Strategies for targeting lentiviral vectors

Vectors derived from retroviruses such as lentiviruses and onco-retroviruses are probably among the most suitable tools to achieve a long-term gene transfer since they allow stable integration of a transgene and its propagation in daughter cells. Lentiviral vectors should be preferred gene delivery vehicles over vectors derived from onco-retroviruses (MLV) since in contrast to the latter they can transduce non-proliferating target cells. Moreover, lentiviral vectors that have the capacity to deliver transgenes into specific tissues are expected to be of great value for various gene transfer approaches in vivo. Here we provide an overview of innovative approaches to upgrade lentiviral vectors for tissue or cell targeting and which have potential for in vivo gene delivery. In this overview we distinguish between three types of lentiviral vector targeting strategies (Fig 1): 1) targeting of vectors at the level of vector-cell entry through lentiviral vector surface modifications; 2) targeting at the level of transgene transcription by insertion of tissue specific promoters into lentiviral vectors; 3) a novel microRNA technology that rather than targeting the 'right' cells will 'detarget' transgene expression from non-target cells while achieving high expression in the target-cell. It is clear that each strategy is of enormous value for several gene therapy approaches but combining these three layers of transgene expression control will offer tools to really overcome several drawbacks in the field such as side-effect of off-target expression, clearance of transgene modified cells by immune response to the transgene and lack of biosecurity and efficiency in in vivo approaches.

Inactivation of the IGF-I receptor gene in primary Sertoli cells highlights the autocrine effects of IGF-I

IGF-I regulates pituitary and gonadal functions, and is pivotal for sexual development and fertility in mammalian species. To better understand the function of autocrine IGF-I in Sertoli cell physiology, we established a system for Cre-mediated conditional inactivation of the IGF-I receptor (IGF-IR) in cultured Sertoli cells. We show here that loss of IGF-IR decreased the number of viable Sertoli cells as a consequence of diminished Sertoli cell proliferation and increased Sertoli cell death. Furthermore, the lack of IGF-IR altered the morphology of cultured Sertoli cells and decreased lactate and transferrin secretions. Collectively, our data indicate that autocrine IGF-I contributes significantly to Sertoli cell homeostasis. The described in vitro system for loss-of-function analysis of the IGF-IR can be readily transposed to study the role of other intratesticular growth factors involved in spermatogenesis.

Targeting retroviral and lentiviral vectors

Retroviral vectors capable of efficient in vivo gene delivery to specific target cell types or to specific locations of disease pathology would greatly facilitate many gene therapy applications. The surface glycoproteins of membrane-enveloped viruses stand among the choice candidates to control the target cell receptor recognition and host range of retroviral vectors onto which they are incorporated. This can be achieved in many ways, such as the exchange of glycoprotein by pseudotyping, their biochemical modifications, their conjugation with virus-cell bridging agents or their structural modifications. Understanding the fundamental properties of the viral glycoproteins and the molecular mechanism of virus entry into cells has been instrumental in the functional alteration of their tropism. Here we briefly review the current state of our understanding of the structure and function of viral envelope glycoproteins and we discuss the emerging targeting strategies based on retroviral and lentiviral vector systems.

MuLV packaging systems as models for estimating/measuring retrovirus recombination frequency

Interaction of retrovirus vectors and endogenous retroviruses present in packaging cell lines and target cells may result in the formation of recombinant viruses. Using sensitive RT-PCR assays, we have investigated human and murine gene therapy packaging cell lines for the incorporation of endogenous retrovirus transcripts into murine leukaemia virus (MLV) vector particles and whether vector genomes are incorporated into human endogenous retrovirus (HERV) particles. VL30 endogenous retrovirus sequences were packaged in particles produced by the murine AM12 packaging system. For every seven MLV-derived -galactosidase beta-Gal vector genomes present in the particles, one copy of VL30 was also packaged. Although human FLY packaging cells expressed HERV transcripts (HERV-K, HuRT, type C, and RTVL-H), none was detectable in the MLV vector particles released from the cells. Non-specific packaging of the MLV gag-pol expression vector transcripts was detected in the FLY virions at a low level (one in 17,000 sequences). In other experiments, gag proteins produced by HERV-K particles present in human teratocarcinoma cells did not appear to package MLV-based vectors that expressed Gal transcripts. These findings indicate that retrovirus vectors interact with human packaging cells to produce retrovirus particles that are far less contaminated by endogenous viral sequences or other types of extraneous particles than murine packaging cells.

Retargeting gene delivery using surface-engineered retroviral vector particles

Retroviral vectors with the capacity to deliver transgenes to specific tissues are expected to be of great value for various gene transfer applications in vivo. Initial attempts to modify vector host-range by the insertion of ligands on their surface glycoproteins have frequently failed, essentially owing to the impairment of the fusogenicity of the vector particles bound to the targeted cell-surface molecules. Several strategies aimed to recover the fusogenic activity of surface-engineered vector particles have recently been explored and have given rise to novel concepts in the field.

Post-mitotic, differentiated myotubes efficiently produce retroviral vector from hybrid adeno-retrovirus templates

We have examined the ability of proliferating myoblasts and post-mitotic, differentiated myotubes to produce retroviral vector using hybrid adeno-retroviral vectors as templates. We show that production of retroviral vector from myoblasts peaks 48 h after adenoviral infection at 4.8 x 10(4) cfu/ml and is scarcely detectable by 96 h. Both fully and partially differentiated myotubes were able to generate a sustained increase in the levels of retroviral vector compared with myoblasts peaking 48 h at 1.4 x 10(5) cfu/ml and 1.8 x 10(5) cfu/ml, respectively. Addition of the cell cycle inhibitor aphidicolin (5 microg/ml) had no effect on the production of retroviral vector from fully differentiated myotubes, but resulted in an 80% increase in vector production from partially differentiated myotubes. Thus indicating that retroviral vector production is more efficient in post-mitotic myotubes and is independent of muscle cell cycle progression.

Organ distribution of gene expression after intravenous infusion of targeted and untargeted lentiviral vectors

Lentiviral vectors represent an attractive technology platform from which to develop a targetable injectable gene delivery system for transduction of specific cell populations in vivo, irrespective of their cell cycle status. Targeted HIV-1-based lentiviral vectors were generated by pseudotyping them with chimeric murine leukemia virus (MLV) envelope glycoproteins displaying N-terminal targeting polypeptides. Vectors displaying an EGF polypeptide were fully infectious on EGF receptor-negative cells, but were inactive on cells with abundant EGF receptors (inverse targeting). Receptor-mediated inactivation of gene transfer was overcome by competing the EGF receptors on the target cells with soluble EGF or by removing the displayed EGF domain from the surface of the vector particles by factor Xa cleavage of a specific protease substrate engineered into its tethering linker (protease targeting). Intravenous infusion of nontargeted HIV-1 vectors led to maximal luciferase activity in liver and spleen with moderate or minimal activity in heart, skeletal muscle, lung, brain, kidney, ovaries and bone marrow. In contrast, intravenous EGF-displaying vectors were expressed maximally in spleen with very low level luciferase expression detectable in liver (EGF-receptor rich). Liver transduction by the EGF-displaying vector was restored by pretreating the animals with soluble EGF suggesting that these vectors are inversely targeted to spleen.

