16 
The working group concluded that descriptive language in the NIH Guidelines 
might be more flexible and thus preferrable over a revision of Appendix B. 
Dr. Gottesman said she might attempt to prepare language for RAC's consideration 
at a subsequent RAC meeting. She said the discussion indicated most retroviruses 
are either: (1) defective when they carry an oncogene; or (2) nondefective 
but they carry no oncogene. Rous sarcoma virus is an exception, however. 
This retrovirus demonstrates that a functional retrovirus can carry an oncogene. 
She felt the discussion indicated that BL2 containment appears in general to 
be appropriate for work with retroviruses. However, consideration should be 
given to higher containment if a broad host range, infectious, amphotropically 
packaged retrovirus carrying an oncogene is being studied. 
Dr. Anderson said Dr. Gottesman should address two separate issues in attempting 
to draft language; the use of retroviral vectors in human gene therapy and the 
use of wild-type viruses in other types of procedures. 
Dr. Mulligan said in human gene therapy the investigator attempts to introduce 
the vector into the cells of human subjects. Two issues should be addressed 
for use of retroviral vectors in human gene therapy: (1) could new viruses be 
generated; and (2) would integration into the human gencme disrupt or activate 
important chromosomal sequences. Dr. Gottesman said consideration should also 
be given to the possibility of genomic rearrangements in human gene therapy. 
[Rapporteur's Note: Retroviral vectors can rearrange their own structure as 
well as exchange sequences and could reccmbine with an endogenous retroviral 
sequence or with other retroviruses. At present, it is thought possible that 
a retroviral sequence might recombine with an endogenous retroviral sequence 
to produce packageable, infectious recombinant virus. The properties such a 
recombinant would possess are unknown, but the potential homology between retro- 
viral vectors and as yet unknown primate retroviruses or human T-cell leukemia 
viruses might be sufficiently close to produce recombinants. There is, however, 
a built-in safety feature when mouse retroviral vectors are used. These mouse 
virus structures have very different sequences from known primate retroviruses, 
and there appears to be little homology between the two. Recombination fre- 
quencies thus would be very lew or negligible. 
The second potential hazard posed by the use of retroviral vectors in human gene 
therapy is mutagenesis as a result of previrus integration into the chromosome. 
Because the provirus carries transcriptional regulatory sequences within the 
LTRs, provirus integration can result in either gene activation (insertional 
mutagenesis) or gene inactivation. Gene activation would be a potential hazard 
particularly if the previrus integrates adjacent to a cellular proto-oncogene.] 
Dr. Gottesman asked the working group what is known about insertional activation 
or inactivation of genes by integration of previrus DNA into chromosomal sites. 
Dr. Mulligan said that certain retroviruses that do not carry an oncogene can 
cause malignancies in animals long after their introduction into the host. 
How these malignancies develop from such infection is not known. It is thought 
in these cases a proto-oncogene is activated by previrus integration. The 
[455] 
