INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Lentivirus vectors based on human immunodeficiency virus type 1 (HIV-1) can transduce nondividing cells in vitro and in vivo (6, 12, 18, 23, 27). This is an advantage over retrovirus vectors derived from oncoretroviruses, such as Moloney murine leukemia virus (MoMLV) because these retrovirus vectors require proliferation of the target cells for integration (8, 25). Several measures have been taken to develop HIV-1 vectors that are safer for gene therapy. First, accessory genes (vif, vpr, vpu, and nef) that are not essential for transduction were eliminated from a packaging construct (12, 13, 30). Second, the use of a three-plasmid expression system that consists of packaging, envelope, and vector constructs has minimized the possibility of generating replication-competent virus through recombination (21, 22, 23). Third, self-inactivating (SIN) vectors are constructed by deletions in the 3' long terminal repeats (LTRs) of the HIV-1 vectors (19, 31). Miyoshi et al. (19) have deleted 133 nucleotides in the U3 region of the 3' LTR, including the TATA box and binding sites for the transcription factors Sp1 and NF-B, in the CS-G vector. Zufferey et al. (31) have generated a SIN vector (SIN-18) that has a more extensive 400-nucleotide deletion in the U3 region of the 3' LTR, thus eliminating all of the transcriptional elements located upstream of the SP1 binding sites. These SIN vectors also enable the regulated expression of genes from internal promoters by eliminating any cis-acting effects of the HIV-1 LTR (19, 31).

The internal cytomegalovirus (CMV) immediate-early promoter has been widely used in lentivirus vectors to control gene expression in mammalian cells. It is expressed efficiently in a variety of cell types (6, 12, 18) but generally less efficiently in lymphoid cells (2). Since dysfunctional T cells are often associated with an inherited or acquired immunodeficiency (11, 16, 26), there are numerous gene therapy applications for such dysfunctional T cells (including HIV-1 disease). We therefore tested other promoters that are expressed well in T cells. We tested the MoMLV (hereafter termed MLV) LTR because overexpression of proto-oncogenes under the control of the MLV LTR has been demonstrated to be one of the common features of MoMLV-induced thymic lymphoma (28). We also tested the LTR of a novel retrovirus, ampho-mink cell focus-forming (AMP/MCF) virus (termed the Rh-MLV LTR), that was identified in the serum of a rhesus monkey that developed rapidly progressive T-cell lymphomas involving the thymus, lymph nodes, liver, spleen, and bone marrow (9, 24, 29). The lymphomas were the result of insertional mutagenesis by the productive retroviral infections in the monkeys that had received autologous transplantation of enriched bone marrow stem cells transduced with a retrovirus vector preparation containing replication-competent viruses (29). The AMP-MCF replication-competent virus arose by two recombination eventsfirst, a recombination between the vector and packaging-encoding sequences in the producer clone used for transduction of bone marrow stem cells (vector-helper recombinant); second, a recombination involving the genome of the vector-helper recombinant and endogenous murine retroviral genomes in the producer clone. The LTR of the AMP-MCF virus also differs from the MLV LTR by a 23-bp insertion (a direct copy of the 23 bp 3' to the site of insertion) in the enhancer region of the U3 segment.

In the present study, we used the gene for enhanced green fluorescent protein (EGFP) as a reporter gene and analyzed the effects of various murine oncoretrovirus LTR promoters in a SIN lentivirus vector. Our results show that the Rh-MLV LTR is a stronger promoter than the MLV and CMV promoters, and the partial gag sequence of MoMLV in the context of an HIV-1-based vector is essential for the high level of gene expression in human T lymphocytes.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of vectors. The HIV-1-based vectors CS-MSV-E, CS-MLV-E, CS-MLV-E, CS-Rh-MLV-E, and CS-Rh-MLV-E were derived from pCS or pCSCG (19). CS-MSV-E was constructed from pCS and SRLEGFP (4). SRLEGFP was digested with BglII, filled in with Klenow fragments, and then excised by NheI. The resulting 2.2-kb fragment, which contains the Moloney murine sarcoma virus (MSV) LTR, a partial gag sequence of MoMLV from LNL6 (gag_803-1612; nucleotides 803 to 1612, as described in GenBank accession number M63653), and EGFP, was ligated to a 7-kb HpaI/XbaI fragment of pCS. To make CS-MLV-E, a 2.2-kb fragment of CS-MSV-E that contains the MSV LTR, gag_803-1612 of MoMLV, and EGFP was subcloned into the XhoI and BstXI sites of the pBlu2SKM vector. The resulting intermediate construct, pBS-MSV-EM, was digested with EcoRI, filled in with Klenow fragments, digested with XmaI to liberate the MSV LTR, and then ligated to a 700-bp fragment of SRLEGFP (which contains the MLV LTR) to give pBS-MLV-EM'. The 2.2-kb BstXI/XhoI fragment of pBS-MLV-EM' was then religated to the same sites in CS-MSV-E. To make CS-MLV-E, pCSCG was digested with EcoRI, filled in with Klenow fragments, and digested with NheI to liberate a 500-bp fragment that contains the CMV promoter. The resulting 8-kb fragment was then ligated to an ~950-bp fragment of pBS-MLV-EM'. The latter fragment was prepared by digesting pBS-MLV-EM' with BstXI, filling in with Klenow fragments, and digesting with SpeI.

