Baculoviruses have recently been shown to be effective gene transfer vectors in mammalian cells. However, very little information is available about their target cell tropism in the central nervous system. We studied transduction efficiency, tropism and biodistribution of baculoviruses after local delivery to rat brain and compared their properties to adenoviruses. It was found that baculoviruses specifically transduced cuboid epithelium of the choroid plexus in ventricles and that the transduction efficiency was as high as 76±14%, whereas adenoviruses showed preference to corpus callosum glial cells and ventricular ependymal lining. Only a modest microglia response was seen after the baculovirus transduction whereas the adenovirus gene transfer led to a strong microglia response. Sensitive nested RT-PCR revealed transgene expression in the hindbrain and in ectopic organs including spleen, heart and lung, which indicates that some escape of both vectors occurs to ectopic organs after local gene transfer to the brain. We conclude that both baculovirus and adenovirus vectors can be used for local intracerebral gene therapy. The knowledge of the cell type specificity of the vectors may offer a possibility to achieve targeted gene delivery to distinct brain areas. Baculoviruses seem to be especially useful for the targeting of choroid plexus cells.
Gene delivery to the central nervous system is crucial for the development of gene therapy for neurological diseases. Baculoviruses have been shown to transduce mammalian cells in vivo, eg in carotid artery, liver, skeletal muscle and murine brain.1234 However, the tropism of baculoviruses in brain has remained unclear. Knowledge about the cell-type-dependent susceptibility for transduction can be utilized for targeting of transgene into certain cell types.
Autographa californica multiple nuclear polyhedrosis virus (AcMNPV) is a baculovirus, which transduces both dividing and nondividing mammalian cells. The virus enters into their natural hosts, ie insect cells via receptor-mediated endocytosis, but a receptor for the baculoviruses has not yet been identified.5,6 Recent studies have suggested that baculoviruses are potentially useful gene therapy vectors, since they can easily transduce human cells.3,7 In addition, baculoviruses are nonreplicative in mammalian cells, nonpathogenic in humans and only low levels of systemic anti-baculovirus antibodies have been detected.8910
In this study, transduction efficiency, tropism and biodistribution of the AcMNPV baculovirus vector was compared to the adenovirus vector using a local intracerebral delivery of the lacZ gene. Distributions of adenoviruses in cerebrospinal fluid and in cerebellum have been previously analyzed in animal models, such as nonhuman primates and rats.111213 However, the distribution and transgene expression in brain and ectopic organs after baculovirus-mediated local gene delivery have not been evaluated. It was found that the lacZ transgene delivered using the baculovirus vector was efficiently expressed in choroid plexus cells and the expression level was comparable to the adenovirus-mediated gene delivery. Our results suggest that baculovirus is a promising tool for gene therapy in the central nervous system and that the natural tropism may have an important impact on the targeted delivery of genes in the brain tissue.
Baculoviruses and adenoviruses showed different in vivo tropism in rat brain
To analyze the gene transfer efficiency of baculovirus and adenovirus vectors, a total of 3×108 pfu viruses were injected into the corpus callosum of adult rats with the stereotaxic apparatus. The expression of transgene was analyzed 5, 10, 14 and 21 days after the gene transfer with X-gal staining and RT-PCR. Representative images of the transgene expression in the forebrain are shown in Figure 1 for baculoviruses and in Figure 2 for adenoviruses. Both viruses lead to transgene expression in endothelial cells of brain microvessels throughout the forebrain (Figures 1a and 2a). CD31 staining of serial sections showed positive cells in the same areas and with similar morphology, suggesting the transduction of endothelial cells (Figure 1b). Baculoviruses showed a strong preference for choroid plexus cuboidal epithelial cells (Figure 1c–g) whereas adenoviruses did not transduce these cells (Figure 2b). In the first part of the experiments the right corpus callosum just above the frontal horn of the lateral ventricle was chosen for the stereotaxic target point. To analyze how the injection site affects transduced cell types, the baculoviruses were also injected into the white matter in close proximity to the cortex (hindlimb area of the cortex) and into the striatum (caudate putamen) as described in the Methods section. As a result, lacZ-marker gene was mostly found in endothelial cells of the microvessels and in choroid plexus cells in the ventricle (Figure 1e, 2 mm from the injection site). Some transgene expression was also seen in subarachnoidal space (data not shown). Injection into the striatum resulted only in a few lacZ-positive cells at the site of injection (Figure 1j), where a strong astroglia positivity was seen as detected with antibody against GFAP (data not shown). A clear difference was seen with the adenovirus vector as injected into the corpus callosum; adenoviruses transduced ventricular ependymal lining (Figure 2b–d) and cells in the corpus callosum (Figures 2e and f) with high efficacy. Injection into the striatum resulted in an effective transduction near the injection site and in the corpus callosum. Cells in the subarachnoidal space were also occasionally transduced (data not shown).
