To test the potential of Rev from HIV infection to control reporter gene expression in an HIV-dependent fashion, we constructed an HIV-like vector as shown in Fig. 1. Overall, we incorporated four separate segments of the HIV genome into the expression vector; however, no HIV gene is expressed from this construct. The 5' end of the vector consists of the HIV 5' LTR, the splice donor 1 site, D1, and a portion of the gag open reading frame that includes the packaging signal. The second HIV segment is from the tat1/rev1 exon that includes splice acceptor site 5, A5, and splice donor site 4, D4. The third segment of HIV DNA is from the env exon and encompasses the RRE, and the splice acceptor site 7, A7. The last segment includes the entire 3' LTR along with a small portion of the nef reading frame 5' to the LTR. In infectious HIV, the joining of splice donor 1 with splice acceptor site 5 (removal of D1/A5 intron) is utilized in transcripts for both Envelope and Nef proteins. In HIV NL4-3 infection of PBMC, the single spliced D1/A5 transcript represents 80% of all Env message . Nef transcript requires a second splicing event, and the most common is removal of the segment between D4 and A7 . The vector has the capacity to express two genes. A multiple-cloning site (MCS) for one of the expressed genes is immediately down-stream of the A5/D4 HIV splice sites; however, this cloning site has not been utilized in this report. This is followed by an internal ribosome entry site (IRES), which is upstream to the reporter gene used in this report, green fluorescent protein or GFP. The reporter gene is adjoined to the RRE-containing env exon segment. Together these segments compose the transfer vector. The full-length transcript generated from the construct possesses efficient HIV splice sites that mediate the removal of the reporter open reading frame (see Fig. 1). The reporter gene is only expressed from the RRE-containing unspliced or singly spliced transcripts, and thus requires the presence of Rev, a process that mimics HIV late gene expression. Co-transfection of this plasmid with an Env-coding plasmid and a packaging construct results in production of an infectious non-replicative lentiviral vector  capable of transferring the Rev-dependent reporter system into cells.
Figure 1 Rev-dependent reporter construct. A. Comparison of the HIV-1 genome and the Rev-dependent vector. The vector contains both LTRs along with the 5' end of the gag gene, splice donors and acceptors, and a portion of the env that includes the Rev response element. No intact HIV genes are present. As used in this report, the lentiviral vector contains an open reading frame encoding green fluorescent protein (GFP). B. In the absence of HIV infection the reporter provirus undergoes basal transcription generating a single message that is rapidly spliced, and results in the removal of the GFP reading frame. In the presence of HIV Rev, singly and non-spliced transcript are delivered to the cytosol, and the reporter gene is expressed.
To permit a quantitative measurement of reporter gene induction from the lentiviral construct in response to HIV infection, we tested the vector in cells infected or not infected with HIV. CEM-SS cells, a human T cell line, were first infected with a modified, non-replicative NL4-3 HIV, where the gene for murine CD24 was inserted into the nef gene . This NL4-3 strain possesses intact tat and rev genes, but lacks an env gene and therefore must be pseudo-typed (see Methods). Staining of cells for murine CD24 thus confirms the presence of HIV in a cell population (compare lack of staining in Fig. 2A to CD24-positive, HIV-infected cells in Fig. 2B). The resultant mixed populations of HIV-infected and non-infected CEM-SS cells were then transduced with a predetermined titer (see Methods) of the VSV-Env pseudo-typed Rev-dependent indicator lentiviral vector (Fig. 2A and 2C). In Fig. 2C, in order to determine whether a single reporter vector was adequately robust, we added one infectious indicator vector per five cells. If we assume that the infection is a Poisson process, then only two percent of the cells should contain two or more integrated indicator vectors. After 3 days, GFP was found expressed only in HIV-infected (CD24-positive) cells (Fig. 2C); HIV-free cells transduced with equivalent levels of indicator viral vector displayed no GFP signal (Fig. 2A). Consistent with the viral input, we found that approximately twenty percent of the CD24-positive cells (HIV+) were also GFP-positive (Fig. 2C; also see legend to Fig. 2). This finding demonstrates that a single infection of the reporter vector was adequate to detect the presence of HIV in the cell. To examine increased dosage of reporter virus, we utilized a second preparation of the lentiviral vector that had been concentrated by ultracentrifugation. As shown in Fig. 2D, the percent of HIV-infected cells (CD24-positive) that expressed GFP increased proportionally with increasing reporter viral vector input. The curve that fits this data is a rectangular hyperbola, defining this as a saturable process, and predicts that a maximum of 80–90% of the HIV-infected cells examined can be labeled by the Rev-dependent reporter virus (see legend Fig. 2 and Discussion).
