EXPERIMENTAL MODELS FOR BRAIN TUMOR

Co-chairs: Eric C. Holland, M.D., Ph.D., and Robert L. Martuza, M.D.

Participants: David Ashley Michael Berens Tom Curran Ronald A. DePinho Glenn Dranoff Henry Friedman

Mark A. Israel Andreas Kurtz Linda M. Liau Luis Parada David H. Rowitch Evan Snyder Carol Wikstrand

STATEMENT OF THE PROBLEM

Models are central to making the transition from scientific concepts to understanding the reality of a tumor in a person. They may be used for therapeutic screens, in preclinical trials, or to study the basic biology of tumors. Unfortunately, the available model systems, whether at the cellular, tissue, or animal level, do not accurately represent the biology of human brain tumors. Established glioma cell lines develop multiple genetic changes over time, so that they no longer reflect the biology of the tumor in the person. In addition, current primary tissue models cannot be maintained in a healthy condition for more than a few days. Finally, implantation animal models do not reflect the interaction between tumor and host that occurs in the human.

CHALLENGES AND QUESTIONS

Existing Models

• Conventionally used glioma-derived cell lines contain genetic and gene expression alterations that are ill defined and do not necessarily reflect the primary tumors from which they were derived. Therefore, data derived from them may not reflect the biology, heterogeneity, or therapeutic response of the primary tumors. The scientific community needs better cell culture and tissue models that more accurately reflect the biology of brain tumors.

• Traditional rodent models do not accurately reflect the growth, invasion, histology, gene expression profiling, vasculature, and stromal interactions of various intracranial tumors.

• No genetically defined brain tumor models exist in species other than in laboratory mice. Such animal models may offer additional insights into the biology of specific brain tumor types or may allow the testing of specific therapeutic modalities that is not possible in mice.

Future Models

• The gene expression profiles, controlling elements, pathways, and cells of origin for brain tumors remain largely unknown.

• Molecular reagents, such as tissue-specific promoters and enhancers, to create genetically accurate models of brain tumors are lacking. Further, it is necessary to identify to oncogenes and tumor suppressors, as well as to develop the technology needed to combine alterations in the appropriate cell types.

• There are no adequate mechanisms to correlate genetic alterations in mouse tumors with their proposed human counterparts.

• There are no readily available and affordable noninvasive techniques to allow investigators to measure changes in tumor volume, growth pattern, gene expression, and other biological parameters of interest.

Model Availability

• Investigators who are not directly involved in the production of animal models lack easy and affordable access to these models.

RESEARCH AND SCIENTIFIC PRIORITIES

Priority 1: Develop tissue and cell culture systems that replicate the biology of human brain tumors more adequately than do the currently available immortalized cell lines.

Specifically, there is a need to develop and validate primary tissue culture systems such as spheroids, brain slice cultures, primary cell cultures, and genetically defined immortalized cells, including stem cells. These systems need to be characterized as to their similarity to the brain tumors that they are designed to model.

Priority 2: Create genetically accurate animal models for brain tumors.

An essential first step is to define and develop reagents, including tissue-specific promoters and enhancers, as well as to expand our capabilities for readily combining multiple genetic alterations in specific cell types.

The development of accurate mouse models is very important. In addition, the development of genetically defined brain tumor models in other species should be encouraged. Such species may include smaller animals, such as Drosophila or fish, as well as larger animals, such as pigs and dogs. These have the potential to be used to further understanding of the biology of the disease. In particular, the anatomical spatial dimensions of larger animal models may better allow for the testing of surgical interventions and novel delivery techniques.

Priority 3: Generate methods to validate, compare, and contrast the animal model with its proposed human counterpart.

These methods should include standard methods such as histology, magnetic resonance imaging, anatomical imaging, and therapeutic response, as well as gene expression and genomic profiling. This will require the development of mouse arrays from the mouse counterparts to the human BTGAP sequences. In addition, methodologies such as BAC arrays or SKY analysis will be required to assess the acquired genomic alterations. Adequate bioinformatics support will be required in order to analyze these data.

Further development and use of noninvasive technology is necessary to measure aspects of tumor biology in these animal models. Specifically, there is a need for image analysis of gene expression, vascularity, tumor size, and invasiveness, and for other noninvasive techniques, such as serum or urine surrogate markers. These may require the generation of specific, genetically altered reporter mice.

Priority 4: Improve access to and make available at reasonable cost to investigators:

• The reagents needed to create new animal models of brain tumors

• Sophisticated technologies used to evaluate and validate those models

• The animal models themselves Expression cDNA microarray and BAC array technologies should be made available for analysis of these murine-derived tumors. Currently, access to these technologies is not readily available to all groups of investigators whose important scientific questions require these animal models. Moreover, investigators who have developed these animal models do not have the resources needed to provide and widely distribute them.

RESOURCES NEEDED

• Because of the lack of adequate cell, tissue, and animal model systems that reflect human brain tumors, and because grants to develop model systems traditionally have not been funded by the National Institutes of Health, a mechanism must be created specifically to fund the development and validation of model systems that more accurately reflect the biology of brain neoplasms. Although the National Cancer Institute (NCI) Mouse Models for Human Cancer Consortium (MMHCC) has been established to fund development of mouse cancer models in general, the National Institute of Neurological Disorders and Stroke (NINDS) needs to emphasize and coordinate with the NCI for the development of additional mouse models of the various brain tumors not addressed through the MMHCC and of models in other animals.

• Resources need to be allocated for the generation of cDNA microarrays based on the mouse equivalent of the human BTGAP sequences.

• A mechanism must be created to ensure affordable access to the reagents and models listed above to investigators. Furthermore, the NCI and NINDS should provide resources to establish a consortium of brain tumor modeling laboratories for the purpose of testing novel therapies.