In Vitro and In Vivo Gene Delivery by Recombinant Baculoviruses

Abstract

Although recombinant baculovirus vectors can be an efficient tool for gene transfer into mammalian cells in vitro, gene transduction in vivo has been hampered by the inactivation of baculoviruses by serum complement. Recombinant baculoviruses possessing excess envelope protein gp64 or other viral envelope proteins on the virion surface deliver foreign genes into a variety of mammalian cell lines more efficiently than the unmodified baculovirus. In this study, we examined the efficiency of gene transfer both in vitro and in vivo by recombinant baculoviruses possessing envelope proteins derived from either vesicular stomatitis virus (VSVG) or rabies virus. These recombinant viruses efficiently transferred reporter genes into neural cell lines, primary rat neural cells, and primary mouse osteal cells in vitro. The VSVG-modified baculovirus exhibited greater resistance to inactivation by animal sera than the unmodified baculovirus. A synthetic inhibitor of the complement activation pathway circumvented the serum inactivation of the unmodified baculovirus. Furthermore, the VSVG-modified baculovirus could transduce a reporter gene into the cerebral cortex and testis of mice by direct inoculation in vivo. These results suggest the possible use of the recombinant baculovirus vectors in combination with the administration of complement inhibitors for in vivo gene therapy.

The baculovirus Autographa californica nuclear polyhedrosis virus is an insect virus possessing a large double-stranded circular DNA genome packaged into a rod-shaped capsid (22). Due to the very strong polyhedrin promoter, baculoviruses have been used as a tool to produce high levels of recombinant protein in insect cells. Several years ago, baculovirus was shown to infect hepatic cell lines and express foreign genes under the control of mammalian promoters (5, 10). We have subsequently shown that recombinant baculoviruses transduce foreign genes into additional cell lines, including those of nonhepatic origin (38). Therefore, baculovirus is now recognized as a useful viral vector not only for abundant expression of foreign proteins in insect cells, but also for gene delivery to mammalian cells (18, 33, 36).

Recently, enhanced gene transfer efficacy was observed in a variety of cell lines with recombinant baculoviruses possessing either other viral envelope proteins, such as vesicular stomatitis virus envelope G protein (VSVG), or excess amounts of its envelope glycoprotein, gp64, on the virion surface (3, 41). Although modification of the virion surface enhanced the efficiency of gene transduction into various cultured cell lines, in vivo gene delivery with recombinant baculoviruses is still unsatisfactory. One obstacle is the inactivation of baculovirus by serum complement (11, 35). In vivo foreign gene transfer with baculovirus vectors into rabbit endothelial cells lining the artery through collar-mediated delivery (1), mouse skeletal muscle cells in the quadriceps by intramuscular injection (32), neural or choroid plexus cells in the rodent brain by intracranial injection (19, 37), and mouse retinal pigment epithelial cells following subretinal injection (8) has been achieved. Few reports exist, however, demonstrating the efficient transfer of genes via baculovirus vectors into internal organs that are directly exposed to serum complement. A recombinant baculovirus possessing decay-accelerating factor, an inhibitor of the various pathways of the complement system, allowed enhanced gene transfer into neonatal rat liver. The level of foreign gene expression, however, was not high, and gene transduction into adult rat liver was not successful (12).

In this study, we examined the efficiency of in vitro and in vivo gene transfer by recombinant baculoviruses possessing rhabdovirus envelope proteins. Recombinant viruses efficiently transferred reporter genes not only into primary neural and osteal cells in vitro but also into the cerebrums and testes of mice in vivo. Addition of a complement inhibitor conferred resistance against serum inactivation of the baculovirus vectors in vitro. The possible application of such a baculovirus vector for future in vivo gene therapy will be discussed.

MATERIALS AND METHODS

Virus construction.

Recombinant baculoviruses were constructed as described previously (41). The AcRVG-CAluc and AcRVG-CAGFP viruses possess rabies virus G glycoprotein (RVG) under the control of the polyhedrin promoter as well as the luciferase and green fluorescent protein (GFP) genes under the control of the CAG promoter (27). The RVG gene was excised from pUCRVG (43) by digestion with EcoRI and HindIII, filled in with Klenow enzyme, and cloned into the BamHI site of pAcYM1 (19) after addition of a BamHI linker.