Activation of membrane fusion by murine leukemia viruses is controlled in cis or in trans by interactions between the receptor-binding domain and a conserved disulfide loop of the carboxy terminus of the surface glycoprotein

Cell entry of retroviruses is initiated by the recognition of cellular receptors and the subsequent membrane fusion between viral and cellular membranes. These two steps are mediated by the surface (SU) and transmembrane (TM) subunits of the retroviral envelope glycoprotein (Env), respectively. Determinants regulating membrane fusion have been described throughout SU and TM, but the processes coupling receptor recognition to fusion are still elusive. Here we establish that a critical interaction is formed between the receptor-binding domain (RBD) and the major disulfide loop of the carboxy-terminal domain (C domain) of the murine leukemia virus SU. Receptor binding causes an alteration of this interaction and, in turn, promotes further events of Env fusion activation. We characterize mutations which, by lowering this interaction and reducing the compatibility between the RBD and C domains of Env glycoprotein chimeras, affect both Env fusogenicity and sensitivity to receptor interference. Additionally, we demonstrate that suboptimal interactions in such mutant Env proteins can be compensated in trans by soluble RBDs in a manner that depends on their compatibility with the C domain. Our results therefore indicate that RBD/C domain interactions may occur in cis, via the proper RBD of the viral Env itself, or in trans, via a distinct RBD expressed by virion-free Env glycoproteins expressed endogenously by the infected cells or provided by neighboring Env trimers.

Concentration of viral vectors by co-precipitation with calcium phosphate

BACKGROUND

The envelope glycoproteins, surface unit (SU) and transmembrane (TM) of the murine leukemia virus (MLV) are not covalently linked and tend to dissociate upon high-speed centrifugation, leading to loss of vector infectivity. This study describes a gentle and simple method to concentrate MLV vectors or HIV vectors pseudotyped with MLV envelopes. Having a fast and inexpensive method to concentrate large volumes of vector supernatant will facilitate in vivo experiments and clinical trials that require high titer vector stocks.

METHODS

The methods employed in the study were co-precipitation of viral supernatant with calcium phosphate, low-speed centrifugation, dialysis, and infection assays with Lac-Z transducing vectors.

RESULTS

Murine leukemia virus vectors and HIV vectors pseudotyped with vesicular stomatitis virus glycoprotein (VSV.G) or MLV envelopes were concentrated successfully using the calcium phosphate co-precipitation method. Parameters that influence virus yield and the reproducibility of the method were investigated. The optimized protocol involves virus harvest in serum-free media, co-precipitation using 60mM calcium chloride, pelleting at 2,000 g, resuspending the pellet in a small volume of 0.1M EDTA-saline, and dialysis against saline to remove EDTA. Volumes were decreased from 300 ml to 10 ml, with 50-100% recovery, and titers can be concentrated up to 1,000-fold.

CONCLUSIONS

The calcium phosphate co-precipitation method to concentrate virus is applicable to retrovirus and lentivirus preparations. It uses simple techniques and does not require expensive equipment. Multiple rounds of co-precipitation can be carried out if required.

T cell-specific expression from Mo-MLV retroviral vectors containing a CD4 mini-promoter/enhancer

BACKGROUND

Gene therapy of various immunological disorders will greatly benefit from improved retroviral vectors (RVs) with T cell specificity. Such vectors can be designed by placing a gene of therapeutic interest under the control of tissue-specific transcriptional elements. However, low titers and loss of specificity are frequently encountered with tissue-specific vectors. The aim of the present study was to develop a T cell-specific RV.

METHODS

We constructed a series of Moloney murine leukemia virus (Mo-MLV)-based RVs expressing enhanced green fluorescent protein (EGFP) under the control of a mini-promoter/enhancer cassette derived from the CD4 gene (CD4pmE) and tested them in cell lines and peripheral blood lymphocytes. Expression of EGFP was monitored by fluorescence microscopy and analyzed by flow cytometry.

RESULTS

The CD4pmE cassette was inserted between the viral long terminal repeats (LTRs) in self-inactivating vectors (SIN vectors) or was substituted to the 3' U3 viral promoter/enhancer (hybrid vectors). High vector titers but poor specific expression of EGFP were achieved when CD4pmE was inserted in sense orientation in SIN vectors. Low titers but high specificity were observed when the CD4pmE cassette was in anti-sense orientation. In contrast, high titers and good T cell specificity were obtained with hybrid vectors.

CONCLUSION

An efficient T cell-specific retroviral vector was obtained.

Characterization of novel safe lentiviral vectors derived from simian immunodeficiency virus (SIVmac251) that efficiently transduce mature human dendritic cells

We describe the generation and the characterization of new lentiviral vectors derived from SIVmac251, a simian immunodeficiency virus (SIV). A methodical approach was used to engineer both efficient and safe packaging constructs allowing the production of SIV viral core proteins. SIV-vectors encoding GFP (green fluorescent protein) were generated as VSV-G-pseudotyped particles upon transient expression of the vector construct and helper functions in 293 cells. The SIV vectors were able to transduce efficiently various target cell types at low multiplicity of infection, including monocyte-differentiated human dendritic cells (DCs) which retained their capacity to differentiate into mature DCs after gene transfer. Transduction of the DCs by the SIV vectors was prevented when infections were performed in the presence of AZT, a reverse-transcriptase inhibitor. After gene transfer, expression of the GFP in the target cells remained constant after several weeks, indicating that the vectors had been stably integrated into the genome of the host cells. Preparations of SIV vectors were systematically checked for the absence of replication-competent and recombinant retroviruses but remained negative, suggesting the innocuousness of these novel gene delivery vectors. Side-to-side comparisons with vectors derived from HIV-1 (human immunodeficiency virus) indicated that the SIV vectors were equally potent in transducing proliferating target cells. Finally, we have determined the infectivity of SIV vectors pseudotyped with surface glycoproteins of several membrane-enveloped viruses.

Development of minimal lentivirus vectors derived from simian immunodeficiency virus (SIVmac251) and their use for gene transfer into human dendritic cells

Lentivirus-derived vectors are very promising gene delivery systems since they are able to transduce nonproliferating differentiated cells, while murine leukemia virus-based vectors can only transduce cycling cells. Here we report the construction and characterization of highly efficient minimal vectors derived from simian immunodeficiency virus (SIVmac251). High-fidelity PCR amplification of DNA fragments was used to generate a minimal SIV vector formed from a 5' cytomegalovirus early promoter, the 5' viral sequences up to the 5' end of gag required for reverse transcription and packaging, the Rev-responsive element, a gene-expressing cassette, and the 3' long terminal repeat (LTR). Production of SIV vector particles was achieved by transfecting 293T cells with the vector DNA and helper constructs coding for the viral genes and the vesicular stomatitis virus glycoprotein G envelope. These SIV vectors were found to have transducing titers reaching 10(7) transducing units/ml on HeLa cells and to deliver a gene without transfer of helper functions to target cells. The central polypurine tract can be included in the minimal vector, resulting in a two- to threefold increase in the transduction titers on dividing or growth-arrested cells. Based on this minimal SIV vector, a sin vector was designed by deleting 151 nucleotides in the 3' LTR U3 region, and this SIV sin vector retained high transduction titers. Furthermore, the minimal SIV vector was efficient at transducing terminally differentiated human CD34(+) cell-derived or monocyte-derived dendritic cells (DCs). Results show that up to 40% of human primary DCs can be transduced by the SIV vectors. This opens a new perspective in the field of immunotherapy.

