INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Gene transfer with potential therapeutic applications represents a very important and exciting challenge in biology and medicine. To this end, several generations of retroviral vectors derived from type C retroviruses have been developed that are capable of delivering a gene into many different cell types. For example, MLV-based dicistronic vectors were recently found to efficiently deliver two genes in human multipotent neuronal precursors (9). However, an important limitation of murine leukemia virus (MLV)-based vectors is their inability to deliver a gene into highly differentiated or growth-arrested cells. On the other hand, MLV vectors can be used in cancer gene therapy (17).

Interestingly enough, lentiviruses are able to infect growth-arrested cells (19) and this appears to be under the control of genetic determinants located in the integrase domain and in the vpx gene encoding a highly basic protein. In support of these genetic findings, the integrase and Vpx proteins are present in the viral particle and are thought to drive the nuclear import of the reverse transcription complex, also called the preintegration complex, probably by interacting with components of the nuclear pore complex (13). This route of viral infection, which appears to be unique to lentiviruses in the retrovirus family, has been exploited for the development of lentivirus vectors derived from human immunodeficiency virus type 1 (HIV-1) (29, 34), HIV-2 (2), feline immunodeficiency virus (30), equine infectious anemia virus (24), and more recently simian immunodeficiency virus (SIV) (37). As expected, these lentivirus vectors can achieve gene transfer into growth-arrested and differentiated cells ex vivo and in vivo (28).

Here we report the construction and characterization of several retroviral vectors based on SIVmac251. The minimal SIV vectors contain less than 20% of the original viral sequences, and recombinant viruses produced by 293T cells were found to carry out very efficient gene transfer into both cycling and growth-arrested cell lines. As a first step toward gene transfer applications, we used the minimal SIV vectors to transduce primary human dendritic cells (DCs). The results show that up to 40% of terminally differentiated DCs can be transduced.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmid DNA construction and amplification. In plasmid pSIV-TGP, which was constructed by Transgene SA and derives from the SIVmac251 genome (GenBank accession no. M19499), a green fluorescent protein (GFP) expression cassette driven by the early cytomegalovirus (CMV) promoter was inserted in place of the env 5' region (see Fig. 1) and an early CMV promoter pCMV replaced U3 in the 5' long terminal repeat (LTR). The pSIV-REV and pMLV-VL30 (GFP) vectors were provided by Transgene SA.

Construction of pSIV-RMES3. (i) The 3' PPT LTR was amplified using primers 1 and 2, and the resulting DNA was cut by KpnI and XhoI and inserted into pBlueScript KS+. (ii) The pCMV-GFP cassette was amplified using primers 3 and 4, and the DNA was cut by ClaI and XhoI and inserted into the plasmid generated in step i. (iii) The RRE region was amplified using primers 5 and 6, and the resulting DNA was cut by MluI and ClaI and inserted into the MluI and ClaI sites in the plasmid generated in step ii. (iv) The 5' pCMV-R-U5-PBS-Leader and the 5' gag region were amplified using primers 7 and 8, and the DNA was cut by XbaI and MluI and inserted into the XbaI and MluI sites of the plasmid generated in step iii, giving rise to pSIV-RMES3, where 57 nt of the 5' end of gag are present (Fig. 1). Note that a stop codon has been engineered 15 nt downstream of the AUG translation initiation codon of gag.

Cell culture and differentiation of human CD34⁺ cells and monocytes. HeLa, sMAGI, and 293T cells were cultured at 37°C in an atmosphere of 5% CO₂ and in Dulbecco's modified Eagle's medium (Life Technologies-Gibco-BRL) complemented with 10% fetal calf serum (FCS), penicillin (100 IU/ml), streptomycin (100 µg/ml), and L-glutamine (2 mM). The selection medium for sMAGI cells contained hygromycin (50 µg/ml) and G418 (200 µg/ml).