To construct the transfer vector pAcRVG-CAG, the CAG cassette was excised from pCAGGS (27) by SalI and BamHI digestion, filled in with Klenow enzyme, and inserted into the EcoRV site of pAcRVG in the opposite direction from the polyhedrin promoter. The luciferase and GFP genes were cloned into the BglII site of pAcRVG-CAG after addition of a BclI linker. The correct orientation and sequence of each transfer vector construct were confirmed by PCR and sequencing. Sf9 insect cells were cotransfected with the transfer vector and baculovirus DNA (Baculo Gold linearized DNA; Pharmingen, San Diego, Calif.). Following homologous recombination, recombinant baculoviruses were isolated and purified as described previously (20). Recombinant baculoviruses AcVSVG-CAluc, Ac64-CAluc, and AcGFP-CAluc, possessing the luciferase gene under the control of the CAG promoter, and VSVG, gp64, and the GFP gene under the control of the polyhedrin promoter were described previously (41). The infectious titers of the recombinant baculoviruses were determined by plaque assay with Sf9 cells.

Expression of foreign genes in insect and mammalian cells.

Expression of RVG protein in insect cells infected with either AcRVG-CAluc or AcRVG-CAGFP was analyzed by immunofluorescence and Western blot analysis. Cell extracts and purified viruses were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred to a polyvinylidene difluoride membrane (Millipore, Tokyo, Japan). An anti-rabies virus rabbit polyclonal antibody (43) was used to detect RVG protein, which was visualized with the alkaline phosphatase assay method as described previously (34). Immunofluorescence analysis was carried out as described previously (45), with cell fixation performed with 4% paraformaldehyde. Expression of either luciferase or GFP proteins in mammalian cells infected with the recombinant baculoviruses was then examined by either the luciferase assay method or fluorescence microscopy, as described previously (41).

Cell cultures and infection with recombinant baculoviruses.

The mammalian neural cell lines BC3H1 (mouse brain, smooth muscle-like tumor), NB41A3 (mouse brain, neuroblastoma), SK-N-MC (human brain, neuroblastoma), IMR32 (human brain, neuroblastoma), and PC-12 (rat adrenal gland, pheochromocytoma) as well as the human hepatoma cell line HepG2 were purchased from Dainippon Pharmaceutical Co., Ltd. (Osaka, Japan). BC3H1, SK-N-MC, IMR32, and HepG2 cells were maintained in Dulbecco's modified Eagle's medium (Gibco Laboratories, Grand Island, N.Y.) containing 2 mM l-glutamine, penicillin (50 IU/ml), streptomycin (50 μg/ml), and 10% (vol/vol) heat-inactivated fetal calf serum (FCS). PC-12 and NB41A3 cell lines were cultivated in RPMI 1640 (Gibco Laboratories) with 10% FCS and Ham's F-12 medium (Gibco Laboratories) supplemented with 2.5% FCS and 15% horse serum, respectively. Recombinant baculoviruses were inoculated into 10⁵ cells in 24-well plates at a multiplicity of infection (MOI) of 50 by a 1-h incubation, facilitating viral adsorption. After washing and the addition of fresh medium, the cells were incubated at 37°C for 24 h.

Primary cell cultures and infection with recombinant baculoviruses.

All animal experiments conformed to the Guidelines for the Care and Use of Laboratory Animals and were approved by the Institutional Committee of Laboratory Animal Experimentation (Research Institute for Microbial Diseases, Osaka University). Primary cerebellar and hippocampal cultures were prepared from Wistar ST rats at embryonic day 19 (CLEA Japan, Tokyo, Japan). Briefly, the whole cerebellum or hippocampus, including the region of the cerebral cortex connecting to the hippocampus, was dissected out from rat brains and treated with 0.1% trypsin. Trypsin-treated tissues were dissociated in 0.05% DNase I and resuspended in seeding medium. Cerebellar cultures were maintained in Dulbecco's modified Eagle's medium-F12 medium (Gibco Laboratories) containing 0.5% FCS, putrescine (100 μM), sodium selenite (30 nM), l-glutamine (4 mM), triiodothyronine (0.5 μg/ml), progesterone (5 nM), bovine insulin (10 μg/ml), transferrin (100 μg/ml), and gentamicin (10 μg/ml). The mixture of cerebral cortex and hippocampal cultures was suspended and grown in neurobasal medium (Gibco Laboratories) containing B27 supplement and l-glutamine (0.5 mM) (Gibco Laboratories).

