Transient vector production

Lentiviral vectors are typically produced in HEK (human embryonic kidney)-293T or HEK-293T-derived cell clones, such as HEK-293T/17 (A.T.C.C. CRL-11268), through transient transfection of multiple vector plasmids. Various transfection methods have been successfully used. Resulting vector titres can vary, depending on the condition of the HEK-293T cells, vector packaging system (second, third or trans-lentiviral systems) and the design of the transfer vector.

As more plasmids are introduced into a transfection, each is proportionately less efficiently delivered into the cells. Therefore four-plasmid systems produce relatively lower titre vectors than three-plasmid systems. Vectors with complex structures, such as multiple expression cassettes, long polycistronic transgene or siRNA (short interfering RNA) expression constructs, tend to produce lower titres of vectors than systems with simpler designs (single expression cassette, short transgene or no siRNA expression).

Inclusion of the cPPT in vector constructs frequently increases vector titres/transduction efficiency. Another factor which can strongly affect vector performance is the internal promoters used to drive trangenes. Tissue-specific promoters can be silent in different types of cells, whereas commonly used constitutively active promoters, such as the CMV promoter, can be less efficient in certain cell types.

Stable packaging cell lines for lentiviral vectors

One can deliver all the elements needed for lentiviral production by transient transfection or engineer stable cell lines to produce one or more viral proteins in trans, reducing the number of plasmids needed for transfection. Stable lentivector packaging cell lines should allow continuous reproducible generation of large vector batches, which would be preferable for the production of clinical-grade lentiviral vectors.

Towards this goal, inducible packaging cell lines for VSV-G-pseudotyped HIV-based lentiviral vectors have been constructed without accessory genes. As an alternative approach, retroviral vector-mediated introduction of vector-packaging elements has been used to establish continous packaging cell lines stably expressing codonoptimized Gag-Pol, Rev, Tat and a γ -retrovirus Env from four independent constructs.

However, those packaging cell lines generate non-SIN vectors or conditional SIN vectors. A clinically applicable SIN lentiviral producer cell line has been generated through retroviral delivery of packaging elements and concatemeric array transfection of a SIN vector genome construct.

Vector concentration

Various concentration methods have been successfully used to increase vector titres, including ultracentrifugation [(1–3)×105 g for 1–4 h at 4 ◦C], high-speed long-term centrifugation (∼104 g for 8–12 h at 4 ◦C using a high-speed centrifuge), precipitation with calcium phosphate, or centrifugal filtration using a 100 kDa molecular-mass cut-off filter.

The efficiency can be affected by the serum concentrations in vector preparations, especially for high-speed centrifugation or centrifugal filtration methods.

Vector titration

Before vector infection of target cells, titration of vectors produced is required to adjust vector doses. Vector titres are also critical to evaluate transduction efficieincy. Various methods have been used, including:

  1. measuring the quantity of vector particle components (physical particle numbers/p24 capsid concentration/RT activity/genome copy numbers) in vector preparations;
  2. measuring proviral vector DNA copy numbers in infected cells; or
  3. measuring transgenes expressed in cells (by flow cytometry or immunostaining). The quantification of viral components in culture supernatants has been broadly used in HIV-1 research.

One major problem of these methods is the equal detection of infectious and noninfectious vectors. Indeed, titration with viral genomic RNA copy numbers and p24 concentrations has been shown to be rather poor in predicting vector transduction efficiency.

In contrast, detecting proviral DNA copy numbers in infected cells by real-time PCR can evaluate infectious vector copy numbers. We and others have successfully used real-time PCR with primer and probe sequences specific to the HIV-1 packaging signal (ψ) that is relatively well-conserved among HIV-1 vectors.

However, it has to be noted that VSV-G-pseudotyped lentiviral vectors can harbour high levels of plasmid DNAs from vector producing HEK-293T cells and only a small proportion of the reverse-transcribed vector DNA can successfully integrate into the host genome. Use of total cellular DNA isolated from cells after one or two passages can minimize the detction of contaminated plasmid DNA and defective non-infectious proviral DNAs in target cells.

This method is suitable if the vector has no marker gene or the transgene is driven by a certain tissue-specific promoter, although high infectious vector copy numbers do not necessarily indicate expression of vector transgenes in target cells. Finally, measuring transduced cells by transgene expression is the most reliable method to evaluate infectious lentiviral titres. This method can be used for vectors with reliable marker genes [GFP (green fluorescent protein), puromycin resistance or neo gene etc.] or when specific antibodies against transgenes are available.

Since this method measures transgene expression, vector titres can vary depending on the vector permissivity of the cells used for vector titration. In addition, this method may not be reliable when measuring the expression of transgenes driven by certain tissue-specific promoters.

Strategies to enhance ex vivo vector transduction

Various strategies have been used to improve vector infection. Although it shows some cytotoxicity, polybrene, which reduces the charge repulsion between virions and the target cell surface, has been widely used to enhance lentiviral vector infection. Retronectin is also used to increase lentiviral transduction of haematopoietic stem cells, particularly those with γ -retroviral Envs.

In addition to those reagents, spinoculation, where vectors are centrifuged down on to monolayers of cells, has been used to increase vector transduction efficiency of various cell types.