5

Production of SIV vectors. Transfections of cells were performed using the calcium phosphate protocol as previously described (8). For large-scale production, we used the calcium phosphate transfection kit (Gibco-BRL). The pSIV vectors were produced in six-well plates by cotransfecting three different plasmids, a SIV vector (2 µg), the helper pSIV3+ coding for Gag-Pol and the regulatory viral proteins except Env and Nef (1.7 µg), and pVSV-G (0.8 µg), into 293T cells seeded the day before transfection at 3.5 · 10⁵ per well. To transfect cells in 10-cm-diameter dishes, amounts of DNAs and volumes were scaled up sixfold. The medium was replaced after 16 h of incubation. Vector-containing media were collected at 48 h or twice at 40 and 48 h after transfection (for large-scale production), clarified by 5 min of centrifugation at 800 × g, and filtered through a 0.45-µm-pore-size-filter. For high-titer preparations, SIV vectors were concentrated by ultracentrifugation at 50,000 × g for 90 min. The viral pellet was resuspended overnight at 4°C in 50 mM Tris · Cl (pH 8)-130 mM NaCl-1 mM EDTA or in phosphate-buffered saline.

Transduction and calculation of the titer in transducing units. For transduction assays, HeLa or 293T cells were seeded at 1 × 10⁵ to 3 × 10⁵ per well in 12-well plates the day before transduction. At day 1, the medium was removed and replaced with transduction medium made of 100 µl of vector, 300 µl of serum-free medium, and Polybrene at 6 µg/ml. After 3 h of incubation, 1 ml of complete medium was added. At day 1, cells from three nontransduced wells were trypsinized and fixed to allow calculation of the rate of cell division. At day 3, cells were trypsinized and GFP expression was analyzed by flow cytometry (FACScalibur; Becton-Dickinson).

Mobilization of the integrated SIV vector. pSIV-TGP, pSIV-T1, pSIV-RMES4, and pSIV-gaMES4sin were produced as described above and used to transduce 293T cells. At 48 h later, most of the cells scored positive for GFP expression (>75%). At 72 h after transduction, the same cells were transfected with a mixture containing Tat-encoding plasmid pSIV3+ (1.8 µg), pSIV-Rev (1.5 µg), and pVSV-G (0.5 µg). The culture medium was changed 1 day later and then collected four times at 8-h intervals, filtered, and used to transduce HeLa cells (400 µl of vector in 400 µl of serum-free Dulbecco's modified Eagle's medium with Polybrene at 12 µg/ml). Three days after transduction, HeLa cells were trypsinized and FACS analyzed for GFP expression.

PCR amplification. To analyze potential transfer of helper functions, 293T cells were transduced with either pSIV-gaMES4sin, pSIV-RMES4, pSIV-T1, or pSIVTGP or left untransduced. For all constructs, transduction efficiencies reached >70% of GFP-positive cells. At 48 h after transduction, 10⁶ cells were trypsinized, pelleted, and resuspended in an excess of lysis buffer containing 50 mM Tris-HCl (pH 8), 30 mM EDTA, 500 mM LiCl, 0.5% sodium dodecyl sulfate, 0.1% -mercaptoethanol, and proteinase K at 100 µg/ml. After 1 h of incubation at 37°C, lysates were phenol extracted and DNAs were alcohol precipitated. DNAs were resuspended in pure water and handled to avoid contamination. PCR amplifications were performed as follows on 100 or 300 ng (for gag) of DNA: 40 cycles of 45 s at 94°C, 45 s at 55°C, and 45 s at 72°C in a 50 µl volume containing 40 µl of H₂O, 5 µl of 10× polymerase buffer (Gibco), 2 µl of 25 mM MgCl₂, 0.8 µl of 25 mM deoxynucleoside triphosphate, 5 U of Taq DNA polymerase (Gibco), and 25 pmol of each primer. The primers used were as follows: pCMV-GFP, primers 3 and 4 (detailed in plasmid construction; LTR-E/DLS, 5'-GGCACGCGTCTTCCCTCACAAGACGGAGTTTC and primer 1; gag, 5'-GCAGATGAATTGAAAAAATTAG-3' and 5'-TTATTATCCCTTCCTGGATAACAAGACAGC-3'.

