A Stable System for the High-Titer Production of Multiply Attenuated Lentiviral Vectors

Transcription

1 doi: /mthe , available online at on IDEAL A Stable System for the High-Titer Production of Multiply Attenuated Lentiviral Vectors Natacha Klages, Romain Zufferey, and Didier Trono 1 Department of Genetics and Microbiology, Faculty of Medicine, University of Geneva, Geneva, Switzerland Received for publication April 4, 2000, and accepted in revised form June 27, 2000 Lentiviral vectors open exciting perspectives for the genetic treatment of a wide array of inherited and acquired diseases, owing to their ability to govern the efficient delivery, integration, and long-term expression of transgenes into nondividing cells both in vitro and in vivo. The genomic complexity of HIV, where a whole set of genes encode virulence factors essential for pathogenesis but not required for gene transfer, allowed a major step toward clinical acceptability through the creation of multiply attenuated packaging systems. Until now, however, vector particles could only be produced by transient transfection because no high-output, stable packaging cell line was available that produced the latest generation of HIV-based vectors. Here we describe such a line, based on the doxycycline-repressible expression of HIV-1 Rev/Gag/Pol and of the vesicular stomatitis virus G envelope (VSV G) in 293 human embryonic kidney cells. Upon induction, the LVG clones can produce 1 to 20 HeLa-transducing units per cell per day for about a week, a yield that compares favorably with that of transiently transfected 293T cells. These virions exhibit functional properties similar to those of viruses produced transiently, in particular the ability to transduce nonmitotic targets. This system will facilitate the further development of lentiviral vectors for gene therapy. Key Words: gene therapy; lentiviral vectors; HIV-based vectors; retroviral vectors; transduction; nondividing cells; VSV G pseudotypes; packaging cell line. INTRODUCTION Lentiviral vectors open new perspectives for the genetic treatment of human diseases. On the one hand, they present all the advantages of their oncoretroviral predecessors: a relatively large capacity, the nontransfer of virus-derived coding sequence, and the integration of the vector genetic cargo into the chromosomes of target cells. On the other hand, they can transduce nondividing cells, a crucial asset in cells such as neurons and hematopoietic stem cells. Lentiviral vectors are therefore particularly promising for gene therapy of neurological and lymphohematological disorders. Human immunodeficiency virus type 1 (HIV-1)-based vector particles are currently generated by coexpressing the virion packaging elements and the vector genome in a so-called producer cell, for instance a 293T human embryonic kidney cell. In a typical protocol, these cells are transiently transfected with three plasmids (1). The first codes for the core and enzymatic components of the 1 To whom correspondence should be addressed at CMU, 1, rue Michel -Servet CH-1211, Geneva 4, Switzerland. Fax: (41 22) didier.trono@medecine.unige.ch. virion, derived from HIV-1. The second is responsible for the envelope, most commonly the G protein of vesicular stomatitis virus (VSV G) because of its high stability and broad tropism. Finally, a third plasmid encodes the genome to be transferred to the target cell, that is, the vector itself. Recombinant viruses with titers of several millions of transducing units per milliliter (TU/ml) can be generated by this technique, and after ultracentrifugation concentrated stocks of approximately 10 9 TU/ml can be obtained. This allows for most in vitro experiments and for in vivo testing in small animal models. The generation of clinically acceptable lentiviral vectors, however, will require stable producer cell lines. This will eliminate the risk of DNA recombination between the plasmids encoding the various components of the vector, a potential source of replication-competent recombinants (RCR). Furthermore, this will facilitate the standardization of the vector production protocol, as well as the performance of all the necessary biosafety controls. Finally, the generation of large batches of vectors will be much easier, a great advantage not only for clinical studies but also for preclinical experiments in large animals. Lentivirus vector packaging cell lines have been previously described, but none was compatible /00 $35.00

