Gene Therapy (2004) 11, 266–275 & 2004 Nature Publishing Group All rights reserved 0969-7128/04 $25.00 www.nature.com/gt RESEARCH ARTICLE Lentiviral vectors pseudotyped with baculovirus gp64 efficiently transduce mouse cells in vivo and show tropism restriction against hematopoietic cell types in vitro

CA Schauber, MJ Tuerk, CD Pacheco1, PA Escarpe and G Veres Cell Genesys Inc., South San Francisco, CA, USA

The envelope glycoprotein from vesicular stomatitis virus found that gp64-pseudotyped lentiviral vectors could effi- (VSV-G) has been used extensively to pseudotype lentiviral ciently transduce a variety of cell lines in vitro, although gp64 vectors, but has several drawbacks including cytotoxicity, showed a more restricted tropism than VSV-G, with potential for priming of immune responses against transgene especially poor ability to transduce hematopoietic cell types products through efficient transduction of antigen-presenting including dendritic cells (DCs). Although we found that gp64- cells (APCs) and sensitivity to inactivation by human pseudotyped vectors are also sensitive to inactivation by complement. As an alternative to VSV-G, we extensively human complement, gp64 nevertheless has advantages characterized lentiviral vectors pseudotyped with the gp64 over VSV-G, because of its lack of cytotoxicity and narrower envelope glycoprotein from baculovirus both in vitro and in tropism. Consequently, gp64 is an attractive alternative to vivo. We demonstrated for the first time that gp64-pseudo- VSV-G because it can efficiently transduce cells in vivo and typed vectors could be delivered efficiently in vivo in mice via may reduce immune responses against the transgene portal vein injection. Following delivery, the efficiency of product or viral vector by avoiding transduction of APCs mouse cell transduction and the transgene expression is such as DCs. comparable to VSV-G-pseudotyped vectors. In addition, we Gene Therapy (2004) 11, 266–275. doi:10.1038/sj.gt.3302170

Keywords: lentivirus; pseudotype; baculovirus; gp64; complement

Introduction expression of the gene to specific cell types. Following systemic administration, the VSV-G pseudotyped vectors The G glycoprotein from vesicular stomatitis virus (VSV- transduce target cells of interest, but can also transduce G) has been used extensively for pseudotyping retroviral other cell types that would have undesirable effects in and lentiviral vectors.1,2 VSV-G is advantageous for gene- gene-therapy protocols. For example, antigen-presenting transfer protocols because of its broad species and tissue cells (APCs) could cause deleterious immune responses tropism, and its ability to confer physical stability and if they are transduced inadvertently. Unfortunately, VSV- high infectivity to vector particles.3 However, the high G-pseudotyped lentivectors can efficiently transduce fusogenicity of VSV-G causes rapid syncytia formation APCs from mice and humans, and can elicit immune and cell death, making it difficult to generate stable cell responses against the transgene product, thus potentially lines expressing the protein.4 A number of groups have negating the therapeutic effects of the protein expres- generated stable packaging and producer cell lines using sion.7–10 VSV-G, but in all cases a repressible or inducible Systemic administration of viral vectors also exposes promoter regulates the expression of VSV-G.4–6 While the particles to possible inactivation by serum comple- regulated expression avoids the problem of VSV-G ment.11 Complement inactivation of vectors is dependent cytotoxicity, it makes stable production of viral vectors on the species derivation of the cell line used to produce more complicated and cell line generation more time- the vectors, as well as the identity of the envelope consuming. glycoprotein used for pseudotyping.12,13 VSV-G pseudo- The broad tissue tropism of VSV-G has its disadvan- types, regardless of producer cell type, will be inacti- tages as well. For certain in vivo gene-therapy applica- vated by human serum complement, while vectors tions, it may be important to restrict the transfer and pseudotyped with other envelope glycoproteins such as the MLV amphotropic envelope (Ampho) or RD114 are resistant to inactivation.14,15 Correspondence: PA Escarpe, Cell Genesys Inc., 500 Forbes Blvd. South Owing to the many disadvantages of VSV-G, there is San Francisco, CA 94080, USA 1Department of Neuroscience, University of Michigan, Ann Arbor, MI, considerable interest in examining alternative envelope USA glycoproteins that might have more suitable properties Received 14 May 2003; accepted 03 September 2003 for use with retroviral or lentiviral vectors in gene- gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 267 transfer applications. Of particular advantage would be transduce three human cell lines. Several independent identification of an envelope that could confer high titer batches of third-generation HIV-1 vector particles carry- in vivo, low cytotoxicity to enable producer cell line ing the gene-encoding green fluorescent protein (GFP) development, resistance to complement and reduced were prepared by transient transfection, and pseudo- ability to transduce APCs. typed with the VSV-G envelope glycoprotein or baculo- We became interested in testing the major envelope virus gp64. The unconcentrated viral supernatants from protein of baculovirus, gp64, for its ability to pseudotype the transfected cells were tested for infectious titer using lentiviral vectors based on the infectivity characteristics 293T, HuH7 human hepatoma and HeLa cells as targets of baculoviral vectors. Baculovirus (Autographa californica (Figure 1a). We found that the gp64-pseudotyped vector nuclear polyhedrosis virus, AcNPV) is an enveloped was infectious in all the three cell lines, with average insect virus that is being developed as a vector for gene titers ranging from 8.1 Â 104 to 1.2 Â 106 transducing therapy.16 Early studies showed that baculovirus vectors units (TU)/ml. The titers of the VSV-G-pseudotyped could efficiently transduce human hepatocyte cell lines vectors ranged from 7.8 Â 105 to 4 Â 106 TU/ml on the and explanted human liver tissue.17,18 Subsequently, it same cell types. The gp64-pseudotyped vector titers were was demonstrated that baculovirus vectors could enter a reduced 25-fold on 293T cells compared to VSV-G, but broader variety of human cell types.19–21 We set out to were only three-fold lower than VSV-G on HuH7 cells. determine whether gp64 could be used to pseudotype The relative titer was the highest on HeLa cells, with other viral vectors and confer high infectivity in cells of gp64 titer (9.3 Â 105 TU/ml) slightly higher than the VSV- hepatic origin, which are primary targets for many gene- G titer (7.8 Â 105 TU/ml). In addition, we found that the therapy applications. Very recently, a report was published that described the use of baculovirus gp64 for pseudotyping HIV-1 a 22 VSVG gp64 vectors. These studies demonstrated that gp64-pseu- 7 dotyped vectors had high titer on several cell lines, and 10 could be concentrated 100-fold by ultracentrifugation. Importantly, it was also shown that gp64 could be 6 constitutively expressed in a stable cell line with no 10 cytotoxic effects.22 The authors showed that this stable cell line could be used to package lentiviral vectors and Titer might enable large-scale production of vectors without 105 regulated expression of the envelope.22 To further explore the utility of gp64 pseudotyping, we carried units/ml) (Transducing out extensive analysis of gp64 for cell and species 4 tropism in vitro, using 24 different target cell lines, and 10 determined the sensitivity of gp64-pseudotyped vectors 293T HeLa 293T HeLa to human complement inactivation. We also scaled up HuH7 HuH7 production of gp64-pseudotyped vectors, and deter- Target cell line mined the binding and elution properties of these vectors b VSVG gp64 during particle purification by anion-exchange chroma- 105 tography. Finally, we carried out two distinct experi- ments to analyze the in vivo transduction efficiency of gp64-pseudotyped vectors. Here we show that gp64 104 confers efficient gene transfer in vivo, and that gp64- pseudotyped vectors are human complement sensitive and have tropism restricted against hematopoietic cells 22 3 including human dendritic cells (DCs). Infectivity 10

