Search

BACKGROUND:Temozolomide (TMZ) is active against glioblastomas (GBM) in which the O6-methylguanine-DNA methyltransferase (MGMT) gene is silenced. However, even in responsive cases, its beneficial effect is undermined by the emergence of drug resistance. Here, we tested whether inhibition of poly (ADP-ribose) polymerase-1 and -2 (PARP) enhanced the effectiveness of TMZ. METHODS:Using patient derived brain tumor initiating cells (BTICs) and orthotopic xenografts as models of newly diagnosed and recurrent high-grade glioma, we assessed the effects of TMZ, ABT-888, and the combination of TMZ and ABT-888 on the viability of BTICs and survival of tumor-bearing mice. We also studied DNA damage repair, checkpoint protein phosphorylation, and DNA replication in mismatch repair (MMR) deficient cells treated with TMZ and TMZ plus ABT-888. RESULTS:Cells and xenografts derived from newly diagnosed MGMT methylated high-grade gliomas were sensitive to TMZ while those derived from unmethylated and recurrent gliomas were typically resistant. ABT-888 had no effect on the viability of BTICs or tumor bearing mice, but co-treatment with TMZ restored sensitivity in resistant cells and xenografts from newly diagnosed unmethylated gliomas and recurrent gliomas with MSH6 mutations. In contrast, the addition of ABT-888 to TMZ had little sensitizing effect on cells and xenografts derived from newly diagnosed methylated gliomas. In a model of acquired TMZ resistance mediated by loss of MMR gene MSH6, re-sensitization to TMZ by ABT-888 was accompanied by persistent DNA strand breaks, re-engagement of checkpoint kinase signaling, and interruption of DNA synthesis. CONCLUSION:In laboratory models, the addition of ABT-888 to TMZ overcame resistance to TMZ.

Despite promising preclinical studies, oncolytic viral therapy for malignant gliomas has resulted in variable, but underwhelming results in clinical evaluations. Of concern are the low levels of tumour infection and viral replication within the tumour. This discrepancy between the laboratory and the clinic could result from the disparity of xenograft versus syngeneic models in determining in vivo viral infection, replication and treatment efficacy. Here we describe a panel of primary mouse glioma lines derived from Nf1 (+/-) Trp53 (+/-) mice in the C57Bl/6J background for use in the preclinical testing of the oncolytic virus Myxoma (MYXV). These lines show a range of susceptibility to MYXV replication in vitro, but all succumb to viral-mediated cell death. Two of these lines orthotopically grafted produced aggressive gliomas. Intracranial injection of MYXV failed to result in sustained viral replication or treatment efficacy, with minimal tumour infection that was completely resolved by 7 days post-infection. We hypothesized that the stromal production of Type-I interferons (IFNα/β) could explain the resistance seen in these models; however, we found that neither the cell lines in vitro nor the tumours in vivo produce any IFNα/β in response to MYXV infection. To confirm IFNα/β did not play a role in this resistance, we ablated the ability of tumours to respond to IFNα/β via IRF9 knockdown, and generated identical results. Our studies demonstrate that these syngeneic cell lines are relevant preclinical models for testing experimental glioma treatments, and show that IFNα/β is not responsible for the MYXV treatment resistance seen in syngeneic glioma models.

Myxoma virus (MYXV) is a well-established oncolytic agent against different types of tumors. MYXV is also known for its immunomodulatory properties in down-regulating major histocompatibility complex (MHC) I surface expression (via the M153R gene product, a viral E3-ubiquitin ligase) and suppressing T cell killing of infected target cells. MHC I down-regulation, however, favors NK cell activation. Brain tumors including gliomas are characterized by high MHC I expression with impaired NK activity. We thus hypothesized that MYXV infection of glioma cells will promote NK cell-mediated recognition and killing of gliomas. We infected human gliomas with MYXV and evaluated their susceptibility to NK cell-mediated cytotoxicity. MYXV enhanced NK cell-mediated killing of glioma cells (U87 cells, MYXV vs. Mock: 51.73% vs. 28.63%, P = .0001, t test; U251 cells, MYXV vs. Mock: 40.4% vs. 20.03%, P .0007, t test). Using MYXV M153R targeted knockout (designated vMyx-M153KO) to infect gliomas, we demonstrate that M153R was responsible for reduced expression of MHC I on gliomas and enhanced NK cell-mediated antiglioma activity (U87 cells, MYXV vs. vMyx-M153KO: 51.73% vs. 25.17%, P = .0002, t test; U251 cells, MYXV vs. vMyx-M153KO: 40.4% vs. 19.27, P = .0013, t test). Consequently, NK cell-mediated lysis of established human glioma tumors in CB-17 SCID mice was accelerated with improved mouse survival (log-rank P = .0072). These results demonstrate the potential for combining MYXV with NK cells to effectively kill malignant gliomas.