RCR detection. To detect a potential replication-competent virus (RCR) in the SIV vector preparation, pSIV-TGP, pSIV-RMES4, and pSIV-gaMES4sin were produced in a six-well plate as already described. Supernatants were clarified and filtered through a 0.45-µm-pore-size filter, and 1 ml of each vector stock was added to 3 × 10⁵ sMAGI cells in a six-well plate. On the same day, a SIVmac251 stock whose titer was predetermined at 3 × 10³ infectious units (IU)/ml (on sMAGI cells) was diluted 10 and 100 times and 100 µl of the stock or of each dilution was used to infect 3 × 10⁵ naive sMAGI cells in a six-well plate. Thus, the numbers of infectious units added for each dilution can be estimated to be 300, 30, and 3. After 24 h, infected and transduced cells were washed twice and allow to grow for 1 week. Titers of the three vector stocks and efficiencies of transduction were determined by FACS analysis of sMAGI cells transduced in a control experiment. After 7 days, all sMAGI cells were fixed in phosphate-buffered saline-4% formaldehyde (15 min at 4°C) and stained for -galactosidase expression (1 h at 37°C). Infection was evident in wells infected with the three dilutions of SIVmac251 as revealed by blue foci of infection. A large number of blue cells were observed upon transduction with pSIV-TGP, which codes for the tat gene, but no blue cells were detected upon transduction with the minimal SIV vectors.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of minimal SIV vectors. Starting from the SIVmac251V molecular clone, the first SIVmac-derived vector (pSIV-TGP) was constructed by inserting a GFP expression cassette in place of env and by replacing the 5' LTR U3 region with the CMV early promoter (pCMV). SIV vector production was achieved by cotransfecting pSIV-TGP together with pVSV-G, encoding the VSV-G envelope, into 293T cells. This first SIV vector was found to have GFP transduction titers on the order of 1.4 × 10⁷ transducing units (TU)/ml on HeLa cells (Fig. 1 and 2). The strategy used to engineer a minimal SIV vector was to keep the essential cis elements, i.e., the LTR, the PBS, the E/DLS, the RRE, and the PPT, on the vector construct while segregating the viral helper functions on other constructs (Fig. 1, vector and helper plasmids). Since the E/DLS packaging signal is thought to extend into the 5' end of gag (22) (33), a functional E/DLS packaging signal was selected by progressively deleting the gag-pol region. Deletion of all but the 5' 218 nt of gag did not result in serious adverse effects on the transduction titer of the SIV vector (pSIV-T1; Fig. 1 and 2) while removing all of gag and the 3' 100 nt of the leader caused a 40-fold reduction in the transduction titer (compare pSIV-T0 to pSIV-T1; Fig. 1 and 2). These data indicate that the 3' end of the leader and the first 5' nucleotide of gag must be included in a minimal SIV vector since they probably are part of the E/DLS packaging signal. In agreement with this result, in vitro SIV RNA dimerization was found to require sequences located from the PBS to the AUG start codon of gag (data not shown).

View larger version (27K): [in this window] [in a new window]   FIG. 1.   Schematic representation of SIV-derived vectors and helper plasmids. pSIV-T0, -T1, -RMES, and -gaMES4sin were constructed as described in Materials and Methods. Viral genes and restriction sites of interest are indicated on the SIV proviral map. gag sequences present in the vector constructs are shown in black. Abbreviations: pCMV, early CMV promoter; LTR SIN, 3' LTR with U3 partially deleted. Note that nef is nonfunctional in all SIV constructs.