As mentioned above, integrating vectors can insert into repressed regions of the chromosome. They can also integrate randomly into regions of the genome that can be problematic. This reality has long been recognized by the field, but did not come to the forefront until the clinical trials that used retroviral gene therapy for XSCID (X-linked severe combined immunodeficiency).

In these trials, patients with X-SCID received CD34 + haematopoietic stem cells that had been genetically modified with γ -retroviral vectors expressing the common γ -chain from IL2RG (interleukin 2 receptor γ ). Although this work demonstrated successful gene therapy in humans, a subset of these patients subsequently developed a leukaemia-like disease that appeared to be due to insertional mutagenesis by the vector.

Given that both γ -retroviruses and lentiviruses integrate semi-randomly into the genome, this side effect is relevant to the lentiviral vectors discussed here. In this section, we summarize ongoing efforts to minimize the risks associated with lentiviral vector integration into the host genome.

Distinct target site preferences between HIV-1 and γ -retrovirus MLV

Characterization of the integration site preferences of HIV-1 and MLV may help to understand the possible risks associated with lentiviral vector integration and to improve vector designs for safer gene therapy to avoid the risk of insertional mutagenesis. Mapping of HIV and MLV integration sites on the human genome sequence found distinct target site preferences between the two viruses.

MLV vectors appear to be biased towards integration near transcription start sites or CpG islands, whereas HIV appears to favour integration within active transcription regions. Cellular lens epithelium-derived growth factor (LEDGF/p75), which binds both chromosomal DNA and HIV-1 IN plays a role in controlling the location of virus integration in human cells.

Low oncogenic potential of lentiviral vector integration

Haematopoietic stem cell gene transfer in a tumour-prone mouse model has revealed that prototypical lentiviral vectors have lower oncogenic potential than conventional retroviral vectors. Similarly, in a sensitive in vitro immortalization assay to quantify the risk of haematopoietic cell transformation, the insertion pattern of lentiviral vectors was found to be approximately 3- fold lower than that mediated by retroviral vectors to trigger transformation of primary haematopoietic cells.

SIN vector design reduces the insertional mutagenesis potential

SIN vector design (deletion of the U3 region in 3 -LTR) can eliminate the enhancers/promoter elements from the both 3 – and 5 -LTRs in integrated proviral DNA. In a mouse model of haematopoietic stem cell gene therapy, introduction of the SIN mutation increased the safety of retroviral and lentiviral vectors. In vitro immortalization assays indicate that the risk of insertional transformation by SIN lentiviral vectors depends on the type of the internal promoter.

Constructing SIN lentivectors with a tissue-specific promoter appears unable to trigger cell transformation in these models. Nevertheless, SIN lentiviral vectors have been shown to induce leukaemia in a mouse model of X-SCID using the IL2RG gene. Woods et al. concluded that the therapeutic IL2RG gene itself can act as a ‘hit’ in leukaemogenesis that may amplify this effect in X-SCID, but perhaps not in other types of genetic diseases. Ginn et al. have also proposed the existence of other risk factors, unrelated to insertional mutagenesis or IL2RG gene overexpression, that may have exacerbated this side effect.

Integration-deficient lentiviral vectors for non-dividing cells

If cell division is prevented, unintegrated lentiviral DNAs can express transgenes. Integration-deficient lentiviral vector therefore represents an interesting alternative approach to reduce insertional mutagenesis. Integration-deficient HIV-1 vectors typically have mutations in the IN gene that amplify the formation of episomal genomes in the nucleus, with minimal vector integration. Integration can be reduced up to 104 -fold with an integration-deficient HIV-1 vector.

An integration deficient HIV-1 vector can transduce non-dividing cells, such as muscle cells, brain cells and ocular cells in vivo. Clearly, integration-deficient lentivectors increase predictable vector safety, although their long-term gene expression is limited to non-dividing cells and their levels of transgene expression are often less than those of conventional integrating lentiviral vectors.

Self-deleting vectors

Self-excising lentiviral vectors have been developed on the basis of the Cre/loxP system. Self-deletion is achieved by expression of the Cre recombinase from vectors with a loxP site into the U3 region of the 3 -LTR. Duplication of the U3 region of the 3 – LTR during reverse transcription generates a proviral DNA with one loxP site in both LTRs.

After integration and subsequent Cre expression, Cre-mediated recombination of these two loxP sites deletes most of the integrated vector genome with the exception of the flanking U sequences and one loxP site. Similarly, lentiviral vectors with loxP sites have been used for conditional deletion of integrated vector sequences (LTR to LTR or transgene-expressing cassette only) in combination with other Cre-expressing vectors.

Site-directed integration using ZFNs (zinc finger nucleases) and non-integrating lentiviral vectors

ZFNs that recognize and cleave unique genomic sequences in living cells can be used for targeted gene editing and mutagenesis. Naldini and co-workers have incorporated the ZFN technology into non-integrating lentiviral vectors to express ZFNs and provide the template DNA for gene correction.

This system achieves high levels of site-directed gene addition in a panel of human cell lines, including embryonic stem cells, allowing rapid selection-free isolation of clonogenic cells with the desired genetic modification.