Production of adeno-associated viral vectors in insect cells using triple infection: optimization of baculovirus concentration ratios

The production of viral vectors or virus-like particles for gene therapy or vaccinations using the baculovirus expression system is gaining in popularity. Recently, reports of a viral vector based on adeno-associated virus (AAV) produced in insect cells using the baculovirus expression vector system have been published. This system requires the triple infection of cells with baculovirus vectors containing the AAV gene for replication proteins (BacRep), the AAV gene for structural proteins (BacCap), and the AAV vector genome (BacITR). A statistical approach was used to investigate the multiplicities of infection of the three baculoviruses and the results were extended to the production of AAVs containing various transgenes. Highest AAV yields were obtained when BacRep and BacCap, the baculovirus vectors containing genes that code for proteins necessary for the formation of the AAV vector, were added in equal amounts at high multiplicities of infection. These combinations also resulted in the closest ratios of infectious to total AAV particles produced. Overexpression of the AAV structural proteins led to the production of empty AAV capsids, which is believed to overload the cellular machinery, preventing proper encapsidation of the AAV vector transgene, and decreased the viability of the insect cells. Delaying the input of BacCap, to reduce the amount of capsids produced, resulted in lower infectious AAV titers then when all three baculoviruses were put into the system at the same time. The amount of BacITR added to the system can be less than the other two without loss of AAV yield.     

Intracellular localization of adeno-associated viral proteins expressed in insect cells

Production of vectors derived from adeno-associated virus (AAVv) in insect cells represents a feasible option for large-scale applications. However, transducing particles yields obtained in this system are low compared with total capsid yields, suggesting the presence of genome encapsidation bottlenecks. Three components are required for AAVv production: viral capsid proteins (VP), the recombinant AAV genome, and Rep proteins for AAV genome replication and encapsidation. Little is known about the interaction between the three components in insect cells, which have intracellular conditions different to those in mammalian cells. In this work, the localization of AAV proteins in insect cells was assessed for the first time with the purpose of finding potential limiting factors. Unassembled VP were located either in the cytoplasm or in the nucleus. Their transport into the nucleus was dependent on protein concentration. Empty capsids were located in defined subnuclear compartments. Rep proteins expressed individually were efficiently translocated into the nucleus. Their intranuclear distribution was not uniform and differed from VP distribution. While Rep52 distribution and expression levels were not affected by AAV genomes or VP, Rep78 distribution and stability changed during coexpression. Expression of all AAV components modified capsid intranuclear distribution, and assembled VP were found in vesicles located in the nuclear periphery. Such vesicles were related to baculovirus infection, highlighting its role in AAVv production in insect cells. The results obtained in this work suggest that the intracellular distribution of AAV proteins allows their interaction and does not limit vector production in insect cells.     

基于TCID50检测AAV9载体制品感染性滴度的方法

目的:建立一种基于半数组织培养感染剂量(median tissue culture infective dose,TCID₅₀)检测9型腺相关病毒(adeno-associated virus type 9,AAV9)载体制品感染性滴度的方法。方法:利用含AAV2 rep和cap基因的1型单纯疱疹病毒(herpes simplex virus type1,HSV1)做为辅助病毒与梯度稀释的AAV9载体制品共同感染HEK-293细胞,培养48 h后用实时荧光定量PCR(quantitative real-time PCR,qPCR)扩增AAV特异性反向末端重复序列(inverted terminal repeats,ITR),根据阳性及阴性感染孔数,利用Kärber法计算样品的TCID₅₀。结果:采用携带增强绿色荧光蛋白报告基因的AAV9载体制品确定辅助病毒HSV1-rc最佳感染复数(multiplicity of infection,MOI)为5,AAV9-101的感染性滴度为1.6×10⁹ TCID₅₀/mL。结论:对AAV9载体制品进行感染性滴度检测,且具有可重复性。     