2 with clinical applications. In several cases the vector titers were too low, and in most the tropism was limited to CD4 + cells because the cognate HIV envelope was used (2 5). One cell line produced VSV G-pseudotyped HIV vector particles at adequate titers, but its design posed unacceptable biosafety risks because the packaging construct was too close to the parental virus (6). Lentiviruses such as HIV-1 can replicate in nonmitotic cells because their so-called preintegration complex, a macromolecular structure comprising the viral genome, a few structural proteins, and the enzymes responsible for reverse transcription and integration, hijacks the cell nuclear import machinery. Fortunately, two-thirds of the HIV genes encode virulence factors that are essential for pathogenesis but completely dispensable for gene transfer (7). This allowed the creation of multiply attenuated lentivector packaging systems, the latest ( third ) generation of which comprises only three of the nine genes of HIV-1: gag, coding for the virion main structural proteins; pol, responsible for the retrovirus-specific enzymes; and rev, which encodes a posttranscriptional regulator necessary for efficient gag and pol expression (8). Cumulated evidence indicates that this multiply deleted system, the biosafety of which is theoretically higher than that of oncoretroviral vectors currently used in clinics, conserves all the properties of its first-generation predecessor, at least in the targets tested so far (8 11). In the retroviral genome, a single RNA molecule contains both the cis-acting elements needed for replication and all the coding sequences. Biosafety of a vector production system is therefore best achieved by distributing the sequences encoding its various components over as many independent units as possible, to maximize the number of crossovers that would be required to recreate an RCR. Accordingly, in the latest version of the HIVbased packaging system, Gag/Pol, Rev, VSV G, and the vector are produced from four separate DNA units (8). Also, the overlap between vector and helper sequences, the ground for homologous recombination, has been reduced to a few tens of nucleotides. Still, the use of this system was so far limited to the production of lentivector by transient transfection. Here we describe a lentivector producer cell line based on the tetracycline-repressible expression of a splitgenome, multiply attenuated HIV-derived packaging construct and the G envelope of VSV. Upon induction, 1 to 20 HeLa-transducing units per cell are produced daily for about a week. These virions exhibit functional properties similar to those of viruses produced transiently, in particular the ability to transduce nonmitotic targets. This system represents an important step toward the clinical use of lentiviral vectors. MATERIALS AND METHODS Cell lines and culture conditions. 293T and HeLa cells were cultured in Dulbecco s modified minimal Eagle s medium (DMEM) supplemented with 10% heat-inactivated fetal calf serum, 2 mm L-glutamine, 100 units/ml penicillin, and 100 µg/ml streptomycin. This medium is designated as 293T medium. 293G cells (a gift from D. Ory) express the VSV G protein under the control of a tetracycline-inducible promoter. 293G cells were generated by cotransfecting pmdtet.g, pbc.tta, and a puromycin resistance gene into 293 cells (12). 293G cells were cultured in the same medium as 293T cells supplemented with 2 µg/ml puromycin and 1 µg/ml doxycycline. This medium will be referred to as 293G medium. The packaging LV G clones were maintained in LV medium which corresponds to 293G medium supplemented with 100 µg/ml hygromycin B. To induce vector production, LV G clones were shifted from LV medium to 293T medium. Medium changes were done every other day in the early experiments and daily in the later experiments. Plasmids. The packaging plasmid pmdlg/prre encoding HIV-1 Gag and Gag/Pol proteins has been described previously (8). In the ptetrev plasmid the sequence encoding the HIV-1 Rev protein is under the control of a tetracycline-inducible promoter. ptetrev was constructed by inserting the HindIII XhoI fragment from prsvrev (8) into ptetsplice digested with HindIII and SalI. For the generation of packageable RNA, the