radiosensitized human glioma cells in vitro.

Experimental design: Using orthotopic xenograft models of GBM, we now show that KU-60019 is also an effective radiosensitizer in vivo. Human glioma cells expressing reporter genes for monitoring tumor growth and dispersal were grown intracranially, and KU-60019 was administered intratumorally by convection-enhanced delivery or osmotic pump.

Results: Our results show that the combined effect of KU-60019 and radiation significantly increased survival of mice 2- to 3-fold over controls. Importantly, we show that glioma with mutant p53 is much more sensitive to KU-60019 radiosensitization than genetically matched wild-type glioma.

Conclusions: Taken together, our results suggest that an ATM kinase inhibitor may be an effective radiosensitizer and adjuvant therapy for patients with mutant p53 brain cancers.

Figure 1

KU-60019 inhibits the ATM kinase and radiosensitizes human glioma cells in vitro. A, Western blotting. U1242 and U87 cells were exposed to 3 μM KU-60019 (−0.5 hr) or left untreated and then irradiated with 5 Gy or not and then processed at 0.5 hr for western blotting with anti-p(S1981)-ATM antibody and normalized to ATM protein with anti-ATM antibody. B, Survival assay. Human glioma U1242 and U87 cells were serially diluted, plated [plating efficiency: 0.20 (U87); 0.22 (U1242)], and exposed to KU-60019 (3 μM) or not and then irradiated with 2, 4, 6, or 8 Gy followed by radiosurvival colony-forming assay. Data points, surviving colonies plotted as fraction of control; error bars, SD; N ≥ 3. Where error bars are not seen they are obscured by symbols. At 8 Gy, KU-60019 radiosensitization of U87 and U1242 cells was highly significant (P < 0.0001). Dose-enhancement ratios - U1242: 3.2; U87: 3.0. Radiobiological parameters - U87 α/β ratio: 11.7 Gy; U87 + KU-60019: 34.8 Gy; U1242: 1.8 Gy; U1242 + KU-60019: 3.4 Gy. C. Survival assay of mouse glioma stem cells. Cells were grown as neurospheres, trypsinized, and seeded in quadruplicate in a microtiter plate followed by exposure to KU-60019 and radiation as described in the Materials and Methods. At 10 Gy, KU-60019 radiosensitization was highly significant (P < 0.0001).

Figure 2

Human U1242/luc-GFP glioma cells form invasive and highly mitotic tumors in the brains of athymic mice. A, U1242 tumors are invasive. Brains were removed and imaged on a Zeiss SV11 stereomicroscope with a Hamamatsu digital CCD camera variable intensity microscopy illuminator system, and a Caliper IVIS-200 system using GFP filters, respectively. Subsequently, brains were embedded in OCT, sectioned in 6-μm sections, and stained with H & E. B, High mitotic index of intra-cranial U1242/luc-GFP tumors. Frozen sections were stained with anti-Ki67 antibody, followed by goat anti-mouse-Alexa 594, and counterstained with DAPI. GFP is the signal from the U1242/luc-GFP cells. The mitotic index (percentage Ki67+/GFP+ vs. Ki67−/GFP+ cells) was estimated at >50%.

Figure 3

Human U1242/luc-GFP gliomas show dose-dependent increases in γ-H2AX, p-(S824)-KAP1, and 53BP1 foci formation after radiation. Athymic female mice growing U1242/luc-GFP intra-cranial tumors were exposed to 0, 5 or 10 Gy followed by perfusion with fixative immediately, 1, or 2 hr post-irradiation. Brains were post-fixed, embedded in OCT and sectioned on a cryostat in 6-μm sections. Sections were stained with anti-p-(S139)-H2AX (A and B), p-(S824)-KAP1 (C), and 53BP1 (D) antibodies, respectively, followed by goat anti-mouse-Alexa 594. Nuclei were counterstained with DAPI. Section (A) shows GFP+ U1242/luc-GFP tumor whereas (B – D) do not. Instead, the characteristic elongated shape of the U1242 cells is clearly seen after DAPI staining.

Figure 4

Radiosensitization of orthotopic U1242/luc-GFP tumors with KU-60019. Athymic female mice were implanted with U1242/luc-GFP cells. Osmotic pumps (ALZET M2002) were filled with KU-60019 or PBS, connected to a cannula (Brain infusion kit 3; ALZET), inserted in the skull at a depth of 3 mm at the site of cell injection, and the pump placed under the skin on the back of the mouse. The M2002 pump delivers the drug at a rate of 0.5 μl/hr over 14 days. A, Mice with tumors were imaged by BLI 7 days post-cell injection followed by pump implantation with KU-60019 (10 μM) (+KU) or not (-KU). Tumor growth was determined at 13 days by BLI followed by 2 Gy of cranial irradiation to both mice. Seven days later (day 20) mice were imaged by BLI, euthanized, and the brains removed for GFP imaging using Caliper IVIS-200. Tumor was evident in the brain receiving radiation alone, but not in the brain receiving KU-60019 and radiation. B, Radiosensitization of U1242/luc-GFP tumors with KU-60019 delivered by intra-tumoral osmotic pump. Cells were injected followed by implantation of osmotic pumps as in (A) except that pumps (ALZET M1007D) were inserted 7 days after cell implantation and irradiation (single dose of 5 Gy) was done 7 days (day 14) later. Tumor growth was monitored by BLI. Two mice (A51 and A53), in the KU-60019 and radiation group that survived for >200 days and were apparently healthy when euthanized. Mean survival: no treatment; 27 ± 4 days (N = 5), 5 Gy; 34 ± 1 days (N = 5), KU-60019; 35 ± 4 days (N = 4), and KU-60019 + 5 Gy; 109 ± 94 days (N = 5). A significant 75 day growth-delay between the KU-60019 and radiation and radiation alone groups was noted (P = 0.00402 vs. all treatments).

