Magnetization transfer and T2-weighted MRI studies are useful for visualizing phenotypic presentations of orthotopic, patientderived xenograft mouse models of glioblastoma
Nagaraja TN, Elmghirbi R, Irtenkauf S, Hasselbach L, Tawil B, Brown SL, Panda S, Cabral G, Mikkelsen T, Ewing JR, and DeCarvalho A. Magnetization transfer and T2-weighted MRI studies are useful for visualizing phenotypic presentations of orthotopic, patientderived xenograft mouse models of glioblastoma. Cancer Research 2017; 77(13 Suppl):3860.
Introduction: Patient-derived xenograft (PDX) models for glioblastoma (GBM) from resected tumor tissues replicate several features of the original tumor. They are considered to be representative models to study tumor progression, and to test responses to putative therapies. Longitudinal noninvasive imaging can be useful in such investigations. To that end, we employed magnetic resonance imaging (MRI) to visualize and measure tumor burden in four different PDX models of GBM. Experimental procedures: Four orthotopic mouse PDX models, HF2587, HF2927, HF3077 and HF3253, developed from neurosphere cultures of four different human glioblastoma samples were used in the study. The neurosphere cells were implanted into the right striatum in immunocompromised nude mice (n=5-8 per model) and allowed to grow for 2-8 weeks, depending on their known growth rates from previous studies. They were imaged in a Varian 7T MRI system with the following weightings: T2 , T1 , magnetization transfer (MT), and contrast enhanced MRI (CE-MRI) with Magnevist as the contrast agent (CA). Following imaging, all the mice were sacrificed and their brains processed for hematoxylin and eosin (H&E) histology and human major histocompatibility complex (MHC) immunohistochemistry. Results: Tumor masses were visible as hyperintense regions on MT and T2-weighted images. The extent of such masses matched the H&E and MHC staining patterns. Ventricle enlargements seen on MRI in several mice were also confirmed by histology. Necrotic cores, when present, were observed on both imaging and on histopathology with good spatial correlations. Surprisingly, post-contrast T1 imaging did not enhance in the tumor mass or peritumorally, except in one mouse in which some intratumoral enhancement was observed. In all other instances from the four PDX models tested, enhancement was observed only when the tumor tissue or parts of it were contiguous with pial or dural vasculature. Conclusions: At 7 Tesla, MT-MRI and T2-weighted imaging, rather than CE-MRI, appear to be of better utility in visualizing these PDX models of GBM if MRI is chosen as the imaging modality. Since the parent tumors imaged at lower field strengths showed contrast enhancement, absence of a similar feature in these models needs additional studies to understand their vascular characteristics. Such properties may include low level of vascularization and/or relatively less leaky tumor vasculature. Another possible reason may be that the models tested represent the invasive features of GBM better than the vascular features, e.g. peritumoral ring enhancement, of larger clinical tumors with increased exposure to hypoxia.