plasmid prrl.gfp.cll (a gift from L. Naldini) was used. This plasmid contains the GFP coding sequence flanked by two chimeric LTRs: the 5 LTR is made of the Rous sarcoma virus (RSV) U3 region fused to the HIV- 1 R and U5 regions, and the 3 LTR is made of the cytomegalovirus (CMV) immediate early promoter fused to the HIV-1 R and U5 regions. The construction of these chimeric LTRs has been described (8). Following reverse transcription, duplication of the 3 LTR U3 region yields a provirus with a CLL.GFP.CLL configuration. Generation of LV G clones. To generate LV G clones, 293G cells ( cells in a 10-cm dish) were stably cotransfected with 20 µg of pmdlg/prre, 20 µg of ptetrev, and 2.5 µg of psv2hygro using the calcium phosphate precipitation method. Cells were cultured in the presence of the precipitate for 16 h, washed, and split at various dilutions (from 1:5 to 1:20). The following day, 293G medium was replaced with selective LV medium. After 14 days under selection LV G clones were picked and expanded individually until enough cells were available for freezing and testing. For each clone, two vials containing cells each were frozen in liquid nitrogen. For a subset of LV clones, the remaining cells were seeded in duplicate into six-well plates. In one well cells were cultured in the absence of doxycycline (293T medium) to induce Rev and Gag/Pol expression. In the other well, control cells were kept in LV medium for maintenance. The induced LV G clones were screened for expression of Gag/Pol by determining the reverse transcriptase (RT) activity in the culture medium after 6 and 10 days of culture without doxycycline. To test for the production of transduction-competent particles, two LV G clones (LV G -1 and LV G -2) releasing high levels of RT activity in the culture medium were transduced with the transiently produced lentiviral vector CLL.GFP.CLL at a multiplicity of infection (m.o.i.) of 2. To increase the number of vector genomes per cell, three additional rounds of transduction were performed under the same conditions. The GFP signal was analyzed after each round of transduction by fluorescence-activated cell sorting (FACS). Vector titration. All vector stocks were titrated on HeLa cells according to our previously described protocol (13). Briefly, HeLa cells (10 5 cells/well) were plated in 2 ml of 293T medium in six-well plates. On the following day, 100, 50, or 25 µl of unconcentrated vector stocks was added to the cells in duplicate. Polybrene was omitted. Medium was not changed until the percentage of GFP fluorescent cells was determined by FACS 2 to 3 days after infection. Titers in TU/ml were calculated by dividing the percentage of GFP-positive cells by the volume of the inoculum (in microliters) and multiplying the result by the starting number of cells and by 10 3 (to convert microliters to milliliters). Since the number of GFP-positive cells increases with the volume of the inoculum according to Poisson s distribution, titer calculations were based on two percentages of GFP-positive cells lower than 15% and related by a factor close to 2. Titers were also expressed on a per producer cell, per day basis. In this case, cells were washed thoroughly after being induced for 1 week or more, that is, when they were actively producing vector particles and had stopped growing. The total number of TU in the supernatant was measured 1 day later and divided by the number of cells on the plate. Vector production by transient transfection. CLL.GFP.CLL vector particles were produced using the usual transient transfection method (1, 8). 171

3 Briefly, 293T cells ( cells in 10-cm tissue culture dishes) were cotransfected with 13 µg of pmdlg/prre, 2 µg of prsvrev, 20 µg of prrl.gfp.cll, and 5 µg of pmd.g. After an overnight incubation in the presence of the precipitate, the culture medium (10 ml) was changed. On the following day, the culture medium was harvested, filtered through 0.45-µm pore-sized polyethersulfone membrane, and stored in 1-ml aliquots at 80 C. Vector titers obtained by this technique averaged 2.5 TU/producer cell. Reverse transcriptase assay. Ninety microliters of vector stocks was mixed with 10 µl of 10% Triton and left for 20 min at room temperature. Ten microliters of the lysate was mixed with 40 µl of RT cocktail (containing [ 3 H]thymidine), and the mixtures were incubated for 90 min