Definition of an amino-terminal domain of the human T-cell leukemia virus type 1 envelope surface unit that extends the fusogenic range of an ecotropic murine leukemia virus

Murine leukemia viruses (MuLV) and human T-cell leukemia viruses (HTLV) are phylogenetically highly divergent retroviruses with distinct envelope fusion properties. The MuLV envelope glycoprotein surface unit (SU) comprises a receptor-binding domain followed by a proline-rich region which modulates envelope conformational changes and fusogenicity. In contrast, the receptor-binding domain and SU organization of HTLV are undefined. Here, we describe an HTLV/MuLV envelope chimera in which the receptor-binding domain and proline-rich region of the ecotropic MuLV were replaced with the potentially corresponding domains of the HTLV-1 SU. This chimeric HTLV/MuLV envelope was processed, specifically interfered with HTLV-1 envelope-mediated fusion, and similar to MuLV envelopes, required cleavage of its cytoplasmic tail to exert significant fusogenic properties. Furthermore, the HTLV domain defined here broadened ecotropic MuLV envelope-induced fusion to human and simian cell lines.

Modifying the host range properties of retroviral vectors

Gene therapy protocols would be greatly facilitated by the availability of targetable injectable vectors which could deliver genes in vivo to specific target cells or to specific disease sites. Efforts to develop such retroviral vectors are therefore a high priority in gene therapy research. In this review, we describe the current state of our understanding of the structure and function of the retroviral envelope glycoprotein complex. We then discuss the results of the various strategies that have been devised to modify the host range of the retroviral envelope glycoproteins with a view to achieving retroviral vectors capable of delivering their genes in a highly specific manner to selected human target cells. The strengths and limitations of these strategies are examined.

A hyperfusogenic gibbon ape leukemia envelope glycoprotein: targeting of a cytotoxic gene by ligand display

An important goal in cancer gene therapy is the development of novel targeted cytotoxic genes. The observation that transfection of a GaLV envelope glycoprotein lacking an R peptide into human cells results in considerable cell-cell fusion and subsequent cell death prompted us to explore the potential for using this fusogenic membrane glycoprotein (FMG) as a targeted cytotoxic gene. As proof of principle, we therefore displayed epidermal growth factor (EGF) on the N terminus of GaLV envelope glycoproteins both with and without an R peptide (GaLV R+ and GaLV R-). Transfection of the GaLVR+ envelope expression plasmids did not cause cell-cell fusion. The GaLV R+ envelopes were incorporated into retroviral vectors whose infectivity was investigated on EGF receptor-positive and -negative cells. The vector incorporating an N-terminally unmodified envelope was able to infect all human cell lines tested. Infectivity of the vector incorporating an envelope on which EGF was displayed was restricted on EGF receptor-positive cells (but not on EGF receptor-negative cells) and could be restored by protease cleavage of the displayed domain or competition with exogenous ligand. The cell-cell fusion capacity of the GaLV R- envelope glycoproteins (N-terminally unmodified and with N-terminal display of both EGF and insulin-like growth factor I [IGF-I]) was investigated by plasmid DNA transfection. While the N-terminally unmodified GaLV R- fused all human cell types tested, fusogenicity of GaLV R- on which EGF or IGF-I was displayed was considerably restricted on receptor-positive cells. "Reciprocal" competition experiments showed that fusogenicity could be restored by competition only with the relevant exogenous ligand. Thus the specificity of cell-cell fusion by a hyperfusogenic GaLV envelope glycoprotein can be regulated by N-terminal display of growth factor ligands. There is therefore significant potential for further development of the targeting of the cell-killing capability of this fusogenic viral glycoprotein by using strategies similar to those we have developed for the targeting of retroviral vectors.

An envelope glycoprotein of the human endogenous retrovirus HERV-W is expressed in the human placenta and fuses cells expressing the type D mammalian retrovirus receptor

A new human endogenous retrovirus (HERV) family, termed HERV-W, was recently described (J.-L. Blond, F. Besème, L. Duret, O. Bouton, F. Bedin, H. Perron, B. Mandrand, and F. Mallet, J. Virol. 73:1175-1185, 1999). HERV-W mRNAs were found to be specifically expressed in placenta cells, and an env cDNA containing a complete open reading frame was recovered. In cell-cell fusion assays, we demonstrate here that the product of the HERV-W env gene is a highly fusogenic membrane glycoprotein. Transfection of an HERV-W Env expression vector in a panel of cell lines derived from different species resulted in formation of syncytia in primate and pig cells upon interaction with the type D mammalian retrovirus receptor. Moreover, envelope glycoproteins encoded by HERV-W were specifically detected in placenta cells, suggesting that they may play a physiological role during pregnancy and placenta formation.

Fusogenic membrane glycoproteins as a novel class of genes for the local and immune-mediated control of tumor growth

We report here the use of viral fusogenic membrane glycoproteins (FMGs) as a new class of therapeutic genes for the control of tumor growth. FMGs kill cells by fusing them into large multinucleated syncytia, which die by sequestration of cell nuclei and subsequent nuclear fusion by a mechanism that is nonapoptotic, as assessed by multiple criteria. Direct and bystander killing of three different FMGs were at least one log more potent than that of herpes simplex virus thymidine kinase or cytosine deaminase suicide genes. Transduction of human tumor xenografts with plasmid DNA prevented tumor outgrowth in vivo, and cytotoxicity could be regulated through transcriptional targeting. Syncytial formation is accompanied by the induction of immunostimulatory heat shock proteins, and tumor-associated FMG expression in immunocompetent animals generated specific antitumor immunity.

Incorporation of simian virus 5 fusion protein into murine leukemia virus particles and its effect on the co-incorporation of retroviral envelope glycoproteins

We describe the generation of murine leukemia virus (MLV) virus particles carrying the paramyxovirus fusion protein F from simian virus 5 (SV5-F). This glycoprotein was expressed in cells providing Moloney MLV (MoMLV) Gag and Pol proteins and a lacZ retroviral vector. SV5-F was correctly expressed, processed, and efficiently incorporated into retroviral particles. SV5-F-bearing retroviruses were not infectious although a weak binding to primate and rodent cells could be detected and SV5-F could mediate cell to cell fusion. We then co-expressed the SV5-F glycoprotein in retroviral particles with chimeric and wild-type MoMLV envelope glycoproteins. Our results show that F strongly inhibited infection via the retroviral envelopes although the mechanism of inhibition was different depending on the retroviral envelope used.

Activation of a cell entry pathway common to type C mammalian retroviruses by soluble envelope fragments

Mutations that negatively or positively affect the fusion properties of murine leukemia viruses (MLVs) have been found within all subdomains of their SU (surface) and TM (transmembrane) envelope units. Yet, the interrelations between these different regions of the envelope complex during the cell entry process are still elusive. Deletion of the histidine residue of the conserved PHQV motif at the amino terminus of the amphotropic or the ecotropic MLV SU resulted in the AdelH or the MOdelH fusion-defective mutant envelope, respectively. These delH mutant envelopes are incorporated on retroviral particles at normal densities and normally mediate virion binding to cells expressing the retroviral receptors. However, both their cell-cell and virus-cell fusogenicities were fully prevented at an early postbinding stage. We show here that the fusion defect of AdelH or MOdelH envelopes was also almost completely reverted by providing either soluble SU or a polypeptide encompassing the receptor-binding domain (RBD) to the target cells, provided that the integrity of the amino-terminal end of either polypeptide was preserved. Restoration of delH envelope fusogenicity was caused by activation of the target cells via specific interaction of the latter polypeptides with the retrovirus receptor rather than by their association with the delH envelope complexes. Moreover crossactivation of the target cells, leading to fusion activation of AdelH or MOdelH envelopes, was achieved by polypeptides containing various type C mammalian retrovirus RBDs, irrespective of the type of entry-defective glycoprotein that was used for infection. Our results indicate that although they recognize different receptors for binding to the cell surface, type C mammalian retroviruses use a common entry pathway which is activated by a conserved feature of their envelope glycoproteins.