Baculovirus gene transfer led to transient gene expression and did not elicit a marked microglia response
Table 1 presents lacZ expression in specific cell types showing the differential tropism of baculoviruses and adenoviruses. The transduction efficiency of baculoviruses was as high as 76.8±14% in choroid plexus epithelial cells. For adenoviruses the transduction efficiencies in corpus callosum and ependymal cells were 71.4±9% and 83.5±11%, respectively. Transgene expression was highest on the fifth day after the baculovirus transduction. The transgene expression decreased rapidly in 2 weeks: after 10 days 30.1±6% of the choroid plexus cells were lacZ-positive, and after 14 days only a few positive cells could be detected. A similar time course was seen with adenoviruses: 5 days after the gene delivery, corpus callosum cells and ependymal cells were strongly lacZ-positive. The transgene expression decreased in 2–3 weeks, but remained detectable at 3 weeks time point (30.6±13% for corpus callosum and 38.1±12% for ependyma).
According to CD11b (OX-42) antibody staining, baculoviruses did not induce any marked microglia response (Figure 1f, Table 1) since only two of the 16 lacZ-positive rats showed positive immunostaining in the brain. On the contrary, adenovirus delivery elicited a marked microglia response within the transduced tissue (Figures 2g and h). Microglia response was seen in all but one of the analyzed rats. The response increased from moderate to strong from day 5 to day 14 and stayed strong for 21 days.
Biodistribution and clinical chemistry
Representative nested RT-PCR analyses of the biodistribution samples are shown in Figure 3. Transgene expression was found in the forebrain and hindbrain after the local delivery and in the spleen, heart and lung from rats tested 5 days after the baculovirus gene transfer (Figure 3a). After adenovirus injections, transgene expression was seen in the forebrain (data not shown) and hindbrain and in the heart (Figure 3b). No major safety problems were found in clinical chemistry analyses (Figure 4), which showed no significant effect of baculoviruses or adenoviruses on acetylaminotransferase, alanineaminotransferase, C-reactive protein, creatinin, biliribin or hemoglobin values.
Baculoviruses are a new promising class of vectors for gene therapy. Previously, it has been shown that they possess the capacity to transduce a variety of mammalian cell types including mesenchymal cells, neural glial cells and hepatocytes.1234,14151617 Our results show that baculoviruses are able to transduce brain cells efficiently in vivo and possess a very specific cell tropism in rat brain. The choroid plexus epithelial cells were transduced with high efficiency (Figure 1, Table 1). Also, modest gene expression in endothelial cells in brain microvessels was detected. Very limited or no expression of the lacZ-transgene was found in neurons, corpus callosum or in the ventricular ependymal lining with X-gal staining. The results differ from previously published studies in nude BALB/c mice and Sprague–Dawley rats,3 where transduced cells were shown to be mainly astrocytes. The reasons for different results are not known but may relate to species differences and differences in stereotaxic coordinates and injection pressure or speed.
The injection site in corpus callosum was selected based on our previous studies in malignant glioma to permit comparison to previous results.18,19 Corpus callosum, which is a major fiber track in the brain, may aid the distribution of virus widely in the brain permitting the virus to contact many different cell types. The results show that adenoviruses transduce cells along the corpus callosum and in the ventricular space. The same distribution pattern can be expected for baculoviruses. Thus, even though baculoviruses most probably distributed along the corpus callosum in a similar manner as adenoviruses, very little gene expression was seen in glial cells as compared to adenoviruses.
