Human Gene Therapy Subcommittee - July 29-30, 1991 
experiments, this could mean that the histocompatibility antigen is over-produced. If 
the construct replicates in mouse cells at 100 copies per cell, that could be the 
explanation for the observed response. This would not be the case in human cells. Dr. 
D. Miller contended that there are two major requirements that must be fulfilled 
before this proposal can be approved: (1) the description of the sequences that are 
included in the vector, which are crucial for proper evaluation of the presented data, 
and its possible utility in humans; and (2) the construction of an animal model with 
tumor present at various sites, where if one site is injected, the effects on the other 
tumors can be determined. 
Dr. Leventhal said these experiments should not be performed merely because the 
patients are terminally ill. It does not allow time to observe the results of the 
treatment. If the patient cannot be examined at five years to see the late effects, there 
is a weakness in the experimental model. This will make it more difficult to move the 
treatment into successively healthier patients. 
Dr. Nabel said the goal of the trial is to determine, in a fairly noninvasive manner, 
whether recombinant gene expression is successful and whether specific immunologic 
responses are generated against the tumor cells in the patients. The growth of local 
tumors, as well as of distant tumors, can be observed. 
Dr. Parkman said that, although a study may be a Phase I trial, the HGTS has to 
consider it in terms of it being a Phase II or a Phase III trial, i.e., is the preclinical data 
adequate to justify the treatment. The efficacy has not been demonstrated in a tumor- 
bearing animal model above what one would observe with Bacillus Calmette-Guerin 
(BCG). For this protocol to be approved, there has to be a demonstration of a 
systemic effect in a preclinical tumor-bearing model. 
Dr. Nabel said BCG induces a marked suppressive response. Clearly one of the 
hallmarks of the transplantation response is that there does not seem to be much 
suppression of that response. In terms of generating a systemic effect and providing 
distant tumor protection, the presence of specific cytolytic T cells in the spleen 
indicates a systemic effect. 
Dr. Parkman noted that the protocol proposes to have biopsies of the lesions 
performed every two to four weeks. This manipulation of the lesion could look like 
regression and produce a false positive result. The investigators should plan to perform 
some control biopsies of non-injected lesions. 
Dr. Mclvor asked which model most closely resembles the treatment proposed for 
humans, the preimmunized animal model or the non-preimmunized animal model. Dr. 
Nabel replied that for the Phase I studies this was the preimmunization animal model. 
Dr. Mclvor said if a liposome-mediated gene transfer is to be performed, there is no 
[896] 
Recombinant DNA Research, Volume 14 