Cells were seeded at 2 × 10⁵ cells/well on poly-l-ornithine-coated coverslips in 24-well plates. These cultures were infected with either AcVSVG-CAGFP or AcRVG-CAGFP at an MOI of 100 at 8 days after incubation. Primary neuronal cultures grown on coverslips were washed with phosphate-buffered saline (PBS), fixed in 3% paraformaldehyde for 20 min, and permeabilized with PBS containing 0.25% Triton X-100 for 10 min at room temperature. Cells were then double labeled with either rabbit anti-GFP (Molecular Probes, Eugene, Oreg.), mouse anti-Calbindin (Swant, Bellinzona, Switzerland), mouse anti-MAP2 (ICN Biomedicals, Costa Mesa, Calif.), or mouse anti- glial fibrillary acidic protein (GFAP) (Zymed Laboratories, Inc., South San Francisco, Calif.) for 30 min to 1 h at 37°C. After incubation, cells were washed with PBS and incubated with Alexa 568-conjugated anti-mouse or Alexa 488-conjugated anti-rabbit immunoglobulin antibody (Molecular Probes) for 30 min at room temperature. Coverslips were mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, Calif.) and subjected to microscopic observation. Images were acquired on a Zeiss LSM510 confocal microscope (Carl Zeiss Inc., Thornwood, N.Y.). The luciferase activity of cerebellar cultures following infection with Ac64-CAluc, AcRVG-CAluc, or AcVSVG-CAluc at an MOI of 50, 100, and 150 was determined. Luciferase activities were determined 2 days after incubation.

Osteoclast or osteoblast cells were isolated from female ddY mice (Japan SLC, Hamamatsu, Japan) and resuspended in Dulbecco's modified Eagle's medium (Gibco Laboratories) containing 10% FCS as described previously (17, 29). After cultivation in a 48-well plate for 3 days, cultures were infected with either AcVSVG-CAluc or Ac64-CAluc at various MOIs. Luciferase activities were determined 2 days after incubation.

Effects of animal serum and complement inhibitor on baculovirus infectivity.

Blood was freshly drawn from animals and healthy volunteers after obtaining informed consent. Sera were isolated by centrifugation at 3,000 × g for 10 min at 4°C. Serum complement was inactivated at 56°C for 30 min. To determine the effects of serum complement on the inactivation of baculovirus, 10 μl of either AcVSVG-CAluc or AcGFP-CAluc (2 × 10⁹ PFU/ml) was incubated with 90 μl of either untreated or heat-inactivated serum for 1 h at 37°C. AcGFP-CAluc was also incubated for 1 h at 37°C in the presence of rat or human serum with various concentrations of FUT-175 (6-amidino-2-naphthyl 4-guanidinobenzoate; Torii & Co., Ltd., Tokyo, Japan), a synthetic protease inhibitor that inhibits C1r or C1 esterase (26, 40) and C3 convertase (13). Residual infectivity was determined by inoculation into HepG2 cells. Luciferase activity was determined 24 h after incubation.

Direct injections of recombinant baculovirus into mouse brain and histological analysis.

BALB/c mice were obtained from CLEA Japan. Three-week-old female mice were injected in the right lateral ventricle with 10 μl of purified AcVSVG-CAGFP (2 × 10⁹ PFU/ml in PBS containing 0.04% trypan blue) with a 28-gauge KN-386 needle (Natsume Co., Ltd., Tokyo, Japan). To assess GFP expression in the brain, fluorescent stereomicroscopic pictures of the whole brain were obtained 24 h after injection. Brains were rinsed with PBS and observed under a Leica WILD M10 fluorescence stereomicroscope (Leica Microsystems, Wetzlar, Germany). Brain samples were then fixed in 4% paraformaldehyde, cryopreserved in 30% sucrose and 5% glycerol, and frozen in Tissue-Tek O.C.T. compound (Sakura Finetechnical Co., Ltd., Tokyo, Japan). Cryostat sections (10 μm) were observed under a fluorescence microscope. Sections were further examined by immunohistochemical analysis after staining with either rabbit anti-GFP (Molecular Probes), mouse anti-MAP2 (ICN Biomedicals), or mouse anti-GFAP (Zymed Laboratories) as described above.