Production and determination of titers of VSVG-pseudotyped lentivirus reporter virus. Vesicular stomatitis virus G (VSVG)-pseudotyped lentivirus vectors were produced by calcium phosphate-mediated transfection of 293T cells as described previously (4). In brief, 20 × 10⁶ 293T cells were transfected with 5 µg of pHCMVG, 12.5 µg of pCMVR8.2DVPR, and 12.5 µg of the lentivirus reporter vector (CSCG, CS-MLV-E, CS-MLV-E, CS-Rh-MLV-E, or CS-Rh-MLV-E). Culture supernatant that contained viruses was collected on days 2, 3, and 4 posttransfection, pooled, and filtered through a 0.22-µm-pore-size filter. To make VSVG-pseudotyped retrovirus vectors, 20 × 10⁶ 293T cells were transfected with 5 µg of pHCMVG, 12.5 µg of SVenvMLV (3, 15), and 12.5 µg of SRLEGFP.

Activation of primary PBMC cultures. Peripheral blood mononuclear cells (PBMC) were separated over a Ficoll-Hypaque gradient (Amersham Pharmacia, Uppsala, Sweden). A culture plate was coated with 10 µg of goat anti-mouse immunoglobulin G antibodies (Cappel, Durham, N.C.)/ml at 37°C for 2 h, washed with sterilized phosphate-buffered saline, and recoated with 1 µg of anti-CD3 monoclonal antibody (MAb) (Coulter Corp., Hialeah, Fla.)/ml at 37°C for 2 h. Nonadherent PBMC were obtained after depleting macrophages by 2 h of adherence to plastic and were cultured in a 6-well plate (at 10⁶ per well) that was coated with immobilized anti-CD3 MAb. The cells were cultured in RPMI 1640 supplemented with 10% AB serum, 100 U of penicillin/ml, 100 µg of streptomycin/ml, and 0.5 µg of soluble anti-CD28 MAb (Pharmingen, San Diego, Calif.)/ml. Activated T cells were collected 48 to 60 h after activation and were used for infection experiments.

Infection of target cells in vitro. CEMX174 and SUPT1 cells were cultured in Iscove's modified Dulbecco's medium (GIBCO) with 10% fetal bovine serum, 100 U of penicillin/ml, and 100 µg of streptomycin (GIBCO)/ml. HeLa and 293T cells were cultured in Dulbecco's modified Eagle medium with 10% calf serum, 100 U of penicillin/ml, and 100 µg of streptomycin/ml. Activated T cells were cultured in RPMI 1640 (GIBCO) supplemented with 10% fetal bovine serum, 100 U of penicillin/ml, 100 µg of streptomycin/ml, and interleukin 2 (1 µg/ml). The cells (10⁵) were infected with VSVG-pseudotyped viruses by incubation with 0.25 ml of virus supernatant (at appropriate dilutions) in the presence of Polybrene (10 µg/ml) at 37°C for 2 h. Three days postinfection, the cells were analyzed by flow cytometry for EGFP expression.