View larger version (39K): [in this window] [in a new window]   FIG. 2.   Transduction titers of SIV vectors. All vectors, except the MLV vector and pSIV-TGP, were produced by cotransfection of 293T cells with the indicated vector, pSIV3+, and pVSV-G. Vector-containing supernatants were clarified and diluted before transduction of HeLa cells. To calculate titers and standard deviations, three distinct vector preparations were analyzed. The SIV vector constructs are shown in Fig. 1.

Analysis of RCR. The helper functions encoding the viral genes, the envelope, and the SIV vector are segregated on distinct plasmid constructs (Fig. 1), and this is thought to prevent the generation of replication-competent SIV-derived viruses. However, 293T cells are transfected with large quantities of plasmid DNAs in order to produce large amounts of SIV vector particles and this may be the source of multiple recombination events giving rise to SIV-derived RCRs potentially carrying helper functions.

View larger version (31K): [in this window] [in a new window]   FIG. 3.   PCR analysis of SIV-directed gene transfer into 293T cells. Production of pSIV-TGP, pSIV-T1, pSIV-RMES4, and pSIV-gaMES4sin is described in Materials and Methods. Virus-containing supernatants were filtered and used to transduce 293T target cells. Three days after transduction, cells were lysed and genomic DNAs were extracted to perform PCR amplifications. (A) PCR assays performed on DNA extracted from transduced 293T cells that corresponds to the pCMV-GFP expression cassette (1,400 bp) and the 5' LTR-E/DLS region (1,100 bp). (B) Amplification of gag in transduced cells. The viral supernatants used to transduce 293T cells were pSIV-gaMES4sin (lane 2), pSIV-RMES4 (lane 3), pSIV-T1 (lane 4), and pSIV-TGP (lane 5). As internal controls, PCR assays were performed on nontransduced cells (lane 1) or H2O (lane 6). DNA extracted from every cell population allowed amplification of -globin (80-bp bands in panel A). (C) PCR assays performed on serial dilutions of control plasmids (pSIV-RMES4 for GFP and pSIV3+ for Gag). Calculated numbers of copies are indicated. Domains subjected to amplification and localization of the primers are depicted. While the GFP and LTR-E/DLS regions were amplified in all of the DNA samples from transduced cells, Gag was detected only in DNA corresponding to cells transduced with pSIV-TGP (B, lane 5). In lane 2, note the 150-nt deletion present only in the LTR DNA from cells transduced with pSIV-gaMES4sin.

5

6

Construction of a minimal self-inactivating SIV vector. Due to the reverse transcription process, replication of the SIV vector will regenerate two copies of the 3' LTR and both of them will be integrated into the host genome. The presence of such possibly functional LTRs in the cellular genome has raised concerns regarding abnormal gene expression driven by the LTR in transduced cells. This prompted us to generate a minimal SIV vector with a large deletion in the 3' LTR U3 domain, which encompasses the binding sites for the Elf-1, NF-B, and SP1 transcription factors and the TATA box (Fig. 4), which are essential for SIV transcription (31). Deletion of this 151-nt region was carried out as described in Materials and Methods and enabled construction of the minimal pSIV-gaMES4sin vector. This minimal SIV sin vector was found to have a high transduction titer on HeLa cells (2 × 10⁶ TU/ml; Fig. 2) and was able to carry out gene transfer in growth-arrested cells with an efficiency comparable to that of other SIV vectors (see Fig. 6; compare pSIV-T1, pSIV-RMES4, and pSIV-gaMES4sin).

View larger version (8K): [in this window] [in a new window]   FIG. 4.   Representation of the SIVmac core enhancer region and flanking sequences. The SIVmac251 sequence is in the GenBank database (accession no. M19499). The major functional elements of the core enhancer and flanking sequences have homology with the same genetic domain of SIVmac239 (31). Purine-rich regions PuB1 and PuB2 were deduced by homology with HIV-2 (18). Important binding motifs are boxed and arrows show the sin deletion (positions 334 to 485).