Results units/ng p24) (Transducing 102 Baculovirus gp64 envelope protein confers high 293T HuH7 HeLa 293T HuH7 HeLa infectivity to third-generation HIV-1 lentiviral vectors Target cell line Kumar et al22 recently reported that the baculovirus gp64 envelope protein could pseudotype viral vectors based Figure 1 Comparison of VSV-G and gp64 pseudotype titers and on HIV-1. While our vector of choice (also derived from infectivities on three target cell lines. (a) Unconcentrated preparations of third-generation lentiviral vectors carrying a GFP transgene were HIV-1) is slightly different from that used in their produced by transient transfection, and were pseudotyped with either the published studies, we were also interested in using VSV-G (left) or gp64 (right) envelope glycoprotein. These preparations gp64 for pseudotyping, based on the ability of baculo- were assayed for end-point titers on 293T (black bars; n¼3), HuH7 human virus vectors to infect human cell types. Our third- hepatoma (gray bars; n¼5) and HeLa cells (hatched bars; n¼2). Titer generation vector differs in that there are three additional values (TU/ml) were calculated based on % GFP-positive cells, as HIV genes deleted (tat, nef and vpu) and the vector is determined by FACS analysis. Standard deviations of titers determined for 23 independent vector preparations are shown as error bars. (b) To normalize produced through expression of Rev protein in trans. To the TU to total particles in the lentiviral vector preparations, p24Gag confirm that gp64 could also pseudotype this more concentrations in ng/ml were determined by ELISA. Infectivity values advanced HIV vector, pseudotyped lentiviral vector (TU/ng p24Gag) of the preparations were calculated by dividing the titer particles were prepared and tested for their ability to values in (a) in TU/ml by the p24Gag concentrations in ng/ml.

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 268 gp64 envelope glycoprotein could also pseudotype an a MLV-based oncoretroviral vector with a titer in the high 105 104 range on 293T cells (data not shown), which is similar to the 293T titer determined for lentiviral vectors 4 10 pseudotyped with gp64 (Figure 1a).

The titer differences between pseudotypes and pre- 3 parations could be due to differences in the level of 10 vector output in the transient transfections. To account 2 for possible variations in total particle production, the Infectivity 10 p24Gag protein levels were quantitated by ELISA and the infectious titer values were normalized to ng of p24Gag to 101 give the infectivity (TU/ng p24Gag), which reflects the units/ng p24) (Transducing ratio of infectious particles to total particles (Figure 1b). 0 Gag 10 The p24 levels in the preparations of gp64-pseudo- XC Vero A549 PC3 HeLa 293T HuH7 Te671 G-361 Ct2 Th Hep3BHepG2 HUAEC Ovear3HT1080NIH3T3 typed particles were generally lower than those for VSV- U118 MG Target cell line G-pseudotyped particles, a phenomenon that seemed to b be specific to transient expression of gp64 protein. On 105 comparing these normalized values for the two vectors, we found that gp64 vector infectivity was eight-fold 104 lower than VSV-G on 293T cells, but there was no

significant difference between their infectivities on HuH7 3 and HeLa target cells. 10 It is possible that transduction using gp64-pseudo- 2 typed vector particles resulted in GFP-positive cells due Infectivity 10 to carryover of GFP protein from the transfected producer cells (pseudotransduction) instead of integra- 101 tion of the transduced GFP vector genome.24,25 However, units/ng p24) (Transducing it was previously reported that integrated transgene 100 DNA could be detected in target cells transduced with Raji FO 293T Molt4 CEM- KG-1 K562 SC-1 gp64-pseudotyped vectors, indicating that gp64 could C8166 TART NHDC mediate true gene transfer.22 To eliminate the possibility Target cell line of pseudotransduction in our experiments, we passaged Figure 2 VSV-G and gp64 pseudotype infectivities on various target cell gp64 vector-transduced cells and examined transgene lines. Preparations of lentiviral vectors carrying the GFP transgene and expression after 2 weeks in culture. Cells that are pseudotyped with VSV-G (hatched bars) or gp64 envelope glycoprotein pseudotransduced should show a decrease over time in (black bars) were prepared by transient transfection, followed by percent GFP-positive cells and in GFP fluorescence concentration by ultracentrifugation. Vectors were serially diluted and used to infect the various target cell lines indicated on the X-axis. FACS intensity, due to degradation of the GFP protein. Instead, analysis was used to determine the percentage of GFP-expressing cells and both VSV-G and gp64-vector-transduced cells main- end-point titer values were calculated. Titer values were normalized to ng tained the same percentage of GFP-positive cells over of p24Gag in the preparations to yield infectivity values in TU/ng p24Gag this time period, and in both cases the GFP fluorescence for each pseudotype and cell line. Cell lines in (a) are as follows: 293T intensity increased over time (data not shown). (human embryonic kidney), Vero (African Green Monkey kidney), HuH7 (human hepatoma), Hep3B and HepG2 (human hepatocellular carcino- mas), Te671 (human rhabdomyosarcoma), U-118 MG (human glioblasto- ma), G-361 (human melanoma), A549 (human lung carcinoma), HUAEC Comparison of gp64 and VSV-G infectivities (human umbilical arterial endothelium), PC-3 (human prostate carcino- on various cell lines in vitro ma), HeLa (human cervical carcinoma), Ovcar-3 (human ovarian To fully exploit their utility in gene-therapy applications, carcinoma), HT-1080 (human osteosarcoma), NIH/3T3 (mouse embryonic it is important to assess the tissue and species tropisms of fibroblasts), Cf2Th (canine thymus) and XC (rat epithelial cells). different envelope glycoproteins. As a step toward Hematopoietic cells analyzed in (b) are: MOLT-4, C8166 and CEM-TART (all human T lymphocytes), KG-1 (human myeloblasts), K-562 (human analyzing the tropism of gp64-pseudotyped lentiviral promyelocytes), SC-1 and Raji (human B lymphocytes), FO (mouse B vectors, we determined the infectivity of these pseudo- lymphocytes) and NHDC. Asterisks indicate instances of undetectable types on a variety of cell lines in vitro. infectivity, with the limit of detection at 1.0 TU/ng p24. Preparations of VSV-G and gp64-pseudotyped lenti- viral vectors carrying a CMV-GFP transgene cassette were generated by transient transfection, and concen- trated 100-fold by ultracentrifugation. As previously (see Figure 1). These small disparities are probably due reported, the gp64-pseudotyped vectors can be ultra- to slight differences in the vector preparations and the centrifuged with no substantial loss of titer or p24Gag and typical variation of the titer assay, and not due to no change in the vector infectivities (data not shown).22 concentration, since the average infectivities of uncon- The concentrated vector preps were analyzed for p24Gag centrated and concentrated preps were similar. The gp64 concentration and then tested for end-point titers on a infectivity was equivalent or better than VSV-G infecti- variety of cell lines. Infectivity values were calculated by vity on a number of cell lines (Figure 2a), including Vero, dividing the titer value by the p24Gag concentration HuH7, U-118 MG, A549, HUAEC, Cf2Th and XC cells. (Figure 2). The infectivity values for VSV-G and gp64 The relative infectivity for gp64 pseudotypes was slightly using 293T, HuH7 and HeLa target cells were slightly reduced (less than 10-fold lower than VSV-G) on 293T, different from those determined in separate experiments Hep3B, G-361, PC-3, HeLa, Ovcar-3 and NIH/3T3 cells.