cis effects in adeno-associated virus type 2 replication

We have utilized deletion mutants of adeno-associated virus (AAV) to investigate which elements of the AAV genome are required in cis for high yields of the wild-type virus in a plasmid transfection assay and in addition whether these elements affect primarily AAV DNA replication or encapsidation. All tested deletions from within the Rep region demonstrated a modest, approximately threefold, decrease in viral production. Deletions within the cap region resulted in markedly less virus. Previous observations suggested that in cells in which recombinant AAV (rAAV) was produced, as in our assay with the helper plasmid pDG, there is a substantial excess of empty capsids. Co-transfections of high- and low-yielding constructs demonstrated that under conditions where Cap is abundant, the constructs with cap deletions did not package efficiently. These observation suggest that the lower yields of rAAV cannot be entirely due to lack of capsids but that elements within the cap region of the wild-type genome are important for efficient encapsidation. The production of virus by the mutants we tested was, however, not consistent with the disruption of a cis-acting packaging signal. Apparently, when Cap is provided "in trans," encapsidation is inefficient. A second observation is that there were equivalent amounts of replicated but unencapsidated viral DNA in cells transfected with each of our constructs. We propose that, in accord with the previously proposed link between DNA replication and encapsidation, the total amount of AAV DNA replication can be limited by the efficiency of encapsidation.     

Efficient generation and amplification of high-capacity adeno-associated virus/adenovirus hybrid vectors

Effective gene therapy is dependent on safe gene delivery vehicles that can achieve efficient transduction and sustained transgene expression. We are developing a hybrid viral vector system that combines in a single particle the large cloning capacity and efficient cell cycle-independent nuclear gene delivery of adenovirus (Ad) vectors with the long-term transgene expression and lack of viral genes of adeno-associated virus (AAV) vectors. The strategy being pursued relies on coupling the AAV DNA replication mechanism to the Ad encapsidation process through packaging of AAV-dependent replicative intermediates provided with Ad packaging elements into Ad capsids. The generation of these high-capacity AAV/Ad hybrid vectors takes place in Ad early region 1 (E1)-expressing cells and requires an Ad vector with E1 deleted to complement in trans both AAV helper functions and Ad structural proteins. The dependence on a replicating helper Ad vector leads to the contamination of AAV/Ad hybrid vector preparations with a large excess of helper Ad particles. This renders the further propagation and ultimate use of these gene delivery vehicles very difficult. Here, we show that Cre/loxP-mediated genetic selection against the packaging of helper Ad DNA can reduce helper Ad vector contamination by 99.98% without compromising hybrid vector rescue. This allowed amplification of high-capacity AAV/Ad hybrid vectors to high titers in a single round of propagation.     

Liver transduction with recombinant adeno-associated virus is primarily restricted by capsid serotype not vector genotype

We and others have recently reported highly efficient liver gene transfer with adeno-associated virus 8 (AAV-8) pseudotypes, i.e., AAV-2 genomes packaged into AAV-8 capsids. Here we studied whether liver transduction could be further enhanced by using viral DNA packaging sequences (inverted terminal repeats [ITRs]) derived from AAV genotypes other than 2. To this end, we generated two sets of vector constructs carrying expression cassettes embedding a gfp gene or the human factor IX (hfIX) gene flanked by ITRs from AAV genotypes 1 through 6. Initial in vitro analyses of gfp vector DNA replication, encapsidation, and cell transduction revealed a surprisingly high degree of interchangeability among the six genotypes. For subsequent in vivo studies, we cross-packaged the six hfIX variants into AAV-8 and infused mice via the portal vein with doses of 5 x 10(10) to 1.8 x 10(12) particles. Notably, all vectors expressed comparably high plasma hFIX levels within a dose cohort over the following 6 months, concurrent with the finding of equivalent vector DNA copy numbers per cell. Partial hepatectomies resulted in approximately 80% drops of hFIX levels and vector DNA copy numbers in all groups, indicating genotype-independent persistence of predominantly episomal vector DNA. Southern blot analyses of total liver DNA in fact confirmed the presence of identical and mostly nonintegrated molecular vector forms for all genotypes. We conclude that, unlike serotypes, AAV genotypes are not critical for efficient hepatocyte transduction and can be freely substituted. This corroborates our current model for AAV vector persistence in the liver and provides useful information for the future design and application of recombinant AAV.     

