The Rev-dependent expression vector displayed highly specific HIV-dependent expression. The application tested here, as a lentiviral reporter for HIV detection, permitted us to address two major concerns: non-specific (HIV-independent) noise and inadequate specific (HIV-dependent) signal. The first concern is based on knowledge that expression from an integrated LTR vector can be affected by the local chromatin activity, resulting in HIV- or Tat-independent high expression . Of course, this is what prevents utilization of LTR-based reporter viruses. With the establishment of LTR-based indicator cells, it's just a matter of removing cells with aberrantly high background reporter expression [4,14]. This problem appeared to be absent in cells transduced with the Rev-dependent vector system, since expression of GFP in both a T cell line and a primary macrophage population occurred only in HIV-positive cells. Furthermore, we found that GFP-encoding transcript was detected by RT-PCR only in HIV-infected populations. Secondly, we wished to know whether this vector, potentially as a single integrant and with the added stringency of Rev-dependence, would generate an adequately robust response. To address this point we examined reporter function under conditions that maximized single reporter integrants. It does assume that infections follow a Poisson function, a central tenet to most virus infection studies. At an input where 80% of the cells do not become infected (approximately one infectious particle per five cells), less than 2% of the cells should possess more than one reporter provirus. Expression of GFP reporter from the HIV-positive cells closely approximated the input reporter viral vector level (Fig. 2 and 3), suggesting that undetected expression of integrated vectors is not prominent. Note, however, that expression of the reporter gene is achieved in cells that are also highly supportive of transcriptional activity from the HIV provirus, that is, LTR activity is supported. Both HIV and the reporter vector use the LTR promoter.
We also examined high doses of reporter viral vector, where a majority of the cells are infected with the Rev-dependent reporter vector. Under these conditions we did not see intact reporter (unspliced) transcript or protein in HIV-negative cells. The nature of the curve that fits the reporter vector detection of HIV-infected cells (Fig. 2D) suggests that the rate-limiting step towards reporter expression follows Michaelis-Menten kinetics. This is characteristic of a single binding-site event, but at this time we are unable to define this rate-limiting process, such as the viral particle binding to the cellular receptor. Additional mechanisms leading to this non-linear function would include the potentially equivalent susceptibility of increasing numbers of infected cells to additional infection, as characterized by a Poisson distribution. Concentration of this viral particle by centrifugation permitted this examination of high vector input, but it also diminished the infectivity of the viral particles. This data also predicts that most of the HIV-infected cells will be susceptible to detection by the reporter virus. Of course, in mixed cell populations or tissues there will be variable susceptibilities to infection by the reporter virus as presently used. The inability to infect all cells by the reporter virus is perhaps the most limiting aspect of this application of the Rev-dependent vector.
In this report we have utilized the term Rev-dependence to distinguish this construct from the existing LTR-based Tat-responsive HIV reporter systems; however, with the intact LTR, we would speculate that this Rev-dependent vector should maintain responsiveness to Tat expression as well. This dual dependency is likely to contribute to the higher stringency seen. In this document, we have characterized this Rev-dependent vector in only one application (viral reporter vector), an application that is challenging and not feasible with the LTR-based reporters. At the same time we view this application as supplementary to existing methods to test cells for the presence of HIV replication. These include the staining or probing of HIV proteins and nucleic acids in fixed cells, and may include amplification with PCR [34-36]. We have also incorporated the vector into continuous cell lines (data not shown). As introduced into a transformed T cell, we find that expression from this Rev-dependent vector is insensitive to cellular activation states, thus maintaining a negative background in the absence of HIV, but has enhanced sensitivity to HIV infection, relative to many of the existing non-T cell indicator cell lines. Although this report has been limited to the expression of reporter genes, the construct's incorporation into a lentivirus and the specificity acquired should also permit its use as an experimental agent to express any gene in HIV-infected cells.