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 269 However, there were several human cell lines such as preparations were then purified by anion exchange HepG2, Te671 and HT-1080, where the gp64 pseudotype chromatography. Both the gp64 and Ampho-pseudo- infectivity was more than 10-fold lower than VSV-G. typed vector particles bound to and eluted from the Surprisingly, the gp64 pseudotypes exhibited low DEAE column under the same conditions as the VSV-G infectivity in cells of hematopoietic origin (Figure 2b). pseudotypes, and 100% of the input vector was Initially, we found no detectable transduced cells when recovered using this process (data not shown). The final gp64-pseudotyped vector was used to infect MOLT-4, vector preparations were characterized by testing for C8166 and CEM-TART cells (all human T-lymphocyte p24Gag concentration, and were confirmed to transduce cell lines). The same vector preps were used for titrations 293T cells in vitro and express human Factor IX protein on all of the cell lines in Figure 2, indicating that the lack (data not shown). For each of the vectors, approximately of transduction on hematopoietic cells was not due to 10 mg of p24Gag (or PBS as a negative control) was lack of active vector particles. Subsequently, we looked at injected into Swiss Nude mice (three mice per vector other hematopoietic cell types and found no detectable pseudotype) via a hepatic portal vein cannula. The doses infectivity for gp64 vector on KG-1 human myeloblasts in 293T TU for each vector were as follows: VSV-G mice or on Raji and SC-1 human B lymphocytes, and received 5.8 Â 107 TU, the gp64 mice received extremely low infectivity on the K-562 human erythro- 1.8 Â 107 TU and the Ampho mice were injected with leukemia cells and FO mouse B lymphocyte cells. 1.7 Â 107 TU. Serum was collected from the mice at Finally, we analyzed the gp64 infectivity on DCs various time points after injection and analyzed for because published reports indicated that VSV-G- human Factor IX protein levels by ELISA (Figure 3). Mice mediated transduction of APCs such as DCs may injected with PBS or the Ampho-pseudotyped lentiviral contribute to immune responses against the transgene vector did not express any detectable Factor IX protein at product or vector.9,26 For this analysis, primary normal any time point. The mice injected with VSV-G-pseudo- human dendritic cells (NHDCs) were obtained and typed vector expressed detectable human Factor IX transduced with the VSV-G and gp64 vectors. Again, protein beginning at Day 7 after injection, and continuing there was no detectable transduction activity of the gp64 at a stable level of 12–17 ng/ml for greater than 9 weeks. vector using these target cells, whereas the VSV-G- Mice injected with the gp64-pseudotyped vector ex- pseudotyped vectors transduced the NHDC at 100 times pressed a stable level of 5–10 ng/ml human Factor IX for the limit of detection. more than 9 weeks. Mice injected with the gp64- Overall, gp64-pseudotyped vectors were inefficient at pseudotyped vector expressed only slightly less Factor transducing cells of hematopoietic origin, including IX than the VSV-G vector-injected mice, despite having human T lymphocytes and primary human DCs. This received approximately three-fold lower TU. profile demonstrates that gp64 has a more restricted set of cell types that can efficiently transduce compared to VSV-G. On the other hand, gp64 infectivity levels were similar to those of VSV-G on many cell lines, suggesting that gp64 pseudotyping might be a useful alternative for transduction of many target tissues. gp64-pseudotyped lentiviral vectors efficiently transduce mouse cells in vivo It is reported that gp64-pseudotyped lentiviral vectors were able to efficiently transduce a number of cell lines in vitro.22 However, it was not known if the gp64- pseudotyped vectors could transduce cells in vivo and confer long-term gene expression. To determine whether gp64 could confer significant infectivity in vivo,we looked for evidence of gene transfer in mice after injection of pseudotyped lentiviral vectors. An initial mouse experiment was designed using human Factor IX Figure 3 Expression of human Factor IX in Swiss Nude mice following as the reporter transgene to facilitate detection of injection with pseudotyped lentiviral vectors. Preparations of lentiviral transduction and transgene expression. Factor IX is vectors carrying a CMV-driven human Factor IX transgene and easily quantified and can be detected in the blood pseudotyped with the VSV-G, gp64 or MLV amphotropic envelope following secretion from many tissues.27 Since the in glycoproteins were produced by large-scale transient transfection followed vivo tissue tropism for gp64 was unknown, we used the by purification and concentration. These preparations were tested for p24Gag concentration and infectivity was confirmed using 293T cells in constitutive CMV promoter to drive Factor IX expression vitro. A volume of 100 ml of each vector preparation or 100 ml of PBS as a and to avoid problems of tissue-specific expression. negative control was injected into each of three NIH Swiss Nude mice. The To generate vectors for injection into mice, vector final doses of TU determined using 293T target cells were: 5.8 Â 107 TU production was scaled up from 10 cm dishes into cell for VSV-G, 1.8 Â 107 TU for gp64 and 1.7 Â 107 TU for the Ampho vector. factories and large-scale transient transfections were Blood was collected from the injected mice at time points after injection, as performed in the 10-stacks. Three lentiviral vector indicated on the X-axis. The concentration of human Factor IX in the mouse serum was determined by ELISA. Data points indicate the average preparations carrying the CMV-human Factor IX trans- human Factor IX concentration in the three mice with the corresponding gene cassette and pseudotyped with VSV-G, gp64 or the standard deviations. VSV-G (squares with dashed line), gp64 (triangles MLV amphotropic envelope (Ampho) were generated in with solid line), Ampho (X’s with dashed line) and PBS control (diamonds this manner. The vector particles from all the three with solid line).

