Field of the Invention

The present invention relates to a retroviral vector in which the cis elements are located downstream of the 3' LTR. These elements are therefore present in the viral RNA genome so that it can be packaged into virions, but are outside of the region of the genome that is reverse transcribed and so are not present in the vector DNA in the target cell. This reduces the disadvantages associated with transfer vector cis element persistence in target cells.

Background to the Invention

Retroviral vectors were among the earliest viral vectors developed for mammalian gene transfer. A number of retroviral species have been developed into vectors, notably alpharetroviruses (1), gammaretroviruses (2), the lentivirus human immunodeficiency virus type 1 (HIV-1) (3), nonhuman lentiviruses (4), and spumaviruses (5).

The most important structural change in moving from retrovirus to retroviral vector is the separation of viral non-coding sequences required in cis on the nucleic acid undergoing gene transfer from the viral protein coding sequences required in trans in the producer cell for the production of virions. This separation renders the vector capable of only one round of infection as no viral proteins are produced in target cells.

A standard third generation HIV-1 -based lentiviral vector such as RRL or CCL (3) requires co-transfection of four plasmids into producer cells during virion production (Fig. 1): a transfer vector containing the essential cis elements and the transgene expression cassette, a packaging plasmid expressing the HIV-1 polyproteins Gag and Gag-Pol, a plasmid for the expression of the HIV-1 Rev protein, and a plasmid for expression of the viral envelope protein. Lentiviral vectors are able to incorporate envelope proteins from other enveloped viruses if they are co-expressed in producer cells, a phenomenon known as pseudotyping. The most commonly used envelope protein is the vesicular stomatitis virus glycoprotein (VSVG) which confers stability and broad tropism upon lentiviral vector virions (6). The essential cis elements contained within the transfer vector include the HIV-1 long terminal repeats (LTRs), the RNA packaging signal (Ψ), and preferably the Rev Response Element (RRE). The LTRs contain sequences required for transcription, reverse transcription, and integration of the vector genome. In self-inactivating (SIN) transfer vectors almost all of the viral 3' U3 region has been removed in order to eliminate its promoter and enhancer activities (3). The mechanism of reverse transcription is such that both proviral U3s originate from the 3' LTR of the RNA genome, so proviruses resulting from infection with this vector lack LTR-driven transcription and cannot transcribe their full genome efficiently in target cells. The RNA packaging signal is thought to extend into the beginning of the gag coding sequence, but a point mutation has been introduced downstream of the gag start codon to prevent translation of the majority of this sequence. The Rev protein interacts with the RRE in producer cells to stabilise transcripts, promote RNA export from the nucleus, and enhance RNA packaging into virions (7;8). Nonessential but commonly used cis elements include an HIV-1 -derived central polypurine tract (cPPT) which enhances transduction of non-dividing cells (9) and a woodchuck hepatitis virus post-transcriptional regulatory element (WPRE) which enhances viral titre and transgene expression through improving the efficiency of polyadenylation (10).

In a standard third generation lentiviral vector these cis elements are reverse transcribed along with the transgene expression cassette in target cells, resulting in a provirus incorporating two SIN LTRs containing 236bp of HIV-1 DNA in each, a primer binding site (PBS) to Gag region containing 490bp of HIV-1 DNA, an RRE containing 858bp of HIV-1 DNA, and a Nef to polypurine tract (PPT) region containing 69bp of HIV-1 DNA, making a total of 1889bp of HIV-1 DNA delivered to target cells. These viral cis elements are covalently attached to the transgene expression cassette and persist in target cells after transduction. If an integrating lentiviral vector is used, they are irreversibly integrated into target cell chromosomes.

Consequences of transfer vector cis element persistence in target cells

The long term persistence of HIV-1 -derived cis elements in target cells creates a number of experimentally observed and theoretically possible problems for the practical application of lentiviral vectors in studying the biological effects of transferred genes on cell culture or animal models, the generation of transgenic animal strains and stable cell lines, and the transfer of therapeutic genes to treat human disease.

Firstly, the transfer vector cis elements contain active splice acceptor sites which are able to splice with host cell genes to create aberrant fusion transcripts (11-13). An event of this type was observed in a gene therapy clinical trial in which splicing between a copy of the patient's growth-promoting HMGA2 gene and an integrated lentiviral provirus caused dysregulation of HMGA2 transcription and a large clonal expansion of a transduced cell (14).