Myocardial injury-induced fibroblast proliferation facilitates retroviral-mediated gene transfer to the rat heart in vivo

BACKGROUND

Efficient and stable transfer of therapeutic DNA into injured myocardium would be an initial step towards a genetic treatment aimed at myocardial repair after myocardial infarction. Proliferating cardiac fibroblasts in the healing myocardium could be a compelling target for retroviral infection. We evaluated the feasibility of direct in vivo gene transfer into injured myocardium using a high-titer, stable retroviral vector.

METHODS

Using the TE-FLY-A-based MFG retroviral vector harboring nlsLacZ reporter, the gene transfer efficiency was assessed first in vitro in rat cardiac fibroblasts, followed by in vivo evaluation in healing rat myocardium after local freeze-thaw injury. A total of 2.5 x 10(7) infectious units of retrovirus were injected into the injured region of a beating rat heart. The transduced cells were identified by X-gal staining and immunohistochemistry.

RESULTS

Highly efficient transduction of cardiac fibroblasts was observed in vitro with 98% of the cells transduced with single infection. The cell proliferation index in the cardiac granulation tissue appeared maximal 3 days after cryoinjury. Retroviral injection into the injured beating heart induced gene expression localized to the wound repair region. One week after retrovirus injection, 14% of the cells in the reparative tissue were beta-gal-positive, while 4% were beta-gal-positive after 4 weeks. The transduced cells were mostly myofibroblasts.

CONCLUSIONS

Local gene transfer to the healing rat heart is feasible by retrovirus in vivo. This observation may serve as a useful guide for the development of gene therapy aimed at myocardial repair after myocardial infarction.

Selective transduction of protease-rich tumors by matrix-metalloproteinase-targeted retroviral vectors

We recently showed that retroviral vectors can be targeted through protease substrate interactions. Infectivity is blocked by a polypeptide fused to the viral envelope glycoprotein (SU) and is restored when a protease cleaves the connecting linker, releasing the inhibitory polypeptide from the viral surface. Protease specificity is achieved by engineering the sequence of the linker. Here, using two different matrix-metalloproteinase (MMP)-activatable vectors, we demonstrated highly efficient and selective transduction of MMP-rich target cells in a heterogeneous cell population. In vivo, the MMP-targeted vectors showed strong selectivity for MMP-rich tumor xenografts. Protease-activatable vectors offer new possibilities for in vivo targeting of gene delivery.

Modifying the host range properties of retroviral vectors

Gene therapy protocols would be greatly facilitated by the availability of targetable injectable vectors which could deliver genes in vivo to specific target cells or to specific disease sites. Efforts to develop such retroviral vectors are therefore a high priority in gene therapy research. In this review, we describe the current state of our understanding of the structure and function of the retroviral envelope glycoprotein complex. We then discuss the results of the various strategies that have been devised to modify the host range of the retroviral envelope glycoproteins with a view to achieving retroviral vectors capable of delivering their genes in a highly specific manner to selected human target cells. The strengths and limitations of these strategies are examined.

Efficient gene delivery to quiescent interleukin-2 (IL-2)-dependent cells by murine leukemia virus-derived vectors harboring IL-2 chimeric envelope glycoproteins

Interleukin-2 (IL-2) is a cytokine that induces the proliferation of certain IL-2 receptor expressing quiescent cells. Human IL-2 was fused to the amino-terminus of amphotropic murine leukemia virus (MLV) envelope glycoproteins. Retroviral vectors were pseudotyped with both the IL-2 chimeric envelope and the wild-type amphotropic MLV envelope. The chimeric IL-2 glycoproteins were incorporated on retroviral vectors and the IL-2-displaying vector particles could bind specifically to cell surface IL-2 receptors. In addition, the IL-2-displaying vectors could infect proliferating cells through amphotropic receptors irrespective of whether the cells expressed the IL-2 receptor. IL-2-displaying vector particles could also transiently stimulate the cell cycle entry and proliferation of several IL-2-dependent cell lines. Finally, retroviral vectors displaying IL-2 could efficiently transduce G0/G1-arrested cells expressing the IL-2 receptor at a 34-fold higher efficiency compared with vectors with unmodified envelopes. This new strategy, whereby C-type retroviral vector particles display a ligand that activates the cell cycle of the target cells at the time of virus entry, may represent an alternative to lentivirus-derived retroviral vectors for the infection of quiescent cells. In addition, upon infection of an heterogeneous population of nonproliferating cells, MLV-retroviral vectors that display cytokines/growth factors will allow the transgene of interest to be integrated specifically in quiescent cells expressing the corresponding cytokine/growth factor receptor.

Retroviral vectors pseudotyped with lymphocytic choriomeningitis virus

Pseudotyping can improve retroviral vector stability and transduction efficiency. Here, we describe a novel pseudotype of murine leukemia virus packaged with lymphocytic choriomeningitis virus (LCMV). This pseudotype was stable during ultracentrifugation and infected several cell lines from different species. Moreover, LCMV glycoproteins were not cell toxic.

Retroviral display of functional binding domains fused to the amino terminus of influenza hemagglutinin

We have previously shown that retroviral vector particles derived from Moloney murine leukemia virus (Mo-MuLV) can efficiently incorporate influenza hemagglutinin (HA) glycoproteins from fowl plague virus (FPV), thus conferring a broad tropism to the vectors. To modify its host range, we have engineered the FPV HA to display four different polypeptides on its N terminus: the epidermal growth factor, an anti-human MHC class I molecules scFv (single-chain antibody), an anti-melanoma antigen scFv, and an IgG Fc-binding polypeptide. All recombinant HA glycoproteins were correctly expressed and processed, and efficiently incorporated into Mo-MuLV retroviral particles, indicating that amino-terminal insertion of large polypeptides did not alter the conformation of HA chimeras. Virions carrying the different chimeras bound specifically to cells expressing the targeted cell surface molecules of each ligand. In addition, all virion types were infectious but exhibited various degrees of specificity regarding the use of the targeted cell surface molecule versus the wild-type FPV HA receptor for cell entry and infection. For some ligands tested, infectivity was significantly increased on cells that express the targeted receptor, compared with cells that express only the wild-type HA receptor. Furthermore, some polypeptides could abolish infectivity via the wild-type FPV HA receptor. Our data therefore indicate that it is possible to engineer the HA envelope glycoprotein by fusing ligands to its amino-terminal end without affecting its fusion activity.

