Gene Therapy for Childhood Brain Tumors

Roger J. Packer, M.D., Chairman, Department of Neurology, Children's Medical Center, Washington DC

by Roger J. Packer, M.D., Founding and Sr. Scientific Advisor, CBTF-MD

Significant progress has been made in the management of some forms of childhood brain tumors. However, for children with malignant gliomas including those of the cerebral cortex, diencephalon, brain stem, or cerebellum, progress has been frustratingly slow. The majority of these children, despite treatment with aggressive surgery, radiation therapy, and chemotherapy, will die of progressive disease within three years of diagnosis. The outcome is even poorer for children with recurrent malignant gliomas, as few of these children will survive for greater than nine months from the time of tumor recurrence.

The recent developments in neuroscience research have opened new potential avenues of therapy for central nervous system tumors. One extremely promising area is the utilization of molecular genetic techniques to develop selective therapies for childhood brain tumors. Organisms, such as viruses, can now be genetically altered to make them carry genetic information into cells within the body. This is the basis of multiple new approaches of gene therapy for a host of neurologic diseases.

There are multiple inherent problems in gene therapy. One is to direct the organism, usually a virus, to transmit information only into tumor cells, while leaving other surrounding brain cells uninfected. A second is the ability for any method to transfect enough tumor cells to affect all of the tumor. A third problem is to determine what information should be integrated into the tumor cell or the host to stop tumor growth.

There is a significant amount of interest for using gene therapy for brain tumors. In contrast to other tumor types, high-grade gliomas are relatively localized and it may be easier to reach the majority of tumor cells. Stereotactic neurosurgical techniques are available which can deliver viral vectors directly to the tumor. If a vector is utilized which only affects rapidly-dividing cells, such as those cells in a brain tumor, the central nervous system would be a relatively ideal site, as few cells in the central nervous system are dividing (other than tumor cells). The preliminary work has been completed in the use of one form of gene therapy for adults with malignant gliomas. This research was initially undertaken at the National Institutes of Health (NIH). Although a variety of types of viruses were considered for use in the treatment protocol, retroviruses were chosen.
Retroviruses are viruses which can only integrate and therefore express vector genes to proliferating cells. In the brain, the tumor is the most mitotically active cell. Other cells such as macrophage-derived cells, blood cells and endothelial cells have some, but a lesser risk, for transduction. In humans, mitotic activity in neurons cease by birth. In addition, the brain is a partially immunologic privileged site so that a producer cell releasing the retroviral vector would theoretically have a longer "life span" than if it were placed in other regions of the body. The retroviral vector chosen in the NIH study was genetically altered to make it replicative-defective. In this way, once the virus integrated into the nucleus of a cell, it would not divide and enter other cells, such as normal brain. In the preliminary adult study, a packaging cell line derived from mouse fibroblasts (murine fibroblast cell line) was utilized to deliver high quantities of virus, over a relatively prolonged time, into dividing cells. This replicative defective viral vector was then altered to insert a gene found in the herpes simplex virus called thymidine kinase. The thymidine kinase gene makes cells specifically sensitive to antiviral treatment.
Thus, the rationale of the initial gene therapy study was that a murine fibroblast packaging cell line could be directly delivered into the region of brain tumor cells. This cell line would release a genetically altered replicative-deficient retrovirus which would only transfect (infect) dividing cells; i.e., tumor cells. The retroviral vector, once inside the cell, would incorporate its genetic information into the nucleus of the dividing cell. This genetic information would include the thymidine kinase gene. After a period of time was allowed for the packaging cells to deliver the viral vector, a drug would be given which would kill all cells carrying the thymidine kinase gene. The only cells in the brain which should carry the thymidine kinase gene should be the tumor cells which took up the viral vector when it was injected. A drug, ganciclovir, would be given intravenously to kill all cells that took up the gene.

In experimental work, it became clear that not all tumor cells would take up the gene for a variety of different reasons, including that not all tumor cells were actively dividing when the virus was present. However, in experimental work, it was noted that despite only a proportion of tumor cells taking up the gene, larger amounts of the tumor were killed when the ganciclovir was given. The reasons for this are unclear but may include the release of various chemicals (cytokines) that would kill adjacent tumor cells. This affect on adjacent tumor cells is called the "bystander" effect.