Secondly, the persistence of cis elements required for RNA packaging enables remobilisation of self-inactivating lentiviral vector genomes in cells expressing viral proteins (15). If used in patients infected with HIV this could result in remobilised lentiviral vector proviruses and recombination with wildtype HIV-1 genomes.

Thirdly, lentiviral cis elements contain untranscribed CpG-rich DNA which is subject to DNA methylation and may contribute to reduction in expression of the delivered transgene through silencing in host cells (16). Fourthly, large untranscribed regions within episomal DNA vectors have been associated with transgene silencing in vivo (17). Reducing the size of the untranscribed region within an episomal integration-deficient lentiviral vector (IDLV) may therefore improve long term expression in applications of these vectors. Fifthly, reducing the size of the reverse transcript may increase the transgene carrying capacity of lentiviral vectors. There are potentially multiple stages of the lentiviral life cycle which are limiting for vector genome size, such as RNA packaging (18), reverse transcription in cell types with low intracellular dNTP concentration (19), and integration into chromosomes.

Minimisation of cis elements within target cell proviruses

Several approaches have been taken towards the goal of minimising the viral cis elements which persist within target cell proviruses. Firstly, a number of authors have investigated the effect of simple deletions and point mutations to remove or inactivate the remaining cis elements. Cui et al reported point mutations to the HIV-1 major splice donor (MSD) positioned upstream of Gag as well as further incremental deletions of the Gag and RRE elements resulting in a transfer vector containing 550bp of HIV-1 cis DNA (or 786bp if both LTRs in the provirus are accounted for) (20). In 293T producer cells these changes resulted in a large decline in expression of unspliced transfer vector RNA, but in TE671 cells this effect was less pronounced due to cell type-specific splicing patterns and the resulting decline in titre was only 2-fold. This system has not been widely adopted within the field, perhaps due to a combination of the reduction in titre and the unusual producer cell line. Kotsopoulou et al reported an attempt to generate a Rev-independent lentiviral vector by codon optimisation of the packaging plasmid and deletion of the RRE from the transfer vector, resulting in a 5 -fold reduction in titre (21). It has since been reported that the Rev/RRE system is required for efficient packaging of transfer vector RNA into virions (8). Koldej et al reported minimisation of the amount of Nef coding sequence upstream of the PPT (22). The entire sequence appears to be dispensable up to the run of 5 thymidine bases immediately upstream of the PPT.

A second approach towards the minimisation of cis elements within target cell proviruses is to design vectors in which these elements are present in the viral RNA genome but subsequently deleted during reverse transcription or integration.

Delviks et al reported a gammaretroviral vector in which the RNA packaging signal was flanked by a 701bp repeat of the herpes simplex virus thymidine kinase gene (HSV-TK) (23). During reverse transcription in target cells, template switching by the reverse transcriptase from one repeat to the other resulted in the deletion of the RNA packaging signal. Since template switching is not an obligatory activity for the reverse transcriptase, the efficiency of this deletion was reported to be 91% of clones. This strategy to delete cis elements has the disadvantage of requiring a new cis element to be introduced (in this case, HSV-TK) which is not itself deleted. Patents and patent applications derived from this vector include US 5741486, US 5714353 and WO 95/032298. A similar strategy was used by Srinivasakumar in a lentiviral vector in which the RRE was flanked by repeat copies of the hygromycin phosphotransferase gene, resulting in deletion of the RRE in 84% of clones (24).

Torne-Celer et al used the ability of retroviral integrases to cleave internal att sites to generate alpharetroviral vectors in which the 5' LTR and the RNA packaging signal are cleaved off the pre-integration complex by the viral integrase enzyme during the 3' end processing stage of integration (25). While 61% of clones in this report carried the expected deletion, internal att site processing appears to produce heterogeneous proviral products and results in titres that are 10 ² to 10 ³ -fold lower than would be expected with a more conventional alpharetroviral vector (26).

A third approach towards the minimisation of lentiviral cis elements is to remove them from target cell proviruses following transduction. Luche et al reported a lentiviral vector in which the RNA packaging signal was flanked with loxP sites so that it could be excised when Cre recombinase was provided in trans in target cells (27). In transduced 293T cells the excision was successful in 12-20% of transduced cells. Fang et al reported a self- minimising lentiviral vector incorporating a Cre recombinase expression cassette which excised the RNA packaging signal, RRE and itself following target cell transduction (28). Successful excision as measured by loss of Cre expression took place in 92% of target cells. Both of these strategies rely on the successful expression and function of Cre recombinase in target cells, and it is highly unlikely that such a strategy would receive regulatory approval for use in patients in the near future.