Retroviral vectors for the expression of two genes in human multipotent neural precursors and their differentiated neuronal and glial progeny

Retroviral vectors allow stable integration of exogenous DNA into genomic DNA and therefore gene transmission to progeny. Multipotent neural precursors and immortal cell lines prepared from them have been demonstrated to integrate into either adult or developing brain in a nontumorigenic, functional manner, without interfering with normal neurobiological processes. These cells thus appear to provide a Trojan horse ideally adapted to directing the expression of transgenes appropriately in a host brain. Here we investigated and optimized the transduction capacity of MuLV-based retroviral vectors in which internal ribosomal entry segments (IRESs) drive coexpression of two heterologous gene products from a single bicistronic mRNA in a human multipotent neural precursor cell line, "Dev," which was prepared from a medulloblastoma. For this, two vectors containing two different combinations of three viral IRESs were used and the capacity of different pseudotyped vectors to permit an efficient and stable transduction of Dev cells was compared. Our data show that (1) the best recombinant vectors for Dev cell transduction are those pseudotyped with the 10A1 MuLV envelope (40% of transduction) and (2) the initial coexpression of neo and plap, observed in transduced undifferentiated Dev cells, is maintained in differentiated Dev cells with a neuronal or glial phenotype. Therefore, these double-IRES vectors may provide an efficient means of transducing the coexpression of two proteins in undifferentiated human neural precursors that is maintained in their differentiated progeny. These data suggest that the double-IRES strategy is well adapted to potential therapeutic situations when coexpression of two different transgenes may be required in the same cell.

High level of retrovirus-mediated gene transfer into dendritic cells derived from cord blood and mobilized peripheral blood CD34+ cells

Dendritic cells (DCs), the most potent antigen-presenting cells, can be generated from CD34+ hematopoietic stem cells and used for generating therapeutic immune responses. To develop immunotherapy protocols based on genetically modified DCs, we have investigated the conditions for high-level transduction of a large amount of CD34+-derived DCs. Thus, we have used an efficient and clinically applicable protocol for the retroviral transduction of cord blood (CB) or mobilized peripheral blood (MPB) CD34+ cells based on infection with gibbon ape leukemia virus (GALV)-pseudotyped retroviral vectors carrying the nls-LacZ reporter gene. Infected cells have been subsequently cultured under conditions allowing their dendritic differentiation. The results show that using a growth factor combination including granulocyte-macrophage colony-stimulating factor plus tumor necrosis factor alpha plus interleukin 4 plus stem cell factor plus Flt3 ligand, more than 70% of DCs derived from CB or MPB CD34+ cells can be transduced. Semiquantitative PCR indicates that at least two proviral copies per cell were detected. Transduced DCs retain normal immunophenotype and potent T cell stimulatory capacity. Finally, by using a semisolid methylcellulose assay for dendritic progenitors (CFU-DCs), we show that more than 90% of CFU-DCs can be transduced. Such a highly efficient retrovirus-mediated gene transfer into CD34+-derived DCs makes it possible to envision the use of this methodology in clinical trials.

Functional characterization of adenoviral/retroviral chimeric vectors and their use for efficient screening of retroviral producer cell lines

We have generated three different E1-deleted replication-defective adenoviral vectors expressing either Moloney murine leukemia virus (Mo-MuLV) Gag-Pol core particle proteins, gibbon ape leukemia virus (GALV) envelope glycoproteins, or an MuLV-derived retroviral vector genome encoding mCD2 antigen, a murine cell surface marker easily detectable by flow cytometry. Each of the three vectors was first characterized individually by infection of cells providing the complementary retroviral function(s) and able to induce the production of retroviral vectors with an efficiency similar to or higher than that of FLY stable retroviral packaging cells [Cosset, F.-L., Takeuchi, Y., Battini, J.-L., Weiss, R.A., and Collins, M.K.L., (1995). J. Virol. 69, 7430-7436]. In small-scale pilot experiments, TE671 cells simultaneously coinfected with the three adenoviral vectors efficiently released helper-free retroviral vectors in their supernatant, with titers greater than 10(6) infectious particles per milliliter by end-point titrations. Our results also indicated that in contrast to retroviral vector-packageable RNAs, the adenovirus-mediated overexpression of both Gag-Pol and Env packaging functions had limited impact on retroviral titers. The primary mechanism suspected is the premature intracellular cleavage of the Pr65gag precursor that we found in gag-pol-expressing cells, which in turn may impair the normal incorporation of high loads of functional Env. Last, the characterization of the adenoviral/retroviral chimeric vectors allowed the screening of various primate cells for retroviral production and we found that three hepatocyte-derived cell lines were highly efficient in the assembly and release of infectious retroviral particles.

Modification of retroviral tropism by display of IGF-I

We have displayed insulin-like growth factor I (IGF-I) as an N-terminal extension of 4070A (amphotropic) retroviral envelope protein. Western blot demonstrated that chimaeric envelope proteins were incorporated into retroviral particles. Interaction between the displayed IGF-I and cell-surface receptors impaired gene delivery. The magnitude of this inhibitory effect was smallest on NIH 3T3 cells, greater on NIH 3T3 cells over-expressing insulin receptor, and greatest on NIH 3T3 cells over-expressing human type-I IGF receptor. Hence, both the number of ligand receptors and their affinity for the displayed ligand influenced the level of gene delivery. The inhibitory effect was abrogated by cleaving the displayed domain from the underlying envelope protein with factor Xa protease, and by the addition of free ligand to the infection. Addition of IGF-I or insulin caused a dose-dependent increase in titre. Possible mechanisms for receptor-mediated inhibition of gene delivery by IGF-displaying vectors are discussed, together with the implications of these results for practical applications of retroviral display and for understanding the mechanism of virus entry.

A proline-rich motif downstream of the receptor binding domain modulates conformation and fusogenicity of murine retroviral envelopes

The entry of retroviruses into cells depends on receptor recognition by the viral envelope surface subunit SU followed by membrane fusion, which is thought to be mediated by a fusion peptide located at the amino terminus of the envelope transmembrane subunit TM. Several fusion determinants have been previously identified in murine leukemia virus (MLV) envelopes, but their functional interrelationships as well as the processes involved in fusion activation upon retroviral receptor recognition remain unelucidated. Despite both structural and functional similarities of their envelope glycoproteins, ecotropic and amphotropic MLVs display two different postbinding properties: (i) while amphotropic MLVs fuse the cells at neutral pH, penetration of ecotropic MLVs is relatively acid pH dependent and (ii) ecotropic envelopes are more efficient than amphotropic envelopes in inducing cell-to-cell fusion and syncytium formation. By exploiting the latter characteristic in the analysis of chimeras of ecotropic and amphotropic MLV envelopes, we show here that substitution of the ecotropic MLV proline-rich region (PRR), located in the SU between the amino-terminal receptor binding domain and the TM-interacting SU carboxy-terminal domains, is sufficient to revert the amphotropic low-fusogenic phenotype into a high-fusogenic one. Furthermore, we have identified potential beta-turns in the PRR that control the stability of SU-TM associations as well as the thresholds required to trigger either cell-to-cell or virus-to-cell fusion. These data, demonstrating that the PRR functions as a signal which induces envelope conformational changes leading to fusion, have enabled us to derive envelopes which can infect cells harboring low levels of available amphotropic receptors.

In vivo selection of protease cleavage sites from retrovirus display libraries

Phage display libraries are widely used for selection and optimization of polypeptide ligands or protease substrates. Because they are expressed and amplified in bacterial hosts, phage are not ideal for displaying eukaryotic polypeptides or for probing mammalian cells. As retroviruses do not suffer from these limitations we constructed plasmids encoding replication-competent murine leukemia viruses displaying a virally encoded epidermal growth factor (EGF) domain at the N-terminus of the envelope glycoprotein. The EGF-displaying viruses replicated freely on EGF receptor-poor cells without deleting the displayed EGF domain but did not propagate on EGF receptor-rich cells because they were sequestered by the EGF receptors. A retrovirus display library was then generated by diversifying the seven-residue linker between the envelope glycoprotein and the displayed EGF domain. Selective pressure for loss of EGF receptor-binding activity was applied to the library by serial passage on EGF receptor-rich HT1080 cells. The selected viruses propagated on these cells with wild-type efficiencies, a phenotype that was conferred by intracellular cleavage of their displayed linker sequences. The selected linker sequences invariably presented arginine-rich motifs matching the consensus cleavage signal for furin-like proteases. Retrovirus display libraries can be used for the selection of polypeptides interacting with components of living mammalian cells.