View larger version (27K): [in this window] [in a new window]   FIG. 5.   Mobilization of the integrated SIV vector. Mobilization of either pSIV-TGP, pSIV-T1, pSIV-RMES4, or pSIV-gaMES4sin was carried out as described in Materials and Methods. Briefly, 293T cells transduced with the indicated vector were cotransfected with pSIV3+, pSIV-Rev, and pVSV-G. Media of transduced and transfected cells were used to transduce naive HeLa cells. GFP expression in HeLa cells was analyzed by FACS 72 h after transduction. Note the high levels of mobilization obtained for pSIV-TGP and pSIV-T1 (91 and 74%, respectively). Mobilization of the integrated pSIV-RMES4 vector was drastically reduced (10% of GFP-positive cells) and <0.1% for the pSIV-gaMES4sin vector.

Transduction of growth-arrested human cells. One major interest in developing lentivirus vectors resides in their ability to transduce growth-arrested cells, and this paves the way to a large number of gene transfer applications impossible to achieve with MLV-derived vectors. A useful protocol by which to examine the ability of lentivirus vectors to transduce growth-arrested cells is to  irradiate the target cells in order to block them in the G₂ phase (19). To validate the present SIV vectors for the transduction of growth-arrested cells, 293T cells were  irradiated 1 day before transduction and then GFP gene transfer by the SIV vectors was measured as before (see Materials and Methods) and compared to that obtained with an MLV vector under the same conditions. Figure 6 reports the percentages of cell transduction obtained with the SIV and MLV vectors. Clearly, the transduction efficiencies of the SIV vectors on growth-arrested cells are comparable to the values obtained on proliferating cells (Fig. 6). On the other hand, transduction with the MLV vector was high on cycling 293T cells but about 100-fold lower on arrested cells (Fig. 6).

View larger version (42K): [in this window] [in a new window]   FIG. 6.   SIV vector transduction of proliferating and growth-arrested human cells. SIV vectors were produced as indicated in Materials and Methods. The vector-containing medium was filtered and added to proliferating 293T cells (filled bars) or to 293T cells  irradiated at 4,000 rads the day before use (open bars). Two days after transduction, cells were trypsinized and analyzed by FACS (see Materials and Methods). Percentages of GFP-positive cells are indicated. As expected, the MLV vector (pMLV-VL30) was unable to transduce growth-arrested cells. This result and the relative efficiencies with which SIV vectors transduced arrested HeLa cells were reproducible in three separate experiments.

Transduction of CD34⁺ cell- and monocyte-derived human DCs. DCs are of great interest for gene transfer, since as antigen-presenting cells they are key actors in the immune response (3). DCs are found in many nonlymphoid tissues such as the skin and mucosa but remain difficult to isolate. In cell cultures, DCs can be established from CD34⁺ progenitor cells or from peripheral blood monocytes by culturing the purified precursors in the presence of a cocktail of cytokines (4, 35). As MLV vectors require cell division to achieve gene transfer, MLV vector-directed transduction into DCs was performed first on CD34⁺ progenitors prior to interleukin-directed differentiation (26, 27, 32).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The data reported here summarize our efforts to construct minimal SIV vectors capable of delivering a gene into simian or human cells in culture. To this end, we took advantage of our growing knowledge of SIV genetic structure and replication (10) and results show that cis elements and trans-acting factors can be segregated while preserving their functions. Regarding lentivirus vector development for gene transfer and biological safety concerns, it should be pointed out that SIVmac provides a very interesting and useful animal model for investigating pathogenesis and prevention of HIV-1 infection (10). It should also be noted that our minimal SIV vector expressing system does not code for the Env and Nef proteins that are required for SIV replication and pathogenesis in rhesus macaques, the natural host of SIVmac (11, 16).