Summary of the Invention

The inventors of the present invention have sought to minimise cis elements within target cell proviruses in an alternative way to reduce the disadvantages associated with transfer vector cis element persistence in target cells.

In existing lentiviral vectors, cis elements such as the RNA packaging signal and the RRE are located between the two viral LTRs and so are within the region of the genome that is reverse transcribed. In the vector described in this invention, these cis elements are located downstream of the 3' LTR. These elements are therefore present in the viral RNA genome so that it can be packaged into virions, but are outside of the region of the genome that is reverse transcribed and so are not present in the vector DNA in the target cell.

Accordingly, in a first aspect of the invention, there is provided a retroviral vector comprising a primer binding site, a long terminal repeat and an RNA packaging sequence, wherein the RNA packaging sequence is located 3' of the long terminal repeat such that reverse transcription initiated at the primer binding site does not lead to reverse transcription of the RNA packaging sequence into vector DNA in a target cell. In other words, reverse transcription of the vector leads to the RNA packaging sequence being excluded from the reverse transcript produced in a target cell. In this vector, no long terminal repeat is located 3' of the RNA packaging sequence.

The retroviral vector can be based on any suitable retrovirus which is able to deliver genetic information to eukaryotic cells. For example, the retroviral vector may be an alpharetroviral vector, a gammaretroviral vector, a lentiviral vector or a spumaretroviral vector. Such vectors have been used extensively in gene therapy treatments and other gene delivery applications. In some embodiments, the retroviral vector is a lentiviral vector. In some instances, the retroviral vector may be based on HIV-1. The vector comprises a primer binding site (PBS). This is a site which binds to a tRNA primer which is responsible for initiating minus strand synthesis during the reverse transcription process. The primer binding site may be located either 5' or 3' of the long terminal repeat. Preferably, this primer binding site is located towards the 5' end of the vector. Preferably, the primer binding site is located on the 5' side of the long terminal repeat.

In some embodiments, the vector comprises a second primer binding site. Preferably, this is located on the 3 ' side of the long terminal repeat (LTR). In some embodiments, these two primer binding sites are referred to as a 5' primer binding site and a 3' primer binding site. In some embodiments, the second primer binding site is located next to the LTR on the 3' side. Preferably, the second primer binding site is located between the LTR and the RNA packaging sequence so that the RNA packaging sequence is located 3' of the second primer binding site. The primer binding sites bind to a tRNA primer which is responsible for initiating minus strand synthesis during the reverse transcription process. The primer binding sites also provide homology for plus strand transfer during the reverse transcription process.

The vector comprises a long terminal repeat (LTR). Preferably, this is located towards the 3' end of the vector (although some components, such as the RNA packaging sequence, may be positioned further towards the 3' end of the vector).

In some embodiments, the vector comprises two long terminal repeats, referred to as a 5' LTR and a 3' LTR. The 5' LTR is located towards the 5' end of the vector (although there may be components further towards the 5' end). Where two LTRs are present, the primer binding site is preferably located on the 3' side of the 5' LTR (but on the 5' side of the 3' LTR). The primer binding site may be located next to the 5' LTR on the 3' side.

Retroviral LTRs are generally segmented into U3, R, and U5 regions. However, in certain LTRs, parts of these regions may be deleted. The term "long terminal repeat" or "LTR" is intended to cover all such variations in LTRs. The LTR participates in the reverse transcription process so that vector DNA is produced in the target cell based on the vector RNA. LTRs can comprise a number of signals required for gene expression such as a transcriptional enhancer, a promoter, a transcription initiation signal and/or a polyadenylation signal.

The LTR may be a self-inactivating (SIN) LTR. In the vector, for enhanced safety the LTR is preferably a self inactivating LTR in which nucleotides in the U3 region have been deleted. This can include the TATA box and binding sites for transcription factors. The deletion is transferred to the 5' LTR after reverse transcription in target cells, resulting in the transcriptional inactivation of the LTR in the proviruses. Where the vector comprises two LTRs, both may be SIN LTRs. The U3 region of the 5' LTR may be replaced with a promoter such as the cytomegalovirus (CMV) or Rous Sarcoma Virus (RSV) promoter. This results in Tat-independent transcription but still maintains high levels of expression. As above, the 3' LTR has had nucleotides deleted from the U3 region. SIN LTRs are well known to those skilled in the art (e.g. see Retroviruses. Edited by Coffin JM, Hughes SH, and Varmus HE. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1997). Preferably, where the vector comprises one LTR, the vector comprises two PBSs. One PBS is located towards the 5' end of the vector (5' PBS) and may be positioned precisely on the transcription start site for the promoter which drives transcription of the vector genome. One PBS is located towards the 3' end of the vector, 3' of the LTR (3' PBS). The 3' PBS acts as the site of initiation of minus strand synthesis during the reverse transcription process. The 5' PBS provides homology for plus strand transfer during the reverse transcription process.