Retrovirus-mediated gene transfer into human CD34+38low primitive cells capable of reconstituting long-term cultures in vitro and nonobese diabetic-severe combined immunodeficiency mice in vivo

Factors that may improve retroviral transduction of primitive human hematopoietic cells were studied using MFG-based vectors containing a LacZ gene and produced either by a murine (psi-Crip) or a human (Tasaf) cell line. Cord blood (CB) or bone marrow (BM) CD34+ cells were stimulated and transduced in the presence of three cytokines (interleukin 3 [IL-3], IL-6, and stem cell factor [SCF; c-Kit Ligand]). In the supernatant infection protocol, hematopoietic progenitor cells as measured by X-Gal staining of colony-forming unit cells (CFU-Cs) were transduced more effectively with Tasaf (20%) than with psi-Crip (8%). In contrast, there was no difference between these two cell lines in a coculture protocol. However, gene transfer into more primitive CD34+CD38- subsets and in LTC-IC-derived colonies was low. The use of a large number of cytokines including FLT3-L and PEG-rhMGDF increased the transduction efficiency into CD34+CD38(-)-derived CFU-Cs (35% by PCR) or LTC-ICs (10%). A virus pseudotyped with gibbon ape leukemia virus (GALV) envelope further improved gene transfer to 60 and 48% for LacZ+ CFU-C- and LTC-IC-derived colonies, respectively. These conditions of transduction allowed multilineage engraftment of primitive cord blood cells in NOD-SCID mice. Moreover, 10% (at least) of the human hematopoietic cells recovered from the marrow of these immunodeficient animals were transduced. These data suggest that the efficiency of transduction of human hematopoietic primitive cells can be significantly improved by judicious combinations of recombinant cytokines and high retroviral titers.

Incorporation of fowl plague virus hemagglutinin into murine leukemia virus particles and analysis of the infectivity of the pseudotyped retroviruses

We describe retrovirus particles carrying the fowl plague virus (FPV) hemagglutinin (HA). When expressed in cells providing Moloney murine leukemia virus (MoMLV) Gag and Pol proteins and a lacZ retroviral vector, FPV HA was found to be efficiently expressed, correctly processed, and stably incorporated into retroviral particles. HA-bearing retroviruses were infectious with a wide host range and were only 10-fold less infectious than retroviruses carrying wild-type MLV retroviral envelopes. We also coexpressed HA proteins in retroviral particles with chimeric MoMLV-derived envelope glycoproteins that efficiently retarget virus attachment but are only weakly fusogenic. Our results suggest that HA can in some cases enhance the fusion ability of these retroviral particles, depending on the cell surface molecule that is used as a receptor.

Packaging of endogenous retroviral sequences in retroviral vectors produced by murine and human packaging cells

Interaction of retrovirus vectors and endogenous retroviruses present in packaging cell lines and target cells may result in unwanted events, such as the formation of recombinant viruses and the mobilization of therapeutic vectors. Using sensitive reverse transcriptase PCR assays, we investigated human and murine gene therapy packaging cell lines for incorporation of endogenous retrovirus transcripts into murine leukemia virus (MLV) vector particles and, conversely, whether vector genomes are incorporated into human endogenous retrovirus (HERV) particles. VL30 endogenous retrovirus sequences were efficiently packaged in particles produced by the murine AM12 packaging system. For every seven MLV-derived beta-galactosidase (beta-Gal) vector genomes present in the particles, one copy of VL30 was also packaged. Although human FLY packaging cells expressed several classes of HERV transcripts (HERV-K, HuRT, type C, and RTVL-H), none was detectable in the MLV vector particles released from the cells. Nonspecific packaging of the MLV Gag-Pol expression vector transcripts was detected in the FLY virions at a low level (1 in 17,000 sequences). These findings indicate that human packaging cells produce retrovirus particles far less contaminated by endogenous viral sequences than murine packaging cells. Human teratocarcinoma cells (GH cells), which produce HERV-K particles, were transduced with an MLV-derived beta-Gal vector. Although both HERV-K and RTVL-H sequences were found in association with the particles, beta-Gal transcripts were not detected, indicating that HERV Gag proteins do not efficiently package MLV-based vectors.

Retroviral vector targeting to melanoma cells by single-chain antibody incorporation in envelope

Two strategies for targeting recombinant retroviruses to melanoma cells were compared. One was to extend the tropism of an ecotropic envelope to human melanoma cells, the other was to enhance the tropism of an amphotropic envelope for melanoma cells. Chimeric retroviral envelopes, incorporating a single-chain antibody (ScFv) directed against high-molecular-weight melanoma-associated antigen (HMWMAA) at the amino terminus are correctly processed and incorporated into virions. ScFv-ecotropic envelope chimeras allow specific, but low-titer, targeting of HMWMAA-positive cells, when co-expressed with ecotropic envelopes. ScFv-amphotropic envelope chimeras bind specifically to HMWMAA-positive cells and allow preferential infection at high titer.

Inverse targeting of retroviral vectors: selective gene transfer in a mixed population of hematopoietic and nonhematopoietic cells

We previously reported that retroviral vectors displaying epidermal growth factor (EGF) as part of a chimeric envelope glycoprotein are sequestered upon binding to EGF receptor (EGFR)-positive target cells, leading to loss of infectivity. In the current study, we have displayed stem cell factor (SCF) on beta-galactosidase-transducing ecotropic and amphotropic retroviral vector particles as a factor Xa protease-cleavable N-terminal extension of the envelope glycoprotein. Viral incorporation of the SCF chimeric envelopes was demonstrated by immunoblotting of pelleted virions and their specific attachment to Kit receptors was demonstrated by flow cytometry. Gene transfer studies showed that when SCF was displayed on an amphotropic envelope, the infectivity of the SCF-displaying vectors was selectively inhibited on Kit-expressing cells, but could be restored by adding soluble SCF to block the Kit receptors or by cleaving the displayed SCF domain from the vector particles with factor Xa protease. The host range properties of EGF-displaying and SCF-displaying vectors were then compared in cell mixing experiments. When EGFR-positive cancer cells and Kit-positive hematopoietic cells were mixed and exposed to the different engineered vector particles, the cancer cells were selectively transduced by the SCF-displaying vector and the hematopoietic cells were selectively transduced by the EGF-displaying vector. Retroviral display of polypeptide growth factors can therefore provide the basis for a novel inverse targeting strategy with potential use for selective transduction of hematopoietic or nonhematopoietic cells (eg, cancer cells) in a mixed cell population.

