Modelling the Meningioma: Insights from in vivo and in vitro approaches

Background. Meningiomas are the most prevalent primary tumor type of the central nervous system. Despite promising results from initial preclinical studies, clinical trials have yet to identify effective medical treatments. A significant issue is that current experimental models do not accurately replicate clinical conditions. Therefore, there is a critical need for meningioma models with high translational value to better understand pathophysiology and evaluate potential treatments. These models should closely resemble clinical meningiomas in various aspects, including phenotypic morphology, immunohistochemistry, copy number variations, and epigenetic profiles. Additionally, factors such as availability and cost should be considered. 

Aim. This thesis aimed to assess meningioma models both in the literature and in the laboratory, exploring their advantages and disadvantages, and evaluating primary patient-derived in vivo and in vitro models. The focus was on the similarities and differences between these models and primary tumors in terms of histology, immunohistochemistry, and epigenetic profiles.

Methods. The thesis employs two distinct methodologies: 1. Literature review and metaanalysis: A systematic review with meta-analyses of meningioma animal models was conducted, assessing tumor take rate, incubation duration for various models, and their advantages and disadvantages. Additionally, a critical appraisal using a self-devised tool, CRItical appraisal of quality of reporting, MEthodological Quality and Risk of Bias in Animal Research (CRIME-Q), was performed and internally validated. 2. Experimental Studies: Patients from the out-patient clinic were included, and meningiomas were extracted during surgery. These tumors were cultured as 2D cell cultures for three passages and implanted in athymic nude rats, either on the convexity or the base of the skull. Also, 3D cell cultures were grown in a xeno-free hydrogel. Both experiments concluded after three months, with xenografts and 3D models being evaluated using histology and immunohistochemistry. Additionally, DNA methylation analysis was performed using the EPICv1 850K or EPICv2 925K array on xenografts 3D models respectively and their corresponding primary patient-derived tumors.

Results. 1. Literature review and meta-analysis: The primary methods for establishing tumor models involved xenografting patient-derived material in immunodeficient mice and using malignant established/commercially available cell lines (immortalized) in most literature. Primary patient-derived material was less common. Small subgroups of genetically engineered models tested various genes. Meta-analyses revealed a consistent high tumor take rate (TTR) with short incubation for immortalized cell lines, with more varied results for primary patientderived cells/material and genetically engineered models. Using the critical appraisal tool, CRIME-Q, we found that included records generally lacked quality reporting. CRIME-Q items on quality of reporting and methodological quality scored high in interrater agreement and Cohen's kappa index – validated internally. 2. Experimental studies: The in vivo xenograft model showed a TTR of 80% for superficially implanted cells and 25% for deeply implanted cells with a high likeness to primary tumors in terms of retainment of morphology, immunohistochemistry and epigenetics with minor differences. Also, no significant differentially methylated regions were found for grouped xenografts compared to their corresponding primary tumors and correlation plots demonstrated good overall correlation between primary tumors and their corresponding xenografts. The 3D in vitro model showed overall similar immunohistochemistry with a higher likeness to the in vivo model compared to 2D in vitro models especially in terms of proliferation index. The cells in the 3D model displayed similar morphology as primary tumors, but morphological features such as fibrous tissue were lost. DNA-methylation analysis showed high congruency with minor differences with correlation plots demonstrating good overall correlation between primary tumors and their corresponding 3D models (R>0.95).

Conclusion, discussion and perspectives. The thesis highlights the lack of quality reporting in many studies within the field of meningioma animal models, which could affect the ability to translate the results to human condition. Furthermore, we have successfully established two types of models that closely represent corresponding primary tumors on an individual level. While xenografts closely represent corresponding tumors, they are unsuitable for large-scale pharmaceutical testing on an individual level. Many animal studies have already failed to translate to human conditions, indicating a need for new pathways and a paradigm shift in drug testing – a personalized approach. Approximately one-third of skull base meningiomas cannot be completely removed and have a high risk of recurrence, necessitating new personalized 3D cell models, like the one presented here. These models have the potential to test multiple drugs based on molecular profiles, potentially utilizing off-label use of already approved drugs with known side effects on an individual level.