Flow cytometric analysis. Infected cells were harvested, run on a FACScan flow cytometer (Becton-Dickinson & Co., San Jose, Calif.), and analyzed by the Cell Quest program (Becton-Dickinson) for the percentage of EGFP⁺ cells and the level of EGFP expression. Five thousand to 10,000 events were acquired for each sample. To determine the expression of CD4 and CD8 molecules on activated T cells, 5 × 10⁵ infected cells were costained with anti-CD4-PE (Coulter) and anti-CD8-PerPC (Becton-Dickinson) MAbs.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction and generation of modified SIN vectors. The lentivirus-based SIN vector CSCG has a 133-bp deletion in the U3 region of the 3' LTR that eliminates the transcriptional interference of gene expression from the internal CMV promoter in the provirus. EGFP expression under the control of the CMV promoter in CSCG is about four times higher in HeLa cells (mean fluorescence intensity [MFI], 248) than in SUPT1 cells (MFI, 57) (Fig. 1). We constructed modified lentivirus vectors that allow higher levels of gene expression in T cells. SRLEGFP is a MoMLV-based vector that contains the MLV LTR, a partial untranslated gag sequence of MoMLV, and the EGFP reporter gene in its proviral form. The 809-bp MoMLV partial gag sequence (gag_803-1612, derived from LNL6 nucleotides 803 to 1612 as described in GenBank accession number M63653) of SRLEGFP was derived from the retrovirus vector LNSX (5, 17), which contains the "extended packaging sequence" essential for the increase in retrovirus vector titers and gene transfer efficiency. In addition, a stop codon was inserted in place of the Pr65 gag start codon to prevent synthesis of Pr65 gag (17). We compared CSCG with SRLEGFP in infected HeLa and SUPT1 cells and found that, in contrast to CSCG, SRLEGFP has a higher EGFP expression in SUPT1 than in HeLa cells. EGFP expression under SRLEGFP in SUPT1 cells is generally 8- to 10-fold higher than that of CSCG (Fig. 1 and Table 1), suggesting that MLV LTR is a stronger promoter than CMV in SUPT1 cells. We therefore tested the promoter activity of various oncoretrovirus LTRs in a SIN vector. The internal CMV promoter of CSCG was replaced by an oncoretrovirus LTR derived either from MLV (CS-MLV-E) or MSV (CS-MSV-E) (Fig. 2). We also tested a novel LTR that has been identified in the AMP-MCF retrovirus found in the serum of a monkey with lymphoma (CS-Rh-MLV-E) (9, 29) (Fig. 2). We hypothesize that since this LTR was derived from a rhesus macaque T-cell tumor it should be expressed efficiently in primate T cells. All three vectors also contain the untranslated partial gag sequence of MoMLV (from the SRLEGFP vector) and the EGFP gene as a reporter. We have also constructed CS-MLV-E and CS-Rh-MLV-E vectors that are devoid of the partial gag sequence. VSVG-pseudotyped vectors were generated by transient cotransfection of each lentivirus vector construct with a packaging construct and a VSVG expression construct in 293T cells. The culture supernatant of the transfectant cells was collected, and the titer was determined by quantitation of the number of EGFP-positive HeLa cells in flow cytometry. The unconcentrated virus supernatants of all vectors, including the CSCG vector, generally yielded a titer of 0.1 × 10⁶ to 2 × 10⁶ IU/ml (see figures below). Thus, the use of murine oncoretrovirus LTR internal promoters in the context of a SIN vector provides vector titers comparable to those of CSCG. SRLEGFP generally yielded a titer of 10⁴ IU/ml.

View larger version (24K): [in this window] [in a new window]   FIG. 1.   Comparison of the control of EGFP gene expression in CSCG and SRLEGFP vectors. HeLa and SUPT1 cells were infected by unconcentrated virus supernatant of CSCG (virus titer, 0.5 × 105 IU/ml; MOI, 0.125) and SRLEGFP (virus titer, 0.64 × 104 IU/ml; MOI, 0.016). The percentage of EGFP+ cells and the MFI of EGFP expression of the infected cells were analyzed by flow cytometry 3 days after infection. FSC, forward scatter; FL, fluorescence.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Maps of various lentivirus-retrovirus hybrid vectors developed from a SIN HIV-1-based CSCG vector. The internal CMV immediate-early promoter was removed from CSCG and replaced with an oncoretrovirus LTR (MLV, Rh-MLV, or MSV) with or without a partial gag sequence of MoMLV.