Preferably, where the vector comprises two LTRs (a 5' LTR and a 3' LTR), the vector comprises one PBS. The PBS is located towards the 5' end of the vector (5' PBS) and is position 3' of the 5' LTR. The PBS acts as the site of initiation of minus strand synthesis and provides homology for plus strand transfer during the reverse transcription process.

In a particular embodiment, the invention provides a retroviral vector comprising: a 5' primer binding site; a 3' long terminal repeat; an RNA packaging sequence; and either a 3' primer binding site or a 5' long terminal repeat, wherein the RNA packaging sequence is located 3' of the 3 ' long terminal repeat such that the RNA packaging sequence is not reverse transcribed into vector DNA in a target cell.

The vector comprises an RNA packaging sequence which is located 3' of the long terminal repeat such that the RNA packaging sequence is not reverse transcribed into vector DNA in a target cell. In embodiments in which the vector comprises two long terminal repeats, the RNA packaging sequence is located 3' the 3' long terminal repeat (i.e. 3' of both LTRs) such that the RNA packaging sequence is not reverse transcribed into vector DNA in a target cell. The RNA packaging sequence is necessary for the essential process of packaging the retroviral RNA genome into the viral particle as it is assembled by the producer cell. The RNA packaging sequence is able to bind to viral proteins within the nascent viral particle. In some embodiments, the RNA packaging sequence comprises the RNA packaging signal (Ψ). In HIV-1, a portion of the gag gene has found to be involved in RNA packaging. The RNA packaging sequence may also comprise the Rev Response Element (RRE). RNA packaging sequences and the components that make this up are well know to those skilled in the art (e.g. see Retroviruses. Edited by Coffin JM, Hughes SH, and Varmus HE. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1997).

In some embodiments, the vector further comprises an exogenous nucleotide sequence for delivery into a target cell. This exogenous nucleotide sequence may be any sequence which someone might want to insert into a target cell. For example, the exogenous nucleotide sequence may be an expressible transgene, an RNA interference cassette or a molecular barcode, for example, for marking the lineage of different cells. The exogenous nucleotide sequence should be located between the PBS (5' PBS where there are two PBSs) and the LTR (3' LTR where there are two LTRs).

In some embodiments, the exogenous nucleotide sequence an expressible transgene. This transgene is a non-retroviral gene and may be any gene the expression of which is desired in a target cell. The expressible transgene should be located between the PBS (5' PBS where there are two PBSs) and the LTR (3' LTR where there are two LTRs). The transgene may encode for a peptide or protein, or a noncoding RNA. In some embodiments, the transgene encodes for a peptide or protein. Preferably, the peptide or protein should be useful in gene therapy. In some embodiments, the transgene is under the control of a promoter (e.g. a PGK or GAPDH promoter).