Gene transduction by recombinant baculovirus into mouse testis.

AcVSVG-CAGFP was delivered into mouse testes via the efferent ductules with an injection pipette, as described by Ogawa et al. (28). Trypan blue (0.02%) was included in the virus suspension to monitor the filling of the seminiferous tubules. Approximately 10 μl of viral suspension (2 × 10¹⁰ PFU/ml) was injected, filling approximately 70% of the seminiferous tubules with the viral suspension. Two days after injection, the testes were observed under a fluorescence stereomicroscope.

RESULTS

Construction of recombinant viruses.

To examine cell surface expression of foreign envelope proteins, Sf9 cells were infected at an MOI of 1 with the recombinant baculoviruses AcRVG-CAluc, Ac64-CAluc, and AcVSVG-CAluc, containing the RVG, gp64, and VSVG genes, respectively, under the control of the polyhedrin promoter. The cell surface expression of foreign envelope proteins on Sf9 cells was detected by using specific monoclonal antibodies (Fig. 1A). Incorporation of rhabdovirus envelope proteins into the viral particles was evaluated by Western blot analysis of purified viruses. The RVG and VSVG envelope proteins were detected in AcRVG-CAluc and AcVSVG-CAluc particles (Fig. 1B). Incorporation of gp64 protein into Ac64-CAluc was higher than that of either AcVSVG-CAluc or AcRVG-CAluc, as reported previously (41).

FIG. 1.

Expression of viral envelope proteins in insect cells and incorporation of envelope proteins into the recombinant baculoviruses. (A) Sf9 cells were infected with AcRVG-CAluc (A to C), Ac64-CAluc (D to F), or AcVSVG-CAluc (G to I) at an MOI of 1 and harvested 72 h after infection. Expression of RVG (A), VSVG (H), and gp64 (C, F, and I) was examined by immunofluorescence analysis after fixation in 4% paraformaldehyde. (B) The incorporation of RVG, VSVG, and gp64 into the purified virions was examined by Western blotting. Lanes 1, 2, and 3, AcRVG-CAluc, Ac64-CAluc, and AcVSVG-CAluc, respectively.

Gene delivery into neural cells.

To investigate the efficiency of the recombinant baculoviruses for the gene transduction of neural cells, five neural cell lines (BC3H1, NB41A3, SK-N-MC, IMR32, and PC-12) were infected with recombinant baculoviruses at an MOI of 50. Reporter gene expression was analyzed at 24 h postinfection. All of the neural cell lines infected with AcRVG-CAluc or AcVSVG-CAluc exhibited 5- to 1,000-fold higher expression than those infected with either Ac64-CAluc or AcGFP-CAluc, which contain the GFP gene under the control of the polyhedrin promoter (Fig. 2). AcVSVG-CAluc and Ac64-CAluc exhibited higher expression than AcRVG-CAluc and AcGFP-CAluc in HepG2 cells. In PC-12 cells, a rat adrenal pheochromocytoma cell line, AcVSVG-CAluc induced about 50 times higher gene transduction than AcRVG-CAluc. These results indicated that the incorporation of rhabdovirus envelope proteins into baculoviruses can promote gene delivery into neural cell lines.

FIG. 2.

Expression of luciferase in various neural cell lines infected with recombinant baculoviruses. Neural cells were infected with AcRVG-CAluc, AcVSVG-CAluc, Ac64-CAluc, or AcGFP-CAluc at an MOI of 50. Luciferase expression was determined 24 h after infection. Values for relative light units (RLU) reflect values obtained for the extracts from 10⁵ cells. The value for mock-infected cells was subtracted from all values. The results shown are the averages of three independent assays, with error bars representing the standard deviation.

