Progress and Prospects: Advancements in Retroviral Vector Design,Generation,and Application

A lthough gammaretroviruses have been hugely important for proof-of-principle demonstrations of efficacy in several specific clinical applications, other retroviral classes, lentiviruses in particular, are emerging as promising alternatives. Of the on-going 35 trials using lentivirus-based vectors (LVs), 24 are in phase I, 9 in phase I/II, and 1 each in phases II and III (source: http://clinicaltrials.gov/). In addition to their ability to stably integrate into the genome of target cells for long-term expression of the gene of interest, LVs enjoy a few significant advantages through their ability to transduce nondividing cells and also to allow robust transgene expression in cells where there is a high probability of epigenetic silencing or position-related variation of activity (Trono, 2000; Schambach and Baum, 2008). Early-generation gammaretroviral vectors have been shown to be particularly prone to insertional mutagenesis through inadvertent long terminal repeat (LTR)-mediated upregulation of proto-oncogene expression (Ott et al., 2006; Hacein-Bey-Abina et al., 2008; Howe et al., 2008; Stein et al., 2010). In part, this risk may be enhanced by the intrinsic preference of these vectors to integrate in and around gene regulatory regions. LVs may therefore have an overall safer integration profile, although both exhibit a semirandom pattern of insertion into the human genome, and this type of risk still cannot be ignored (Bushman et al., 2005; Cavazzana-Calvo et al., 2010). The safety of LVs has been enhanced in several ways including self-inactivating (SIN) vector design, tat independence, use of insulators, and employment of internal tissue- or lineage-specific promoter and enhancer sequences, thereby reducing risks of insertional mutagenesis and vector mobilization (Matrai et al., 2010). Even so, longer term observations in the clinical arena will be needed to define the risk for newer integrating vectors with more clarity. Concurrently, the development of a stable packaging cell line to ensure useful batch consistency and size during large-scale clinical vector production and purification under strict Good Manufacturing Practice (GMP) conditions has been attempted by various groups. Inducible systems for generating stable cell lines have been developed to eliminate cytotoxicity from viral gene products required for packaging, such as the HIV-1-derived protease and Rev, and the heterologous VSV-G envelope protein (Xu et al., 2001; Ikeda et al., 2003; Cockrell et al., 2006; Broussau et al., 2008; Kutner et al., 2009; Throm et al., 2009). It is also noteworthy that replication-competent LVs (generated by homologous recombination between packaging and vector constructs) have not been detected in large-scale production batches based on latest generation packaging systems and vector design (Merten et al., 2011). Here, we comment on six original research articles addressing three separate areas relating to continuing efforts in improving retroviral design, production, and application for preclinical research and clinical applications.

In this issue of Human Gene Therapy, Marc Giry-Laterrière and colleagues describe a modified tetracycline (TET)-inducible, polycistronic vector system encoding a selectable marker and allowing for simple transgene cloning. As noted in their paper, the individual features incorporated in the “polyswitch vector” have been previously described by several groups. The novelty lies in the combination of these in a single vector to facilitate stable, inducible, and reversible transgene expression from a single integration event. Since the description of the first tetracycline-inducible system by Gossen and Bujard (1992), various small-molecule inducer systems have been employed to control the levels and timing of transgene expression both in vitro and in vivo (Vigna et al., 2002; Sirin and Park, 2003; Galimi et al., 2005). However, many have suffered from either low levels of inducibility or leakiness in the absence of the specific inducer. The modified pTF promoter used in this current study was shown to have low basal activity and high inducibility. The authors also separated a selectable marker (regulated by a ubiquitous human elongation factor-1α promoter) from the inducible transgene cassette for enrichment of transduced cells before activating the transgene of interest, and were able to demonstrate regulated expression from their vector in cell lines as well as primary human cells. Of course, it will now be important to evaluate the performance of this vector in preclinical animal models, especially in terms of drug inducibility and reversibility on withdrawal. Although this is likely to be useful for a number of applications, there will also be restrictions on the nature of the transgene dictated purely by the packaging capacity of LVs, which usually ranges between 8 and 10 kilobases.

The majority of laboratories around the world generate lentiviral vectors by following a transient transfection protocol rather than employing a packaging cell line. This is due primarily to the difficulty in maintaining stable, high-level expression of separated packaging components. Although this is somewhat problematic for repeated batch production and scale-up, transient transfection allows considerable flexibility in shuffling packaging (integrating vs. nonintegrating), envelope (various pseudotypes), and vector plasmids in a combinatorial way without necessitating the time-consuming and arduous production of stable packaging cell lines for each vector. In general, 293T cells are widely used as producer cells owing to their high transfectability (Soneoka et al., 1995). In their report, Gama-Norton and colleagues have investigated the feasibility of obtaining a stably transfected single-copy vector integrant in a suitable chromosomal locus in HEK293 cells, which they argued would provide an enhanced level of safety by eliminating any potential risk posed by the presence of SV40 large T antigen in 293T cells. After screening more than 100 independent HEK293 cell clones containing integrated LVs, the authors concluded that none generated titers comparable to the 293T controls. Moreover, their results indicated that the presence of a single copy (rather than transfected multiple copies) of integrated vector in both 293T and HEK293 cells reduced the ability of the cell lines to produce high-vector titers when supplied with packaging proteins in trans. Restoration of large T-antigen expression partially overcame this restriction, apparently through indirect effects on the cell rather than increased transcriptional activity of the vector. With the issue of SV40 large T antigen toxicity in mind, in a previous paper published in Human Gene Therapy, Merten and colleagues described large-scale manufacture and characterization of an LV produced for ex vivo clinical gene therapy of the Wiskott-Aldrich syndrome (WAS). In their report, they showed that it is possible to detect the presence of SV40 large T antigen by Western blotting in concentrated LV preparations generated by ultracentrifugation. However, they were also successful in completely eliminating the contaminants when the vector preparations were purified through multistep membrane and chromatographic processes. Further, by employing a sensitive indicator cell line and probing with sequence-specific primers, they concluded that the small amounts of residual SV40 large T antigen DNA that might have been released from the producer cells during the scale-up was highly unlikely to transform target cells (Merten et al., 2011).