at 37 C and spotted on DE 81 filter paper (2.3 cm in diameter). Filters were washed three times in 2 SSC and once in ethanol, air-dried, immersed in scintillation liquid, and counted in a β counter. Western blots. Proteins were extracted from the vector-producing cells immediately after the collection of vector particles. LV G cells induced for 13 days and transiently transfected 293T cells producing equivalent titers were washed twice with PBS and lysed in 500 µl of lysis buffer (0.5% Triton, 10 mm Tris, 10 mm NaCl, and protease inhibitors). Protein concentration in the extracts was determined using the Bio-Rad protein assay (Bio-Rad). Four hundred microliters of the protein extract was mixed with 200 µl of 3 loading buffer (30% glycerol, 6% SDS, 187 mm Tris) and boiled for 5 min at 95 C. Particles were concentrated by microcentrifuging 1.5 ml of vector stocks at 14,000 rpm for 90 min. The pellet was resuspended in 50 µl of lysis buffer and mixed with 25 µl of 3 loading buffer. Proteins (33 µg for cell lysates and 10 5 cpm of RT activity for virion lysates) were loaded on a 15% polyacrylamide gel and separated by electrophoresis. Proteins were electrotransferred to a PVDF membrane (NEN Life Sciences). The membrane was hybridized successively with a mouse monoclonal antibody against HIV-1 p24 Gag (1:500), a rabbit polyclonal antibody against VSV G (1:2000, gift from M. Matsuda), and a β- tubulin-specific monoclonal antibody (1:400, Boehringer), stripping with 0.2 M NaOH for 5 min between probings. Transduction of nondividing cells. Hela cells were treated with aphidicolin (15 µg/ml) and analyzed for cell cycle progression as previously described (1). Intracellular DNA and RNA concentrations were quantified using 7-AAD and pyronin, respectively. Cells were analyzed using a Becton Dickinson FACScan apparatus and the WindMDI software program. RESULTS Creation of Lentivector Packaging Cell Lines The tetracycline-repressible system was previously used for the generation of a stable packaging cell line producing VSV G-pseudotyped oncoretroviral or lentiviral vector particles (4, 6, 12, 14, 15). Based on these precedents, a method for the controllable expression of HIV-1 gag/pol was designed (Fig. 1). Advantage was taken of the fact that HIV-1 gag/pol transcripts are retained and quickly degraded in the nucleus unless sufficient amounts of the viral Rev posttranscriptional regulator are present in the cell. The plasmid pmdlg/prre induces the Rev-dependent expression of HIV-1 gag and pol from the immediate early CMV promoter. A ptet-rev plasmid was constructed, which expresses rev from a transcription module that contains the tet operator fused to a minimal CMV promoter. In the absence of doxycycline tta, a fusion product of the amino-terminal DNA-binding domain of the tet repressor and the carboxy-terminal activation domain of VP-16 from herpes simplex virus (16) binds the tet operator of ptetrev and Rev is produced. 293G cells express tta constitutively and VSV G under the control of doxycycline. ptet-rev and pmdlg/prre were cotransfected into 293G cells together with psv2.hygro, a plasmid conferring resistance to hygromycin. Two hundred FIG. 1. Steps to the generation of stable packaging cell lines for the production of multiply attenuated HIV-1-based vectors. 293 cells expressing a constitutive tta activator and a doxycycline-repressible VSV G (293 G) were transfected with a CMV-driven Gag/Pol plasmid and a doxycycline-controllable Rev construct to yield LV G. These cells were then transduced with the GFP-expressing CLL.GFP.CLL lentiviral vector for titer determination. 172

4 from the CMV promoter were placed in the U3 region to allow the production of vector genomic transcripts in the absence of Tat, the main HIV transactivator (Fig. 1). Lentivector particles containing this genome were produced by transient transfection of 293T cells, and the resulting virions were used to transduce the LV G -1 and LV G -2 clones once a day for 4 consecutive days to ensure that all cells expressed high levels of packageable vector genomes. The presence of the vector did not significantly modify the yield and the apparent protein composition of the viral particles released in the supernatant of these cells (data not shown). FIG. 2. Western blot analysis of HIV-1 Gag and VSV G proteins in producer cells and vector particles. (A) Immunodetection