CS-MLV-E has higher EGFP expression in T cell lines than CSCG. The CS-MLV-E vector differs from CSCG only in the internal promoter (MLV LTR instead of CMV) and the inclusion of the gag_803-1612 of MoMLV. Thus, this lentivirus-based vector contains a transcriptional promoter-enhancer closely resembling that of the SRLEGFP provirus. CS-MLV-E was used to infect HeLa, 293T, and T-cell lines (SUPT1 and CEMX174) as target cells in order to analyze the promoter activity of the MLV LTR in the SIN vector. CSCG and SRLEGFP were used as controls. The percentage of EGFP-positive cells and the MFI of EGFP expression of the infected cells were analyzed 3 days after infection. The CSCG vector has a greater MFI of EGFP expression in HeLa and 293T (MFIs, 160 and 572, respectively) than in the CSCG-infected T-cell lines (Fig. 3). The MFIs of EGFP in the CSCG-infected SUPT1 and CEMX174 cells were usually 4- to 20-fold lower than those in HeLa or 293T cells (Fig. 1 and 3). In contrast, the MFI of EGFP expression in SRLEGFP-infected SUPT1 or CEMX174 cells is always two- to sixfold higher than that in HeLa or 293T cells (Fig. 1 and 3). The EGFP expression of SRLEGFP in SUPT1 or CEMX174 is generally 6- to 10-fold higher than that of CSCG, despite the fact that SRLEGFP has a lower virus titer than CSCG (Fig. 3 and Table 1). We compared the EGFP expression of our CS-MLV-E vector to that of CSCG and SRLEGFP. The EGFP expression of our CS-MLV-E vector in these four cell lines resembles that of the SRLEGFP vector, i.e., the CS-MLV-E vector has a higher EGFP expression in T-cell lines (SUPT1 and CEMX174) than in HeLa and 293T cells. The MFI of EGFP expression of the SUPT1 or CEMX174 cells which were infected by unconcentrated virus supernatant of CS-MLV-E is 3- to 10-fold higher than that of the CSCG-infected cells (Fig. 3).

View larger version (39K): [in this window] [in a new window]   FIG. 3.   EGFP expression of CS-MLV-E in lymphoid- and non-lymphoid-cell lines. Lymphoid cells (SUPT1 and CEMX174) and non-lymphoid cells (HeLa and 293T) were infected by unconcentrated virus supernatant of CSCG (virus titer, 0.8 × 105 IU/ml; MOI, 0.2), CS-MLV-E (virus titer, 0.96 × 105 IU/ml, MOI, 0.24), and SRLEGFP (virus titer, 2 × 104 IU/ml; MOI, 0.05). The percentage of EGFP+ cells and the MFI of EGFP expression of the infected cells were analyzed by flow cytometry 3 days after infection. FSC, forward scatter; FL, fluorescence.

View larger version (36K): [in this window] [in a new window]   FIG. 4.   EGFP expression of CS-MLV-E in SUPT1 cells is higher than that of CSCG at low MOIs. SUPT1 cells (0.1 × 106) were infected by virus supernatant of CSCG (virus titer, 3 × 105 IU/ml), CS-MLV-E (virus titer, 1.2 × 106 IU/ml), and SRLEGFP (virus titer, 0.7 × 105 IU/ml) at the indicated dilutions (undiluted [1×, MOI = 0.8] and 1/10× and 1/20× dilutions for CSCG; 1× (MOI = 3.0), 1/20×, 1/50×, and 1/100× dilutions for CS-MLV-E; 1× (MOI = 0.18), 1/10×, and 1/20× dilutions for SRLEGFP). The percentage of EGFP+ cells and the MFI of EGFP expression of the infected cells were analyzed by flow cytometry 3 days after infection. FSC, forward scatter; FL, fluorescence.

Rh-MLV LTR, a novel MoMLV-derived LTR, has stronger promoter activity than MLV LTR. A novel retrovirus (AMP-MCF) involving sequences derived from endogenous murine MCF-type viruses was isolated from the serum of one rhesus monkey that developed rapidly progressive T-cell lymphomas (29). It arises by recombination involving the genome of the vector, packaging-encoding sequences, and endogenous murine retroviral genomes in the producer clone, followed by further mutation in the rhesus monkey. We designated the 5' LTR of this AMP-MCF virus the Rh-MLV LTR. The U3 segment of the Rh-MLV LTR has a 23-bp insertion, unlike the sequence of MLV LTR. The inserted segment is a direct copy of the 23 bp 3' to the site of insertion and is inserted within the 5' member of the two 75-bp direct repeats which compose the enhancer region of U3. We have replaced the internal CMV promoter of CSCG with this Rh-MLV LTR and the partial gag sequence (gag_871-1612) of MoMLV (CS-Rh-MLV-E). Virus supernatant of CS-Rh-MLV-E was used to infect HeLa, 293T, SUPT1, and CEMX174 cell lines, and EGFP expression in the infected cells was analyzed.