A new system for stringent, high-titer vesicular stomatitis virus G protein-pseudotyped retrovirus vector induction by introduction of Cre recombinase into stable prepackaging cell lines

We report here on stable prepackaging cell lines which can be converted into packaging cell lines for high-titer vesicular stomatitis virus G protein (VSV-G)-pseudotyped retrovirus vectors by the introduction of Cre recombinase-expressing adenovirus. The generated prepackaging cell lines constitutively express the gag-pol genes and contain an inducible transcriptional unit for the VSV-G gene. From this unit, the introduced Cre recombinase excised both a neomycin resistance (Neo(r)) gene and a poly(A) signal flanked by a tandem pair of loxP sequences and induced transcription of the VSV-G gene from the same promoter as had been used for Neo(r) expression. By inserting an mRNA-destabilizing signal into the 3' untranslated region of the Neo(r) gene to reduce the amount of Neo(r) transcript, we were able efficiently to select the clones capable of inducing VSV-G at high levels. Without the introduction of Cre recombinase, these cell lines produce neither VSV-G nor any detectable infectious virus at all, even after the transduction of a murine leukemia virus-based retrovirus vector encoding beta-galactosidase. They reproducibly produced high-titer virus stocks of VSV-G-pseudotyped retrovirus (1.0 x 10(6) infectious units/ml) from 3 days after the introduction of Cre recombinase. We also present evidence that VSV-G-producing cells are still fully susceptible to transduction by VSV-G pseudotypes. However, in this vector-producing system, which regulates VSV-G pseudotype production in an all-or-none manner, the integration of vector DNA into packaging cell lines would be minimized. We further show that heparin significantly inhibits retransduction of VSV-G pseudotypes in the culture fluids of packaging cell lines, leading to a two- to fourfold increase in the yield of the pseudotypes after induction. This vector-producing system was very stable and should be advantageous in human gene therapy.

Highly efficient retrovirus-mediated gene transfer into rat hepatocytes in vivo

We have used high-titer (10(8) ffu/ml) recombinant retroviral vectors to transfer the beta-galactosidase (beta-Gal) gene to rat hepatocytes in vivo. In animals injected twice in the portal blood stream the next day after partial hepatectomy, half of the hepatocytes (46 +/- 17%) expressed the marker at the end of liver regeneration. The number of positive cells closely correlated with the viral titer as well as with beta-Gal enzymatic activity present in the whole liver. Because genes transferred via retroviral vectors in the liver are known to be expressed permanently, our present results open new possibilities for the development of gene therapy protocols for hereditary liver diseases using recombinant retroviral vectors.

Masking of retroviral envelope functions by oligomerizing polypeptide adaptors

We have constructed chimeric retroviral envelopes displaying N-terminal polypeptides that are known to form homotrimeric associations. The amphotropic receptor (RAM-1) binding domain from the trimeric surface (SU) glycoprotein of 4070A murine leukemia virus (MLV)-inhibited ecotropic receptor (Rec-1) mediated infection by the SU glycoprotein of Moloney MLV when grafted to its N-terminus. The block to Rec-1-mediated infection was reversed when the RAM-1 binding domain was cleaved from the vector particles using an engineered factor Xa protease-sensitive cleavage signal between the envelope glycoprotein and its N-terminal extension. Trimeric leucine zipper peptides and the trimeric C-terminal domain of CD40 ligand were shown to inhibit RAM-1-mediated infection of NIH3T3 cells by the 4070A envelope when fused to its N-terminus, whereas monomeric helical peptides and the monomeric epidermal growth factor domain did not. The block to RAM-1-mediated infection was reversed when the trimeric polypeptides were cleaved from the vector particles by addition of factor Xa protease. Envelope binding assays using cleaved and uncleaved chimeric 4070A envelopes revealed that binding to RAM-1 receptors on mammalian cells was hindered by trimeric, but not by monomeric, N-terminal polypeptides. These results have important implications for the design of protease-activatable vectors for targeted gene delivery.

Defective herpes simplex virus type 1 vectors harboring gag, pol, and env genes can be used to rescue defective retrovirus vectors

A retroviral packaging transcription unit was constructed in which the Moloney murine leukemia virus (MoMLV) gag-pol and env genes are expressed under the control of herpesvirus regulatory sequences. This transcription unit, lacking long terminal repeats, primer binding sites, and most of the retrovirus packaging signal but retaining both retroviral donor and acceptor splice sites, was cloned into a herpes simplex virus type 1 (HSV-1) amplicon plasmid, and amplicon vectors (the gag-pol-env [GPE] vectors) were generated by using a defective HSV-1 vector as helper virus. The GPE vector population was used to infect human TE671 cells (ATCC CRL 8805), harboring a lacZ provirus (TE-lac2 cells), and supernatants of infected cells were collected and filtered at different times after infection. These supernatants were found to contain infectious ecotropic lacZ retroviral particles, as shown both by reverse transcription-PCR and by their ability to transduce a beta-galactosidase activity to murine NIH 3T3 cells but not to human TE671 cells. The titer of retroviral vectors released by GPE vector-infected TE-lac2 cells increased with the dose of infectious amplicon particles. Retrovirus vector production was inhibited by superinfection with helper virus, indicating that helper virus coinfection negatively interfered with retrovirus production. Induction of retrovirus vectors by GPE vectors was neutralized by anti-HSV-1 but not by anti-MoMLV antiserum, while transduction of beta-galactosidase activity to NIH 3T3 cells by supernatants of GPE vector-infected TE-lac2 cells was neutralized by anti-MoMLV antiserum. These results demonstrate that HSV-1 GPE amplicon vectors can rescue defective lacZ retrovirus vectors and suggest that they could be used as a sort of launching ramp to fire defective retrovirus vectors from within virtually any in vitro or in vivo cell type containing defective retroviral vectors.

A gene delivery system activatable by disease-associated matrix metalloproteinases

We are developing protease-activatable gene delivery vehicles for selective gene delivery to protease-expressing cells. Angiogenesis, inflammation, and cancer invasion are linked to the overexpression of matrix metalloproteinases (MMPs), which destroy the extracellular matrix. Therefore, the MMPs are promising targets for therapy. We have displayed epidermal growth factor (EGF) on retroviral vector particles as an MMP-cleavable amino-terminal extension of the 4070A murine leukemia virus (MLV) envelope glycoprotein. This was achieved by engineering an MMP-cleavage signal (PLGLWA) into the linker between the EGF domain and the 4070A SU. The chimeric envelope was expressed and incorporated into viral particles, and the EGF domain could be cleaved from the surface of the viral particles by gelatinase A (MMP-2). The MMP-sensitive vector and control MMP-insensitive vectors could bind, via their displayed EGF domains, to EGF receptors on A431 cells but were unable to infect them because the EGF receptor (EGFR) does not support postbinding steps required for retroviral entry. In the presence of exogenous MMPs, the infectivity of the MMP-sensitive vector, but not of the MMP-insensitive vectors, was restored on A431 cells, and this cleavage activation could be partially blocked by MMP inhibitors. Endogenous MMPs produced by EGFR-positive HT 1080 cells could selectively activate the MMP-sensitive vector giving rise to a titer that was 1,000-fold higher on HT 1080 cells than on MMP-negative A431 cells. Inhibitor studies and gelatin zymograms indicated that the membrane-associated MT-MMP expressed on the HT 1080 cells played an important role in cleavage activation of the vector. When presented simultaneously with both EGFR-positive cell lines A431 and HT 1080, the vector could efficiently discriminate between the two different cell types, infecting the MMP-positive HT 1080 cells in preference over the A431 cells.