Initially, in clinical studies, a group of adults with recurrent malignant brain tumors were treated in this fashion. In the initial study, the viral vectors were delivered into the brain after the tumor had been partially removed surgically. The edge of the remaining tumor was injected with the viral packaging cells and 14 days later ganciclovir was given to try to eradicate the cells which took up the genetic information. As was predicted by experimental work, there was evidence of antitumor affect in several patients in this trial. Also, it was determined that all of the foreign cells injected were destroyed by either the host immune system or the ganciclovir. The therapy was surprisingly well tolerated and seemed to be of some benefit.

Based on this information and the known poor survival rate of children with recurrent malignant gliomas, a study was developed and completed at the Children's National Medical Center between 1996 and 1999. Other institutions participating in the study included the Mayo Clinic and the Children's Hospital of Los Angeles. The study showed surprisingly little in the way of toxicity and one child who was treated with the gene therapy approach, who had previously failed multiple different therapies, remains alive and free of disease three years from treatment. However, the treatment was not uniformly effective. There was hope that the approach could be utilized for newly-diagnosed patients with malignant gliomas; however, the drug company supporting the study was merged with another company who was not interested in pursing further studies in pediatric patients or adults.


At the present time, there are no gene therapy studies open for children with brain tumors. Researchers across the world are looking at this and other gene therapy-related approaches. A variety of different viral vectors are presently being explored including the use of a herpes vector, adenovirus vectors, and adeno-associated virus vectors. In the study performed at the Children's National Medical Center, a surgical resection was undertaken to remove as much of the tumor as possible. After resection, at the time of surgery, the viral packaging cells which released the retroviral vector were directly injected into the rim of removed tumor in an attempt to kill any remaining tumor cells.
Obviously, much still needs to be known about the potential efficacy of gene therapy for children with malignant brain tumors. At present, retroviruses are predominantly being used in therapeutic trials, but researchers are working with a variety of different types of viruses which may be more effective. Adenoviruses are presently being evaluated in experimental work. They have the advantage of transecting a higher number of tumor cells, but they have the disadvantage of affecting all cells, not only mitotically active cells. Other gene therapy products are also being evaluated for their possible ability to turn off tumor cell growth once they are taken up by the tumor cell. Factors which genetically control cell growth or specifically affect a variety of different growth factors, on which cells are dependent, are all potential candidates for incorporation into such gene therapy trials. The safety of any of these approaches on the brain, especially the developing brain of the child, need to be closely monitored.

Other delivery techniques are being looked at for gene therapy. One such technique is the direct infusion of the gene into the tumor cell, without prior surgical debulking. This infusional technique has the advantage of avoiding another surgery, but has a disadvantage of asking the gene approach to work on a larger tumor bulk and to have the gene therapy vector reach the leading edge of the tumor without the aid of direct injection. As other gene products and viruses are utilized, their effect on the nervous system, especially long-term effects, need to be carefully monitored. Viruses have the capacity, if they can divide, to cause significant central nervous system infection and inflammation.
Despite all these drawbacks, gene therapy holds significant promise for the treatment of childhood brain tumors. Although this therapy is clearly in its infancy, it is the belief of those investigators exploring this technique that it should be utilized as soon as possible in children with malignant gliomas. At the present time, as stated previously, there is little in the way of effective therapy, especially for those patients with recurrent disease. It is hoped that if this therapy is found to be promising in children with recurrent disease, it can be incorporated into trials for patients with newly-diagnosed disease as a means to, at the bare minimum, improve local disease control.

Roger J. Packer, M.D. is Executive Director of the Center for Neuroscience and Behavioral Medicine and Chairman, Department of Neurology at the Children's National Medical Center, Washington, D.C.; Professor of Neurology and Pediatrics, The George Washington University; and Professor in Neurosurgery, University of Virginia, Charlottesville, Virginia and a consultant to the Pediatric Neuro-Oncology Program at the National Cancer Institute. He chairs the Medulloblastoma of the Children's Cancer Group and serves as a member of the Pediatric Brain Tumor Consortium Steering Committee, Brain and CNS Committee of Commission on Cancer of the American College of Surgeons, and is the Senior Medical/Scientific Advisor to the Childhood Brain Tumor Foundation. He graciously donated his time to write this article for www.childhoodbraintumor.org.

This article was written for the Childhood Brain Tumor Foundation, Germantown, Maryland.

Thanks to Our Sponsors