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 270 The expression of human Factor IX in mice after injection of less than half as many 293T-TU in the gp64 injection with pseudotyped lentiviral vectors showed vector dose. Finally, very few GFP-positive cells were that the VSV-G and gp64 envelopes could confer efficient observed in the livers of the Ampho vector-treated mice, transfer of a gene into mouse cells in vivo. To determine though virtually all were hepatocytes (Figure 4d). In the whether liver cells in the mice were being transduced, Factor IX and GFP experiments, VSV-G and gp64 showed we injected mice with pseudotyped vectors carrying a similar and significant expression, with notably lower or CMV-GFP transfer vector and examined liver sections for undetectable expression in the Ampho vector-treated GFP expression. A total of 12 Swiss Nude mice (three mice. mice per group) were injected via a hepatic portal vein cannula with PBS or with lentiviral vectors carrying the CMV-GFP gene cassette pseudotyped with VSV-G, gp64 gp64-pseudotyped vector particles are sensitive to or MLVAmpho envelope protein. The mice were injected inactivation by human serum complement with approximately 10 mg of p24Gag, which equated to Lentiviral vectors produced from human cells pseudo- 3.5 Â 108 TU (using 293T cells as target cells) for the VSV- typed with VSV-G but not Ampho envelope are sensitive G vector, 1.3 Â 108 TU for the gp64 vector and to inactivation by human serum complement.14 To 1.7 Â 108 TU for the Ampho vector. At day 11 postinjec- determine whether gp64-pseudotyped lentiviral particles tion, all 12 of the mice were killed and their livers were are resistant or sensitive to complement, we tested the removed, fixed and sectioned. GFP expression was effects of human, rat and mouse serum on the infectivity examined in the sections via fluorescence microscopy, of gp64-pseudotyped vector preparations. VSV-G- and and representative sections for each vector pseudotype gp64-pseudotyped lentiviral vectors carrying the GFP are shown (Figure 4). The patterns of GFP expression transgene were produced by transient transfection in were consistent throughout each liver and between the 293T cells and concentrated 100-fold by ultracentrifuga- three mice in each group (data not shown). In both the tion. Titers were determined using 293T cells as targets mice injected with VSV-G-pseudotyped vector and mice and the vector preps were normalized to a concentration injected with gp64 vector, approximately 3–5% of liver of 106 GFP TU/ml. The vector suspensions were diluted cells were positive for GFP. The majority of the positive 1:5 into normal or heat-inactivated serum or into cells in both experimental groups were nonparenchymal medium as a no-serum control. After incubation at cells (possibly Kupffer or sinusoidal cells), while a few 371C for 1 h, the suspensions were diluted and used to were hepatocytes (Figure 4b). The gp64-vector-injected infect 293T target cells to determine end-point titers. The mice also showed a slightly greater number of GFP- titer of the no-serum control samples were calculated positive hepatocytes compared to VSV-G (Figures 4b, c). and set as 100% recovery of titer for the reaction. The titer Approximately the same number of liver cells was of samples exposed to serum is reported as a fraction of positive in the VSV-G and gp64 mice, despite the the 100% control sample. VSV-G pseudotypes were sensitive to inactivation, with complement-active serum causing 30–100-fold reduction in titer (Figure 5a, left). The gp64-pseudotyped lentiviral vector was also sensi- tive to inactivation by the human sera. In all, 20–50-fold reductions in titer for gp64 pseudotypes were observed in the serum from the same donors (Figure 5b, right). The human serum samples were tested for hemolytic

complement activity (CH50), and all were found to be in the normal range (data not shown). The inactivation of gp64-pseudotyped particles by serum from mice and rats was also evaluated. It has been reported that baculovirus vectors produced in insect cells are sensitive to substantial inactivation in vitro by human, rat and guinea-pig serum, and that the vectors have extremely poor infectivity in vivo in mice and rats due to complement inactivation.28–30 By testing gp64- pseudotyped lentiviral vectors, whether the complement sensitivity of baculovirus vectors in rodent serum is due Figure 4 Expression of GFP in mouse livers following injection with to recognition of the specific envelope glycoprotein or the pseudotyped lentiviral vectors. Preparations of lentiviral vectors carrying a insect-derived lipid envelope could be evaluated. In this CMV-GFP transgene cassette and pseudotyped with the VSV-G, gp64 or study, the gp64-pseudotyped lentiviral vectors were not MLV amphotropic envelope glycoproteins were produced by large-scale inactivated by serum from C57BL/6 mice or Nude rats transient transfection, followed by purification and concentration. These preparations were tested for p24Gag concentration and infectivity was (Figure 5b). Neither the pooled serum from C57BL/6 determined using 293T target cells in vitro. A volume of 100 ml of each mice (two mice sera per pool) nor the serum from a vector preparation or 100 ml of PBS as a negative control was injected into single donor rat caused more than a four-fold reduction each of three NIH Swiss Nude mice via portal vein cannula. The final doses in titer, unlike the more severe inactivation observed in of TU determined using 293T target cells were: 3.5  108 TU for VSV-G, 8 8 human serum (Figure 5a). VSV-G- and Ampho-pseudo- 1.3  10 TU for gp64 and 1.7  10 TU for the Ampho vector. At day 11 typed vectors were also resistant to inactivation by both after injection, mice were killed and their livers were removed, fixed and sectioned. Liver sections were mounted on slides and imaged by the mouse and rat sera (data not shown). Thus, like VSV- fluorescence microscopy at  20 magnification. (a) PBS-injected; (b) G, the gp64 envelope glycoprotein makes lentiviral VSV-G-vector-injected; (c) gp64-vector-injected and (d) Ampho-vector- vectors sensitive to inactivation by human serum, but injected mouse liver. not by mouse or rat serum.

