Development and Applications of Retroviral Vectors

The ability of retroviruses to integrate efficiently into the genomic DNA of animal cells and be stably replicated and transmitted to all of the progeny of these cells provided a strong incentive for the development of retroviral gene transfer vectors. The discovery that acutely oncogenic retroviruses often arise as the result of acquisition of sequences derived from cellular proto-oncogenes provided an additional stimulus. It was clear from many studies that retroviral genomes could accommodate extensive alterations, and even though these changes often resulted in defects in replication, the altered viruses could be propagated in the presence of replication-competent or “helper” virus. Indeed, early synthetic retroviral vectors were produced by using helper virus (Shimotohno and Temin 1981; Wei et al. 1981; Tabin et al. 1982). However, the presence of helper virus precluded many types of experiments where viral spread after infection was unacceptable, especially for many genetic studies or for human gene therapy.

A major advance in retroviral vector design came with the development of retroviral packaging cells that provide all of the retroviral proteins in trans but did not produce replication-competent virus (Mann et al. 1983; Watanabe and Temin 1983). Many of the first generation of packaging cell lines produced helper virus as a result of recombination events, especially after the introduction of retroviral vectors or after prolonged cultivation, but evolution in design has greatly reduced this frequency. The host range of vectors produced by packaging cell lines has also been extended by the development of packaging cell lines based on a variety of mammalian and avian retroviruses.

Retroviral vectors have also been improved by the elimination of all viral coding regions and by the reduction of the remaining viral elements to the minimum required for high-efficiency transfer (Miller and Rosman 1989). This is possible because the early steps in the viral life cycle, retroviral entry into cells, reverse transcription of the viral genome into DNA, and integration of the viral genome into the host genome, do not depend on viral protein synthesis. As a result, retroviral vectors can be used to transfer only the genes of interest.

Pseudotyping of vectors with the surface protein of vesicular stomatitis virus (VSV) was an important advance in retroviral vector technology. This change has expanded the host range of retroviral vectors to include fish, insect, and amphibian cells and, because the VSV G protein is more stably associated with virions than retroviral envelope proteins, allows the efficient concentration of virions to produce high-titer vector stocks.

In contrast to typical vectors derived from oncoviruses, lentiviruses such as human immunodeficiency virus (HIV) can infect nondividing cells. This property would be especially useful for gene therapy applications, especially those involving in vivo gene transfer or transfer to slowly dividing cells such as hematopoietic stem cells. Because HIV is a relatively complex retrovirus, packaging cells and vectors based on HIV have been difficult to develop and often produce only low-titer viral stocks. Recently, the situation has been improved by construction of HIV vectors with HIV core proteins and the envelope protein of VSV (Akkina et al. 1996; Naldini et al. 1996). Vectors of this type are able to infect neurons in rats (Naldini et al. 1996).

Retroviral vectors have advantages when compared to many other gene transfer systems. These include the ability to transduce a wide range of cell types from different animal species, to integrate genetic material carried by the vector into recipient cells precisely, to express the transduced genes at high levels, and the lack of vector spread or production of viral proteins after infection. Retroviral vectors have been profitably employed in the study of viral replication. It is usually much simpler to follow the replication of, and measure the titer of, vectors containing selectable (or scorable) genetic markers than it is to obtain such information from unmarked viruses. In addition, the ability to limit viral replication to a single round improves the measurement of rates of mutation and recombination. In the early days of retrovirology, naturally occurring vectors, i.e., acutely transforming viruses that had acquired cellular oncogenes, were often used to study viral replication; in most cases, these naturally occurring vectors have been superseded by viral vectors created in the laboratory. The use of retroviral vectors as tools to investigate the retroviral life cycle is discussed in other chapters. This chapter focuses on vector design issues and other experimental applications of retroviral vectors, including their recent application to the treatment of human disease.