I-Mei Siu, Vafi Salmasi, Brent A. Orr, Qi Zhao, Zev A. Binder, Christine Tran, Masaru Ishii, Gregory J. Riggins, Christine L. Hann and Gary L. Gallia
Chordomas are rare tumors arising from remnants of the notochord. Because of the challenges in achieving a complete resection, the radioresistant nature of these tumors, and the lack of effective chemotherapeutics, the median survival for patients with chordomas is approximately 6 years. Reproducible preclinical model systems that closely mimic the original patient's tumor are essential for the development and evaluation of effective therapeutics. Currently, there are only a few established chordoma cell lines and no primary xenograft model. In this study, the authors aimed to develop a primary chordoma xenograft model.
The authors implanted independent tumor samples from 2 patients into athymic nude mice. The resulting xenograft line was characterized by histopathological analysis and immunohistochemical staining. The patient's tumor and serial passages of the xenograft were genomically analyzed using a 660,000 single-nucleotide polymorphism array.
A serially transplantable xenograft was established from one of the 2 patient samples. Histopathological analysis and immunohistochemical staining for S100 protein, epithelial membrane antigen, and cytokeratin AE1/AE3 of the primary patient sample and the xenografts confirmed that the xenografts were identical to the original chordoma obtained from the patient. Immunohistochemical staining and western blot analysis confirmed the presence of brachyury, a recently described marker of chordomas, in the tumor from the patient and each of the xenografts. Genome-wide variation was assessed between the patient's tumor and the xenografts and was found to be more than 99.9% concordant.
To the best of their knowledge, the authors have established the first primary chordoma xenograft that will provide a useful preclinical model for this disease and a platform for therapeutic development.
I-Mei Siu, Betty M. Tyler, James X. Chen, Charles G. Eberhart, Ulrich-Wilhelm Thomale, Alessandro Olivi, George I. Jallo, Gregory J. Riggins and Gary L. Gallia
Diffuse brainstem tumors are the most difficult type of pediatric CNS malignancy to treat. These inoperable lesions are treated with radiation alone or in combination with chemotherapy, and the survival rate is less than 10%. It is therefore essential to develop a reliable animal model to screen new therapeutic agents for the treatment of this type of tumor.
A multipotent human glioblastoma stemlike neurosphere line, 060919, was established from a surgically resected glioblastoma specimen; when cells were implanted intracranially into athymic nude mice, they formed invasive, vascular tumors that exhibited the features of glioblastoma. Ten female Fischer 344 rats received an injection of 75,000 F98 rat glioma cells and 10 female athymic nude rats received an injection of 75,000 060919 human glioblastoma stemlike cells in the pontine tegmentum of the brainstem. A control group of 5 female Fischer rats received an injection of saline in the same location as the animals in the tumor groups. Kaplan-Meier curves were generated for survival, and brains were processed postmortem for histopathological investigation.
Both F98 cells and 060919 cells grew in 100% of the animals injected. Median survival of animals injected with F98 was 15 days, consistent with the authors' previous reports on the establishment of the brainstem tumor model using the F98 rat glioma line. Median survival of animals injected with 060919 was 31 days. Histopathological analysis of the tumors confirmed the presence of brainstem lesions in animals that received brainstem injections of F98 and in animals that received brainstem injections of 060919. The 060919 brainstem tumors histologically resembled glioblastoma.
Tumor take and median survival were consistent for animals injected in the brainstem with either the established F98 rat glioma cell line or the 060919 human glioblastoma stemlike neurosphere line. Histopathological features of the 060919 brainstem tumors resembled glioblastoma. Establishment of this human glioblastoma stemlike brainstem animal model will improve the evaluation and identification of more efficacious agents for the treatment of high-grade brainstem tumors.