Repurposing the DNA Labeling Agent EdU for Therapy against Heterogeneous Patient Glioblastoma

Abstract

Glioblastoma (GBM) is the most common type of malignant central nervous system tumor, and more than 300,000 people are diagnosed with GBM worldwide annually. Based on its recognition as damage by nucleotide excision repair, we now repurpose the DNA labeling agent 5-ethynyl-2'-deoxyuridine (EdU) as a treatment for GBM. We tested the efficacy of EdU in several different model systems, including not only GBM cell lines in in vitro cell culture and in vivo orthotopic mouse models of GBM, but also against living, uncultured tumor tissues of patients with GBM grown within our organotypic brain slice culture (OBSC) ex vivo platform. When compared with the standard-of-care drug temozolomide (TMZ) in in vitro GBM cell survival assays, EdU displayed ED50 values orders of magnitude lower than those of TMZ in all five GBM tumor lines tested. Against two in vivo orthotopic brain tumor models, EdU significantly extended survival relative to controls. EdU efficacy against a panel of patient GBMs largely correlated with the clinical Ki-67 status of each tumor, save for one tumor that remained unresponsive to treatment with both EdU and TMZ. Overall, these data suggest that (i) EdU has potential to be repurposed as an anticancer therapeutic and is especially adept at killing rapidly proliferating cells with low off-target toxicity; (ii) the OBSC platform can measure nuanced differences in efficacy of experimental therapeutics on heterogeneous patient tumor tissues; and (iii) OBSCs can continue to help identify potential responders and nonresponders to EdU treatment via functional precision testing of patient tumors ex vivo.

Figure 1:. EdU shows better therapeutic effects than TMZ in 5 of 5 glioma cell lines in vitro.

Comparison of dose-response of U87 (A), LN229 (B), U251 (C) using bioluminescent (BLI) assay (left) and 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide (MTT) assay (middle) and the corresponding IC50 values (right). (D) Comparison of dose-response of GBM8 using BLI assay (left) and the corresponding IC50 values (right) (E) Comparison of dose-response of CT2A cell lines using MTT assay (left) and the corresponding IC50 values (right). Unless otherwise stated, all dose-response readouts were t= 3 days post-treatment with single therapy EdU and TMZ. n=3 biological replicates per cell line per drug for each assay. NR represents “not reached” indicating that the IC50 values could not be estimated in the dose range investigated. GBM8 cell clusters grown in suspension were not successfully assayed with MTT. CT2A could not be successfully transduced to express luciferase for BLI assays. All data are shown as mean ± SEM.

Figure 2:. EdU provides survival benefits for mice bearing intracranial human glioblastoma or mouse glioma grafts.

(A) Overview of in vivo intracranial allograft study with CT2A mouse glioma cell line and EdU treatment. (B) The Kaplan-Meier survival of mice in in vivo intracranial allograft study with CT2A (n=10 mice per Thymidine Control and 200 mg/kg EdU conditions, n=5 mice in Untreated condition). Day 0 marks the day of the intracranial inoculation of 50,000 CT2A cells. The black arrows indicate the days 10 total i.p. EdU injections were administered. (C) Mean body weights of the mice over time in Untreated, Thymidine Control, and 200 mg/kg EdU conditions in in vivo intracranial allograft study with CT2A. “Rebound” indicates the resulting weight rebound following the change in the EdU dosing regimen in 200 mg/kg EdU condition. (D) Overview of in vivo intracranial xenograft study with U251 human glioblastoma cell line and EdU treatment. (E) The Kaplan-Meier survival plot of in vivo intracranial xenograft study with U251 (n=8 mice per condition). Day 0 marks the day of the intracranial inoculation of U251 cells. The black arrows indicate the days 15 total i.p. EdU injections were administered. (F) Mean body weights of the mice over time in Thymidine Control, 50 mg/kg EdU, and 200 mg/kg EdU conditions in in vivo intracranial xenograft study with U251.

Figure 3:. Testing EdU on an Organotypic Brain Slice Culture Model of Glioblastoma.

(A) An illustration of an organotypic brain slice culture of tumor-seeded healthy 8-day old rat brain slices suspended in media or drug by a cell culture insert. (B) Dose-response curve of OBSC-seeded GBM8 (red) and non-tumor bearing OBSCs (grey) treated with increasing doses of EdU. (C) Semi-quantitative image showing the preferential localization of 1000 μM EdU in GBM8, a patient-derived glioblastoma cell line. (D) Normalized Z-stack signal (lowest signal is 0%, highest signal is 100%) from the tumor-air to the tumor-slice interface showing lesser tumor and increasing EdU localization towards the tumor-air interface.

Figure 4:. EdU performs as well as or better than TMZ on GBM patient tissue regardless of MGMT and recurrence status.

(A) Overview of the organotypic brain slice culture (OBSC) model of Glioblastoma showing the resection and mechanical dissociation of human brain tissue engrafted as microtumors on healthy rat brain slices. (B-F) Dose-response curves (left) and quantitative bioluminescent images (right) of GBM patient tissues PT1, PT2, PT3, PT4, and PT5 t=3 days post-treatment with single therapy EdU and TMZ. (G) Dose-response curves (left) and quantitative bioluminescent images (right) of toxicity profile of brain slices treated with single therapy EdU and TMZ measured with a propidium iodide assay for cell death (n=6 technical replicates of brain slices per dose), one-way ANOVA analysis (**p = 0.0025). 100% dead slices were treated with 70% ethanol and frozen for 24 hrs. (H) Comparisons of DSS of TMZ & EdU accounting for therapeutic windows between tumor kill and off-target toxicity. All data are shown as mean ± SEM.

Figure 5:. Clinical Correlations of Study Outcomes.

(A) Charts showing the therapeutic windows across 11 parameters accounted for by the DSS of patient tissue tested against EdU and TMZ. Normalized therapeutic windows range from −1 to +1 with the more negative value indicating greater off-target toxicity compared to tumor kill and vice versa. (B-D) Relationship between ED10, ED25, and ED50 values of EdU and the Ki-67 percentages assigned to patient tumors. PT5, an unresponsive tumor to both EdU & TMZ is identified by the red dot. PT5 is not considered in the calculation of the R² values for correlation between ED10, ED25, ED50 and the Ki-67 indices. Non-Responsive (NR), Not Applicable (N/A).