To analyze if the cell preference was due to injection site, baculovirus vectors were injected using three different stereotaxic coordinates: one close to the horn of the lateral ventricle, one deeper in the parenchyma and one in the striatum (caudate putamen). Injections into the corpus callosum and into the white matter in close proximity to the cortex lead to similar results. Injection into the striatum lead only to a few positive cells at the injection site, suggesting the preference of baculoviruses to choroid plexus epithelial cells and microvascular endothelial cells. The observed transduction pattern of baculoviruses may be due to the expression of a putative, yet unidentified baculovirus receptor in these cells or to events subsequent to the viral entry or expression. It has been previously suggested that the block in transgene expression may occur during viral uncoating or transportation to the nucleus.7
A different pattern of transduction was seen with adenoviruses. Using the same stereotaxic coordinates, adenoviruses preferentially transduced glial cells in the corpus callosum and ependymal layer in the ventricles (Figure 2). The results are in line with previously published findings, which have shown that adenoviruses can lead to efficient transduction of ependymal cells and meninges covering the brain with some transduction of the cerebral blood vessels.11121320 Both viruses were shown here to transduce endothelial cells in microvessels throughout the brain.
Baculoviruses are regarded as safe and nonpathogenic in humans.8910 However, limited information is still available about their safety and elicited immunological responses in mammalians. Our results indicate that, as measured by immunohistochemistry with OX-42 antibody, baculoviruses elicit much less microglia response than adenoviral vectors.2122 However, this may be partly explained by the different tropism of the viruses. Clinical chemistry analyses did not reveal any major safety issues. Both virus vectors lead to transient transgene expression, the peak expression being 3–5 days after the gene transfer. The gene expression was diminished from 80% of the cells (at day 5) to 30% at day 10 and was almost completely gone 14 days after the baculovirus gene transfer. The reasons for the short-term expression are not known. However, our results show that baculoviruses did not markedly elicit microglial response, which was seen after adenovirus transduction.
Biodistribution analyses indicated that both vectors were found in ectopic organs as detected with a sensitive nested RT-PCR 5 days after the gene transfer (Figure 3). These results are in line with previous reports of the systemic leakage of adenovirus vector.23,24 Baculoviruses seemed to have a more pronounced biodistribution, which may be related to the fact that they efficiently transduce choroid plexus cells, which are part of the blood–brain barrier. Current results underline the need to develop approaches that reduce the leakage of gene transfer vectors to ectopic tissues. The transgene expression was also occasionally seen in the subarachnoidal space, indicating that viruses were distributed along the neuraxis via cerebrospinal fluid. The viruses found in spleen and heart most probably escaped through the arachnoid villi, which is the principal absorption site of the cerebrospinal fluid into the systemic circulation. These findings indicate that safety and biodistribution of the baculovirus vectors need to be carefully evaluated in any future experiments. However, since choroid plexus cells are involved in the production of cerebrospinal fluid they could become an excellent target for the production of secreted therapeutic proteins in the brain. It is concluded that utilizing the naturally restricted cell tropism, baculoviruses may provide an efficient tool for gene delivery to cerebral choroid plexus cells and may become a useful vector for gene therapy of several brain disorders.
Materials and methods
Production of recombinant viruses
Viruses containing nuclear-targeted β-galactosidase cDNA under the cytomegalovirus promoter were constructed as previously described.2,25 Briefly, baculoviruses were constructed using pFASTBac1-plasmid (Gibco BRL, Life Technologies, Gaithersburg, MD, USA) and a Bac-To-Bac Baculovirus Expression system. First-generation E1/partial E3-deleted LacZ-adenoviruses were generated by homologous recombination and produced in 293 cells. Baculoviruses were titered on sf9 insect cells and adenoviral titers were determined on 293 cells. The absence of toxicity, wild-type viruses, microbiological contaminants and lipopolysaccharide were tested as described.2,26
In vivo injections of viruses
Inbred female BDIX rats (n=38)27 were used for the studies. Results were confirmed in Wistar rats (n=11) (data not shown). Rats (200–250 g) were anesthetized intraperitoneally with a solution (0.150 ml/100 g) containing fentanyl-fluanisone (Janssen-Cilag, Hypnorm®, Buckinghamshire, UK) and midazolame (Roche, Dormicum®, Espoo Finland), placed into stereotaxic apparatus (Kopf Instruments), and 1 × 108 plaque forming units of the virus in PBS/0.1% sucrose was injected by Hamilton syringe with a 27-gauge needle18 into a depth of 2.5, 1.7 or 5 mm (see coordinates below). The procedure was repeated in three consecutive days. Injections of the viral vectors intracranially in the right corpus callosum were performed at the following coordinates: (A) 1 mm caudal to bregma, 2 mm right to sutura sagittalis (n=14 for baculovirus and n=27 for adenovirus),(B) 2 mm caudal to bregma, 2.5 mm right to sutura sagittalis (n=8 for baculovirus); and (C) 0.7 mm rostral to bregma, 3 mm right to sutura sagittalis and 5 mm ventrally to dura mater (depth) (n=11 for baculovirus).