Integrating Adenovirus–Adeno-Associated Virus Hybrid Vectors Devoid of All Viral Genes

Recently, we demonstrated that inverted repeat sequences inserted into first-generation adenovirus (Ad) vector genomes mediate precise genomic rearrangements resulting in vector genomes devoid of all viral genes that are efficiently packaged into functional Ad capsids. As a specific application of this finding, we generated adenovirus-adeno-associated virus (AAV) hybrid vectors, first-generation Ad vectors containing AAV inverted terminal repeat sequences (ITRs) flanking a reporter gene cassette inserted into the E1 region. We hypothesized that the AAV ITRs present within the hybrid vector genome could mediate the formation of rearranged vector genomes (DeltaAd.AAV) and stimulate transgene integration. We demonstrate here that DeltaAd.AAV vectors are efficiently generated as by-products of first-generation adenovirus-AAV vector amplification. DeltaAd.AAV genomes contain only the transgene flanked by AAV ITRs, Ad packaging signals, and Ad ITRs. DeltaAd.AAV vectors can be produced at a high titer and purity. In vitro transduction properties of these deleted hybrid vectors were evaluated in direct comparison with first-generation Ad and recombinant AAV vectors (rAAVs). The DeltaAd.AAV hybrid vector stably transduced cultured cells with efficiencies comparable to rAAV. Since cells transduced with DeltaAd.AAV did not express cytotoxic viral proteins, hybrid viruses could be applied at very high multiplicities of infection to increase transduction rates. Southern analysis and pulsed-field gel electrophoresis suggested that DeltaAd.AAV integrated randomly as head-to-tail tandems into the host cell genome. The presence of two intact AAV ITRs was crucial for the production of hybrid vectors and for transgene integration. DeltaAd.AAV vectors, which are straightforward in their production, represent a promising tool for stable gene transfer in vitro and in vivo.     

利用多基因缺失型杆状病毒在昆虫细胞中生产腺相关病毒

腺相关病毒(adeno-associated virus, AAV)是基因治疗领域最常使用的病毒载体之一,产量低、成本高是该产业面临的关键瓶颈问题。本研究旨在基于多基因缺失型杆状病毒,建立双病毒感染昆虫细胞以生产AAV的技术体系。首先,进行AAV生产用多基因缺失型重组杆状病毒的构建和扩增,并检测杆状病毒滴度及其感染细胞的效果;然后,使用双杆状病毒共感染昆虫细胞,并优化感染条件;最后,基于优化条件进行AAV生产,并检测评估产量、质量等相关指标。结果表明,AAV生产用多基因缺失型杆状病毒滴度较野生型无差异,感染后细胞存活率下降明显减缓。使用双病毒路线进行AAV优化生产,Bac4.0-1的基因组滴度为1.63×10¹¹ VG/mL,Bac5.0-2的基因组滴度为1.02×10¹¹ VG/mL,较野生型产量分别提升了240%和110%。电镜下,3组均具有正常的AAV病毒形态,且转导活性相近。本研究建立了基于多基因缺失型杆状病毒感染昆虫细胞的AAV生产体系,显著提高了AAV产量,具有一定的应用价值。     

