20 
almost exclusively in the thymus. It does not occur in fibroblast cells in 
culture. Same of these recombinations occur in the env region with resulting 
alterations in host range. Viruses with extended host ranges occur through 
recombination between endogenous sequences, which are expressed during high 
level replication of a spreading virus, and infecting virus. Generally the 
host range of the resulting recombinant virus can be predicted from knowledge 
of the host and the virus. If the host range of two infecting viruses are 
known, predictions of the host range of recombinants can be made. 
Dr. Gottesman asked whether a unique virus could be generated in mice through 
recombination of a viral vector and an endogenous virus or viral sequence. 
Dr. Hartley replied the probability of obtaining a recombinant is lew for one 
cycle of replication. The probability rises with the number of replication 
cycles. Such a situation would not be unique to recombinant DNA derived viruses 
Dr. Liberman asked whether it would be possible to certify vectors for human 
gene therapy as had been done for other host-vector systems. (See Appendix E 
of the NIH Guidelines.) He did not foresee a need for a great number of vectors 
and thought certification of the vectors most likely to be used in human gene 
therapy protocols might simplify and speed-up review of protocols. 
Dr. Gottesman said in the past NIH has certified vectors; some problems, however 
developed when better vectors than those which had been certified were sub- 
sequently developed. Each protocol for human gene therapy will have to be 
carefully evaluated on a case-by-case basis by the Working Group on Human Gene 
Therapy and the RAC. As part of this review, the retroviral vector constructs 
will be carefully scrutinized. 
Dr. Liberman asked whether data on the vectors could be generated in advance 
of submission of the first human gene therapy protocol. Prior review of such 
data would also speed-up the review process. He suggested the Working Group 
on Viruses should also review the biology of seme of the retroviral vectors 
most likely to be used in human gene therapy. 
Dr. Gottesman said the Working Group on Human Gene Therapy could receive and 
review iri vitro data on the vectors before submission of clinical protocols. 
Dr. Anderson said investigators are usually developing, constructing, and 
testing several vectors simultaneously. Investigators do not wish to submit 
data on every vector; most of these vectors will not be used in actual human 
gene therapy protocols. Dr. Gottesman said investigators would not be 
required to submit such data, rather they may submit the data if they wished. 
Dr. Simpson asked Dr. Walters whether RAC will Drovide guidance to IBCs in 
reviewing the retroviral vectors used in human gene therapy. He said retro- 
virology itself is a complex subject and still greater expertise is required 
to evaluate the vector constructs. 
Dr. Anderson said retroviral vectors are very intricate; the type of expertise 
and information necessary to evaluate these vectors may not be available to 
IBCs and Institutional Review Boards (IRBs). He asked Dr. Walters if the 
[459] 