Alongside the remarkable advances in retrovirus-mediated immunotherapy through antigen-redirected T cells (Park et al., 2011; Porter et al., 2011), LVs are promising agents for a host of vaccination strategies targeting cancer and infectious diseases, often through transduction of potent antigen-presenting cells such as dendritic cells (DCs). LVs have been employed by various groups to transduce DCs following which the encoded antigen is processed and presented in association with suitable MHC molecules to provide long-lasting antigen presentation, theoretically leading to a more effective and sustained T cell-mediated immune response (Schroers et al., 2000; Lizee et al., 2004; Froelich et al., 2010). To some extent, this effect may be enhanced by the ability of LVs to stimulate DC maturation through Toll-like receptor-3 (TLR3) and TLR7, although the degree to which nonvector components of the LV preparation contribute to this process is not yet clearly defined (Zarei et al., 2002; Metelo et al., 2011). The ability to target DCs in vivo is an important consideration, and various approaches have been investigated including incorporation of DC ligands in the LV envelope or molecular bridges (Yang et al., 2008). As an example, Ageichik and colleagues (in this issue of Human Gene Therapy) describe the use of chimeric measles virus (MV) hemagglutinin (H) to target antigen-presenting cells (APCs). The MV H was mutated to prevent binding to specific MV receptors (although allowing fusion at neutral pH and therefore obviating the need for endocytosis of the targeted receptor) and incorporated a single-chain antibody that recognizes murine MHC class II. They demonstrate effective transduction of B cells and DCs in vitro, and also a robust immune response to a model antigen in vivo. Similarly, Froelich and colleagues report, in this issue of Human Gene Therapy, a novel envelope glycoprotein isolated from a species of alpharetrovirus, the mosquito-borne Aura virus. LVs pseudotyped with this envelope had a stronger preference for transducing cells expressing DC-SIGN, C-type lectin receptors that are expressed on the surface of macrophages and DCs and that have an affinity for mannose-containing glycoproteins found normally on the surface of invading pathogens. Under optimal conditions, the authors were able to achieve a modest transduction titer in 293T cells expressing DC-SIGN. For cells deficient in DC-SIGN, the transduction efficiency was 10-fold lower, indicating some degree of specificity. Previously, LVs pseudotyped with Sindbis virus (another alpharetrovirus)-derived glycoprotein have been shown to bind specific cell types in conjunction with other targeting antibody moieties (Morizono et al., 2010). Similar to Sindbis virus glycoprotein-pseudotyped LVs, the release of the viral core after infection of cells by Aura virus G-pseudotyped LVs is pH dependent and associated with the maturation of the endosomes after virus–endosome fusion (Joo and Wang, 2008). A clearer understanding of the mechanisms of Aura virus-mediated cell entry is therefore warranted before this system can compete with more conventional LVs in terms of efficiency.

The discovery of RNA interference (RNAi) by small interfering RNAs (siRNAs) (Fire et al., 1998) has paved the way for therapeutic silencing of mRNAs encoding disease-causing proteins. Various methods have been developed to harness the potential of RNAi for both in vitro and in vivo applications. An increasing number of preclinical studies and clinical trials employing RNAi strategies are being conducted for the treatment of inherited and acquired diseases including cancer (source: http://clinicaltrials.gov/). Viral vector-mediated delivery of short hairpin RNAs (shRNAs), which are processed in the target cell to generate the siRNAs of choice, has proved promising (Brummelkamp et al., 2002). However, there are several important areas of investigation for optimization of activity including length and sequence of encoded shRNA, vector copy number (high numbers may be cytotoxic), off-target effects, and choice of regulatory sequence for directing appropriate shRNA expression (Fish and Kruithof, 2004; Kim et al., 2005). Besides these, each target organ, cell type, and gene target presents unique challenges. In their report in this issue, Papadaki and colleagues employ a foamy viral vector (FV) system for stable and efficient expression of shRNA for in vitro and in vivo gene silencing. They demonstrate the efficacy of their vector system in targeting endogenous BCR-ABL transcripts, resulting in loss of apoptotic resistance in a human cancer cell line, and also provide experimental evidence of gene silencing in murine hematopoietic stem cells in vivo. In another study in this issue, Li and colleagues demonstrate the utility of LV-mediated RNA interference in downregulating the expression of Bax inhibitor-1 (Bi-1), an antiapoptotic gene, to slow down cell proliferation in two human nasopharyngeal carcinoma (NPC) cell lines coupled with enhanced susceptibility to apoptosis. Furthermore, they provide evidence that intratumoral injection of the same vector significantly suppresses tumor growth in an immunodeficient nude mouse xenograft model.

The road ahead for state-of-the-art gene-based therapies is undoubtedly promising with continuing advances in our understanding of the basic biological principles determining cellular interaction (with viral vectors), cellular processes (transcriptional and translational mechanisms governing expression of exogenously introduced genes), and cellular integrity (cytotoxicity arising from insertional mutagenesis or ectopic gene expression). The development of sophisticated viral vector systems can confidently be predicted to increase the number of tractable clinical applications, but also to address existing concerns regarding safety, efficacy, and manufacture as medicinal products.