Methods Of Production And Testing

A. Viral Life Cycle

As for C-type viruses, the life cycle of lentiviruses can be conceptually divided into 2 stages for the purposes of using it as a vector. The entry, reverse transcription, transport into the nucleus and integration correspond to the use of the vector to carry heterologous genes into target cells. The subsequent provirus transcription synthesis of viral proteins, packaging of the viral genome, maturation of the viral particle, and export from the cell correspond to the production of the vector particles.

B. Transient and Stable Vector Producer Systems

There are 2 general categories of vector production procedures at present. These are transient transfection of highly transferable cells (Fig. 3A), usually 293 or 293T cells with most or all the plasmids encoding the packaging and vector components (15-17,45-49), and production from stable cell lines already carrying some or all the necessary molecular components. The stable cell lines normally carry an inducible expression cassette for the VSV-G protein so that production of vector is triggered by induction of VSV-G expression (Fig. 3B). Because the VSV-G protein is toxic, these types of cells produce vector for at most around 4 days. This contrasts unfavorably with MLV-based packaging cells that use nontoxic envelopes and can produce vector for 2 weeks or more with 1 run.

No publications describe FIV packaging cell lines, although there are investigators working on these. There is 1 report about an early EIA V-basedpackaging cell (80). Several publications (81-85) have described HIV-1 packaging cell systems along these lines, but they have so far been put to very little use. This is probably the case for several reasons. First, these cell lines at present produce at best equivalent, more often lower, titers than the transient procedures (in the range of 105-106 TU/mL). Second, the packaging cell lines are more cumbersome to use in the research laboratory because it is likely that cellular clones of producer cell lines, carrying the vector of interest, would need to be isolated and tested, whereas the transient technique allows vector preparation to be rapidly generated. Third, packaging and producer lines usually would offer the prospect of continued harvest of vector for 2 weeks or more, but with inducible VSV-G, producer lines do not last more than a few days. Fourth, the real advantage of producer cell lines lies within the ability to characterize the vector-producing source in detail and this is of real advantage for late preclinical and clinical experiments, not for smaller lab or small animal experiments. It is not clear how much effort has gone into optimizing these systems, and there is certainly scope for some variation in approach. For example, it is known fromMLV systems that limited induction of VSV-G is better than high levels (86). An SIV packaging cell line has also been described (87) and performs more or less like the HIV-based lines. In addition, cell lines for making true HIV vectors with HIV envelopes have been described and made in monkey cell lines (88-90), but the titers from these were 10e4/mL or less. One packaging cell line (82) was used in conjunction with a vector that initially is driven by a CMV promoter at the 5' end, but after 1 round of replication the 5' end acquires a tetracycline-inducible promoter (from the original 3'LTR). The attraction of this maneuver is that it allows transduction of vector genomes into the packaging cell, rather than transfection, but will also allow an effectively ''SIN'' configuration in target cells with an inactive (unin-duced) promoter at the 5'LTR end. It is known from studies in MLV vectors that transduction of the vector, as opposed to transfection, into packaging cells allows higher vector genome expression (a limiting factor in making high titer producer lines) and clean insertion of multiple copies of the genome

The components of the transient transfection system are those shown in Fig. 3a, or some variation thereof. In this situation, the expression of VSV-G does not have to be directed off an inducible promoter system as it is expressed transiently anyway. Because of the need for efficient transfection to make vector, almost all systems use the 293 or 293T

(92) cell lines as the cell substrate. The major technical innovations here to make life easier have been the use of sodium butyrate to boost expression (93,94), and the use of standard transfection agents that are sold commercially (e.g., 95,96) to try and avoid the variability inherent in calcium phosphate precipitation methods. However, for clinical use it is likely that the calcium phosphate method will be preferred because, it seems that either most of the transfection agents have undefined components, the components are trade secret, and/or the manufacturer is unwilling to file a master file with the Food and Drug Administration (FDA).

C. Processing and Purification

The most common processing method by far for these vectors is high-speed centrifugation and resuspension (11). Done once, this can concentrate the vectors about 1000-fold and the resulting preparations show little loss of activity and are surprisingly clean (i.e., the level of obvious contaminants from the tissue culture process such as BSA and DNA are not excessive) and with titers (see below) of around 109 TU/mL. This is, in fact, one of the reasons that the use of these vectors has become quite widespread. Such preparations are not toxic to animals in general, and the high potency allows administration to small animals such as mice and to tissues such as the brain and eye (see below), where it is impossible to effectively administer more than a few microliters. The resuspension buffer is typically ''isotonic'' PBS or Tris buffer with a cryoprotec-tant sugar and sometimes human serum albumin. The formulation of the vector preparations has not received much attention and investigators have usually extrapolated from useful ret-roviral vector formulations (91). There are no reports of stability studies.

It is possible to reconcentrate the 1000-fold concentrated material by further centrifugation, but beyond about 3000fold total concentration results have not been good and difficulties in handling the material (viscosity etc.), toxicity, and significant losses begin to occur.

Alternate concentration methods are desirable because (1) centrifugation is a method that is notoriously difficult to scale up and use for manufacturing significant quantities of material for clinical trials and eventual product marketing; (2) alternate purification methods could be used with centrifugation to achieve higher concentrations; and (3) if other envelopes apart from VSV-G are to be used, it is likely that losses from centrif-ugation will be significant. There is a good deal of guidance for alternate procedures in the retroviral vector literature (e.g., 97) and column chromatography and other methods are starting to be used (98-100)

D. Titering

The titer or potency of vector preparations is typically measured by a dilution series of vector applied to target cells growing on tissue culture plates, allowing the cells to take up and express the vector, then stain for gene expression. This is descended from the classic phage plaque methods originally introduced by Max Delbruck and extended to mammalian viruses by Dulbecco (101). This type of measurement has served the viral and vector community well, but there are several issues with it. First, this works best with a marker that is easy to measure like beta-galactosidase or green fluorescence protein. Second, it is well known that this does not measure all the vector that is supplied to the plate in question, and that a similar ''titer'' can be observed by removing the applied supernatant and applying to a second plate. Third, along the same lines, it is possible to ''increase'' titers by methods such as ''spinoculation'' (102), where the cells are centrifuged during transduction. Fourth, the titer is very dependent on the cells used as target cells, the strain of the particular cell type, and seemingly minor changes in conditions. However, under highly standardized conditions this assay can be used to assess the potency of vector preparations.