View larger version (35K): [in this window] [in a new window]   FIG. 5.   Partial gag sequence of MoMLV in CS-MLV-E and CS-Rh-MLV-E vectors is involved in enhanced EGFP expression in CEMX174 and SUPT1. CEMX174 and SUPT1 cells were infected by virus supernatant of CSCG (virus titer, 0.3 × 106 IU/ml; MOI, 0.075), CS-MLV-E (virus titer, 1.1 × 106 IU/ml; MOI, 0.055), CS-MLV-E (virus titer, 1.3 × 106 IU/ml; MOI, 0.065), CS-Rh-MLV-E (virus titer, 1.9 × 106 IU/ml; MOI, 0.095), and CS-Rh-MLV-E (virus titer, 1.2 × 106 IU/ml; MOI, 0.06). The percentage of EGFP+ cells and the MFI of EGFP expression of the infected cells were analyzed by flow cytometry 3 days after infection. FSC, forward scatter; FL, fluorescence.

The partial gag sequence of MoMLV in CS-MLV-E and CS-Rh-MLV-E is essential for enhanced EGFP expression in CEMX174 cells. We investigated whether the partial, untranslated gag sequence of MoMLV present in both the CS-MLV-E and CS-Rh-MLV-E vectors affects EGFP expression in infected cells. The partial gag sequence has previously been shown to encode packaging signal for MoMLV retrovirus that markedly increases the titer of in vitro-packaged retroviral vectors (5, 17). We deleted this partial gag sequence of MoMLV (gag_871-1612) in CS-MLV-E (CS-MLV-E) and CS-Rh-MLV-E (CS-Rh-MLV-E) and compared their EGFP expressions in HeLa, 293T, SUPT1, and CEMX174 cell lines with those of CS-MLV-E and CS-Rh-MLV-E, respectively. This sequence should have no effect upon packaging of lentivirus vectors, and as indicated, we observed no change in these lentivirus vector titers. Deletion of the partial gag sequence in CS-MLV-E and CS-Rh-MLV-E did not significantly decrease the MFI of EGFP in HeLa or 293T cells, possibly due to the relatively low EGFP expression under the control of MLV or Rh-MLV LTRs in these cells (Table 1). However, an approximately twofold decrease in EGFP expression was observed in the infected SUPT1 and CEMX174 cells when the partial gag sequence was deleted (Fig. 5). These results demonstrated that (i) the partial gag sequence of MoMLV in these vectors is essential for the enhanced EGFP expression in these two T-cell lines; (ii) MLV or Rh-MLV LTR alone gave rise to a twofold increase in EGFP expression in SUPT1 cells compared to that of the CMV promoter in CSCG; (iii) the two- to threefold increase in EGFP expression of the CS-MLV-E- and CS-Rh-MLV-E-infected CEMX174 cells, in comparison to that of CSCG, could be largely attributed to the presence of this partial gag sequence.

CS-MLV-E and CS-Rh-MLV-E have higher EGFP expression in infected, primary PBMC culture. The above-mentioned results show that, in infected SUPT1 cells, the MLV LTR is a stronger promoter than CMV (comparing CS-MLV-E with CSCG), the Rh-MLV LTR is a stronger promoter than the MLV LTR (comparing CS-Rh-MLV-E with CS-MLV-E), and the partial untranslated gag sequence of MoMLV (gag_871-1612) is required for further enhancement of gene expression (comparing CS-MLV-E with CS-MLV-E and CS-Rh-MLV-E with CS-Rh-MLV-E). We tested our vectors in a primary PBMC culture that was activated by CD3 and CD28 antibodies. Activated T-cell blasts were then infected by virus supernatants of CSCG, CS-MLV-E, and CS-Rh-MLV-E and analyzed for EGFP expression 3 days postinfection. As shown in Fig. 6A, we found that CS-MLV-E and CS-Rh-MLV-E had four- and eightfold-higher EGFP expression, respectively, than CSCG, which was fully consistent with the results obtained in SUPT1 and CEMX174 cells. The expression of EGFP remained high and stable when we analyzed the same infected culture 8 days postinfection. These results confirm that the CS-MLV-E and CS-Rh-MLV-E vectors have higher gene expression levels in T lymphoid cells than CSCG and that Rh-MLV is a stronger promoter than MLV in T cells. In addition, we have shown that control of EGFP expression under these vectors is similar in both CD4⁺ and CD8⁺ T cell subsets (Fig. 6B).