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 271 also found that gp64 could be used to pseudotype an MLV-based oncoretroviral vector and could confer titers in the same range as the pseudotyped lentiviral vectors (data not shown). This result extended the published report that gp64 can also pseudotype vectors based on HIV-2, showing that gp64 pseudotyping may be feasible with any member of the retrovirus family.22 Analysis of the in vitro tropism of gp64 suggested that gp64 pseudotyping might be advantageous for applica- tions where tissues such as the liver, lung, skin, glia, endothelium or thymus are targets for transduction, because gp64 infectivity was high on cell lines derived from these tissues. In addition, reasonable infectivity (greater than 103 TU/ng p24) for gp64-pseudotyped vectors was observed in monkey (Vero), dog (Cf2Th), rat (XC) and mouse cells (NIH/3T3), indicating that gp64 could transduce cell types in several species that would be useful preclinical animal models. Kumar et al22 examined six target cell lines and observed similar titers for both gp64 and VSV-G pseudotypes.22 Extensive Figure 5 Complement sensitivity of gp64-pseudotyped lentiviral vectors examination of 24 target cell types indicated that gp64 in human, mouse and rat serum. (a) Concentrated preparations of infectivity varies more widely than that of VSV-G. The lentiviral vectors carrying a GFP transgene and pseudotyped with VSV-G disparities in infectivity may be due to differences in the (left) or gp64 (right) envelope glycoprotein were normalized to a ability of the vector particles to bind to the target cells, concentration of 106 TU/ml. The vector was then diluted 1:5 into medium (the no-serum control) or human serum that was normal (hatched bars) or implicating possible variations in the abundance of gp64 heat-inactivated (black bars) from five single human donors (denoted # 41– receptors on the different cell types. Different reports 45). After a 1-h incubation at 371C, the vectors were tested for end-point have suggested that phospholipid molecules including titers on 293T cells. The titer from vector incubated with the medium phophatidylinositol and phosphatidic acid as well as control was set as 100% recovery of titer. Bars represent the percentage of proteins are possible gp64 receptors.31–34 Alternatively, titer remaining in the vector samples after incubation with normal or heat- there may be differences in the membrane composition inactivated serum, compared to vector incubated with the medium control. (b) gp64-pseudotyped lentiviral vectors were tested as in (a) for titer that affect the fusion of the viral membrane with the remaining after incubation with normal (hatched bars) or heat-inactivated target cell membrane after binding has occurred. Indeed, (black bars) serum from C57/BL6 mice (three pools of serum each from two membrane lipid composition has been shown to affect mice; left) and one Nude rat (right). fusion mediated by gp64, probably as a result of differences in the shape of the fusion intermediate structures in the envelope.31 Unfortunately, because the variations in membrane lipid composition and protein Discussion levels between these different cells has not been explored, it is unclear whether the pattern and range of In this study, we report the additional characterization of gp64 pseudotype infectivity in the different cell types lentiviral vectors pseudotyped with baculovirus gp64 as correlate with modulations in receptor attachment or an alternative to VSV-G. Kumar et al22 reported recently other membrane fusion factors. that gp64 could pseudotype an early-generation HIV-1 Interestingly, gp64-pseudotyped lentiviral vectors vector and confer titers greater than 106 TU/ml. The were extremely inefficient at transducing human studies reported here confirmed the ability of gp64 to T-lymphocyte cell lines and other hematopoietic cell pseudotype HIV vectors almost as efficiently as the VSV- types. This restricted transduction of hematopoietic cell G envelope protein, and also extended the previous work lines was also reported in published studies using into in vivo studies using third-generation lentiviral baculovirus vectors, where K-562 and MOLT-4 cells as vectors lacking the HIV tat gene as well as all four HIV well as several other B- and T-lymphocyte cell lines were accessory genes. The infectivity of the gp64-pseudotyped poorly infected.17,20,21 This restricted tropism of gp64 vectors on 293T cells was several-fold lower than that of may preclude the use of this envelope protein for gene- VSV-G pseudotypes, while it was very near or higher therapy applications targeting the blood or immune than the VSV-G infectivity on HuH7 and HeLa cells. The system. However, because the gp64-pseudotyped vectors absolute titers and relative infectivities of VSV-G versus could not transduce NHDC to any detectable level, the gp64 pseudotypes reported here differ slightly from the selective tropism of gp64 might prove to be beneficial for results of Kumar et al.22 The slight discrepancies may be some cases. Lentiviral vectors pseudotyped with VSV-G due to the differences in the viral vectors and target cells can transduce human and mouse DCs with high or variation in the method of titer determination. While efficiency and induce humoral immune responses we found that the absolute titers (TU/ml) of gp64- against transgene products.7,9,26 Also, VSV-G itself has pseudotyped lentiviral vectors were slightly lower than been demonstrated to elicit antibody responses.14,35 The for VSV-G vectors, this was generally due to lower inability of gp64 pseudotypes to transduce DCs suggests particle production (lower p24Gag concentration) in the that there could be a reduction in immune responses transiently produced gp64 vector preparations. This against the vector components or against the transgene phenomenon could be overcome by producing vector products in vivo when gp64-pseudotyped vectors are particles in cells that were stably expressing gp64. We used.