of VSV G shows that LV G -1 and -2 clones producing very little envelope in the presence of doxycycline (lanes 3 and 6, respectively) can be induced to express VSV G (lanes 4 and 7, respectively) in amounts comparable to 293T cells transiently transfected with pmd.g (lane 2). The VSV G protein is also detectable on vector particles produced by the LV G -1 and -2 clones (lanes 5 and 8, respectively). Lane 1, untransfected 293 cells. (B) Immunodetection of HIV-1 p24 (capsid) and processing intermediates of p55 Gag. As for VSV G, the synthesis of these proteins shows the same stringent regulation by doxycycline in both LV G clones. In the particles, p55 Gag has been fully processed by the HIV-1 protease, while the majority of p55 Gag still appears uncleaved in the producer cells. (C) Immunodetection of β-tubulin demonstrating that comparable amounts of protein from all cellular extracts were loaded on the gel. Protein amounts from vector particles were standardized for RT activity. hygromycin-resistant clones were selected and 73 of them were screened for the release of RT activity in the supernatant in the presence or absence of 1 µg/ml doxycycline. Two clones exhibiting very low levels of RT in the presence of the drug and a high degree of inducibility were analyzed further by Western blot, using antibodies against HIV-1 p24 capsid and VSV G. In the presence of doxycycline, LV G -1 (Figs. 2A and 2B, lane 3) and LV G - 2 (lane 6) expressed very low amounts of VSV G and HIV-1 Gag antigen. When the drug was removed, the production of viral proteins was induced (lanes 4 and 7), and particles of the predicted immunoreactivity were released in the supernatant (lanes 5 and 8). As expected, HIV-1 Gag was incompletely processed in the producer cells, whereas only fully cleaved p24 was detected in the particles. The LV G -1 and LV G -2 clones and transiently transfected 293T cells exhibited similar patterns of HIV- 1 gag expression. Establishment of Lentivector-Producing Clones To test the ability of the packaging clones to produce transduction-competent lentiviral vector particles, the proviral genome of an HIV-derived vector expressing the green fluorescent protein (GFP) was introduced in the cells. This provirus, CLL.GFP.CLL, contained modified long terminal repeats in which transcriptional elements Growth and Survival of Lentivector-Producer Cells Significant levels of cytotoxicity accompany the expression of several HIV proteins (17) and of the G protein of VSV (14). The growth kinetics of the LV G -1 clone, with or without packageable lentivector genome, were thus examined. For this, cells were plated at a density of /cm 2 and counted every day for more than 2 weeks while performing regular medium changes but without splitting. Figure 3 shows the results for the clone LV G -1. In the presence of doxycycline, cells reached a stationary phase after 7 days at a density of /cm 2 and maintained this density for 9 days before detaching from the plate. In the absence of the drug, the cells became slightly less confluent and detached slightly sooner, but could still be maintained in culture for approximately 2 weeks. Identical growth curves were obtained for the clone LV G -1/GFP containing a packageable vector. FIG. 3. Growth kinetics of LV G -1 cells in the presence or absence of doxycycline. LV G -1 cells were plated at a density of /cm 2 and counted at indicated intervals with regular medium changes but without splitting. 173

5 nondividing targets (Fig. 5) and to be concentrated to high titers (not illustrated). DISCUSSION FIG. 4. Kinetics of lentivector production by the LV G -1/GFP clone. LVG- 1/GFP cells were shifted to doxycycline-free culture medium on day 0. Reverse transcriptase activity (open area, cpm/ml) and vector titer (filled area, TU/ml) were determined in culture medium (10 ml) collected at the indicated time points. Samples were frozen at 80 C until analyzed in parallel. The histogram shows mean values from four experiments. Bars indicate standard errors. In this experiment, TU/ml corresponds to 2.5 TU per producer cell per day. Kinetics of Production of Transduction-Competent Lentivector Particles The levels of RT activity and the number of HeLa TU released in the extracellular milieu were used to monitor the production of vector particles from the two stable packaging clones. For this, LV