View larger version (364237K): [in this window] [in a new window]   FIG. 6.   CS-Rh-MLV-E vector allows a high level of EGFP expression in primary activated T cells. (A) Flow cytometry analysis of CSCG-, CS-MLV-E-, and CS-Rh-MLV-E-infected T cells on day 3 or day 8 postinfection. (B) EGFP expressions in the infected CD4+ and CD8+ T-cell subsets were similar. Human PBMC were activated by plate-bound anti-CD3 and anti-CD28 MAbs for 60 h. Activated T-cell blasts were infected by virus supernatant of CSCG (virus titer, 0.3 × 106 IU/ml; MOI, 0.75), CS-MLV-E (virus titer, 1.1 × 106 IU/ml; MOI, 0.55), CS-MLV-E (virus titer, 1.3 × 106 IU/ml; MOI, 0.65), CS-Rh-MLV-E (virus titer, 1.9 × 106 IU/ml; MOI, 0.48), and CS-Rh-MLV-E (virus titer, 1.2 × 106 IU/ml; MOI, 0.3). The percentage of EGFP+ cells and the MFI of EGFP expression of the infected cells were analyzed by flow cytometry on days 3 and 8 postinfection. On day 8, aliquots of the infected cells were also stained for surface expression of CD4 and CD8 molecules and analyzed by fluorescence-activated cell sorter. (C) Flow cytometry analysis of the CSCG-, CS-RhMLV-E-, and SRLEGFP-infected SUPT1 and activated primary T cells. SUPT1 cells were infected by virus supernatant of CSCG (virus titer, 1 × 106 IU/ml; MOI, 2.5), CS-RhMLV-E (virus titer, 1.4 × 106 IU/ml; MOI, 0.07), and SRLEGFP (virus titer, 7 × 104 IU/ml; MOI, 0.18) at the following dilutions: undiluted (1×) for CSCG, 1/50× dilution for CS-Rh-MLV-E, and 1× for SRLEGFP. Activated T-cell blasts were infected by undiluted virus supernatant of CSCG (virus titer, 1 × 106 IU/ml; MOI, 2.5) and CS-RhMLV-E (virus titer, 1.4 × 106 IU/ml; MOI, 3.5), except for SRLEGFP, where 140×-concentrated virus supernatant (virus titer, 9.8 × 106 IU/ml; MOI, 25) was used in the infection. The percentage of EGFP+ cells and the MFI of EGFP expression of the infected cells were analyzed by flow cytometry on day 3 postinfection. FSC, forward scatter; FL, fluorescence.

Further enhancement of EGFP expression by the Rh-MLV LTR and the partial gag sequence (gag_871-1612) in a SIN vector that has further deletions in the U3 region. In the CS-G SIN vector setting, 133 nucleotides in the U3 region of the 3' LTR, including the TATA box and binding sites for transcription factors Sp1 and NF-B, were deleted (19). A more extensive deletion in the U3 region of the 3' LTR (eliminating all of the transcriptional elements located upstream of the SP1 binding sites) was used in the SIN vector (SIN-18) used in the construction of RRL-PGK-EGFP-SIN18 (31).

