Results Encouraging In First Trial of Gene Therapy in Brain Ca

NASHVILLE--Researchers at the National Institutes of Health haveseen objective responses in some of the 15 patients treated todate in the first clinical trial of gene therapy in brain tumors,Michael Blaese, MD, said at the scientific subcommittee sessionon gene therapy at the American Society of Hematology (ASH) meeting.

Although the results are encouraging, "we're not curing anyonewith this strategy at the moment," said Dr. Blaese, chiefof the Clinical Gene Therapy Branch, National Center for HumanGenome Research (NCHGR). "But I think we're getting closer,and eventually this might be an effective treatment for a varietyof localized cancers," he continued.

The trial, launched about 2 years ago, involves an approach designedto make brain tumors sensitive to the antiviral drug ganciclovir(Cytovene). In the protocol (developed by Dr. Blaese and his colleagues,Drs. Kenneth Culver, Edward Oldfield, and Zvi Ram), geneticallyaltered mouse cells that produce retroviral vectors carrying theherpesvirus thymidine kinase gene are injected into the tumor.The tumor cells are genetically modified by the retroviral vectorsto produce herpes thymidine kinase and thus become targets forganciclovir.

The patients in the study had all failed surgery and radiotherapy,and most had also failed chemotherapy. Responses were evaluatedin a variety of ways, including PET scans and MRI scans priorto therapy and 2 weeks following therapy.

Dr. Blaese showed MRI scans from one of the first patients inthe study who had relapsed with two brain lesions following primarysurgery and radiotherapy. "We have been following this patientnow for 23 months after treatment with vector-producer cells andganciclovir without having seen a recurrence," he said. Threeof the responses seen were sustained for at least a few months,but in many patients who showed an initial response, the tumorquickly returned, and some patients had no response at all, Dr.Blaese said.

He noted that the results appear to be due to the actual transferof the gene and not just the initial injection of altered cells.In a study of a patient on another arm of the protocol in whichtumors were injected with vector-producer cells and resected onday 7 without ganciclovir therapy, the researchers showed thatthe cells in the tumor had been transduced, "so one can transfergenes effectively to tumor cells in the vicinity of the injectionusing this kind of strategy," he said.

Dr. Blaese emphasized the importance of the "bystander effect,"in which the ganciclovir treatment kills both tumor cells containingthe gene and neighboring tumors that had not been gene modified.He believes there may also be an immune component to this effect,because in some clinical situations, the researchers have seenvery substantial reductions in volume in tumors in which lessthan 1% of tumor cells were effectively transduced with the gene.

Improving the spread of the transgenes throughout the tumor isan important research goal. Biopsies show that gene transfer occursonly to tumor cells within 5 to 25 cell diameters of the pointwhere the altered cells are planted.

"Retroviral vector particles are large; they simply do notdiffuse like soluble proteins . . . so if this is going to bean effective treatment, we have to work very seriously at developinggene transfer systems that will allow genes to be diffused throughoutthe tumor," Dr. Blaese said.

One approach is to use different kinds of vectors, such as adenovirusvectors, that might produce higher enzyme levels. Another is touse a vector-producer cell that is motile, like a glioma, thatmight be able to "crawl around inside the tumor and spreadthe vector throughout a wider area. That may not work out, butwe're testing the idea," Dr. Blaese said.

The 'Ideal System'

The ideal system would be to use a replication complement vectorwith viral replication restricted to the tumor. "It may seemfar fetched, but in fact there is a history going back at least40 years of treating cancer with various viral infections, specificallytrying to take advantage of the fact that certain viruses growfar more luxuriously in tumor tissue than in normal tissues,"Dr. Blaese said.

At the NCHGR, researchers are working with replication complementviruses that have been engineered to remove certain viral genesso that they can no longer divide within normal tissue. But, hesaid, when they are inserted into malignant cells, they are complementedby the gene mutations in the tumor and can proliferate specificallywithin the tumor.

"It is our hope that by using this general approach of modifyingviruses that take advantage of the mutations in the tumor to allowthem to divide, we will be able to get the genes to spread muchmore effectively to the tumor cells," he said.