By means of a systematic approach using high-fidelity PCR amplification to generate specific SIV DNA sequences, we have been able to construct the minimal vectors SIV-RMES3 and SIV-RMES4, which combine small size and the ability to efficiently transfer a gene to cells in culture. As observed for other lentivirus vectors, the minimal and efficient SIV-RMES vectors contain 20% or less of the original SIV sequences. Interestingly, the functional E/DLS packaging sequence present in SIV-RMES vectors does not extend more than 57 nt into gag (Fig. 1 and 2). This situation is actually much different from that of MLV-based vectors, in which an efficient E/DLS packaging sequence contains 400 to 500 nt of the 5' gag region (1). This feature of minimal lentivirus vectors should minimize possible recombination events between the vector and helper RNAs and thus improve their biological safety. In agreement with this notion, the pSIV-T1 and minimal pSIV-RMES4 vectors were unable to transfer gag-pol sequences into target cells, as measured by PCR amplification (Fig. 3).

To further improve the predictive safety of the minimal pSIV-RMES vector, essential cis transcription elements, such as the binding sites for the Elf-1, NF-B, and SP1 transcription factors and the TATA box, have been deleted from the 3' LTR (Fig. 4). Vector mobilization assays showed that this U3 deletion completely prevented LTR-driven transcription of the integrated SIV sin vector (Fig. 5). At the same time, the sin vector pSIV-gaMES4 was still very efficient on both cycling and growth-arrested cells (Fig. 6).

As opposed to that of MLV-derived vectors, the production of large amounts of SIV vectors, and of lentivirus vectors in general, suffers from the fact that no high-producer helper cell lines are available. This is thought to be due, at least in part, to the presence of toxic viral regulatory proteins, such as Rev, which is essential for vector and helper function expression (23, 25). Other viral factors, such as Vif and Vpr, may also interfere with the establishment of efficient producer helper cell lines, but our preliminary data indicate that they are dispensable for SIV vector production (29a). Large amounts of specific lentivirus vectors will most probably be required for biomedical purposes but at the same time will necessitate established and well-characterized cell lines providing in trans all of the necessary packaging functions, and this is why the development of inducible SIV helper and producer cell lines is in progress.

DCs are present in lymphoepithelial tissues, including the mucosa and skin, and are mandatory not only for the initiation of the primary immune response to invasive pathogens but also in the control and maintenance of tolerance (36). Immunotherapy using DCs has been shown to induce immunity against tumors (38, 39) and can be envisioned as a strategy to treat autoimmune disease and to induce transplantation tolerance.

Previous studies have shown that a VSV-G-pseudotyped HIV-1 vector could transduce proliferating CD34⁺ progenitors (20). However, no evidence that a lentivirus vector could transfer a gene into mature nondividing DCs has been reported. We show here that the minimal SIV vector is highly efficient at delivering a gene into terminally differentiated human DCs derived from two distinct differentiation pathways. Transduction gave rise to up to 40% transduced DCs while preserving the viability and expression of DC-specific molecules, including CD1a, and of the costimulatory molecule CD80 (Table 2).

Monocyte-derived DCs are highly phagocytic compared to CD34⁺ cell-derived DCs (15), which are more efficient at inducing antigen presentation and T-cell activation. (12). The observations that transduction efficiencies were similar in monocyte-derived DCs and CD34⁺ cell-derived DCs and that an MLV vector was unable to transduce monocyte DCs indicate that transduction results from infection and not from phagocytosis. Moreover, these data suggest that the SIV vector may be valuable in vivo to target DCs of epithelial tissues and mucosa specialized in antigen processing, as well as DCs in secondary lymphoid organs specialized in T-cell priming. Finally, since the SIV vector is able to deliver genes without transferring the viral sequence, it is not supposed to elicit an immune response in the host. This advantage should allow repeated transductions with a limited risk of development of immunity to the vector. Minimal SIV vectors may thus open new avenues in many clinical applications requiring antigen targeting into DCs, such as antitumor therapy or vaccination.