View larger version (35K): [in this window] [in a new window]   FIG. 7.   Further enhancement of EGFP expression by the Rh-MLV LTR and the partial gag sequence in a SIN vector that has further deletions in the U3 region. SUPT1 cells were infected by virus supernatant of CSCG (virus titer, 0.5 × 106 IU/ml; MOI, 0.025), CS-RhMLV-E (virus titer, 0.5 × 106 IU/ml; MOI, 0.025), RRL-PGK-EGFP-SIN18 (virus titer, 0.3 × 106 IU/ml; MOI, 0.075), and SIN18-Rh-MLV-E (virus titer, 0.6 × 106 IU/ml; MOI, 0.03) at the following dilutions: 1/50× for CSCG, CS-RhMLV-E, and SIN18-Rh-MLV-E and 1/10× for RRL-PGK-EGFP-SIN18. For T-cell blast infection experiments, 1× undiluted virus supernatants of CSCG (MOI, 1.25), CS-Rh-MLV-E (MOI, 1.25), RRL-PGK-EGFP-SIN18 (MOI, 0.75), and SIN-18-Rh-MLV-E (MOI, 1.5) were used. The percentage of EGFP+ cells and the MFI of EGFP expression of the infected cells were analyzed by flow cytometry 3 days after infection. FSC, forward scatter; FL, fluorescence.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We studied gene expression under the control of an internal CMV immediate-early promoter and MSV, MLV, and Rh-MLV LTRs in the SIN lentiviral vector. Transduction of SUPT1 cells by the vector that contains the internal MLV LTR resulted in a twofold increase in EGFP expression over the CMV promoter. The EGFP expression can be further increased by twofold when a gag sequence (gag_803-1612) derived from MoMLV is included in the CS-MLV-E construct. Interestingly, replacement of MLV by Rh-MLV LTR (CS-Rh-MLV-E) allowed a twofold-higher EGFP expression than CS-MLV-E and a five- to eightfold-higher EGFP expression than the CMV promoter. There was a further 2- to 3-fold increase in EGFP expression (thus, a 10-fold-higher gene expression than CMV) when the Rh-MLV promoter and gag_871-1612 were used in a SIN vector setting (SIN-18) that has a further deletion in the U3 region of the HIV-1 LTR. Thus, the combination of the Rh-MLV LTR and the gag sequence (gag_871-1612) of MoMLV in SIN-18-Rh-MLV-E in the context of the SIN-18 vector represented the strongest promoter activity in the T-cell lines and activated primary T cells tested.

The CS-MLV-E vector differs from CSCG only in the internal promoter (MLV LTR instead of CMV) and the inclusion of a partial gag sequence of MoMLV. Thus, this lentivirus-based vector contains a transcriptional promoter-enhancer unit closely resembling that of the murine oncoretrovirus SRLEGFP provirus. Perhaps not surprisingly, we found that the control of EGFP expression under the CS-MLV-E vector in HeLa, 293T, SUPT1, and CEMX174 cells resembles that under the SRLEGFP vector. Notably, CS-MLV-E maintained two- to threefold-higher EGFP expression than CSCG independent of the MOI. However, we found that EGFP expression of CS-MLV-E, despite having a higher percentage of infectivity (EGFP-positive cells), is still three- to fivefold lower than that of SRLEGFP. It is not clear why SRLEGFP expresses EGFP at a higher level than CS-MLV-E. There are three mutually nonexclusive possibilities. First, the size of the SRLEGFP proviral genome is smaller than that of CS-MLV-E, thus allowing a more efficient transcription and translation. Second, there exists a DNA sequence downstream of the 3' MLV LTR promoter in the SRLEGFP construct that we did not clone into CS-MLV-E, and the context of the MLV LTR may influence expression. Third, the transcriptional elements located upstream of the SP1 binding sites in the U3 region of the 3' LTR of the CS-G SIN vector may interfere with the activity of the internal MLV promoter. In this regard, it is of interest to compare the levels of gene expression of CS-Rh-MLV-E and SIN-18-Rh-MLV-E because the vectors contain identical Rh-MLV internal promoters and the gag_871-1612 sequence. One major difference between the two vectors is the SIN settingCS-Rh-MLV-E vector contains a 133-nucleotide deletion in the U3 region of the 3' LTR that includes the TATA box and binding sites for transcription factors Sp1 and NF-B, whereas SIN-18-Rh-MLV-E has a more extensive 400-nucleotide deletion in the U3 region of the 3' LTR that has eliminated all of the transcriptional elements located upstream of the SP1 binding sites. The observation that SIN-18-Rh-MLV-E has two- to threefold-higher EGFP expression than CS-Rh-MLV-E supports the hypothesis of residual promoter interference in the CS-Rh-MLV-E vector.

We compared the efficiency of T-lymphoid-cell transduction by SRLEGFP and our lentivirus-oncoretrovirus hybrid vector. Despite the high level of gene expression in T cells by SRLEGFP, we reasoned that this retrovirus vector is undesirable for gene therapy of T cells for two reasons. First, SRLEGFP virus titer is consistently 10- to 100-fold lower than that of the lentivirus (or lentivirus-oncoretrovirus hybrid) vector, making it more difficult to produce a large quantity for gene therapy purposes. Second, although SRLEGFP is expressed efficiently in T cells, it requires a much higher MOI to achieve a low level of gene transduction in primary T-cell blasts compared to that required by CS-Rh-MLV-E (Fig. 6C). SIN-18-Rh-MLV-E expressed EGFP at a level comparable to that of SRLEGFP in T lymphoid cells with relatively high transduction efficiency (Fig. 7). Thus, SIN-18-Rh-MLV-E may prove useful in T-lymphoid-cell gene therapy because of its relatively high virus titer, ability to infect T cells efficiently, and high level of gene expression in T cells.