For the reasons listed above, plus the fact that most vectors do not encode a gene product that is straightforward to detect in tissue culture, several other assays have been developed and used.

One common assay is the polymerase chain reaction (RT-PCR) measurement of the level of vector RNA in the preparation (103). In this assay, an aliquot of the vector preparation is disrupted and the level of some vector sequence (usually the packaging signal, although care must be taken to avoid competing gag sequences from the packaging construct) is measured by priming with oligomers for a first reverse transcription step followed by regular PCR amplification, usually followed in real time and measuring the number of PCR cycles necessary to observe a signal. Often this is compared in the same experiment to a beta-galactosidase vector preparation and a titer/mL obtained by comparing the two. Alternatively, an internally consistent measurement can be made against a known RNA standard. This assay is simple and rapid and samples the whole aliquot, so it is widely used. The drawbacks are that the relationship to actual in vitro and in vivo potency is not well characterized in general and this is probably vector dependent.

A third assay (the so-called ''DNA transduction'' assay) (96,103,104) is to perform the limiting dilution assay on tissue culture cells, as described above, then measure the level of the vector DNA as compared with cellular DNA directly by PCR (usually by the real-time, PCR cycle-to-detection method used for the RT-PCR assay). This in principle is an absolute measurement and is suitable as a potency assay, but uncertainties around the level of contamination and the rate of disappearance of DNA used in the transient transduction process, have to be dealt with. In addition, it is possible that all the lentiviral vector DNA may not integrate immediately (12), so the timing of the measurement may make a difference.

A fourth assay used with HIV-based vectors is simply to measure the levels of the p24 capsid protein (e.g., 35) using inexpensive widely available commercial kits (used for blood testing). This assay is very robust but measures only the level of p24 in the preparation. The link between this number and actual transduction potency is not obvious, and may be different with different vector preparations. Such assays form part of the basic characterization of vector preparations, but it is recognized that, at present, these preparations have properties that are not necessarily consistent from batch to batch. Such inconsistency may be linked to issues such as the ratio of active to inactive particles, the extent of encapsidation of non-viral RNA, the cell line from which the particles were made, and the variations in the preparation methods in general. Further characterization of these issues is being undertaken and will be necessary as these vectors become more commonly used in the clinic. (See, for example, ''Guidance for Human Somatic Cell and Gene Therapy,'' U.S. Department of Health and Human Services, FDA, Center for Biologics Evaluation and Research, March 1998, somgene.pdf 1998.)

E. Replication-competent Lentivirus and Other Safety Issues in Manufacturing

Biological entities, in terms of FDA definitions of product types, are defined by the process by which they are made, such that robust, reliable manufacturing processes will need to be developed further. However, an important issue that has attracted a good deal of attention is the possibility of regenerating some form of replicating lentiviral entity. This attention was engendered by both the precedent of replication-competent retrovirus being generated in MLV-type production systems and its link to lymphoma in monkeys (5), and the fact that HIV is a well-known and potentially lethal human pathogen. This issue was one force driving the development of the stripped-down, accessory protein-minimized, third-generation production systems. It is likely that these efforts have made generation of replication-competent lentivirus (RCL) (105) in third-generation systems, in the manner seen for MLV type systems, close to impossible, and that any replicating agent may need to use or acquire by recombination, cellular components as well as the viral system machinery. For example, an envelope from a human endogenous retrovirus can pseudotype HIV-1 (66), when the gene is supplied in trans. This potential acquisition of cellular components through some form of nonhomologous recombination will likely need to be even more extensive to generate a hypothetical replicating entity from cells making vectors based on viruses such as FIV and EIAV that are not capable of replication in human cells. Nevertheless this issue is perceived as significant and the likelihood of generating RCI will be judged by the data generated in vector production and testing. Therefore, tests for RCL form a key element of testing of vector preparations. The major difficulty is that, as it is unclear exactly what the properties of the replicating entity would be, designing a reliable test and spike control has been difficult. In general, a putative RCL needs to be able to replicate in amplifier or test cells to be detected. It has been agreed generally that the risk involved to construct, test, and use experimentally some of the postulated hybrids outweighs any possible general benefit to testing of vector preparations. For example, a version of HIV without accessory proteins and a VSV-G type envelope could have extremely unpredictable properties if a human infection occurred. No investigator wants to use this.

Therefore, various test methods have been proposed and developed. These include the use of cell lines that express an envelope and the use of an HIV gag-pol vector as the spike with read-out being the production of p24 (84), as well as the use of the ''product-enhanced reverse transcriptase assay'', which assays for the RT activity on a defined template (96), and then amplifies the defined RT product by PCR and detects using real-time PCR cycle-dependent assay. The spike in this case could be almost any replicating retrovirus such as MLV. Both these assays require some cellular passaging to get rid of vector signal and hopefully amplify or conserve the RCL signal. Alternative assays look directly for postulated recombination products by PCR but have the drawbacks that this represents a guess at what the entity would be and also it does not actually test for replicating agents.

0 0

Post a comment