We next examined the efficiency of gene transfer by recombinant baculoviruses into primary rat cerebellar cells (Fig. 3). While the efficiencies of gene transfer into primary rat cerebellar cells with recombinant baculoviruses were 10- to 100-fold lower than those into mouse and human neural cell lines, these transduction rates were similar to those for rat PC-12 cells. Although AcRVG-CAluc and Ac64-CAluc exhibited lower gene transduction into the PC-12 cell line than AcVSVG-CAluc, AcRVG-CAluc exhibited higher gene transduction into primary rat cerebellar cells than AcVSVG-CAluc and Ac64-CAluc. To examine foreign gene expression in primary neural cells in greater detail, we examined GFP expression in primary cerebellar and hippocampal cultures infected with AcVSVG-CAGFP with the confocal microscope. Immunostaining of GFP-expressing cells with a Calbindin purkinje cell marker, a GFAP glial marker, and a MAP2 neuronal marker confirmed the phenotype of the infected cells. AcVSVG-CAGFP transferred GFP to purkinje cells (Fig. 4A to C), astrocytes (Fig. 4D to F), and pyramidal cells (Fig. 4G to I). A similar level of GFP expression in these primary neuronal cells was demonstrated following infection with AcRVG-CAGFP (data not shown).

FIG. 3.

Expression of luciferase in primary rat cerebellar cells infected with recombinant baculoviruses. Rat cerebellar cells were infected with Ac64-CAluc, AcVSVG-CAluc, or AcRVG-CAluc at MOIs of 50, 100, and 150. Luciferase expression was determined 24 h after infection. Values for relative light units (RLU) reflect values obtained for the extracts from 10⁵ cells. The value for mock-infected cells was subtracted from all values. The results shown are the averages of three independent assays, with error bars representing the standard deviation.

FIG. 4.

Gene transduction into primary rat cerebellar cells with AcVSVG-CAGFP. Primary rat cerebellar and hippocampal cultures were infected with AcVSVG-CAGFP. Immunofluorescence was examined by confocal microscopy. (A, D, and G) Anti-GFP immunochemistry. (B) Anti-Calbindin immunochemistry as a purkinje marker. (E) Anti-GFAP immunochemistry as a glial marker. (H) Anti-MAP2 immunochemistry as a neuronal marker. C, F, and I are merged images.

Despite comparable gene delivery into human and mouse neural cell lines and primary rat neural cells of the recombinant baculoviruses possessing RVG protein (AcRVG-CAGFP and AcRVG-CAluc) and those possessing VSVG protein, gene transduction into HepG2 and PC-12 cell lines was low. We therefore used AcVSVG-CAGFP and AcVSVG-CAluc for further in vitro and in vivo experiments.

Gene transfer into osteal cells.

Recently, the efficient gene transduction of osteogenic sarcoma cell lines by recombinant baculoviruses possessing cytomegalovirus or Rous sarcoma virus promoters was reported (39). To determine the efficiency of gene transfer by recombinant baculoviruses into primary osteal cells, primary mouse osteoblasts and osteoclast cells were infected with either AcVSVG-CAluc or Ac64-CAluc (Fig. 5). AcVSVG-CAluc transduced the luciferase gene into osteal cells more efficiently than did Ac64-CAluc, in a dose-dependent manner. For transduction into osteoclast cells, 100- to 1,000-fold more virus than was needed for osteoblast cells was required. It is worth noting that no cytopathic effects were observed even at high MOIs. The necessity for increased viral titers may reflect the size of the cells (osteoclasts are 10- to 20-fold bigger than osteoblasts) or the expression of cellular receptors specific for the recombinant viruses.

FIG. 5.

Gene transduction into primary mouse osteoblast (left) and osteoclast (right) cells with recombinant baculoviruses. Osteal cells (10⁵) were infected with either Ac64-CAluc or AcVSVG-CAluc at various MOIs. Following harvest at 24 h after infection, luciferase expression was determined. Values for relative light units (RLU) reflect values obtained for the extracts from 10⁵ cells. The value for mock-infected cells was subtracted for all values. The results shown are the averages of three independent assays, with error bars representing the standard deviation.

Effects of serum on gene transduction by recombinant baculoviruses.

Previous studies suggest that baculoviral gene transfer into hepatocytes is strongly reduced in the presence of serum complement (11, 35). Pseudotype retroviruses possessing VSVG protein exhibited more resistance to serum components than those lacking the protein (30). To compare the susceptibility of the recombinant baculoviruses to inactivation with serum complement, AcVSVG-CAluc and AcGFP-CAluc were incubated for 1 h at 37°C with either untreated or heat-inactivated serum from various animals at a final concentration of 90%. Residual infectivity was determined following inoculation into HepG2 cells (Fig. 6). Significant reductions in luciferase expression following AcGFP-CAluc infection were observed following incubation with serum from human, guinea pig, and rat; only moderate or slight reductions in infectivity were observed following incubation with rabbit, hamster, and mouse serum. AcVSVG-CAluc, however, exhibited resistance to the serum treatment, with the exception of rat serum. Reductions in inactivation were observed in human and guinea pig serum. The reduction of gene expression following serum treatment was abolished by inactivation of the serum by incubation at 56°C for 30 min. These results indicate that incorporation of VSVG protein into baculoviruses confers resistance against inactivation by serum complement; the serum from rabbit, hamster, and mouse possesses relatively weaker inactivation effects against baculoviruses than that from human, guinea pig, and rat.