Characterization of AAV vector particle stability at the single-capsid level

Virus families have evolved different strategies for genome uncoating, which are also followed by recombinant vectors. Vectors derived from adeno-associated viruses (AAV) are considered as leading delivery tools for in vivo gene transfer, and in particular gene therapy. Using a combination of atomic force microscopy (AFM), biochemical experiments, and physical modeling, we investigated here the physical properties and stability of AAV vector particles. We first compared the morphological properties of AAV vectors derived from two different serotypes (AAV8 and AAV9). Furthermore, we triggered ssDNA uncoating by incubating vector particles to increasing controlled temperatures. Our analyses, performed at the single-particle level, indicate that genome release can occur in vitro via two alternative pathways: either the capsid remains intact and ejects linearly the ssDNA molecule, or the capsid is ruptured, leaving ssDNA in a compact entangled conformation. The analysis of the length distributions of ejected genomes further revealed a two-step ejection behavior. We propose a kinetic model aimed at quantitatively describing the evolution of capsids and genomes along the different pathways, as a function of time and temperature. This model allows quantifying the relative stability of AAV8 and AAV9 particles.     

Mutational analysis of narrow pores at the fivefold symmetry axes of adeno-associated virus type 2 capsids reveals a dual role in genome packaging and activation of phospholipase A2 activity

Adeno-associated virus type 2 (AAV2) capsids show 12 pores at the fivefold axes of symmetry. We mutated amino acids which constitute these pores to investigate possible functions of these structures within the AAV2 life cycle. Mutants with alterations in conserved residues were impaired mainly in genome packaging or infectivity, whereas few mutants were affected in capsid assembly. The packaging phenotype was characterized by increased capsid-per-genome ratios. Analysis of capsid-associated DNA versus encapsidated DNA revealed that this observation was due to reduced and not partial DNA encapsidation. Most mutants with impaired infectivity showed a decreased capability to expose their VP1 N termini. As a consequence, the activation of phospholipase A2 (PLA2) activity, which is essential for efficient infection, was affected on intact capsids. In a few mutants, the exposure of VP1 N termini and the development of PLA2 activity were associated with enhanced capsid instability, which is obviously also deleterious for virus infection. Therefore, PLA2 activity seems to be required on intact capsids for efficient infection. In conclusion, these results suggest that the pores at the fivefold axes function not only as portals for AAV2 single-stranded DNA packaging but also as channels for presentation of the PLA2 domain on AAV2 virions during infection.     

Determination of the baculovirus transducing titer in mammalian cells

Baculovirus has emerged as a promising vector for in vivo or ex vivo gene therapy. To date, the infectious titer and multiplicity of infection (MOI) based on the ability of baculovirus to infect insect cells are commonly adopted to indicate the virus dosage. However, the infectious titer and MOI do not reliably represent the baculovirus transducing ability, making the comparison of baculovirus-mediated gene transfer difficult. To determine the baculovirus transducing ability more rapidly and reliably, we developed a protocol to evaluate the transducing titers of baculovirus stocks. The virus was diluted twofold serially and used to transduce HeLa cells. The resultant transduction efficiencies were measured by flow cytometry for the calculation of transducing titers. Compared to the infectious titer, the determination of transducing titer is more reproducible as the standard deviations among measurements are smaller. Also, the transducing titers can be obtained in 24 h, which is significantly faster as opposed to 4-7 days to obtain the infectious titer. More importantly, we demonstrated that baculoviruses with higher transducing titers could transduce cells at higher efficiency and yield stronger and longer transgene expression, confirming that the transducing titer was representative of the baculovirus transducing ability. This finding is particularly significant for ex vivo gene delivery whereby unconcentrated viruses are used for transduction and long-term transgene expression is desired. In this regard, our titration protocol provides a simple, fast, and reliable measure to evaluate the quality of virus stocks during virus production and purification, and is helpful to predict the performance of vector supernatants and ensure reproducible gene delivery experiments.     