Expression of the transgene may aid normal growth of the cell or maintain the health of a subject. In some embodiments, the transgene encodes for a peptide or protein which is absent or underexpressed in a subject. Alternatively, the transgene may encode for a peptide or protein which helps to prevent or ameliorate a medical condition. The peptide or protein may be one which is useful in treating diseases such as cancer, atherosclerosis, sickle-cell anaemia, infection, metabolic disorders, neurological illness and the thalassemias. Examples of such peptides and proteins are haemoglobin, hematopoietic growth factors such as granulocyte-macrophage colony stimulating factor (GM-CSF), macrophage colony stimulating factor (M-CSF), granulocyte colony stimulating factor (G- CSF), erythropoietin (EPO), common gamma chain, Wiskott Aldrich Syndrome protein (WASp), GP91phox, and ABCD1. Another example is tumour necrosis factor (TNF), which is a molecule that can be used to treat cancer, and in particular, tumours. The tumour suppressor p53 and retinoblastoma (RB) are also contemplated. Various cytokines such as mast cell growth factor (MGF) and interleukins 1-11 are also proteins which are contemplated by the present invention. A multidrug resistance gene (mdR) encoding p- glycoprotein is also contemplated as the transgene. The peptide or protein may also be a selectable marker for antibiotic resistance in eukaryotes. Other types of selectable markers such as adenine phosphoribosyl transferase (APRT) in APRT-deficient cells, a fluorescent protein or the firefly luciferase gene are also included. The peptide or protein can be a protein that will provide the host with an additional or altered enzymatic activity, such as the herpes simplex virus thymidine kinase protein for 'suicide therapy' of reactive transplants, or a toxin, such as the diphtheria toxin protein for treatment of cancer. The transgenes encoding these proteins can be provided by any of a variety of methods, such as routine cloning procedures (Sambrook et al. (1989), Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY), excision from a vector containing the gene of interest, or chemical or enzymatic synthesis based on published sequence information. In many instances the DNA encoding the protein of interest is commercially available. In another embodiment, the transgene encodes a protein which enables experimental manipulation of the cell, for example a toxin or a fluorescent or drug-selectable marker. In another preferred embodiment, the transgene is capable of being transcribed into a noncoding RNA molecule. In some embodiments, the transgene may encode a noncoding RNA which can alter the level of expression of genes within the cell or is a component of a ribonucleoprotein complex with enzymatic activity. Examples of such noncoding RNAs are short interfering RNAs (siRNAs), microRNAs, small nucleolar RNAs (snoRNAs), small nuclear RNAs (snRNAs), piwi-interacting RNAs (piRNAs), long noncoding RNAs, transfer RNAs (tRNAs) and ribosomal RNAs (rRNAs). In some embodiments, the transgene may encode a noncoding RNA which is sufficiently complementary to hybridize to an mRNA or DNA of interest. Such an RNA molecule is an antisense RNA, and has utility in preventing or limiting the expression of over-produced, defective or otherwise undesirable molecules or to investigate the function of a gene. The vector of the present invention can comprise, as the transgene, a sequence encoding an antisense RNA which is sufficiently complementary to a target sequence such that it binds to the target sequence. For example, the target sequence can be part of the mRNA encoding a polypeptide such that it binds to and prevents translation of mRNA encoding the polypeptide. In another embodiment, the target sequence is a segment of a gene that is essential for transcription such that the antisense RNA binds the segment (e.g. a promoter or coding region) and prevents or limits transcription. Hence, the antisense RNA must be of sufficient length and complementarily to prevent translation of its target mRNA or transcription of its target DNA. One of ordinary skill in the art can determine antisense molecules having sufficient complementarily to a target sequence such that the antisense molecule is capable of binding to the target and thereby inhibiting translation or transcription. The transgene sequence can be provided, for example, by chemical or enzymatic synthesis, or from commercial sources.

The vector may comprise further elements which help in transduction and expression of the vector. These elements are generally located between the PBS (5' PBS where there are two PBSs) and the LTR (3' LTR where there are two LTRs).

For example, the vector may further comprise a 3' polypurine tract (3 ' PPT). This is a site which is resistant to the RNAseH activity of the retroviral reverse transcriptase when present in RNA. Preferably, this is located 5' of the LTR (the 3' LTR if there are two). The PPT may be located next to the LTR (the 3' LTR if there are two) on the 5' side.

The vector may further comprise a central polypurine tract (cPPT). The vector may further comprise a post-transcriptional regulatory element (PRE) such as a woodchuck hepatitis virus post-transcriptional regulatory element (WPRE). In some embodiments, the vector further comprises a central polypurine tract (cPPT) and a post-transcriptional regulatory element (PRE).

The vector may further comprise a polyadenylation (polyA) signal such as the simian virus 40 (SV40) early or late polyadenylation (polyA) signal. Preferably, the polyA signal is located 3' of the RNA packaging sequence. Preferably, the vector further comprises a SV40 late polyA signal. Preferably, this is located at the 3' end of the vector after the other components such as the RNA packaging sequence. The vector preferably comprises a promoter to drive transcription of the vector genome. This may be any suitable promoter, including a Rous Sarcoma Virus (RSV) promoter, a human cytomegalovirus (CMV) immediate early promoter, a spleen focus forming virus (SFFV) promoter or an HIV-1 U3 promoter. This promoter is preferably positioned at the 5' end of the vector.

Preferably, the primer binding site of the vector is positioned precisely on the transcription start site for the promoter which drives transcription of the vector genome. Where there are two PBSs, the 5' PBS is preferably positioned precisely on the transcription start site for the promoter which drives transcription of the vector genome.

Preferably, a retroviral major splice donor (MSD) site is located near the 5' end of the vector close to the promoter. This MSD may be located 3' of the PBS (the 5' PBS where there are two). This has been found to increase titre levels of the vector.