Animals were killed with CO2 5, 10, 14 and 21 days after the gene transfer. Rats were perfused with 1 × PBS by transcardiac route for 10 min followed by fixation with 4% paraformaldehyde/0.15 M sodium phosphate buffer (pH 7.4) for 10 min. For the RNA analysis, the fixation with 4% paraformaldehyde was omitted. Brains were removed and divided at the injection site into two coronal pieces. Samples from forebrain and hindbrain, liver, kidney, heart, spleen, lung and skeletal muscle (psoas major) were taken. Tissue samples were rinsed in 1 × PBS and embedded in OCT compound (Miles, Elkhart, IN, USA) for cryo-sectioning. For the RT-PCR, unfixed tissue samples were snap-frozen in liquid nitrogen and stored at −70°C for the RNA isolation. The LacZ activity of the sections (14 μm in thickness) was analyzed with 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal; MBI Fermentas) for 18 h to identify β-galactosidase positive cells.25 Gene transfer efficiency was calculated from 5–8 brain sections adjacent to the injection site from each animal. The percentage of the β-galactosidase positive nuclei, indicating the nuclear targeted transgene, was calculated from the total number of nuclei of the specific cell types (ie choroid plexus epithelial cells for baculoviruses, ependymal cells lining in the ventricles and glial cells in corpus callosum for adenoviruses) from multiple areas of 200×200 μm. Monoclonal antibodies CD31 (1:200, Dako), anti-fibrillary acidic protein (GFAP 1:400, Boehringer Mannheim) and CD11b (OX-42 1:200, Serotec) were used to identify endothelial, astrocytic and microglial cells, respectively. Avidin–biotin–HRP system and biotinylated secondary antibodies with DAP staining were used for signal detection (Vector Elite, Vector Laboratories, Burlingame, CA, USA). Sections were counterstained with Mayer's Carmalum or hematoxylin and data were collected with Image-Pro Plus software with an Olympus AX70 microscope (Olympus Optical, Japan). Controls for immunostainings included incubations with class- and species-matched immunoglobulins and incubations without primary antibodies.
Total RNA was extracted from snap-frozen spleen, liver, kidney, lung, heart, skeletal muscle and transduced brain samples using TRIZOL reagent (Gibco-BRL). Samples were subsequently treated with RQ1 RNase free DNase (Promega, Madison, WI, USA) to eliminate DNA contamination. M-MuLV reverse transcriptase (MBI Fermentas) was used for cDNA synthesis. RT-PCR protocol has been described previously.2 Dynazyme DNA polymerase (Finnzymes, Espoo, Finland) was used to amplify cDNA template. Primer (20 pM/reaction) sequences for LacZ gene were 5′- TTG GAG GCC TAG GCT TTT GC -3′ for adenovirus and 5′-TTG GCC TAG AGT CGA CGG AT -3′ for baculovirus as forward primers and 5′- TGA GGG GAC GAC GAC AGT AT -3′ for both viruses as a reverse primer. Thirty-nine cycles with 1 min denaturation (95°C), 2 min annealing (57.5°C) and 3 min extension (72°C) times were used after a hot start (95°C, 5 min, 57.5°C, 3 min), followed by 10 min final extension at 72°C. A 5 μl aliquot of the first PCR product was used for the second PCR with forward primers 5′- GGT AGA AGA CCC CAA GGA CTT T -3′ for adenovirus and 5′- CCA AGA AGA AAC GCA AAG TG -3′ for baculovirus. A reverse primer for both viruses was 5′-CGC CAT TCG CCA TTC AG -3′. The same protocol was used as in the first PCR, but with 19 cycles. Bands were visualized on 1% agarose gel using ethidiumbromide staining.
Clinical chemistry analyses
Clinical chemistry analysis from serum samples was done in at Kuopio University Hospital Central Laboratory using routine clinical chemistry assays with Delta Pro V 5 equipment (Kone Instruments Corporation).
This study was supported by grants from the Finnish Academy, Sigrid Juselius Foundation and Ark Therapeutics, Ltd. We would like to thank Ms Seija Sahrio, Ms Mervi Nieminen and Mr Tommi Heikura for excellent technical assistance.