Optimization of reovirus production from mouse L-929 cells in suspension culture

Reovirus serotype 3 Dearing (T3D) has shown potential as a novel cancer therapy. To support the increasing demand for reovirus, a two-stage perfusion mode scheme is proposed for cell growth and reovirus production. Mouse L-929 cells were used as the host for reovirus infection due to their ability to grow well in suspension culture. Several L-929 cell growth and reovirus infection characteristics were investigated and optimized in spinner flask batch cultures. For the growth of L-929 cells, a balanced nutrient-fortification of SMEM medium increased the maximum cell density by 30%, compared to normal SMEM; however, ammonia and lactate accumulations were found to inhibit further cell growth. For the production of reovirus, approximately 90% increase in viral yield resulted when the infection temperature was reduced from 37 to 33 degrees C. Infectious reovirus particles were shown to be stable in conditioned medium at 37 and 33 degrees C. The final virus titer was dependent on the multiplicity of infection (MOI) and the host cell density at the time of infection. A combination of an MOI of 0.1 pfu/cell and an initial host cell density of 1.0 x 10(6) cells/mL in fortified medium resulted in a maximum virus titer of (4.59 +/- 0.16) x 10(9) pfu/mL and a specific yield of (2.34 +/- 0.08) x 10(3) pfu/cell. At an optimal harvest time of the infection process, 99% of the virus was associated with the cellular debris. Finally, the presence of 5.0 mM ammonia in the culture medium was shown to seriously inhibit the reovirus yield, whereas lactate concentrations up to 20 mM had no effect.     

Surface loop dynamics in adeno-associated virus capsid assembly

The HI loop is a prominent domain on the adeno-associated virus (AAV) capsid surface that extends from each viral protein (VP) subunit overlapping the neighboring fivefold VP. Despite the highly conserved nature of the residues at the fivefold pore, the HI loops surrounding this critical region vary significantly in amino acid sequence between the AAV serotypes. In order to understand the role of this unique capsid domain, we ablated side chain interactions between the HI loop and the underlying EF loop in the neighboring VP subunit by generating a collection of deletion, insertion, and substitution mutants. A mutant lacking the HI loop was unable to assemble particles, while a substitution mutant (10 glycine residues) assembled particles but was unable to package viral genomes. Substitution mutants carrying corresponding regions from AAV1, AAV4, AAV5, and AAV8 yielded (i) particles with titers and infectivity identical to those of AAV2 (AAV2 HI1 and HI8), (ii) particles with a decreased virus titer (1 log) but normal infectivity (HI4), and (iii) particles that synthesized VPs but were unable to assemble into intact capsids (HI5). AAV5 HI is shorter than all other HI loops by one amino acid. Replacing the missing residue (threonine) in AAV2 HI5 resulted in a moderate particle assembly rescue. In addition, we replaced the HI loop with peptides varying in length and amino acid sequence. This region tolerated seven-amino-acid peptide substitutions unless they spanned a conserved phenylalanine at amino acid position 661. Mutation of this highly conserved phenylalanine to a glycine resulted in a modest decrease in virus titer but a substantial decrease (1 log order) in infectivity. Subsequently, confocal studies revealed that AAV2 F661G is incapable of efficiently completing a key step in the infectious pathway nuclear entry, hinting at a possible perturbation of VP1 phospholipase activity. Molecular modeling studies with the F661G mutant suggest that disruption of interactions between F661 and an underlying P373 residue in the EF loop of the neighboring subunit might adversely affect incorporation of the VP1 subunit at the fivefold axis. Western blot analysis confirmed inefficient incorporation of VP1, as well as a proteolytically processed VP1 subunit that could account for the markedly reduced infectivity. In summary, our studies show that the HI loop, while flexible in amino acid sequence, is critical for AAV capsid assembly, proper VP1 subunit incorporation, and viral genome packaging, all of which implies a potential role for this unique surface domain in viral infectivity.     