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 272 After determining that the gp64-pseudotyped lentivir- over the VSV-G envelope that it can be stably expressed al vectors showed significant infectivity on a number of without cytotoxic effects, allowing a generation of a cell lines in vitro, we set out to confirm that the envelope packaging cell line without inducible expression. Finally, could also confer infectivity in vivo. We demonstrated gp64 can be a useful alternative to VSV-G for many that gp64-pseudotyped lentiviral vectors efficiently applications because it transduces cells in vivo with a transfer the human Factor IX gene or the GFP gene to similar efficiency but has a more restricted tropism mouse cells following portal vein injection, contrary to profile, especially with respect to cells of the hemato- several published studies that baculovirus vectors are poietic system, including DCs. not infectious if injected into the vasculature of mice.18,32 This indicates that the baculovirus infection is blocked at the postentry step and not at the receptor-binding or Materials and methods entry stage. The selectivity of the Ampho-pseudotyped lentiviral vectors for transduction of hepatocytes was Plasmid constructs interesting and novel. Previous studies showed that the The VSV-G and baculovirus (AcNPV) gp64 envelope Ampho receptor Pit2 should be highly expressed in liver genes were subcloned into a mammalian expression 23 tissue, but the exact cell types of expression were not vector (pMD-2) that carries a CMV promoter, a b-globin delineated.36 Published reports have shown that retro- intron upstream of the ORF and a b-globin poly A signal. viral vectors carrying the amphotropic envelope can The VSV-G ORF fragment was removed by EcoRI 4 transduce hepatocytes, but in most circumstances liver digestion of pMD.G with EcoRI and ligated into EcoRI- 23 proliferation has been induced to overcome the cell cycle digested pMDLg/pRRE to produce pMD2-VSVG-Env requirement of these vectors; so the transduction (pCG909). The gp64 ORF fragment was prepared by PCR patterns are not entirely dependent on the envelope.37,38 amplification of the AcNPV envelope region from Thus, this study is the first reported use of Ampho- BacPak6 (Clontech, Palo Alto, CA, USA) with addition pseudotyped lentiviral vectors in vivo, and demonstrates of EcoRI sites at each end. The primers used for 0 the transduction of mouse hepatocytes in the absence of amplification are as follows: forward primer 5 -ac- 0 any liver proliferation. ctgggaattcgccgccaccatgctactagtaaatcagtcacaccaa-3 (adds A drawback of baculovirus vectors has been the an EcoRI site and a Kozak consensus sequence 0 inability to transduce cells in animal models in vivo upstream of the ATG); reverse primer 5 - 0 0 because of complement inactivation. It has also been actgacgaattcttaatattgtctattacggtttcta 3 (adds an EcoRI site reported that the VSV-G envelope glycoprotein-pseudo- downstream of the stop codon). PCR amplification using typed lentiviral vectors are sensitive to human comple- these primers produced a product of 1.6 kb. The PCR ment inactivation, whereas Ampho-pseudotyped vectors product was digested with EcoRI and ligated into EcoRI are resistant. Gp64 does indeed confer sensitivity to cut pMDLg/p.RRE to produce pMD2-AcNPV-Env inactivation by human complement, but there was no (pCG917). For both VSV-G and gp64, the resulting sensitivity to complement from mice or rats. Since all of plasmids were confirmed by sequencing. the vector pseudotypes (VSV-G, gp64 and Ampho) are produced in the same cell type, the resistance or Cell culture sensitivity of the vectors is dependent on the identity The following cell lines were obtained from ATCC and of the envelope glycoprotein, and not a glycosylation passaged in the specified medium: 293T (CRL-11268), characteristic of the membrane envelope derived from HeLa (CCL-2), HepG2 (HB-8065), K-562 (CCL-243), PC-3 the producer cell. The sensitivity of gp64- and VSV-G- (CRL-1435), U-118 MG (HTB-15), NIH:Ovcar-3 (HTB- pseudotyped lentiviral vectors to human complement 161), MOLT-4 (CRL-1582), Hep3B (HB-8064), G-361 inactivation will be a significant problem for the use of (CRL-1424), A549 (CCL-185), Te671 (CRL-8805), HT- these particles in human gene-transfer protocols that 1080 (CCL-121), NIH/3T3 (CRL-1658), Cf2Th (CRL- require vascular injections. Fortunately, the problem 1430), Vero (CCL-81), FO (CRL-1646), KG-1 (CCL-246), might be avoided by directly injecting the vectors into SC-1 (CRL-8756), Raji (CCL-86) and XC (CCL-165). C8166 specific tissues or through the use of complement (# 404) and CEM-TART (# 1944) were obtained from the regulatory proteins such as DAF (C Schauber, manu- NIH AIDS Research and Reference Reagent Program. script in preparation). HuH7 cells were obtained from Dr Jing-Hsuing (James) In rodent serum, we found that gp64-pseudotyped Ou, at the University of Southern California. HUAEC vectors are resistant to complement inactivation. These (#CC-2520) and NHDC (CC-2701) were both obtained data are further confirmed by our observation in mouse from Cambrex (East Rutherford, NJ, USA) and cultured studies that gp64 efficiently transduces murine cells in the specified Clonetics media. Cells were grown at 1 following injection into the vasculature, contrary to the 37 Cin5%CO2 at 85% humidity. published reports that baculovirus vectors are sensitive to rodent sera and do not transduce cells in vivo after Small-scale production of pseudotyped lentiviral vascular injection. Therefore, it is likely that baculovirus vectors by transient transfection sensitivity to rodent complement is a result of recogni- Third-generation HIV-1 vectors were prepared by tran- tion of elements derived from the insect producer cell sient transfection, according to published procedures,23 membrane or other baculovirus envelope proteins and using a mixture of four plasmids in the following not specific to the gp64 glycoprotein itself. amounts per 10 cm dish: pMDLg/pRRE (Gag/Pol In summary, our findings suggest that the gp64 expression plasmid) at 6.5 mg, pRSV-Rev23 at 2.5 mg, envelope glycoprotein is a useful component of the Transfer Vector (pRRL.sin.CMV.eGFP.ppt.pre or lentiviral vector system because it confers high transduc- pRRL.sin.CMV.eGFP.pre)39 at 10 mg and 3.5 mg of envel- tion efficiency in vitro and in vivo. It has the advantages ope plasmid (pMD2.VSVG-Env or pMD2.AcNPV.Env or