Figure 5

KU-60019 delivered by CED radiosensitizes human orthotopic gliomas. A, KU-60019 radiosensitization. Human glioma U1242/luc-GFP cells were injected intra-cranially and tumor growth determined by BLI. Six days after cell injection, KU-60019 (250 μM) in 12.5 μl was infused by CED directly into the tumor followed immediately by 3 Gy of radiation to the head. This procedure was repeated again on day 9 and 12. Mean survival: no treatment; 35 ± 8 days (N = 5), 3 × 3 Gy; 38 ± 8 days (N = 7), 3 × KU-60019; 38 ± 8 days (N = 6), and 3 × (KU-60019 + 3 Gy); 60 ± 19 days (N = 6). The KU-60019 and radiation group survived significantly longer (P = 0.00238 vs. all treatments) with a growth delay of 22 days over radiation alone. Note that radiation alone did not increase survival compared to no treatment. (B and C) KU-60019 blocks repair foci formation in tumor and mouse brain. Human glioma U1242/luc-GFP cells were injected intra-cranially and allowed to form tumors as determined by BLI. At ~21 days the mice were infused with KU-60019 (250 μM) and immediately exposed to either 0 or 5 Gy of radiation followed by cardiac perfusion with fixative. Brains were post-fixed, imbedded in OCT and sectioned in 6-μm sections. Sections were stained with anti-p(S139)-H2AX antibody followed by goat anti-mouse-Alexa 594, counterstained with DAPI, and imaged by confocal microscopy at 63×. Insets show individual nuclei. B, KU-60019 inhibits DDR in tumor. Glioma cells (elongated nuclei) show γ-H2AX foci after irradiation, which is blocked by KU-60019. C, Radiation response in mouse brain is inhibited by KU-60019. Mouse brain cells (small, round nuclei with distinct, large nucleoli) forms γ-H2AX foci, which is also inhibited by KU-60019.

Figure 6

Genetically matched human glioma differing in p53 status show a significantly different response to KU-60019. U87/luc-DsRed (p53 wild type) cells were infected with a mouse retrovirus expressing the p53-281G allele. A, Western blotting for p53 shows the presence of p53-281G in U87-281G cells but not in the parental U87 cells. B, Cell cycle analysis (BrdU and 7-AAD flow cytometry) of untreated and irradiated (10 Gy) cells shows that the G1/S checkpoint is abrogated in U87-281G whereas parental cells have an intact checkpoint. Both cell populations have intact G2/M checkpoints. C, Growth and survival assay (trypan blue exclusion) after 7 days shows that U87-281G cells grow faster and are also more sensitive to KU-60019 (3 μM) than parental U87 cells. D, Survival assay (luciferase) shows that U87-281G cells are also more radiosensitive than the parental U87 cells (P = 0.0154 at 9 Gy). E, Survival assay (colony forming) shows that both U87 (P =0.0005) and U87-281G (P =0.0014) cells are significantly radiosensitized at 2 Gy by KU-60019. Furthermore, U87-281G cells shows a trend toward being more radiosensitive than parental U87 cells but this difference was not significant (P =0.166). Importantly, U87-281G cells were significantly more sensitive to KU-60019 and 2 Gy than U87 cells (P <0.0001). Plating efficiency: 0.19 (U87); 0.12 (U1242). F, U87 glioma with mutant p53-281G (top panel) is more sensitive to KU-60019 and radiation than parental p53 wild type U87 glioma (bottom panel). Human glioma U87/luc-DsRed (p53 wild type) and matched p53-281G U87/luc-DsRed cells were injected intra-cranially into athymic mice. Mice were treated as described in the legend to Fig. 5. Mice injected with p53-281G U87/luc-DsRed cells showed >160 days of tumor growth-delay after KU-60019 and radiation relative to radiation alone (P = 0.00011 vs. all treatments). All animals had evidence of tumors on day 5 (Supplementary Fig. S8). Mean survival: no treatment; 42 ± 3 day (N = 5), 3 × 3 Gy; 44 ± 6 days (N = 5); 3 × KU-60019; 51 ± 15 days (N = 6), and 3 × (KU-60019 + 3 Gy): >160 ± 49 days (N = 8). One mouse in the KU-60019 and radiation group developed an infection and was euthanized on day 181. Mice injected with the parental U87 cells only showed 16 days of growth-delay after KU-60019 and radiation relative to radiation alone (P = 0.0448). However, KU-60019 and radiation was not significantly different from KU-60019 alone. Mean survival: no treatment; 46 ± 6 days (N = 6), 3 × 3 Gy; 49 ± 10 days (N = 6); 3 × KU-60019; 70 ± 8 days (N = 5), 3 × (KU-60019 + 3 Gy); 65 ± 16 days N = 7).