Disulfide bond formation in the herpes simplex virus 1 UL6 protein is required for portal ring formation and genome encapsidation

The herpes simplex virus 1 (HSV-1) UL6 portal protein forms a 12-subunit ring structure at a unique capsid vertex which functions as a conduit for the encapsidation of the viral genome. We have demonstrated previously that the leucine zipper region of UL6 is important for intersubunit interactions and stable ring formation (J. K. Nellissery, R. Szczepaniak, C. Lamberti, and S. K. Weller, J. Virol. 81:8868-8877, 2007). We now demonstrate that intersubunit disulfide bonds exist between monomeric subunits and contribute to portal ring formation and/or stability. Intersubunit disulfide bonds were detected in purified portal rings by SDS-PAGE under nonreducing conditions. Furthermore, the treatment of purified portal rings with dithiothreitol (DTT) resulted in the disruption of the rings, suggesting that disulfide bonds confer stability to this complex structure. The UL6 protein contains nine cysteines that were individually mutated to alanine. Two of these mutants, C166A and C254A, failed to complement a UL6 null mutant in a transient complementation assay. Furthermore, viral mutants bearing the C166A and C254A mutations failed to produce infectious progeny and were unable to cleave or package viral DNA. In cells infected with C166A or C254A, B capsids were produced which contained UL6 at reduced levels compared to those seen in wild-type capsids. In addition, C166A and C254A mutant proteins expressed in insect cells infected with recombinant baculovirus failed to form ring structures. Cysteines at positions 166 and 254 thus appear to be required for intersubunit disulfide bond formation. Taken together, these results indicate that disulfide bond formation is required for portal ring formation and/or stability and for the production of procapsids that are capable of encapsidation.     

Effect of temperature on recombinant protein expression in Semliki Forest virus infected mammalian cell lines growing in serum- free suspension cultures

The firefly luciferase gene was introduced into the Semliki Forest virus (SFV) vector and high titer recombinant SFV particles generated. The broad host range of SFV allowed efficient infection and high level expression of four mammalian cell lines growing in serum-free suspension cultures. The incubation temperature had dramatic effects on the level and duration of recombinant protein expression. For example, the luciferase activity was significantly higher in the rodent BHK and CHO cell lines incubated at 33 °C compared to 37 °C when harvested 19 h post-infection. At 33 °C the specific expression levels increased 10–20 fold during prolongation of the post-infection time up to 50 h. In contrast, a significant decrease in luciferase activity was observed from 26 h post-infection for cell cultures incubated at 37 °C. Only a slight temperature effect on luciferase expression was seen in the human cell line HEK293 and no effect was observed for the subclone293(EBNA). This revised version was published online in August 2006 with corrections to the Cover Date.     

Rapid uncoating of vector genomes is the key to efficient liver transduction with pseudotyped adeno-associated virus vectors

Transduction of the liver with single-stranded adeno-associated virus serotype 2 (AAV2) vectors is inefficient; less than 10% of hepatocytes are permissive for stable transduction, and transgene expression is characterized by a lag phase of up to 6 weeks. AAV2-based vector genomes packaged inside AAV6 or AAV8 capsids can transduce the liver with higher efficiency, but the molecular mechanisms underlying this phenomenon have not been determined. We now show that the primary barrier to transduction of the liver with vectors based on AAV2 capsids is uncoating of vector genomes in the nucleus. The majority of AAV2 genomes persist as encapsidated single-stranded molecules within the nucleus for as long as 6 weeks after vector administration. Double-stranded vector genomes packaged inside AAV2 capsids are at least 50-fold more active than single-stranded counterparts, but these vectors also exhibit a lag phase before maximal gene expression. Vector genomes packaged inside AAV6 or AAV8 capsids do not persist as encapsidated molecules and are more biologically active than vector genomes packaged inside AAV2 capsids. Our data suggest that the rate of uncoating of vector genomes determines the ability of complementary plus and minus single-stranded genomes to anneal together and convert to stable, biologically active double-stranded molecular forms.     