Figure 2 Specific expression of reporter by the Rev-dependent lentivirus in HIV infected T cells. CEM-SS cells were not infected (A) or were infected (B, C) with HIV-1 NL4-3.HSA.R+E-, an HIV-1 NL4-3 clone with the murine heat stable antigen (HSA; CD24) gene inserted into the nef gene. Both uninfected (A) and infected (C) cells were then transduced with the Rev-dependent reporter virus vNL-GFP-RRE(SA) (5 ng p24/106 cells; VSV-G envelope) or not transduced (B). At 72 hrs post transduction, cells were stained with R-phycoerythrin conjugated rat-anti-mouse CD24 antibody and analyzed by flow cytometer for CD24 and GPF expression. GFP was expressed only in HIV-1 infected (C), but not in HIV-1 uninfected (A) cells. Fifty thousand cells were examined in each run. In this run 61% of total cells were CD24-positive and of the CD24-positive population 17% were GFP-positive. Thresholds for CD24-positive and GFP-positive were set at a fluorescent intensity of 10 on the x- and y-axis respectively. Two independent analyses yielded the same result. D. Percent HIV-infected cells that express GFP versus reporter vector input. A concentrated preparation of reporter particles was used at varied dosages in HIV-infected CEM cells, as above. Flow cytometry studies yielded percent CD24-positive cells that became GFP-positive after 3 days (y-axis) versus reporter virus input (ng p24). Non-linear curve fitting (rectangular hyperbola) yielded R2 = 99.4; maximum GFP-positive = 86.1 ± 6.3% (best-fit values ± std. error); KD = 128.1 ± 21.5 ng p24. Dashed lines are 95% confidence bands.
Lentiviral vectors possess the capacity to infect differentiated macrophages, a known target cell for in vivo HIV replication. However, regulation of HIV LTR transcriptional activity in macrophages, compared to T cells, involves a different set of cellular factors [30,31]. We thus examined the capacity of the lentiviral expression vector to respond to HIV infection of macrophages. Human primary monocyte-derived macrophages were first infected with the CCR5-utilizing HIV-1AD8 followed 24 hours later by addition of the reporter lentiviral vector under conditions similar to the above CEM cell experiment. After 5 days the cells were fixed and stained for the presence of intracellular HIV p24 (Fig. 3). HIV-infected macrophages, stained for anti-p24, display cytosolic, perinuclear areas of high intensity due to accumulated viral particles, along with formation of large cells [32,33]. High intensity focal staining with anti-p24 (red fluorescence) was seen in multiple cells (see Fig. 3B, cells 1–9). Staining of cells in non-infected control populations was absent for both anti-p24 and GFP (data not shown), and GFP was expressed only in cells that were p24-positive (compare Fig. 3B and 3C). Of 308 cells examined here, 97 were p24-positive, and 22 of these were also GFP-positive. Under the conditions used in this experiment, and consistent with the viral input, approximately one fifth of the p24-positive macrophages displayed GFP production. As with the CEM experiment (Fig. 2), this finding suggests that a single infecting indicator viral vector in macrophages is capable of reporting the presence of HIV.
Figure 3 Specific expression of reporter by the Rev-dependent lentiviral vector in HIV-infected primary macrophages. M-CSF-treated monocyte-derived macrophages were infected or were not infected with HIVAD8, followed by Rev-dependent reporter virus vNL-GFP-RRE(SA). (A). Bright-field image of fixed cells with black bar representing 50 micrometers. Nuclei are prominent. Cells that were found to stain positive for HIV p24 are numbered in this image. B. Red fluorescence of identical field. All cells were stained for anti-p24 antibody, followed by a secondary red fluorescent antibody to detect productively infected cells. Intracellular red fluorescent foci identify productive infection; nine cells were identified in this field (B). We examined macrophage populations with four concentrations of HIV input (36, 72, 169, and 361 ng p24 HIVAD8/106 cells) followed by a constant input of reporter lentiviral vector (5 ng p24 vNL-GFP-RRE(SA)/106 cells; VSV-G envelope). Independent of the input of HIV, approximately one fifth of the p24-positive cells expressed GFP (23.0 ± 2.8%; mean ± SD; n = 4), as shown by green fluorescence in this same field (C). Magnification was identical on all three images. The reporter virus was demonstrated to function in macrophages from three donors.
The Rev-dependent system outlined in Fig. 1 operates under the assumption that in the absence of HIV there is elimination, through splicing, of GFP-encoding transcript. Upon Rev expression, the non-spliced RRE-containing GFP transcript would be rescued. To test this hypothesis, we performed RT-PCR on control and HIV-infected CEM-SS cells that were secondarily exposed to the reporter GFP-encoding viral vector at high input levels (see Fig. 4 legend). High levels of reporter vector were used to challenge the cell's capacity to remove GFP-encoding non-spliced transcript in the absence of HIV, that is, to remove false-positive reporter message. We used two sets of primers that amplified the spliced or non-spliced transcripts from the lentiviral construct (Fig. 4A), and compared spliced to unspliced reporter transcript in four cell populations: uninfected cells, cells infected with HIV only, cells infected with reporter virus only, and cells infected with both virus. Note that the vectors used to detect spliced transcript will also recognize HIV transcripts. As shown in Fig. 4B, HIV-specific transcript, consistent with the size of fully spliced transcript amplified with these primers (0.94 kb), appeared in the two cell populations infected with HIV (lanes 2 and 4). In cells exposed to the reporter vector but lacking HIV, we observed basal levels of transcription from the LTR promoter of the reporter vector (lane 3). This transcriptional activity represents HIV-independent LTR-mediated signal (the leakiness seen with LTR-based systems); however, with this Rev-dependent system the transcripts were all spliced (lane 3; spliced RNA vs unspliced RNA) and thus lacked the capacity to generate the GFP reporter. Following exposure to the reporter viral vector, the generation of a non-spliced, reporter-encoding message (seen in lane 4) was dependent on the presence of HIV infection. These data confirmed that expression of GFP-encoding transcript was HIV-dependent, and that the dependency corresponded to the prevention of transcript splicing, consistent with Rev activity.
Figure 4 Reverse transcriptase PCR analyses of gene expression from lentiviral vector NL-GFP-RRE-(SA) in response to HIV-1 infection. (A) A diagram of PCR primers used. Two sets of primers were used to detect the spliced (black arrows) and unspliced (green arrows) transcripts by reverse transcriptase PCR (RT-PCR) as described in Methods. The mRNA molecules from infected and uninfected cells were extracted at day 3 after lentiviral vector infection and analyzed by RT-PCR for the spliced transcripts and unspliced transcripts as shown in (B). (B) CEM-SS cells were either uninfected (lane 1, control), infected with HIV alone (lane 2, HIV, VSV-G pseudo-typed, 564 ng p24 per million cells), infected with NL-GFP-RRE-(SA) alone (lane 3, 1764 ng p24 virion per million cells), or infected with HIV first, then NL-GPF-RRE(SA) 24 hours later (lane 4). The cellular β-actin transcript was also amplified to ensure that comparable numbers of cells were used for the analyses.