Variants selected by treatment of human immunodeficiency virus-infected cells with an immunotoxin

An immunotoxin has been made by coupling anti-human immunodeficiency virus (HIV) envelope antibody 907 to ricin A chain (907-RAC). 907 recognizes an epitope within the immunodominant PB-1 loop of gp120. Variant cells were selected by cloning persistently infected H9/human T lymphocyte virus IIIB cells in the presence of the immunotoxin. Clones resistant to 907-RAC arose at a frequency of 0.1-1.0%. Seven clones were selected for intensive analysis. When studied, these clones fell into two distinct groups, members of which appeared to be identical, suggesting that the variation arose before the selection process. In contrast to the parent cells, none of the cloned variants produced infectious HIV. The first set of clones, designated the "E" variants, expressed decreased levels of the HIV envelope on the cell surface. However, levels of intracellular HIV antigens and reverse transcriptase were equal to or greater than that of the parental cell line. Radioimmunoprecipitation demonstrated that the gp160 was truncated to 145 kD (gp120 was normal length), capable of binding to CD4, and, unlike normal gp160, was released in its unprocessed form into the cellular supernatant. Sequence analysis demonstrated that a deletion at codon 687 of the envelope gene resulted in the production of this truncated protein. Ultrastructural analysis of E variants demonstrated some budding forms of virus, but also large numbers of HIV within intracellular vesicles. The second set of variants, the "F" series, produced no HIV antigens, reverse transcriptase, nor was there ultrastructural evidence of virus. However, proviral DNA was present. Virus could not be induced with agents known to activate latent HIV. These cells also lacked cell surface CD4 and could not be infected with HIV. These studies demonstrate that variation in HIV can affect the phenotype of the cells carrying the altered virus, allowing for escape from immunologic destruction. The E variants may serve as prototypes for attenuated HIV, which could be used as a vaccine. We have reconstructed the mutation found in the E variants within the infectious HIV clone HXB-2 and demonstrated that the resulting virus retains its noninfectious phenotype.