FIG. 6.

Effect of VSVG incorporation into virions on inactivation of baculovirus by animal serum. Luciferase expression in HepG2 cells infected with either AcVSVG-CAluc or AcGFP-CAluc was examined after incubation with either untreated or heat-inactivated animal serum. Values for relative light units (RLU) reflect values obtained for the extracts from 10⁵ cells. The results shown are the averages of three independent assays, with the error bars representing the standard deviation.

Effects of FUT-175 on serum inactivation of baculovirus.

Baculovirus activates the classical complement system, leading to viral inactivation through the assembly of late complement components (11, 35). To circumvent serum inactivation, we examined the effects of a complement pathway inhibitor, FUT-175, on gene transduction by AcGFP-CAluc in the presence of rat and human serum. AcGFP-CAluc was incubated with 40% rat or human serum in the presence of various concentrations of FUT-175 for 1 h at 37°C. The remaining infectious titer was determined by infection of HepG2 cells (Fig. 7). Treatment with FUT-175 suppressed serum inactivation of AcGFP-CAluc in a dose-dependent manner. Particularly, complete restoration of infectivity to control levels was obtained at a concentration of 1 μg/ml in human serum.

FIG. 7.

Effect of FUT-175 on inactivation of baculovirus by rat and human serum. AcGFP-CAluc was incubated with medium containing 40% animal serum with various concentrations of FUT-175. Remaining baculovirus infectivity was evaluated by luciferase expression following infection of HepG2 cells. Values for relative light units (RLU) reflect values obtained for the extracts from 10⁵ cells. The results shown are the averages of three independent assays, with the error bars representing the standard deviation.

Gene transfer in vivo.

Sarkis et al. demonstrated GFP expression in mouse and rat neural cells after immunostaining of brains injected with an unmodified baculovirus containing the GFP gene under the control of the cytomegalovirus promoter together with cobra venom factor, an inhibitor of the complement system, with a stereotaxic apparatus (37). Recently, Lehtolainen et al. showed that unmodified baculovirus transduced choroid plexus cells in rat ventricles (19). To examine the ability of the VSVG-modified baculovirus to mediate more efficient gene transduction into the cerebrum following ordinary injection routes, AcVSVG-CAGFP was injected directly into the mouse brain. GFP expression at the brain surface of the injected areas was examined by fluorescence stereomicroscopic observation 2 days after injection. GFP expression was clearly detected in areas of the brain corresponding to the injection route of AcVSVG-CAGFP with a fluorescence stereomicroscope (Fig. 8A). Cross sections of the mouse cerebrum following injection with the recombinant baculovirus clearly revealed GFP expression in the cerebral cortex. Immunohistochemical staining of sections with antibodies against GFAP and MAP2 revealed that both astrocytes and pyramidal cells were transduced, expressing GFP following injection with the baculovirus (Fig. 8B).

FIG. 8.

GFP expression in mouse brains after cerebral injections of AcVSVG-CAGFP. Mice were injected with 4 × 10⁷ PFU of AcVSVG-CAGFP in the right lateral ventricle. GFP expression in the brain was examined by fluorescent stereomicroscopy 2 days after injection. (A) Panels A to D are stereomicroscopic images of whole brain (A and B) and brain cross sections (C and D). Panels A and C are bright-field views, while panels B and D are fluorescent views. Arrows and dark staining indicate the injection route, as the infiltrated viral inoculum contained 0.04% trypan blue. (B) Immunohistochemical staining of the cryostat sections was examined by fluorescence microscopy following staining with antibodies specific for GFP (A and D), GFAP as a glial marker (B), or MAP2 as a neuronal marker (E). Panels C and F are merged images.

We also inoculated AcVSVG-CAGFP into mouse testes via the efferent ductules. GFP expression in the testes was examined by fluorescence stereomicroscopy 2 days postinfection. High levels of GFP expression were observed in the whole mouse testes and in the seminiferous tubules (Fig. 9). In sections, we observed clear GFP expression in the basal and Sertoli cells, but not in spermatocytes or sperm cells.

FIG. 9.

GFP expression in mouse testis after inoculation of AcVSVG-CAGFP. Mice were injected with 2 × 10⁸ PFU of AcVSVG-CAGFP via the efferent ductules. GFP expression in the testes was examined by fluorescent stereomicroscopy 2 days after injection. Panels A and B are stereomicroscopic images of whole testis. Panels C and D are images of the efferent ductules, while panels E and F are microscopic images of cross sections of the seminiferous tubules. The left-hand panels are bright-field views. The right-hand panels are fluorescent views. Panel E is the same section as panel F stained with hematoxylin. The scale bars in the upper, middle, and lower panels represent 1 mm, 200 nm, and 50 nm, respectively.

DISCUSSION

Viral vectors derived from naturally occurring viruses, such as adenoviruses, adeno-associated viruses, retroviruses, and herpesviruses, are highly adapted to their natural hosts, providing the means for efficient gene transfer into cultured cells, animal models, and possibly patients (24). Baculovirus vectors, which can incorporate large DNA inserts, efficiently infect not only insect cells but also various mammalian cell lines without apparent viral replication or cytopathic effects (18, 33, 36). Furthermore, injection of baculovirus into the brain did not reveal any major safety issues, as evaluated by several clinical chemistry analyses (19). The emergence of replication-competent virus breakthroughs is a major concern discouraging the in vivo use of replication-incompetent viral vectors, such as replication-deficient adenovirus. No evidence of viral replication and transcription of the baculovirus genome, however, has been detected in mammalian cells (42).

Baculovirus vectors are used for a multitude of applications, including the production of virus-like particles and viral display systems (6). Similar to retroviral vectors, the efficiency of gene delivery into mammalian cells by baculoviruses was enhanced by the incorporation of foreign envelope proteins into virions (3, 41). In this study, the VSVG-modified baculovirus delivered reporter genes efficiently not only into neural cell lines, but also into primary rat neural cells. We also constructed an RVG-modified baculovirus for use in gene transduction into neural cells, as rabies virus is known to utilize the nicotinic acetylcholine receptor (9) and the low-affinity nerve growth factor receptor P75NTR (44) for entry. The RVG-modified virus exhibited a 10- to 500-fold-higher efficiency of gene transduction into neural cell lines than the unmodified control baculovirus. The RVG and VSVG recombinants, however, gave similar infectivities on neuronal cells in vivo. This result suggests that neural cells may express similar levels of receptors for both rabies virus and VSV.

Gene delivery to osseous tissue is essential for genetic treatment of bone diseases. Osteoblast and osteoclast cells are involved in bone formation and resorption, respectively. Disorders of these cells lead to bone diseases, such as osteopetrosis, osteosclerosis, and osteoporosis (15). Adenovirus and VSVG pseudotype retrovirus vectors can transduce foreign genes into some osteocytes in vitro and in vivo (2, 14, 21). The infectivities of VSVG-modified baculoviruses are also higher than those of the recombinant baculovirus possessing excess gp64 envelope protein. Although the infectivity of the VSVG-modified baculovirus to primary mouse osteoblast and osteoclast cells was lower than those exhibited for other cell types, these cells did not demonstrate any cytopathic effects, even at a high MOI. Furthermore, luciferase expression did not increase following infection at an MOI higher than 10⁴ (data not shown), suggesting that the receptors for the VSVG-modified baculovirus on osteal cells were saturated at this point.

Baculoviruses are thought to be inactivated by serum complement in organs in direct contact with complement components (11, 35). Hüser et al. reported that the incorporation of human decay-accelerating factor into the viral envelope together with gp64 confers resistance to inactivation by serum complement, suggesting that it may be possible to circumvent complement inactivation with appropriate genetic strategies (12). The VSVG-modified baculovirus exhibited greater resistance to animal serum inactivation than the unmodified control baculovirus, similar to other pseudotype retrovirus systems (30). Barsoum et al. hypothesized that VSVG recombinant baculovirus conferred resistance to complement, imparting the ability to perform gene transduction into mouse hepatocytes following tail vein injection (3). Pieroni et al. demonstrated increased gene delivery into mouse quadriceps after direct intramuscular injection of VSVG-modified baculovirus, which partially bypasses the complement system (32). Although mouse serum had only a small effect on VSVG-modified baculovirus infectivity (Fig. 6), we could not detect either luciferase or GFP expression following injection of baculoviruses into mice by the intravenous, intraperitoneal, or intrahepatic route (data not shown). These results suggest that host factors other than the serum complement system inactivate baculoviruses in mice.

We employed the synthetic protease inhibitor FUT-175 to prevent complement activation during baculovirus infection. FUT-175, which inhibits the complement pathways, has been used clinically for more than 20 years in Japan for treatment of patients with acute pancreatitis and disseminated intravascular coagulation. No serious side effects have been associated with its use (13, 23, 31). FUT-175 specifically binds the Bb fragment of factor B, an important enzyme in the alternative complement pathway. FUT-175 is also incorporated into the active site of the intermediary C1r form, inhibiting both the alternative and classical pathways of the complement system (13, 26). This compound also protects retroviral vectors against serum inactivation (23). The infectivity of the unmodified control baculovirus could be recovered in the presence of FUT-175 in a dose-dependent manner, suggesting that FUT-175 prevents activation of complement-mediated inactivation. Although it is difficult to prevent the activation of the entire complement system in vivo with only FUT-175, due to its short half-life, in vivo clinical application of FUT-175 in combination with the VSVG-modified baculovirus system may prove highly effective.

Previous studies demonstrated that recombinant baculoviruses, including envelope-modified viruses, can transfer reporter genes into human liver segments ex vivo (35) and rat hepatocytes (12), rabbit carotid artery (1), mouse skeletal muscle (32), rodent glial cells (37), and mouse retinal cells in vivo (8). Gene transfer by baculovirus vectors in vivo has not been successful, however, in organs directly exposed to the complement system. In this study, we demonstrated efficient gene transduction of GFP into mouse brain and testes by direct injection of a GFP-encoding VSVG-modified baculovirus. For gene transfer into testes, foreign gene delivery has been mediated by previous viral vectors, such as adenovirus (4), and with nonviral vectors through both lipofection (16) and electroporation (25). Furthermore, we demonstrated that baculovirus vectors are capable of delivering foreign genes to the interstitial compartment of the adult mouse testes. Although histochemical studies of the infected testis indicated that expression of the introduced gene extended into the innermost region of the testes, the GFP-expressing cells were confined to the spermatogenic cells and Sertoli cells within the seminiferous tubules.

Although the reason underlying the lack of gene expression in sperm cells is not known, construction of a recombinant baculovirus possessing a ligand specific for sperm cells will determine the ability of baculoviral vectors for gene transduction into these cells. For gene transfer into the central nervous system, Sarkis et al. demonstrated reporter gene expression in the brains of mice and rats following direct injection with unmodified baculovirus. Using a stereotaxic apparatus as a precaution to avoid hemorrhage, they obtained results suggesting that the baculovirus was not inactivated by the complement system within the brain (37). In this study, we could detect GFP expression in the mouse cerebral cortex by fluorescence stereomicroscopy following injection of the VSVG-modified baculovirus into the brain. These results indicate that the VSVG-modified baculovirus is a promising vector for gene delivery into the brain.

Gronowski et al. demonstrated that baculovirus is able to stimulate interferon production from both human and mouse cells in vitro. Pretreatment with baculovirus also confers protection against lethal challenge of mice with encephalomyocarditis virus (7). Inhibition of activation by either antibodies against gp64 or UV inactivation suggested that virus-dependent processing, in addition to the interaction of gp64 with cell surface molecules, is required for the reaction. Induction of the innate immune response following administration of baculovirus is important for future applications of baculovirus vectors in vivo, not only for gene therapy but also for vaccine trials.

In summary, we have investigated the feasibility of gene transfer with recombinant baculoviruses in vitro and in vivo. Further studies examining the transcription of baculovirus genes in mammalian cells will be required for certification of safety for in vivo use. In addition, the development of a vector capable of targeting specific organs is needed for future in vivo applications of the baculovirus vector in the treatment of acquired or inherited diseases in humans.