Production methods for gene transfer vectors based on adeno-associated virus serotypes

Vectors derived from adeno-associated virus serotype 2 (AAV-2) represent a most promising tool for human gene transfer because these vectors are neither pathogenic nor toxic to the target cell, and allow long-term gene expression in a large variety of tissues. However, they are rather inefficient at infecting a number of clinically relevant cell types, and transduction by these vectors is likely hampered by neutralizing antibodies that are highly prevalent in the human population. Therefore, an increasing number of researchers are currently turning their attention to the five other serotypes of AAV, to try and develop these as novel vectors for human gene transfer, hoping to overcome the problems associated with AAV-2 vectors. Here I describe and discuss the methodology to produce these alternative AAV vectors in tissue culture. In detail, two strategies are compared that rely on transfection of cells in culture with either two or three plasmids, containing the AAV vector genome and encoding AAV and adenoviral helper functions. Either of these protocols can be used to package a recombinant AAV genome into capsids of its own serotype (generation of "real" serotypes) or to "cross-package" this vector DNA into capsids derived from another AAV serotype ("pseudotyping"). As these approaches are still in their early stages, the existing limitations of current technology are discussed, and possible further improvements proposed.     

trans-Encapsidation of a Poliovirus Replicon by Different Picornavirus Capsid Proteins

A trans-encapsidation assay was established to study the specificity of picornavirus RNA encapsidation. A poliovirus replicon with the luciferase gene replacing the capsid protein-coding region was coexpressed in transfected HeLa cells with capsid proteins from homologous or heterologous virus. Successful trans-encapsidation resulted in assembly and production of virions whose replication, upon subsequent infection of HeLa cells, was accompanied by expression of luciferase activity. The amount of luciferase activity was proportional to the amount of trans-encapsidated virus produced from the cotransfection. When poliovirus capsid proteins were supplied in trans, >2 × 10⁶ infectious particles/ml were produced. When coxsackievirus B3, human rhinovirus 14, mengovirus, or hepatitis A virus (HAV) capsid proteins were supplied in trans, all but HAV showed some encapsidation of the replicon. The overall encapsidation efficiency of the replicon RNA by heterologous capsid proteins was significantly lower than when poliovirus capsid was used. trans-encapsidated particles could be completely neutralized with specific antisera against each of the donor virus capsids. The results indicate that encapsidation is regulated by specific viral nucleic acid and protein sequences.     

Production of recombinant adeno-associated viral vectors using a baculovirus/insect cell suspension culture system: from shake flasks to a 20-L bioreactor

Production of recombinant adeno-associated viral vectors using a baculovirus/insect cell system at various scales is presented. Shake flask studies were conducted to assess conditions to be used in bioreactors. Two insect cell lines, Trichoplusia ni (H5) and Spodoptera frugiperda (Sf9), were compared for their ability to produce rAAV-2 after infection with recombinant baculoviruses coding for the essential components of the vector. The effect of varying the ratio between individual baculoviruses and the effect of the overall multiplicity of infection (MOI), as well as the cell density at infection, were also examined. Infectious rAAV-2 particles were proportionally produced when increasing the individual MOI of BacRep virus up to 1.6. When equal amounts of each virus were used, a leveling effect occurred beyond an overall MOI of 5 and a maximum titer was obtained. Increasing the cell density at infection resulted in higher yields when infecting the cells in fresh medium; however, for the production of bioactive particles, an optimal peak cell density of approximately 1 x 10(6) cells/mL was observed without medium exchange. Infection in 3- and 20-L bioreactors was done at an overall MOI of 5 with a ratio of the three baculoviruses equal to 1:1:1. Under these conditions and infecting the cells in fresh medium, a total of approximately 2.2 x 10(12) infectious viral particles (bioactive particles) or 2.6 x 10(15) viral particles were produced in a 3-L bioreactor. Without replacing the medium at infection, similar titers were produced in 20 L. Our data demonstrates the feasibility of rAAV-2 production by BEVS at various scales in bioreactors and indicates that further optimization is required for production at high cell densities.