Preview

PDF, English Download (6MB) | Terms of use

Abstract

Treatment of pediatric tumors remains challenging and even though research has made a lot of progress and about 80% of diagnosed patients can nowadays be cured, for 20% of patients curative therapy is lacking and this ratio has not much improved in the last >20 years. For several pediatric tumors of the central nervous system, which account for 20-25% of all cancers in children and which are the second most common group of tumors after leukemia, outcome is explicitly poor. To improve survival of pediatric patients with high-risk brain tumors, new treatment strategies need to be developed and tested. In the thesis described here, which consists of three projects, not only new treatment strategies were tested but also already published research findings were evaluated. In the first project a target actionability review (TAR) was prepared. A systematic literature search for literature published between 2014 and 2021 was performed to evaluate the process of replication stress as a therapeutic target for treatment of 16 different solid pediatric tumor entities. By using pre-defined search terms that were either general keywords of geno- and phenotypes observed with replication stress or specific genes that play a major role for replication stress, 319 papers were identified and included for further review based on abstract and title. The papers were evaluated by two reviewers independently and findings related to target activation in clinical series, from in vitro and in vivo preclinical experiments or from clinical trials were summarized. Data was scored for quality and outcome and reviews documented in the web portal R2. After evaluation of papers with discrepant scores by a third reviewer, in total 145 publications addressing 37 different drug targets were included for analysis and scores were visualized by heatmaps that are publicly available within the R2 TAR platform (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi?option=imi2_targetmap_v1). Besides identification of 31 alternative potential targets to target replication stress in pediatric solid tumors, the targets ATM, ATR, CHK1, DNA-PK, PARP, and WEE1, were analyzed in more detail. The analysis revealed that the targets ATR, CHK1, PARP, and WEE1, were the most promising targets for monotherapy or combination therapies with chemo-/radiotherapy for treatment of neuroblastoma, osteosarcoma, high-grade glioma and medulloblastoma. The evidence scores for the targets ATM and DNA-PK were positive for treatment of high-grade glioma or neuroblastoma and osteosarcoma, respectively, however, the results were based on a limited amount of literature and need to be studied in more detail for a comprehensive analysis. An intensively studied module with 114/401 evidence entries (28%) was “combinations” and within this the most studied strategy was combining a PARP-inhibitor with chemo- or radiotherapy, which yielded positive appraisal scores for treatment of neuroblastoma, rhabdomyosarcoma, Ewing sarcoma, osteosarcoma, nephroblastoma, high-grade glioma and ependymoma. The second project was performed to target replication stress in vitro and in vivo in cell lines and patient-derived xenograft (PDX) models, respectively of medulloblastoma (MB), ependymoma (EPN) and embryonal tumors with multilayered rosettes (ETMR). With an in vitro drug screen on the MB Group 3 cell line HD-MB03 and the ETMR cell line Bt183 the most synergistic combination partner for Irinotecan, respectively the active metabolite SN-38, was evaluated using a drug library with 76 compounds, which revealed the PARP-inhibitors Olaparib and Talazoparib as the most synergistic combination partners. The synergistic effect was also confirmed when using the brain penetrant PARP-inhibitor Pamiparib, while fetal Astrocytes, as a control cell line, did not respond effectively to the combination treatment. To verify the effect of the treatment with Pamiparib and Irinotecan in vivo, five different PDX models representing ETMR (Bt183), Sonic Hedgehog (SHH) MB (med-1712FH and BT084), Group 3 MB (nch2194) and ZFTA-fusion positive ependymoma (Bt165) were injected subcutaneously or orthotopically into immunodeficient NSG mice. Subcutaneous tumors of the ETMR model showed complete regression when treated with Irinotecan and Pamiparib in combination, however, for the other models no synergistic effect was observed when injected orthotopically. Nevertheless, Irinotecan alone was able to induce a significant survival benefit and tumor growth inhibition for the MB Group 3 and the ZFTA-fusion positive EPN model. Refinements of the treatment strategy including dose adaptations of Irinotecan and Pamiparib and using a nanoformulated version of SN-38 (peg-SN-38), which is characterized by a longer half-live and accumulation in the tumor, showed no significant differences when applied to the MB Group 3 model nch2194. The observed effect on tumor growth and survival was solely based on Irinotecan or peg-SN-38 and adding the PARP-inhibitor Pamiparib could not further increase the effect. The third project focused on the generation and molecular characterization of SHH MB PDX models that are resistant to the treatment with the Smoothened (SMO) inhibitor Sonidegib. For treatment of patients diagnosed with a SHH MB, especially adult patients, use of inhibitors that target the transmembrane protein SMO, a key component of the SHH pathway, is promising. However, even though patients typically show initial tumor regression, they often develop resistance to the treatment. To understand the mechanisms of resistance and to develop new treatment strategies that overcome resistance, preclinical models that are resistant to SMO inhibition with the same molecular characteristics as seen in patients are needed. For generation of resistant models, mice harboring tumors of a PTCH1-mutated SHH MB PDX model that is sensitive to SMO inhibition were treated in vivo with the SMO-inhibitor Sonidegib using intermitted treatment cycles until tumors became resistant to therapy. Vehicle-treated and the nine generated resistant tumors were analyzed with whole genome and RNA sequencing to evaluate the underlying mechanism of resistance and confirmed target engagement of Sonidegib. Eight models acquired resistance due to a missense mutation in SMO and one model became resistant due to an inactivating point mutation in MEGF8, which is a negative regulator of the SHH signaling pathway. For development of further treatment strategies, an in vitro drug screen with 76 drugs was performed with a sensitive, treatment-naïve and a resistant model and revealed the XPO1-inhibitor Selinexor as one of the top hits being effective in both models. To confirm efficacy in vivo, a sensitive, vehicle-treated and two resistant models, one SMO-mutated and the MEGF8-mutated, were intracranially injected into NSG mice and treated with Selinexor. Treatment of the two resistant models resulted in tumor growth inhibition and significant survival benefit. The results of the three projects described in this thesis will support the improvement of treatment strategies for high-risk pediatric brain tumor entities. Evaluation of published literature helps to streamline the best treatment strategies and identify knowledge gaps. Applying one of the most promising approaches, a combination of a topoisomerase-inhibitor with a PARP-inhibitor, in in vitro and in vivo preclinical studies revealed further information on responding and non-responding tumor entities. In addition, the generation of new, clinically relevant preclinical medulloblastoma models may improve translational research for medulloblastoma patients and can be considered as a paradigm how to predict resistance to targeted monotherapy in clinically relevant models.