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 273 pMD2.Ampho.Env).40 Vector particles were harvested in temperature. Standards were made using serial dilutions conditioned medium in one 24-h collection. Vectors were of 20% naı¨ve mouse serum in BBM spiked with human concentrated by ultracentrifugation in a SW28 swinging Factor IX (Calbiochem, La Jolla, CA, USA, 25 ng/ml bucket rotor at 19 500 r.p.m. for 2 h 20 min. Titer analysis starting concentration). Mouse serum samples were and determination of p24Gag concentration were per- diluted 1:5 or 1:10 in BBM and 50 ml/well was loaded formed as previously described.23 in duplicate. In 1:10 dilutions, naı¨ve mouse serum was added to the dilution to make a final solution of 20% Large-scale production and purification mouse serum in BBM. After a 2-h incubation at room of pseudotyped lentiviral vectors temperature, the plates were washed and incubated for m Large-scale preparations of pseudotyped vectors were 90 min with 50 l/well horseradish peroxidase (HRP)- produced by transient transfection in 10-stack cell conjugated goat anti-human Factor IX polyclonal anti- factories (Nunc, Rochester, NY, USA) as follows: low- body (Affinity Biologicals, Ancaster, ON, Canada) that passage 293T cells were expanded and seeded at was diluted 1:100 in BBM. After a final wash step, the m 5.75 Â 108 cells per cell factory the day prior to transfec- plates were developed with 50 l/well o-phenylenedia- tion. The following day, the medium on the cells was mine dihydrochloride peroxidase substrate (Sigma, St changed 2 h prior to transfection. Four plasmids were Louis, MO, USA). After 25 min, the color reaction was m cotransfected in the following amounts per cell factory: stopped by adding 50 l/well 2 M H2SO4, and the optical pMDLg/p.RRE at 747.5 mg, pRSV-Rev at 287.5 mg, Trans- density was determined by reading in a plate reader/ fer Vector (pRRL.sin.CMV.eGFP.ppt.pre or pRRL.sin.CM- spectrophotometer at 490 nm. The limit of detection for V.eGFP.pre or pRRL.sin.CMV.hFIX.ppt.pre)41 at 1.15 mg the ELISA is approximately 1 ng/ml. and 402.5 mg of envelope plasmid (pMD2.VSVG-Env or pMD2.AcNPV.Env or pMD2.Ampho.Env). For each cell Liver histology factory, the DNA was mixed with 0.1 Â TE to a final Livers were processed as previously described for in situ hybridization of mouse embryos.42 The medial and left volume of 51.75 ml. Next, 5.75 ml of 2.5 M CaCl2 was added and mixed. This mixture was then added to lateral lobes of the mouse livers were analyzed sepa- 57.5 ml of 2 Â HBS and added to the 10-stack. The rately. Samples were cyrosectioned at 5–10 mm and medium was changed 16 h later and conditioned directly visualized for GFP fluorescence using a Zeiss medium containing vector particles was harvested in Axioplan microscope (Carl Zeizz, Thornwood, NY, two collections at 24 and 48 h after the medium change. USA). Vectors were purified by DEAE ion exchange (ToyoPearl DEAE-650C) chromatography by loading the condi- Determination of complement sensitivity tioned medium at a 20:1 ratio of load volume to bed Serum samples from single human donors, Nude rats or volume. The column was then washed with 5 volumes of pooled sera from C57BL/6 mice were thawed and half of PBS and eluted with 2.5 volumes of PBS containing each serum sample was heated at 561C for 1 h to 0.75 M NaCl. The eluate was further concentrated by inactivate the complement. All sera were made by ultracentrifugation as for small-scale vector preparations. collecting fresh whole blood and allowing it to clot For the human Factor IX vectors that were injected into completely for several hours. Clots were pelleted by mice, TU were calculated based on the p24Gag concentra- centrifugation and the serum supernatants were re- tion and the average infectivity (in TU/ng p24Gag) moved and frozen immediately in aliquots. CMV-GFP determined for GFP vectors with the same pseudotypes. vector samples with different envelope pseudotypes were diluted to 106 TU/ml. A volume of 30–75 ml of the In vivo mouse studies vector was diluted 1:5 into normal or heat-inactivated All animal work was approved by the Institutional serum (or medium containing 10% heat-inactivated FBS as the no-serum control) and the mixture was incubated Animal Care and Use Committee at Cell Genesys. 1 Female NIH Swiss Nude, 6–8-week old (Taconic, at 37 C for 1 h. Following the incubation, medium was Germantown, NY, USA), mice were cannulated in the added to the reaction for an additional 1:5 dilution and hepatic portal vein 1 to 2 days prior to injection. Vectors then serially diluted twice at a 1:10 ratio, and used to m were injected into the cannula at a total volume of 100 ml. infect 293T cells in the presence of 8 g/ml Polybrene for Blood was collected from the retro-orbital plexus at the end-point titer determination. The titer value was indicated time points and processed into serum for divided by the titer determined for the vector mixed human Factor IX analysis. For GFP analysis, animals with medium (the no-serum control), and reported as the were euthanized 11 days after injection, and the medial % recovery of titer compared to this control. All serum and left lateral lobes of the liver were collected and samples were analyzed for hemolytic complement processed for GFP analysis, as described below. activity (CH50) using kits from Quidel (CH50 Eq Enzyme Immunoassay; San Diego, CA, USA) or Diamedix (E-Z CH ; Miami, FL, USA). ELISA determination of human Factor IX concentration 50 Flat-bottomed 96-well plates (Nunc, Rochester, NY, USA) were coated with 100 ml/well of mouse monoclonal anti- Acknowledgements human Factor IX antibody (Roche Applied Science, Indianapolis, IN, USA) diluted to 2 mg/ml in boric We thank Minh Nguyen for experimental assistance and acid-buffered saline (BBS) and left overnight at 41C. comments on the manuscript. We also thank Jeanette Following coating, the plates were washed with BBS Dilley, Trini Arroyo, Debbie Farson and Hend Omran for containing 0.025% Tween-20, and blocked with 200 ml/ cell lines, as well as Alanna Weissman-Ward, Jim Liu and well BBM (1% nonfat milk in BBS) for 2 h at room Pratima Kundu for vector production and purification.