In one embodiment in which the vector comprises one LTR and two PBSs, the vector comprises, in 5' to 3' direction, the following components:

5' - promoter - PBS - expressible transgene - LTR - PBS - RNA packaging sequence - 3' . In one embodiment in which the vector comprises two LTRs and one PBS, the vector comprises, in 5' to 3' direction, the following components:

5' - promoter - LTR - PBS - expressible transgene - LTR - RNA packaging sequence - 3' .

In one embodiment in which the vector comprises one LTR and two PBSs, the vector comprises, in 5' to 3' direction, the following components:

5' - promoter - PBS - MSD (optional) - cPPT (optional) - expressible transgene - WPRE (optional) - PPT - LTR - PBS - MSD (optional) - RNA packaging sequence - polyA signal (optional) - 3 ' . In one embodiment in which the vector comprises two LTRs and one PBS, the vector comprises, in 5' to 3' direction, the following components: 5' - promoter - LTR - PBS - MSD (optional) - cPPT (optional) - expressible transgene - WPRE (optional) - PPT - LTR - MSD (optional) - RNA packaging sequence - polyA signal (optional) - 3 ' . In another embodiment, there is provided a host cell containing the vector described above. The host cell may be any suitable eukaryotic cell into which the vector may be introduced.

Plasmids encoding the retroviral vectors of the present invention are transfected into suitable host cells (or packaging cells) by standard methods known to one of ordinary skill in the art. Suitable packaging cells are defined herein as cells that contain helper virus sufficient to allow the packaging of RNA transcribed from the retroviral vector and the release of vector virus particles, or virions. Generally additional plasmids encoding transacting viral sequences but lacking the cis-acting sequences required for packaging are co- transfected. These supply the required structural and enzymatic proteins to package and produce the expressed viral backbone RNA. Such packaging cells are known and available to one of ordinary skill in the art, and include, for example, HEK293T cells.

Recombinant retrovirus produced from the transfected cells is harvested by standard methods. The harvested retrovirus, in the form of virions, is used to transduce a permissive target cell by standard techniques. A target cell is defined herein as any cell that is permissive to infection by the virus produced by the retroviral vector of the present invention. The target cell can be in vivo or ex vivo. Representative target cells include, for example, bone marrow stem cells, hepatocytes, muscle cells, tumour cells, neurons, retina and airway epithelial cells. The provirus that is formed in the target cell can then express the transgene. Because the provirus contains no RNA packaging sequence, any endogenous helper proteins present cannot trigger production of an infectious virus from the provirus.

In a further embodiment, there is provided a virion containing the vector described above.

Also provided is a pharmaceutical composition comprising the vector or virion described above. The pharmaceutical composition may further comprise one or more pharmaceutically acceptable excipients. Additionally, there is provided a vector or virion for use in therapy, in particular, gene therapy. A vector or virion for use in delivering a transgene to a subject in gene therapy is also provided.

The invention provides a method of delivering a gene to a target cell, the method comprising administering an effective amount of a vector or virion to the target cell. This method can be used to create a cell line which expresses a gene of interest. For example, the gene can encode for a biotherapeutic so that the cell line produces the bio therapeutic, for example, a therapeutic protein or antibody.

The invention also provides a cell produced by the above method.

Furthermore, there is provided a method of delivering a gene to a target cell in a subject, the method comprising administering an effective amount of a vector or virion to the subject. The subject may be human or animal. Where the subject is human, the gene can be delivered to the subject as part of gene therapy so that the gene is expressed in the subject. Where the subject is an animal, the above method can be used to produce a transgenic animal in which the gene is expressed. The invention also provides a transgenic animal produced by the above method.

The invention will now be described in detail, by way of example only, with reference to the following figures. Brief Description of the Figures

Figure 1. Schematic representation of the HIV-1 genome and an RRL/CCL third generation lentiviral vector system derived from it. U3, unique in 3' region' R, repeat region; U5, unique in 5' region; PBS, primer binding site; Ψ, RNA packaging signal; PPT, polypurine tract; LTR, long terminal repeat; RSV, Rous Sarcoma Virus U3 promoter; CMV, human cytomegalovirus immediate early promoter; Agag, truncated gag sequence; fs, frameshift mutation; RRE, Rev Response Element; cPPT, central polypurine tract; wPRE, woodchuck hepatitis virus posttranscriptional regulatory element; AU3, self- inactivating U3 region; pA, polyadenylation signal; VSV-G, vesicular stomatitis virus glycoprotein.