G -1/GFP and LV G -2/GFP cells were plated at a density of to /cm 2, and the culture medium was changed and subjected to analysis every other day. Vector production by the clone LV G -1/GFP is shown in Fig. 4. Although kinetics varied slightly for the two clones, levels of RT induction between 30- and 40-fold were observed in both cell lines after 7 to 10 days following the removal of doxycycline. The yield of transducing particles increased even more dramatically, with a daily production that rose from less than 10 4 to 1 20 TU per producer cell. Under these experimental conditions, vector production above 1 TU/cell lasted for about 7 days, after which the cells exhibited signs of suffering and detached from the plate. Toward the end of production, an increasing number of packaging cells became GFP negative by FACS analysis, an indication of reduced transcriptional activity. Eventually, the evolution toward cell death was confirmed by the inability of GFP-negative cells to exclude propidium iodide. Importantly, lentivector particles produced from the stable packaging clones exhibited properties in all points identical to those of virions generated by transient transfection of 293T cells, including the ability to transduce In just a few years, considerable progress has been accomplished in the design of lentivectors that have a high degree of biosafety and performance. Exciting new avenues are opened for gene therapy, in particular for situations where the long-term expression of foreign genes is desired. Based on this premise, it is likely that a lentivector will soon be proposed for a clinical application. This move forward would be greatly facilitated by the availability of high-output stable packaging cell lines producing the latest generation of lentiviral vectors. The present work demonstrates that such cell lines can be generated, paving the way to the large-scale production of clinical-grade vector stocks. Under the experimental conditions described in this paper, the vector titers recovered from the LV G producer clones fluctuated between 1 and 20 HeLa-transducing units per producer cell per day for approximately 1 week, comparing favorably with those obtained by transient transfection of 293T cells (1 to 3 TU/cell/day for 2 to 3 days). Since such numbers were measured in the supernatant of unselected cell populations, it is reasonable to assume that even higher titers could be recovered by screening for the best producer clones. Also, efforts aimed at optimizing the protocol for vector production are likely to result in further improvements. Of note, the addition of sodium butyrate, a transcriptional stimulator previously found to be necessary for the production of lentiviral vector particles from a first-generation packaging line (6), was neither required nor beneficial for the full induction of the LV G clones (not illustrated). Also in contrast to other packaging systems, the LV G cell lines have been used to produce vectors for more than 6 months without any loss of the vector yield. We previously reported that the HIV-1 accessory genes and tat are unnecessary for the generation of lentiviral vector particles fully efficient for transduction of dividing as well as nondividing cells, both in vitro and in vivo (8). This allowed for the creation of a so-called third-generation packaging system based on three of the nine genes of HIV-1 only, gag, pol, and rev (8). From such an extensively deleted packaging system, the parental virus cannot be reconstituted, since some 60% of its genome has been completely eliminated, including several of the genes coding for critical HIV virulence factors. The biosafety of the vector production system is further increased by the split genome design, since Gag/Pol, Rev, VSV G, and the vector are produced from four separate DNA units (8). Finally, in the clones described here, the stable integration of these sequences alleviates the risk of emergence of episome-derived recombinants, a real concern when high-copy plasmids are introduced in cells for the transient production of vectors. Correspondingly, we 174