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 274 In addition, we are grateful to Melissa Gonzalez, who 17 Hofmann C et al. Efficient gene transfer into human hepatocytes performed the liver histology, Sandy Sanchez for DNA by baculovirus vectors. Proc Natl Acad Sci USA 1995; 92: sequencing and oligo synthesis, Satya Yendluri for 10099–10103. providing the Factor IX ELISA protocol and especially 18 Sandig V et al. Gene transfer into hepatocytes and human liver to members of the Cell Genesys animal facility including tissue by baculovirus vectors. Hum Gene Ther 1996; 7: 1937–1945. Melinda vanRoey, Tammy Langer, Sheila Tanciongo, 19 Shoji I et al. Efficient gene transfer into various mammalian cells, Amanda Koehne and Jacki Gire. Final thanks go to all including non-hepatic cells, by baculovirus vectors. J Gen Virol the members of the lentiviral vector group for helpful 1997; 78: 2657–2664. discussions and Andy Simmons and Tom Harding for 20 Condreay JP, Witherspoon SM, Clay WC, Kost TA. Transient and critical reading of the manuscript. stable gene expression in mammalian cells transduced with a recombinant baculovirus vector. Proc Natl Acad Sci USA 1999; 96: 127–132. 21 Boyce FM, Bucher NL. Baculovirus-mediated gene transfer into References mammalian cells. Proc Natl Acad Sci USA 1996; 93: 2499–2504. 22 Kumar M, Bradow BP, Zimmerberg J. Large-scale production of 1 Emi N, Friedmann T, Yee JK. Pseudotype formation of murine pseudotyped lentiviral vectors using baculovirus gp64. Hum leukemia virus with the G protein of vesicular stomatitis virus. Gene Ther 2003; 14: 67–77. J Virol 1991; 65: 1202–1207. 23 Dull T et al. A third generation lentiviral vector with a 2 Burns JC et al. Vesicular stomatitis virus G protein pseudotyped conditional packaging system. J Virol 1998; 72: 8463–8471. retroviral vectors: concentration to very high titer and efficient 24 Gallardo HF, Tan C, Ory D, Sadelain M. Recombinant gene transfer into mammalian and nonmammalian cells. Proc retroviruses pseudotyped with the vesicular stomatitis virus G Natl Acad Sci USA 1993; 90: 8033–8037. glycoprotein mediate both stable gene transfer and 3 Yee JK, Friedmann T, Burns JC. Generation of high-titer pseudotransduction in human peripheral blood lymphocytes. pseudotyped retroviral vectors with very broad host range. Blood 1997; 90: 952–957. Methods Cell Biol 1994; 43: 99–112. 25 Liu ML, Winther BL, Kay MA. Pseudotransduction of 4 Ory DS, Neugeboren BA, Mulligan R. A stable human-derived hepatocytes by using concentrated pseudotyped vesicular packaging cell line for production of high titer retrovirus/ stomatitis virus G glycoprotein (VSV-G)-Moloney murine vesicular stomatitis virus G pseudotypes. Proc Natl Acad Sci USA leukemia virus-derived retrovirus vectors: comparison of VSV- 1996; 93: 11400–11406. G and amphotropic vectors for hepatic gene transfer. J Virol 1996; 5 Farson D et al. A new-generation stable inducible packaging cell 70: 2497–2502. line for lentiviral vectors. Hum Gene Ther 2001; 12: 981–997. 26 Gruber A et al. Dendritic cells transduced by multiply deleted 6 Kafri T, van Praag H, Ouyang L, Gage FH. A packaging cell line HIV-1 vectors exhibit normal phenotypes and functions and for lentivirus vectors. J Virol 1999; 73: 576–584. elicit an HIV-specific cytotoxic T-lymphocyte response in vitro. 7 Dyall J, Latouche JB, Schnell S, Sadelain M. Lentivirus- Blood 2000; 96: 1327–1333. transduced human monocyte-derived dendritic cells efficiently 27 High KA. Gene transfer as an approach to treating hemophilia. stimulate antigen-specific cytotoxic T lymphocytes. Blood 2001; Circ Res 2001; 88: 137–144. 97: 114–121. 28 Hofmann C, Strauss M. Baculovirus-mediated gene transfer 8 Chinnasamy N et al. Efficient gene transfer to human peripheral in the presence of human serum or blood facilitated by blood monocyte-derived dendritic cells using human inhibition of the complement system. Gene Therapy 1998; 5: immunodeficiency virus type 1-based lentiviral vectors. Hum 531–536. Gene Ther 2000; 11: 1901–1909. 29 Sarkis C et al. Efficient transduction of neural cells in vitro and in 9 Metharom P, Ellem KA, Schmidt C, Wei MQ. Lentiviral vector- vivo by a baculovirus-derived vector. Proc Natl Acad Sci USA mediated tyrosinase-related protein 2 gene transfer to dendritic 2000; 97: 14638–14643. cells for the therapy of melanoma. Hum Gene Ther 2001; 12: 30 Huser A, Rudolph M, Hofmann C. Incorporation of decay- 2203–2213. accelerating factor into the baculovirus envelope generates 10 VandenDriessche T et al. Lentiviral vectors containing the human complement-resistant gene transfer vectors. Nat Biotechnol immunodeficiency virus type-1 central polypurine tract can 2001; 19: 451–455. efficiently transduce nondividing hepatocytes and antigen- 31 Chernomordik L, Leikina E, Cho MS, Zimmerberg J. Control of presenting cells in vivo. Blood 2002; 100: 813–822. baculovirus gp64-induced syncytium formation by membrane 11 Welsh RMJ, Cooper NR, Jensen FC, Oldstone MB. Human serum lipid composition. J Virol 1995; 69: 3049–3058. lyses RNA tumour viruses. Nature 1975; 257: 612–614. 32 Tani H et al. Characterization of cell-surface determinants 12 Takeuchi Y, Porter C, Strahan K. Sensitization of cells and important for baculovirus infection. Virology 2001; 279: 343–353. retroviruses to human serum by alpha(1–3)galactosyltransferase. 33 Wickham TJ et al. Equilibrium and kinetic analysis of Auto- Nature 1996; 379: 85–88. grapha californica nuclear polyhedrosis virus attachment 13 Takeuchi Y, Liong SH, Bieniasz PD. Sensitization of rhabdo-, to different insect cell lines. J Gen Virol 1992; 73 (Part 12): lenti-, and spumaviruses to human serum by galactosyl(alpha1– 3185–3194. 3)galactosylation. J Virol 1997; 71: 6174–6178. 34 Wang P, Hammer DA, Granados RR. Binding and fusion of 14 DePolo NJ et al. VSV-G pseudotyped lentiviral vector particles Autographa californica nucleopolyhedrovirus to cultured insect produced in human cells are inactivated by human serum. Mol cells. J Gen Virol 1997; 78 (Part 12): 3081–3089. Ther 2000; 2: 218–222. 35 Charan S, Hengartner H, Zinkernagel RM. Antibodies against 15 Sandrin V et al. Lentiviral vectors pseudotyped with a modified the two serotypes of vesicular stomatitis virus measured by RD114 envelope glycoprotein show increased stability in sera enzyme-linked immunosorbent assay: immunodominance of and augmented transduction of primary lymphocytes and serotype-specific determinants and induction of asymmetrically CD34+ cells derived from human and nonhuman primates. cross-reactive antibodies. J Virol 1987; 61: 2509–2514. Blood 2002; 100: 823–832. 36 Kavanaugh MP et al. Cell-surface receptors for gibbon ape 16 Kost TA, Condreay JP. Recombinant baculoviruses as leukemia virus and amphotropic murine retrovirus are inducible mammalian cell gene-delivery vectors. Trends Biotechnol 2002; sodium-dependent phosphate symporters. Proc Natl Acad Sci 20: 173–180. USA 1994; 91: 7071–7075.

Gene Therapy gp64-pseudotyped lentiviral vectors in vivo CA Schauber et al 275 37 Menoret S et al. lacZ transgenic rats tolerant for beta- 40 Finer MH et al. kat: a high-efficiency retroviral transduc- galactosidase: recipients for gene transfer studies using lacZ as tion system for primary human T lymphocytes. Blood 1994; 83: a reporter gene. Hum Gene Ther 2002; 13: 1383–1390. 43–50. 38 Xu L et al. Transduction of hepatocytes after neonatal delivery of 41 Tsui LV et al. Production of human clotting Factor IX without a Moloney murine leukemia virus based retroviral vector toxicity in mice after vascular delivery of a lentiviral vector. Nat results in long-term expression of beta-glucuronidase in Biotechnol 2002; 20: 53–57. mucopolysaccharidosis VII dogs. Mol Ther 2002; 5: 141–153. 42 Simmons AD, Horton S, Abney AL, Johnson JE. Neurogenin2 39 Follenzi A et al. Efficient gene delivery and targeted expression expression in ventral and dorsal spinal neural tube progenitor to hepatocytes in vivo by improved lentiviral vectors. Hum Gene cells is regulated by distinct enhancers. Dev Biol 2001; 229: Ther 2002; 13: 243–260. 327–333.

Gene Therapy