Receptor co-operation in retrovirus entry: recruitment of an auxiliary entry mechanism after retargeted binding

We have constructed Moloney murine leukemia virus (MoMLV)-derived envelope glycoproteins (AMO) displaying an amino-terminal Ram-1-binding domain in which a variety of different amino acid spacers have been inserted between the displayed domain and the MoMLV surface (SU) subunit. Titres of retroviruses generated with these chimeric envelopes were enhanced on cells expressing both Ram-1 and Rec-1 receptors compared with the titres on cells expressing only one or other receptor type. The absolute viral titres and the degree of titre enhancement due to receptor cooperativity were highly variable between the different chimeric envelopes and were determined primarily by the properties of the interdomain spacer. An extreme example of receptor co-operativity was encountered when testing Ram-1-targeted AMOPRO envelopes with specific proline-rich interdomain spacers. AMOPRO viruses could not enter cells expressing only Rec-1 or only Ram-1 but could efficiently infect cells co-expressing both receptors. The data are consistent with a model for receptor co-operativity in which binding to the targeted (Ram-1) receptor triggers conformational rearrangements of the envelope that lead to complete unmasking of the hidden Rec-1-binding domain, thereby facilitating its interaction with the viral (Rec-1) receptor which leads to optimal fusion triggering.

Retroviral display of antibody fragments; interdomain spacing strongly influences vector infectivity

Five different single-chain antibody fragments (scFv) against human cell-surface antigens were displayed on murine ecotropic retroviral vectors by fusing them to the Moloney SU envelope glycoprotein. The spacing between the scFv and the SU glycoprotein was varied by fusing the scFv to residue +7 or to residue +1 of Moloney SU and by inserting linker sequences of different lengths between the domains. All of the chimeric envelopes were efficiently incorporated into vector particles and could bind to human cells through their displayed antibody fragments, but did not infect them. The spacing between the scFvs and the SU glycoproteins had no significant effect on the efficiency of envelope expression or viral incorporation and did not affect the binding properties of the chimeric envelopes, nor did it influence the efficiency of targeted gene delivery to human cells by scFv-displaying vectors. However, on murine fibroblasts the infectivity of vectors incorporating the chimeric envelopes was strongly influenced by the length of the interdomain spacer. The titers were very low when the single-chain antibodies were fused through a tripeptide linker to SU residue +7 and were greatly enhanced (up to 10(5)-fold) when they were fused to SU residue +1 through a heptapeptide linker. These results point to the importance of steric interactions between the domains of chimeric envelope glycoproteins and may have implications for retroviral vector design for human gene therapy.

Targeting retrovirus entry

Most of the clinical gene therapy trials that have been initiated to date have employed ex vivo strategies in which cells are genetically modified outside the body and reimplanted. The ability to deliver genes accurately and efficiently to selected target cell populations in vivo would greatly expand the scope of gene therapy, but current vectors are not well suited to this task. Here we review recent attempts to develop retroviral vectors incorporating engineered envelope glycoproteins that are capable of delivering their genes in a highly specific manner to selected human target cells.

Homologous and nonhomologous retroviral recombinations are both involved in the transfer by infectious particles of defective avian leukosis virus-derived transcomplementing genomes

We previously described avian leukosis virus-based packaging cell lines that produce stocks of retroviral vectors in which replication-competent viruses were not detectable. However, following infection of target cells with these retroviral stocks, we recently obtained colonies resulting from the transmission of recombinant genomes. Here, we have analyzed their genetic structure and shown that (i) each of them results from recombination between the packaging- and integration-defective transcomplementing genomes and the retroviral vector; (ii) recombination probably occurred during the reverse transcription step, involving strand switching of the reverse transcription growing point from the infectious retroviral vector to the transcomplementing RNA; and (iii) sequence identity and nonhomologous sequences were both used for the strand switching.

Transposable elements behavior following viral genomic stress in Drosophila melanogaster inbred line

To analyze the behavior of endogenous transposable elements under genomic stress, a Drosophila melanogaster inbred line was submitted to three kinds of viral perturbations. First, a retroviral plasmid containing the avian Rous Associated Virus type 2 (RAV-2) previously deleted for the viral envelope coding gene (env) was introduced by P element transformation into the Drosophila genome. An insertion of this avian retroviral sequence was detected by in situ hybridization in site 53C on polytene chromosome arm 2R. Second, Drosophila embryos were injected with RAV-2 particles produced by cell culture after transfection with the retroviral plasmid. Third, the Drosophila melanogaster inbred line was stably infected by the sigma native virus. It appears that neither the offspring of the flies in which the viral DNA was found integrated nor those from the infected sigma flies showed copia or mdg1 element mobilization. Injection of the avian RAV-2 particles led, however, to the observation of somatic transpositions of mdg1 element on the 2L chromosome, the copia element insertion pattern remaining stable. Thus, endogenous transposable elements show more instability in sublines injected with exogenous viral particles than in a transgenic subline containing a foreign viral insert, all transposable elements not being equally sensitive to such genomic stress.

TM domain swapping of murine leukemia virus and human T-cell leukemia virus envelopes confers different infectious abilities despite similar incorporation into virions

We investigated the influence of transmembrane protein (TM) domains on incorporation of retroviral envelopes into virions and on infectivity. We introduced complete, truncated, or chimeric Friend murine leukemia virus (F-MuLV) and human T-cell leukemia virus type 1 (HTLV-1) envelopes into an MuLV particle-producing complementation cell line. As shown previously for HTLV-1 envelopes containing extracellular domains of F-MuLV TM (C. Denesvre, P. Sonigo, A. Corbin, H. Ellerbrok, and M. Sitbon, J. Virol. 69:4149-4157, 1995), reverse chimeric F-MuLV envelopes containing the extracellular domain of HTLV-1 TM were not processed. In contrast, a chimeric MuLV envelope containing the entire HTLV membrane-spanning and cytoplasmic domains (FHTMi) was efficiently processed, fusogenic as tested in a cell-to-cell assay, and efficiently incorporated into MuLV particles. However, these MuLV particles bearing FHTMi envelope proteins could not infect mouse or rat cells which are susceptible to wild-type F-MuLV. Therefore, envelopes which are readily fusogenic in cell-to-cell assays and also efficiently incorporated into virions may not necessarily confer virus-to-cell fusogenicity. HTLV envelopes, whether parental, chimeric (containing the MuLV cytoplasmic tail) or with a truncated cytoplasmic domain, were incorporated into MuLV particles with equal efficiencies, indicating that the cytoplasmic tails of these envelopes did not determine their incorporation into virions. In contrast to FHTMi envelope, HTLV-1 envelopes with F-MuLV membrane-spanning and cytoplasmic domains, as well as wild-type HTLV-1 envelopes, conferred virion infectivity. These results help to define requirements for envelope incorporation into retroviral particles and their cell-free infectivity.

Comparison of efficiency of infection of human gene therapy target cells via four different retroviral receptors

The relative efficiency of transduction of gene therapy target cells was measured for retroviruses bearing the envelopes of amphotropic murine leukemia virus (MLV-A), xenotropic murine leukemia virus (MLV-X), gibbon ape leukemia virus (GALV), feline leukemia virus subgroup B (FeLV-B), and the feline endogenous virus RD114. These viruses use various cell-surface receptors. Activated peripheral blood lymphocytes (PBL) and primary melanoma cultures were infected relatively poorly by MLV-X pseudotypes. RD114 pseudotypes infected PBL relatively well, whereas bone marrow progenitor cells were efficiently infected by all viruses. Helper-free virus bearing the envelopes of MLV-A, RD114, or GALV was similarly tested. All infected melanoma or bone marrow progenitor cells efficiently, whereas MLV-A was relatively inefficient for infection of PBL. The general utility of RD114 pseudotyped virus for gene delivery coupled with its resistance to inactivation by human serum makes this envelope the most suitable choice for in vivo gene therapy.