6 FIG. 5. Transduction of nondividing cells by LV G -1/GFP-produced vector particles. Control and aphidicolin-treated HeLa cells were transduced with a GFPexpressing lentiviral vector produced from the LV G -1/GFP packaging clone or from transiently transfected 293T cells or with a GFP-expressing MLV-based vector. GFP expression was measured 2 days later by FACS. (A) Cell cycle analysis showing the efficacy of aphidicolin to block cells at the G1/S transition. (B) Lentiviral vectors produced by stable cell lines transduced nondividing cells as efficiently as vectors produced by transient transfection. As expected, transduction by MLV-based vector was strictly dependent on cell division. 175

7 failed to detect the transfer of packaging sequences in target cells transduced with LV G -derived vectors, even through the use of a sensitive polymerase chain reactionbased assay (not illustrated). Although it will be essential to exert the highest degree of scrutiny to verify the absence of transfer of helper virus sequences when using large batches of high-titer vector stocks, our preliminary analyses are extremely encouraging. In conclusion, the hereby-described inducible packaging cell lines allow for the safe and efficient production of high-titer, VSV G-pseudotyped, multiply attenuated lentiviral vectors. They will facilitate efforts aimed at using this gene delivery system for the genetic treatment of human diseases. ACKNOWLEDGMENTS We thank P. Salmon for the MLV-derived vector, D. Ory for the 293G cell line, L. Naldini for the plasmid prrl.gfp.cll, and M. Matsuda for the polyclonal anti-vsv G antibody. This work was supported by grants from the Swiss National Science Foundation and the Gabriella Giorgi Cavaglieri Foundation. REFERENCES 1 Naldini, L., et al. (1996). In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector. Science 272: Carroll, R., Lin, J. T., Dacquel, E. J., Mosca, J. D., Burke, D. S., and St. Louis, D. C. (1994). A human immunodeficiency virus tape 1 (HIV-1)-based retroviral vector system utilizing stable HIV-1 packaging cell lines. J. Virol. 68: Yu, H., Rabson, A. B., Kaul, M., Ron, Y., and Dougherty, J. P. (1996). Inducible human immunodeficiency virus type 1 packaging cell lines. J. Virol. 70: Kaul, M., Yu, H., and Dougherty, J. P. (1998). Regulated lentiviral packaging cell line devoid of most viral cis-acting sequences. Virology 249: Corbeau, P., Kraus, G., and Wong-Staal, F. (1996). Efficient gene transfer by a human immunodeficiency virus type 1 (HIV-1)-derived vector utilizing HIV packaging cell line. Proc. Natl. Acad. Sci. USA 93: Kafri, T., van Praag, H., Ouyang, L., Gage, F. H., and Verma, I. M. (1999). A packaging cell line for lentivirus vectors. J. Virol. 73: Zufferey, R., Nagy, D., Mandel, R. J., Naldini, L., and Trono, D. (1997). Multiply attenuated lentiviral vector achieves efficient gene delivery in vivo. Nat. Biotechnol. 15: Dull, T., et al. (1998). A third-generation lentivirus vector with a conditional packaging system. J. Virol. 72: Han, J. J., et al. (1999). Transgene expression in the guinea pig cochlea mediated by a lentivirus-derived gene transfer vector. Hum. Gene Ther. 10: Deglon, N., et al. (2000). Self-inactivating lentiviral vectors with enhanced transgene expression as potential gene transfer system in Parkinson s disease. Hum. Gene Ther. 11: Park, F., Ohashi, K., Chiu, W., Naldini, L., and Kay, M. A. (2000). Efficient lentiviral transduction of liver requires cell cycling in vivo. Nat. Genet. 24: Ory, D. S., Neugeboren, B. A., and Mulligan, R. C. (1996). A stable human-derived packaging cell line for production of high titer retrovirus/vesicular stomatitis virus G pseudotypes. Proc. Natl. Acad. Sci. USA 93: Zufferey, R., Donello, J. E., Trono, D., and Hope, T. J. (1999). Woodchuck hepatitis virus posttranscriptional regulatory element enhances expression of transgenes delivered by retroviral vectors. J. Virol. 73: Yang, Y., et al. (1995). Inducible, high-level production of infectious murine leukemia retroviral vector particle pseudotyped with vesicular stomatitis virus G envelope protein. Hum. Gene Ther. 6: Chen, S. T., Iida, A., Guo, L., Friedmann, T., and Yee, J. K. (1996). Generation of packaging cell lines for pseudotyped retroviral vectors of the G protein of vesicular stomatitis virus by using a modified tetracycline inducible system. Proc. Natl. Acad. Sci. USA 93: Gossen, M., and Bujard, H. (1992). Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89: Kaplan, A. H., and Swanstrom, R. (1991). The HIV-1 gag precursor is processed via two pathways: Implications for cytotoxicity. Biomed. Biochim. Acta 50: