-Virus DNA

-Basic DNA

binding protein

Figure 2 Schematic illustration of the budded form of AcMNPV. (Modified from Ref. 8.) See the color insert for a color version of this figure.

tion of so-called peplomer structure at one end of the virion (Fig. 2). The nucleocapsid of OV and BV are similar, but their envelopes differ in the composition (Fig. 2). The differences in lipid and protein composition of BV and OV envelopes reflect different origin and functions during the virus life cycle (46). Basic DNA-binding protein, p6.9, is shown to be present in the virions of several baculoviruses, including AcMNPV. It may become phosphorylated upon entry into the insect cells, which may result in unpackaging of the viral DNA (47,48).

AcMNPV enters cells by adsorptive endocytosis (34). The role of the major envelope glycoprotein, gp64, is essential for viral entry. It mediates pH-dependent escape of the AcMNPV capsid from the endosomes (49). gp64 is necessary and sufficient for virus preparation because it is needed for efficient virion budding from the insect cells (50). The cellular surface molecules for AcMNPV attachment and entry are not known, but the large range of target cells suggests that the molecules are common cell surface components such as integrins, phos-pholipids, or heparan sulfate proteoglycans (51-53).

The range of transfer plasmids and parent viruses available for the AcMNPV-based system, and the characteristics (growth, expression level) of the cell lines supporting AcMNPV, have made it a common choice for eukaryotic protein production. The most commonly used cell lines for

AcMNPV are Sf-9 and Sf-21AE. They both originate from IPLB-Sf-21 cells, which were derived from Spodopterafrugi-perda pupal ovarian tissue (54). A new cell line derived from Trichoplusia ni egg cell homogenates (BTI-TN-5B1-4 = High Five®), however, has become a popular alternative to Sf cell lines due to the fact that these cells have been shown to produce up to 28-fold more secreted proteins than any other insect cell lines (55-57). These cell lines grow well in both monolayer and suspension cultures, and a serum-free culture medium can be used if desired.

To enable expression of a recombinant protein in insect cells, the gene for the desired protein is usually placed under a strong polyhedrin promoter (polh) of AcMNPV (4). For other cells a promoter active in the target cells, such as CMV (cytomegalovirus), RSV, or CAG (chicken p-actin promoter) must be used because polh is inactive in these cells. The polh is normally responsible for the synthesis of polyhedrin, which can constitute up to 50% of the total protein of the infected cell. Fortunately, polyhedrin is not essential to virus replication or infection in the cell culture (58). However, the use of the polh promoter may be restricted in some cases by the fact that it is activated very late in the infection cycle at a point where the host cell machinery for posttranslational modifications is no longer functional. Problems with the polh promoter have been encountered, especially with proteins whose biological activity depends on proper glycosylation (59). In such cases, the use of an alternative strong viral promoter that initiates transcription earlier in the infection cycle (while the host posttranslational modification pathways are still functional) could be useful. Indeed, Chazenbalk and Rapoport (60) were able to produce a more glycosylated and functional form of extracellular domain of the human thyrotropin receptor under a late basic protein promoter (late). Other promoters, which will be activated earlier than polh, include promoters for the p10 gene (p10), the major capsid protein gene (vp39), the basic 6.9-kD protein gene (cor), and the viral ie1 gene (ie1). These promoters are available in a variety of baculovirus plasmids (4,61-63).

B. Preparation of Recombinant Baculoviruses

Due to the large size of baculovirus genome (80 kbp-200 kbp) a homologous recombination procedure was originally adopted to insert foreign genes into baculovirus genome, instead of conventional plasmid cloning techniques (64). In practice, the target gene is subcloned into a transfer vector containing a suitable promoter, flanked by baculovirus DNA derived from a nonessential locus, such as the polyhedrin gene of AcMNPV. The viral DNA and transfer plasmid are then cotransfected into insect cells or yeast cells where the recombination events take place. Typically, 0.1 % to 1 % of the resulting progeny is recombinant, which complicates their identification. Because the target gene is inserted into the polyhedrin locus, altered plaque morphology of the recombinant viruses can be used for the identification of recombinant viruses. The cells, in which the nuclei do not contain PIBs, contain a recombinant virus. However, the detection of the desired PIB-minus plaque phenotype against the background of greater than 99% wild-type parental viruses is difficult. Viral identification may be facilitated by the introduction of a lacZ cassette (p-galac-tosidase) along with the foreign gene, which enables the detection of the recombinant viruses according to a blue color (4). Drug selection may also be used (65).

Recently, several techniques have been developed to further facilitate and speed up the construction of recombinant baculoviruses (63,66). By using a unique restriction site (Bsu361) at the polyhedrin locus, Kitts and coworkers (67) were able to linearize the double-stranded circular genome of AcMNPV. The linearization reduced the background of wildtype viruses and, as a result, 10% to 25% of the progeny viruses were recombinant. To obtain an even higher proportion of recombinants (85%-99%), Kitts and Possee (68) further modified the AcMNPV genome to enable Bsu361 digestion to also remove an essential gene (ORF 1629) from the AcMNPV genome. Infective viruses will only be reconstituted by recombination with the transfer vector carrying the gene of interest, whereby an intact ORF 1629 will be restored to the genome. The system also enables a blue/white color selection of the recombinant viruses. However, it suffers from the need of time-consuming plaque assays to purify the recombi-

nant virus. A version of this system has been developed in which the target gene is amplified with specific primers using the polymerase chain reaction. This enables the ligase-free coupling of the linearized transfer vector and amplified gene in the mixture, which then can be used directly to transfect insect cells with cut viral DNA. The avoidance of cloning steps in Escherichia coli speeds up the construction of the recombinant virus (69). To ease the manipulation of ORF 1629-deleted baculovirus genome, Je et al. (70) recently described a bacmid form of this virus that can be maintained in E. coli.

Construction of the recombinant baculovirus genomes by traditional cloning techniques was reported by Ernst and coworkers (71). They introduced the I-Scel meganuclease site into the AcMNPV genome by homologous recombination. The new virus genome, called Ac-omega, can be cut with I-Scel meganuclease, and the target gene bearing compatible ends can be ligated straight to the linearized Ac-omega DNA under a polyhedrin promoter. This method is simple and less time consuming than conventional homologous recombination, but owing to the normal background problems encountered with traditional cloning techniques, the need of plaque purification cannot be entirely avoided. In addition, handling of the large AcMNPV genome (134 kbp) makes the method inconvenient. Another in vitro systembased on the preparation of recombinant baculoviruses by versatile Cre-loxP recombi-nase has also been described (72). The advantage of this system is a possibility to recover cloned inserts from baculovirus genome into Cre/loxP compatible plasmids. However, only up to 50% of the viral progeny are recombinants.

To avoid laborious and time-consuming plaque purification process, the genetic material can be introduced into the baculovirus genome outside the insect cells. Patel and cowork-ers (73) reported a novel method of propagating the viral genome by homologous recombination in the yeast Saccharo-myces cerevisae, where the appropriate recombinants can be more easily selected. Viruses are then obtained by transfecting insect cells. This method is rapid (pure recombinant virus within 10-12 days) and efficient, and it ensures that there is no parental virus background. It also eliminates the need for time-consuming plaque assays, and multiple recombinants can be readily isolated. The major disadvantages of this system are the need of experience in yeast culturing and the incompatibility of traditional transfer vectors with the system. An even faster approach for generating recombinant baculoviruses uses site-specific transposition with Tn7 to insert foreign genes into bacmid DNA (virus genome) propagated in E. coli cells. The E. coli clones containing recombinant bacmids are selected by color (p-galactosidase), and the DNA purified from a single white colony is used to transfect insect cells (74). The system has the same advantages as the yeast system but is faster (pure recombinant virus within 7-10 days) and easier to work with for those not familiar with the yeast cells. The poor selection features of the original system have been overcome by a modified donor vector (pBV boost) and an improved selection scheme of the baculovirus bacmids in E. coli with a mutated lavansucrase gene from Bacillus amyloliquela-ciens. The new selection schema bypasses the disadvantages associated with the original transposition-based generation of baculovirus genomes in E. coli while retaining the simple, rapid, and convenient virus production (75).

C. Baculoviruses and Nontarget Vertebrate Cells

Baculoviruses have been studied in the past for their ability to infect nontarget cells for safety considerations with regard to their use as biological pesticides. Volkman and Goldsmith demonstrated that baculoviruses were able to enter certain cell lines derived from vertebrate species (12). Thirty-five nontarget host cell lines, 23 of human and 12 of nonhuman vertebrate origin, were exposed to AcMNPV. However, no evidence of viral gene expression was obtained. This study was in accordance with the earlier reports showing uptake of AcMNPV by several vertebrate cell lines with no evidence of viral replication (11,76). Carbonell et al. (13) were the first to show that by constructing a recombinant baculovirus bearing a suitable promoter, marker gene expression in nontarget cells can be detected. Because a low level of marker gene expression was observed in the studied cells, it was claimed that marker protein was carried into the cells with the virus and not actively expressed in the cells (15,16). Several other publications, however, have confirmed Carbonell and coworkers' (13) initial findings and demonstrated that AcMNPVs containing mammalian expression cassettes can enter mammalian cells and express reporter genes under the control of strong viral promoters. Hofmann et al. (14,77) demonstrated that recombinant AcMNPV containing the luciferase gene under a CMV promoter can efficiently infect human hepatocytes (Huh7 and HepG2), as well as primary hepatocytes of human and rabbit origin. Boyce et al. (78) confirmed these results and showed that a virus carrying lacZ reporter gene under the control of RSV promoter led to a high-level expression of the marker gene in human hepatocellular carcinoma line HepG2, as well as in primary rat hepatocytes. More recently, primary rat and human hepatic stellate cells (HSCs) were shown to be highly susceptible to baculovirus-mediated gene delivery (79). These findings favor the potential of baculovirus vectors for liver-directed gene therapy.

Further studies have indicated that a high-level expression of marker gene can be achieved not only in hepatic, but also in other cell lines (77,80-84). Shoji et al. (80) transduced a panel of mammalian cells with a baculovirus vector carrying a marker gene in comparison with a replication-defective ade-novirus vector. High-level luciferase activity was detected in this study not only in human hepatocytes, but also in other cell lines such as monkey kidney cells (COS7), porcine kidney cells (CPK), and human cervix carcinoma cells (HeLa). Furthermore, the same level of marker gene expression was observed in these cells by both viruses, but as an advantage for the baculovirus, much lower cytotoxicity was associated with it than adenovirus at a high multiplicity of infection (moi). The use of CAG promoter in this study resulted in a 10-fold higher level of luciferace gene activity than in the previous study with a CMV promoter (14). This indicates the impor tance of the promoter and the expression cassette per se for successful transduction of target cells not only by baculovi-ruses, but also by any other gene delivery vector (20,85-88). High-level expression of T7 RNA polymerase was also directed by CAG in HepG2 and CPK cells, and lower expression was observed in some other cell lines (81). Condreay et al. (84) demonstrated that recombinant baculoviruses containing green fluorescent protein (GFP) gene under the CMV promoter can transduce a wide range of mammalian cell types originating from different tissues. Cell lines of hepatic origin were transduced efficiently, as described earlier, but notable gene expression was also detected in cell lines derived from kidney tissue (Cos-7, BHK, CV-1, 293) and other nonhepatic hepatic cell lines such as keratinocytes (W12, primary human keratinocytes), bone marrow fibroblasts, and osteosarcoma cells (MG-63). The lowest efficiency of transduction and level of GFP expression was seen in cell lines of hematopoietic origin, such as THP-1, U937, K562, Raw264,7 and P388D1. Efficient gene transfer has also been obtained in primary mouse and human pancreatic islet cells (89). Recombinant baculovirus expressed the glycoprotein gB of pseudorabies virus (PrV) under the CAG promoter in various mammalian cell lines and produced specific antibodies in mice against PrV (90). High levels of expression of PrV gB were observed in many porcine kidney cell lines such as CPK, SK-H, and CPK-NS, and in hamster kidney cells (BHK-21). The potential of baculovirus for therapeutic applications was tested by a virus containing a p53 tumor suppressor gene under the control of the CMV promoter together with an anticancer drug, adriamycin (91). Greater than 95% of Saos-2 cells were killed by the combination of the recombinant virus (moi of 100) and adriamycin (35 ng/mL), suggesting that the combination treatment greatly enhanced apoptosis of the tumor cells. Table 2 summarizes some common cell lines found to be good and poor targets for baculovirus transduction by several independent authors.

Baculoviruses have been pseudotyped by a vesicular stomatitis virus G transmembrane glycoprotein (VSV-G) to further broaden the host range. VSV-G has an extremely broad host range because its entry into the cells seems to not be dependent on the presence of any specific receptor but a phospholipid component of the plasma membrane (92). VSV-G mediates pH-dependent membrane fusion in endosomes (93), and has earlier been used to pseudotype and stabilize other enveloped viruses such as murine retroviruses and lentiviruses (94-96). VSV-G has increased the efficiency of baculovirus transduction in a range of mammalian cells that have been studied (82,97). In these studies, VSV-G was cloned under the control of the polh promoter so that it was expressed in infected insect cells but not in mammalian cells. This circumvents the toxicity of VSV-G in mammalian cells. Indeed, the inactivity of baculoviral promoters in mammalian cells is a very useful property of baculovirus vectors. The level of marker gene expression in HepG2 cells treated with a pseu-dotyped virus was 10-fold higher than in the same cells treated with a wild-type virus. Pieroni et al. (98) showed that VSV-G enhances the transduction efficacy of pseudotyped baculovirus also in vivo. VSV-G can complement the gp64 and there-

Table 2 Common Cell Lines Found to be Good, as Well as Poor Targets for Baculovirus Transduction In Vitro by More Than One Independent Author

Good targets

Poor targets

Cell lines


Cell lines


Liver cells Huh7

HepG2 Primary hepatocytes Kidney cells 293 BHK COS-7 CPK 12 Others A594 CHO Hela

97,98,107,112,119,178) (14,52,77-82,91,97,112,178) (14,77,78,115)

(51,78,84,98,106,107,112) (52,84,90) (20,78,80,112,178) (52,80,81,90,178)

(14,77,78,82,97) (14,52,82,84,139) (14,20,28,52,78,80-84,97,98, 112-114,178)

Blood cells HL-60

fore allows productive infection, replication, and propagation of the gp64 deleted baculovirus in Sf9 insect cells (99). However, the virus propagation appeared to be inefficient and delayed as compared with the wild-type virus, but virions were similar in morphology to the wild-type viruses. Interestingly, the use of mouse hepatitis virus S protein (MHVS) pseu-dotyped baculoviruses can also result in a 100- to 500-fold higher marker gene expression than the wild-type virus in many mammalian cells at a low moi (52).

Agents that inhibit histone deacetylation, such as trichos-tatin A (TSA) and sodium butyrate, can have a significant effect on the gene expression in cells (100,101). The effect is supposed to be mainly due to the more exposed chromatin structure (102). These agents also enhance adenovirus- and retrovirus-mediated transgene expression in vivo and in vitro (5,103,104). Thus, it is not surprising that sodium butyrate and TSA also increase the expression of marker genes significantly in the baculovirus-treated cells in vitro (5,84).

D. Baculovirus Hybrid Vectors

The extraordinary capacity of AcMNPV to carry foreign DNA (no known limit for foreign DNA, but at least 50 kbp tolerated) allows construction of recombinant viruses bearing large expression cassettes (4,105,106). This is a valuable property for many purposes, including hybrid vector construction, as shown by recent reports. Palombo and coworkers (107) constructed a baculovirus-adeno-associated virus (Bac-AAV) hybrid vector to prolong the transient nature of baculovirus-me-diated transgene expression. The idea was to use the natural integration capacity of AAV to carry the transgene cassette into a defined region (chromosome 19) of the host cell genome

(107). p-galactosidase and hygromycin resistance gene excision from the baculovirus backbone vector and a subsequent integration into the genome of 293 cells was shown to occur with significant frequency. However, integration into the desired region occurred only in a fraction of the clones, and nonspecific integration and multiple insertions of the marker gene were detected.

A system for the production of gutless adenovirus vectors (FD-AdVs) that does not require helper adenoviruses was recently described (106). The helper virus was replaced by the baculovirus-adenovirus hybrid vector (Bac/Ad) containing a Cre recombinase-excisable copy of the packaging-deficient adenovirus genome. 293-Cre cells were transfected with the FD-AdV plasmid containing a transgene cassette, packaging signals, and 2 copies of the inverted terminal repeats, and were followed by transduction by the Bac/Ad. High titer FD-AdV virus preparations (108 pfu/mL) were attained. However, the system has to be further improved to avoid generation of replication-competent viruses during a large-scale production.

Yap et al. (81) constructed a recombinant baculovirus carrying a cDNA of the bacteriphage T7 RNA polymerase under the control of the CAG promoter. High-level expression of this enzyme in various mammalian cell lines was observed after the baculovirus transduction. A plasmid bearing the entire poliovirus genome under the T7 promoter yielded a high-titer of infectious poliovirus in the HeLa cells after prior trans-duction with the baculovirus.

To efficiently propagate and study hepatitis B virus (HBV) and hepatitis C virus (HCV) in cultured cells, recombinant baculoviruses carrying the HBV (108-111) and HCV (105) cDNAs under mammalian promoters have been prepared. The control of the gene expression in both the HCV minigenome and the full-length HCV construct was also investigated (112). In addition, Tet-off and ecdysome/ponasterone-inducible (pon) systems were compared to control the gene expression. The tetracycline-controlled system gave a low basal activity and was highly inducible in almost 100% of HepG2 cells. Hepatitis-baculovirus hybrid vectors represent a simple and highly flexible system for studying the effects of antivirals and/or cytokines on HBV and HCV production, and for the understanding of their replication and pathogenesis at the molecular level.

To investigate biology of the human cytomegalovirus (HCMV), genes encoding immediate early proteins of the HCMV were cloned into baculoviruses under the CAG promoter (113). These viruses provided a new strategy for efficient isolation of the HCMV viruses with mutations in essential genes.

Finally, a recombinant baculovirus expressing Ebola virus nucleoprotein under the CMV promoter was used to transduce HeLa cells, which were subsequently used as an antigen to detect Ebola virus IgG antibodies from serum samples (114).

E. Entry Mechanisms into the Mammalian Cells

Cell surface receptors responsible for baculovirus attachment and entry are unknown, although the dogma that baculovirus enters the insect cells by an adsorbtive endocytosis has been confirmed by several studies (34). In contrast, the mechanisms involved in the baculovirus entry into mammalian cells are still poorly known. Indeed, several contradictory reports have been published concerning essential cellular motifs responsible for baculovirus uptake into mammalian cells. All these studies, however, agree on that the uptake occurs via an endo-cytotic route. Early reports suggested that the degree of virus uptake was variable, depending on the cell, incubation time, temperature, and viral phenotype (12). The effect of the titer and virus competition studies in the cell culture have suggested that gene transfer into hepatocytes might be due to the presence of a specific receptor (14). A candidate receptor was postulated to be an asialoglycoprotein receptor. Inhibition tests with choloroquine also indicated that endosomal maturation was essential for virus transport into the nucleus in agreement with the mechanism of virus uptake in the insect cells (Fig. 1) (14,78). Contrary to a supposed specific receptor theory, baculoviruses were reported to enter several cell lines by absorptive endocytosis, which does not require any interaction with a high-affinity receptor but rather interaction with a heterogeneous cell surface motif(s) (51). Electrostatic charges were shown to be important. Neutralization of negatively charged epitopes at the cell membrane appeared to be critical for baculovirus-cell interactions and the subsequent entry. Heparan sulfate proteoglycan seemed to be an important docking motif. Enzymatic removal of heparan sulfate groups from the cell surface caused a significant reduction in transduction (51).

The mode of baculovirus entry into the target mammalian cells has been further studied by constructing recombinant baculoviruses expressing gp64, VSV-G, MHVS, or GFP to compare susceptibility of various cell lines to these recombinant baculoviruses (52). Increased amounts of gp64 or foreign envelope protein (VSV-G, MHVS) on the virus surface (envelope) caused higher expression than the control virus in various mammalian cell lines. Furthermore, this study indicated that phospholipids, such as phosphatidic acid or phosphatidyl-inositol, on the cell surface played an important role in the transduction of mammalian cells by baculovirus, whereas hep-arin and heparan sulphate did not.

The basolateral surface has shown to be important in bacu-lovirus-mediated transduction when the hepatocytes acquire intercellular junctions and form islands in the cell culture. Disruption of the cell-cell junctions masking the baculovirus motifs by a calcium chelator, ethylene glycol-bis(p-amino-ethylether)-N,N,N',N'-tetraacetic acid (EGTA), significantly improved the transduction efficiency (115). Baculovirus entry into cells or tissues with tight cellular junctions may thus require transient breakage of intercellular linkages to allow viral contact with the basolateral surface.

Van Loo et al. (28) showed that baculoviruses can also transduce nondividing cells and that the mechanism of the viral capsid transport into the nucleus before uncoating is apparently identical to that of insect cells (Fig. 1) (116). Pig kidney cells (Pk1) were arrested in S phase with aphidicolin, a reversible blocker of DNA polymerase, to show that the arrested cells can be transduced similarly as untreated mitotic cells. An electron microscopy study of the viral uptake showed that baculoviruses enter mammalian cells via endocytosis and are released into the cytoplasm by an acid-induced fusion of the viral envelope with the endosomal membrane in accordance to the entry into the insect cells. Cytochalasin D, which causes reversible depolymerization of actin filaments, inhibited marker gene expression, suggesting that actin filaments are important for viral capsid transport into the nucleus. Similar results were recently obtained also in HepG2 cells and it was further suggested that the transduction block in the nonsusceptible vertebrate cells lies in the cytoplasmic transport or nuclear entry of the virus capsid (117).

F. Gene Delivery In Vivo

Current evidence suggests that baculoviruses also provide an effective tool for in vivo gene delivery. The first attempts of in vivo gene transfer with baculovirus were performed into the liver parenchyma of rats and mice (17). Several attempts were also undertaken to inject baculoviruses directly into systemic and intraportal circulation (17). These experiments resulted in undetectable transgene expression, suggesting that the virus was somehow inactivated by serum components. This led to several studies in immune-compromised animals. When Hofmann et al. (77) injected a p-galactosidase expressing baculovirus directly into the liver parenchyma of C5-defi-cient immunocompromised mice, a few transduced hepato-

cytes were detected around the injection site. They also injected recombinant baculoviruses into the Huh7-derived human hepatocarcinomas generated in nude mice (T cell deficient) and got low gene transfer efficiency. A systemic bacu-lovirus gene vector delivery into complement deficient Neu-ro2a tumor-bearing A/J mice resulted in transgene expression primarily in liver, spleen, and kidney, but significant expression was also found in the tumor (118). Injection of decay acceleration factor (DAF)-modified complement-resistant ba-culovirus vector into the liver parenchyma of complement-sufficient neonatal Wistar rats caused an enhanced expression of the marker gene suggesting that generation of complement-resistant vectors could improve the gene transfer efficiency in vivo (119). However, direct intrahepatic injection of the complement resistant AcMNPV-DAF-Pgal vector in adult rats resulted only in single positive cells distant from the injection site.

Delivery methods that allow gene transfer in the absence of serum, or to the sites where viruses are not exposed to the complement, have led to more successful experiments in immune-competent animals. BALB/c mice, nude mice, and Sprague-Dawley rats were injected with recombinant bacu-lovirus directly into the striatum of brain (18,20). Marker gene expression was detected in the striatum, the corpus callosum, and the ependymal layer, indicating the ability of baculovi-ruses to transduce neural cells in vivo. Transduced cells were identified mainly as astrocytes with only a few positive neurones. No difference was detected between the 3 species or between the cobra venom factor CVF (an inhibitor of the complement system) treated and untreated groups, suggesting immune-privileged nature of the brain. Transduction efficiency, tropism, and biodistribution of the baculoviruses after local delivery into the brain have been studied in comparison to adenoviruses also in BDIX and Wistar rats (18). In this study, baculoviruses were found to transduce cuboid epithelium of the choroid plexus cells very efficiently (76% ┬▒ 14). A clear difference was observed with the adenovirus vector when injected into the corpus callosum; adenoviruses did not transduce the choroid plexus cells, whereas ventricular epen-dymal lining and cells in the corpus callosum were transduced with a high efficacy. Injection into the striatum resulted in an effective transduction near the injection site and in the corpus callosum. Both viruses lead to transgene expression in endo-thelial cells of brain microvessels throughout the forebrain (18).

In a different approach, carotid arteries of New Zealand White rabbits were successfully transduced by recombinant baculoviruses using a collar device. This system allowed gene delivery with minimal exposure to complement (5). Transient expression in the adventitial cells was observed with an efficacy and duration comparable to adenoviruses. Recombinant baculoviruses have also been tested for direct administration into a mouse eye by subretinal injections. A strong expression of the marker gene in retinal pigment epithelial cells was reported by this study (19). Intravitreal injection of the virus resulted in the marker gene expression in the corneal endothe-

lium, lens, retinal pigment epithelial cells, and retina. The ocular tissue contains areas where antigens are not subjected to the complement pathway and therefore makes it a good target for baculovirus-mediated gene therapy.

Direct injection of recombinant baculoviruses into the quadriceps femoris muscle of BALB/c and C57BL6 mice resulted in a transient expression of p-galactosidase. Expression levels were 5- to 10-fold higher when VSV-G pseudotyped baculoviruses were used (98). The authors also used C5-defi-cient mice where a higher and more sustained gene expression (up to 178 days) was observed than in BALB/c and C57BL6 mice, suggesting a different gene transfer efficiency between the different mouse strains and the importance of the complement activity.

Baculoviruses have also shown be potentially useful as vaccines. Induction of anti-PrV gB antibody was observed in mice that were inoculated intranasally or intramuscularly. Higher levels of antibodies against PrV gB were detected in serum after intramuscular rather than intranasal inoculation (90). In a related study, Lindley et al. (120) reported a novel use of the baculovirus gp64-display system (121) as a rapid method to produce monoclonal antibodies directed against gp64-fusion proteins. This method might also be useful for vaccination with desired antigens, as suggested by Tami et al. (122). Vaccination with either recombinant baculoviruses or infected cells may prove to be a safe alternative for vaccination approaches (123,124). Table 3 summarizes the current data of the baculovirus-mediated gene transfer studies in vivo.

G. Safety

Since the 1950s, extensive safety testing of baculoviruses has been conducted. These studies have revealed that nuclear polyhedrosis viruses are harmless to and unable to replicate in microorganisms, noninsect invertebrate cell lines, vertebrate cells, vertebrates, plants, and nonarthropod invertebrates. These trials have included long-term carcinogenicity and tera-togenicity tests, tests in primates, and tests in humans (10). Ten different mammalian species have been studied, including rats, mice, dogs, guinea pigs, monkeys, and humans. In these tests baculoviruses were administered by a variety of routes, including orally, intravenous injection, intracerebral injection, intramuscular injection, and topically. No toxicity, allergic responses, or pathogenicity associated with the baculoviruses were detected (125). The safety of baculoviruses is also underscored by the fact that we are exposed daily to baculovirus particles present in large numbers in our environment and food. Yet no diseases have been linked to baculoviruses (126).

The safety of baculoviruses is secured at several levels of restrictions in baculovirus infectivity (127). The polyhedrin matrix of OV is essential to horizontal transfer of the virus by protecting it from the environment. The alkaline midgut of insects facilitates the dissolution of OV matrix, leading to primary infection of the larva (Fig. 1). Organisms such as birds and mammals that lack such alkaline conditions in their digestive tract or other potential points of entry, such as respiratory tract, are not infected by OV (128-130). Tissue or cell-

type specificity may create the second level of restriction, although the baculovirus host range does not appear to be limited at the point of entry into the target cells (12,52). This is particularly true for AcMNPV. The subsequent steps following the virus entry by adsorbtive endocytosis may be much more important (28,34,117). Indeed, it was recently suggested that the block in the entry of the AcMNPV into the unsusceptible vertebrate cells lies in the defective transport or entry of the nucleocapsid into the nucleus (117). Although escape from the endosomes certainly creates some barriers, it was found that nucleocapsids seemed to enter into the cytoplasm even in the cells in which no marker gene expression was detected. This is in agreement with the known ability of gp64 to mediate pH-dependent membrane fusion in endosomes (49). Finally, if the nuclecapsid reaches the nucleus, strictly guided molecular mechanisms that cover the expression of the baculovirus genome remain in place. Indeed, baculoviruses propagate only in restricted insect cells and are inherently unable to replicate or express their genes in nontarget mammalian cells (6,131,132). This is a great advantage when baculoviruses are used for gene therapy because risks related to the rise of replication-competent viruses during the virus production can be avoided. Replication-competent viruses are the major concern with most of the current main stream gene delivery vectors based on natural human pathogens (133).

The large size of the baculovirus genome raises a theoretical concern about the possibility of homologous recombination of its genome (or parts of it) into the target cell genome. In the worst case, this might lead to malignancy of the target cell as reported recently in the case of naturally integrating retrovirus and AAV vectors (134-138). Integration of AcMNPV genomic fragments into the mammalian genome has shown to take place in vitro in the presence of selection pressure (84,139). In these studies, Chinese hamster ovary (CHO) cells, which were stably transduced by selection pressure with recombinant baculoviruses, expressed GFP at least 25 passages (84). Analysis of the baculovirus-derived DNA indicated that at least 12 kbp of DNA derived from the viral vector had stably integrated into the transduced CHO cells (84). Integration into the CHO genome occurred as small fragments (5-18 kb) via illegitimate recombination (139). Two of the clonal cell lines maintained starting levels of the GFP expression over a 5-month period with and without selection. The 2 remaining clones, however, showed a loss of the marker gene expression. Baculoviruses can thus be used for the preparation of stable cell lines. However, no evidence of Integration of AcMNPV genomic fragments into the target cell genome has been found without concomitant antibiotic selection pressure (140,141).

One of the biggest challenges in gene therapy is the immune response of the host. The host defense mechanisms function both at the cellular level by generating cytotoxic T cells and at the humoral level by generating antibodies against foreign antigens. Cellular immunity eliminates the transduced cells, whereas humoral immunity protects against the repeated administration of the vector (98,142,143). The host may recognize not only the vector and the transgene product, but also the foreign DNA, which makes the vector design even more challenging (144). However, the eye, brain, and reproductive organs possess immune-privileged regions (145) and therefore may provide good targets for the baculovirus-mediated gene therapy (19,20).

As discussed earlier, immune responses against baculovi-ruses were suggested as a reason for no detectable transgene expression when mice and rats were treated with systemic, intraportal, or direct injections into the liver parenchyma (17). Therefore, the influence of untreated and heat-inactivated serums from different species was tested for baculovirus trans-duction efficiency. The results indicated that the classical complement (C) cascade inactivates baculoviruses rapidly and the extent of inactivation varies from one species to another (17,146). Incubation of baculoviruses in the serum of either C3- or C4-deficient guinea pigs did not cause any significant neutralization of the baculovirus, suggesting that C3 and C4 components are essential (17,146). In line with these findings, the direct injection of recombinant baculoviruses into C5-defi-cient mice resulted in a higher and longer-lasting expression of the marker gene than that in mice, which were not complement deficient (98). In ex vivo experiments that excluded the C system, detectable levels of the marker gene expression were found in human liver segments perfused by the baculovirus vectors (17,77). More detailed investigations have demonstrated that the classical pathway of the C system and assembly of the very late C components are essential for the inactivation of the baculovirus in human serum, indicating the presence of IgM or IgG antibodies against baculoviruses (146). These antibodies are most probably part of the innate self/nonself pattern recognition immune system detecting antigens without a known history of immunization (147,148).

A couple of efforts have been taken to protect baculovirus from inactivation by the complement system (149,150). The treatment of human serum with functional blocking agents (CVF, anti-C5 monoclonal antibodies) against the components of the C-cascade increased vector survival significantly in a dose-dependent manner in vitro (146). Soluble complement receptor type 1, a potent inhibitor of both the classical and alternative C pathways, increased baculovirus survival in human serum in a dose-dependent manner (151). A complete baculovirus survival was achieved by the soluble complement receptor at concentration of 100 ^g/mL of serum (151). As mentioned earlier, incorporation of the complement-regulatory protein DAF into the viral envelope was also shown to improve gene transfer efficiency in neonatal Wistar rats in vivo (119).

Baculoviruses have been shown to stimulate antiviral activity in mammalian cells by promoting cytokine production. Baculovirus exposure resulted in the activation of TNF-a, IL-1a, and IL-1p expression in the primary hepatocyte cultures and production of interferons (IFNs) in some mammalian cells (152,153). Kupffer cells were present in the primary hepatocyte cultures and were most probably responsible for the cyto-kine production (152). The IFN-stimulating activity of the baculoviruses required live virus and was not due to the presence of viral RNA, DNA, or bacterial endotoxin (153). The detailed mechanism for the observed baculovirus-mediated interleukin induction remains to be studied, but the data suggest a unique process involved in the baculovirus-mediated stimulation of the mammalian IFN. In accordance to these studies, baculovirus-mediatedperiadventitial gene transfer was found to result in mild immune responses (5). Interestingly, however, only a modest microglia response was seen in ratbrain after the baculovirus transduction whereas the adenovirus gene transfer led to a strong microglia response (18).

Despite the profound existing knowledge about the bacu-lovirus safety, further studies are still needed before clinical trials with baculovirus vectors can be launched. Particularly the biodistribution of the virus in vivo is still largely unknown and the immune responses in vivo must be studied more carefully. Further investigations are also required to determine whether some of the viral genes are expressed in mammalian cells. The current data strongly suggest that the AvMNPV genome is strictly silent (6,131,132), but some immediate early genes whose promoters are recognized by host RNA polymerases may well prove to be active (154).

0 0

Post a comment