Figure 2. Schematic of the vector plasmid developed in this invention and schematic of reverse transcription of LTR1 vectors showing that packaging sequences (Gag + RRE) at the 3' end of the RNA genome are not reverse transcribed. SIN LTR, self-inactivating LTR; SV40pA, simian virus 40 polyadenylation signal; GAPDH, human glyceraldehyde 3- phosphate dehydrogenase promoter; eGFP, enhanced green fluorescent protein.

Figure 3. Detailed schematic of third generation lentiviral vector reverse transcription showing that all sequences are reverse transcribed. PGK, human phosphoglycerate kinase promoter.

Figure 4. Detailed schematic of LTR1 reverse transcription showing exclusion of the PBS- Ψ-RRE RNA sequence from the reverse transcript.

Figure 5. Maps of vector plasmids described in this invention. SV40EpA, simian virus 40 early polyadenylation signal; SV40LpA, simian virus 40 late polyadenylation signal; DIS, HIV-1 dimerisation signal; pA ^" , AATAAA to AACAAA mutant polyadenylation signal. AmpR, Ampicillin Resistance gene; ori, bacterial origin of replication.

Detailed Description of the Invention

Summary

Lentiviral transfer vector cis elements such as the HIV-1 RNA packaging signal (Ψ) and Rev Response Element (RRE) are essential for viral RNA genome packaging into virions in producer cells. In standard lentiviral vectors, these cis elements are reverse transcribed into DNA along with the transgene expression cassette and persist in target cells after transduction. This persistence creates several known and potential problems for lentiviral vector gene therapy applications. Splice sites within cis elements have been shown to splice with nearby host genes, creating aberrant fusion transcripts. The CpG island within the RNA packaging sequence undergoes DNA methylation in some target cells, potentially contributing to transgene silencing. The RNA packaging sequence also enables remobilisation of lentiviral vector genomes in cells expressing lentiviral proteins, which could be problematic in HIV-positive patients. Large packaging sequences within the reverse transcript may reduce the size of the transgene cassette which can be accommodated.

In standard lentiviral vectors, essential cis elements are located between the two viral long terminal repeats (LTRs) and so are within the region of the vector that is reverse transcribed. The inventors have developed a novel transfer vector in which the RNA packaging signal and the RRE are located downstream of the 3 ' LTR so are present in the RNA genome during virion packaging but are outside of the region of the genome that is reverse transcribed into DNA in the target cell. These vectors can be produced to high titre (2.6 x 10 ⁸ TU/ml by eGFP flow cytometry of 293Ts, pCCL parallel preparation 1.3 x 10 ⁹ TU/ml) and eGFP expression is maintained to 14 days post-transduction. It is suggested that the use of this configuration to eliminate most of the remaining viral DNA from target cell proviruses could be a feature of the next generation of gene therapy vectors based on HIV-1 and other retroviruses.

Introduction

In existing lentiviral vectors, cis elements such as the RNA packaging signal and the RRE are located between the two viral LTRs and so are within the region of the genome that is reverse transcribed. In the LTRl vector described in this invention, the cis elements are located downstream of the 3' LTR (Fig. 2). These elements are therefore present in the viral RNA genome so that it can be packaged into virions, but are outside of the region of the genome that is reverse transcribed and so are not present in the vector DNA in the target cell. The name "LTRl" was chosen because these vectors contain only one LTR in the transfer vector plasmid rather the two LTRs present in a conventional lentiviral vector.

To understand the mechanism behind this invention it is useful to outline with the classical model of retroviral reverse transcription employed by third generation lentiviral vectors such as RRL or CCL (Fig. 3). In these vectors (as in those from previous generations), the substrate for reverse transcription is a viral RNA genome located in the target cell. A host- derived tRNA molecule bound to the PBS acts as a primer for the initiation of minus strand DNA synthesis by the viral reverse transcriptase enzyme. Once initiated, minus strand synthesis proceeds to the end of the 5' R region. As the minus strand is synthesised, the RNaseH activity of reverse transcriptase degrades the RNA strand of the resulting RNA- DNA double stranded hybrid. Upon reaching the end of the 5' R region, minus strand transfer occurs via homology between the R regions on the minus strand DNA and the 3' end of the RNA genome. Minus strand synthesis continues from this region. RNaseH degradation continues, but a short PPT within the RNA genome immediately upstream of the U3 region is not degraded. The PPT acts as a primer for the initiation of plus strand DNA synthesis. Plus strand synthesis continues into the tRNA primer and stops after the first 18bp due to the presence of a modified RNA nucleotide at the next position in the tRNA which reverse transcriptase cannot use as a template. Plus strand transfer then occurs via homology between the complementary PBS sequences on each DNA strand, and synthesis of both DNA strands continues until a full length double stranded DNA provirus is created.