The partial gag sequence has previously been shown to encode a packaging signal for MoMLV retrovirus that markedly increases the titer of in vitro-packaged retrovirus vectors. However, the MoMLV-packaging sequence should not affect HIV-1 packaging. In fact, we show that inclusion of this sequence in CS-MSV-E, CS-MLV-E, or CS-Rh-MLV-E does not significantly affect the virus titer when compared to the parental CSCG vector. Interestingly, we found that deletion of this extended packaging sequence of MoMLV resulted in an approximately twofold decrease in EGFP expression observed in the infected SUPT1 and CEMX174 cells. These results demonstrated a novel, unappreciated feature of this extended packaging sequence in gene expression. There are two possible explanations for the enhanced gene expression in these vectors. First, the extended packaging sequence stabilizes the mRNA transcript and therefore allows higher gene expression. Second, it may allow higher gene expression through a more efficient mRNA nuclear export. Recently, King et al. reported that incorporation of this extended packaging sequence into Rex-dependent expression vectors based on the human T-cell leukemia virus allows Rex-independent gene expression (14). Using fluorescent in situ hybridization combined with confocal microscopy, they showed that the 312-nucleotide element can replace Rex-mediated nuclear export and expression of transcripts containing HTLV-1 cis-acting repressive elements. Furthermore, it is interesting to note that introduction of this partial gag sequence (gag_871-1612) downstream of the CMV promoter (CS-CMV-gagE) or PGK promoter (SIN-18-PGK-gagE) diminished gene expression when compared with the CMV (CSCG) and PGK (RRL-PGK-EGFP-SIN18) promoters alone, suggesting that the activity is dependent on being in the context of an MLV LTR.

In our present study, we found that a combination of the novel MoMLV-derived LTR (Rh-MLV) and the partial gag sequence of MoMLV allowed the strongest EGFP expression in the T-cell lines and activated primary T cells. This primate-derived murine retrovirus LTR, Rh-MLV LTR, is identified in the AMP-MCF retrovirus that was isolated from the serum of one rhesus monkey that developed rapidly progressive T-cell lymphomas (29). It is important to note that multiple copies of this AMP-MCF virus genome were only present in tumor DNA of one of the three animals that developed lymphoma and that tumor cells are of clonal origin with respect to the retroviral integration sites. These data are consistent with a pathogenic mechanism in which cell transformation and clonal tumor evolution arise from insertional mutagenesis of critical growth control genes by retrovirus integration. As with oncogenesis by weakly oncogenic retroviruses (those of birds, for example), cancer is induced by rare integration events resulting from a high level of chronic replication of the retroviruses. The use of such a primate-derived LTR, therefore, would not result in a new risk for the use of such vectors in clinical applications. Future experiments will be necessary to address other safety issues (such as the potential recombination between the endogenous retroviral sequences and the gag_871-1612 sequence of MuLV).

Genetic modifications of hematopoietic progenitor cells (1, 3, 7, 20) could lead to expression of genes of interest in lymphoid cells, thus allowing potential treatment of dysfunctional T cells, which are often disrupted by an inherited or acquired immunodeficiency, such as HIV-1 disease (10). Therefore, it is desirable to obtain SIN HIV-1-based vectors that allow a high level of gene expression in T lymphocytes. CS-Rh-MLV-E and SIN-18-Rh-MLV-E show 5- and 10-fold-higher EGFP expression, respectively, than the CMV promoter (CSCG). In addition, a direct comparison of SIN-18-Rh-MLV-E with a PGK-containing vector, RRL-PGK-EGFP-SIN18, shows that there was a 6- to 10-fold-higher MFI of EGFP expression with SIN-18-Rh-MLV-E (Fig. 7 and Table 2), demonstrating that the combination of the Rh-MLV LTR and gag_871-1612 sequence performs better than the housekeeping PGK gene promoter. In summary, we have generated lentivirus vectors that have combined useful features of HIV-1 and MoMLV. These should be useful in allowing a higher level of therapeutic gene expression in lymphoid cells in gene therapy.