The LTRl vector described in this invention uses a different mechanism for reverse transcription (Fig. 4). In this vector, the tRNA primer is predicted to be bound to a PBS located closer to the 3' end of the RNA genome. A tRNA molecule bound to the 5' PBS cannot initiate minus strand synthesis because there is no template upstream from this position. After minus strand synthesis initiates at the 3' PBS, it proceeds to the end of the 5' PBS without a minus strand transfer step taking place. Plus strand synthesis initiates at the PPT as before and plus strand transfer occurs via homology between the complementary PBS sequences before synthesis of the double stranded DNA provirus. The 3' end of the RNA genome including the RNA packaging signal and RRE is predicted not to be reverse transcribed as it lacks homology with both the minus and plus strands, so cannot be primed for reverse transcription. Instead it is predicted to be separated from the reverse transcription complex and eventually degraded by target cell RNases.

A number of reports confirm the novel configuration of the LTRl genome is competent for packaging into virions and reverse transcription in target cells. Firstly, it has been shown that the HIV- 1 RNA packaging signal and RRE remain capable of efficient packaging into virions when relocated close to the centre of heterologous RNA molecules (8). Secondly, it has been shown that a PBS flanked by the appropriate local HIV secondary structure remains capable of efficient initiation of minus strand synthesis when relocated to the centre of a lentiviral vector genome (29).

The practical development of a lentiviral vector which employs this novel reverse transcription mechanism to eliminate cis elements such as the RNA packaging signal and the RRE from target cell proviruses is described herein. 1. Materials and Methods

1.1. Plasmids

The parental plasmids for the constructs described in this invention were the pRRL.SIN and pCCL.SIN lentiviral transfer vector plasmids for HrV-1 components, the PGK promoter, the eGFP cDNA and the wPRE (3), the pCI plasmid for the SV40 late polyadenylation signal (Promega), and gene synthesis of the GAPDH promoter (Life Technologies) based on the University of California, Santa Cruz (UCSC) human genome build hgl9 between the primer sequences described in (30). The sequences for all plasmids are provided in the Appendix. Plasmids pCMV-dR8.74 and pMDG2 were generated and distributed by Didier Trono at the Ecole Polytechnique Federale de Lausanne.

1.2. Preparation of lentiviral vectors

1.2 x 10 ⁷ HEK293T (293T) cells were seeded in T175 flasks in 20ml Dulbecco's Modified Eagle's Medium supplemented with 10% fetal calf serum and 1% pencillin-streptomycin (complete DMEM) one day before transfection to reach >90% confluence. For each flask, 50μg vector plasmid, 32.5μ packaging plasmid pCMV-dR8.74, and n.5μg vesicular stomatitis virus envelope plasmid pMDG2 was added to 5ml Optimem (Gibco) and 0.22μιη filtered. Ιμΐ lOmM PEI (Sigma- Aldrich 40872-7) was added to 5ml Optimem and 0.22μπι filtered. The two mixtures were combined and allowed to complex for 20min. The complex was added to the T175 flask containing 293T cells and incubated at 37°C with 5% C0 ₂ for 4 hours. The complex was removed and replaced with 20ml complete DMEM. 24 hours later, the medium was replaced. 48 hours after transfection the medium was removed, 0.22μπι filtered, and centrifuged at 50,000g for 2 hours. Virus pellets were resuspended in 150μ1 Optimem and aliquots were stored at -80°C. 1.3. Titration of lentiviral vectors by flow cytometry

10 ⁵ 293T cells were seeded in 250μ1 of complete DMEM into each well of a 24-well plate one day before transduction. For transduction, a 5-fold dilution series of virus was performed in 50μ1 aliquots of Optimem. Each aliquot of diluted virus was added to the medium in one well. Transgene expression was measured by flow cytometry 14 days post- transduction. Expression titre was calculated by selecting the well in which 5-15% of transduced cells expressed the transgene of interest and dividing the number of transduced cells in this well by the volume of virus used to transduce them.

1.4. Plasmid rescue

Briefly, HEK 293T cells were transduced at high MOI with LTR1.20/AmpR-ori, containing an Ampicillin resistance gene and bacterial origin of plasmid replication, and left to culture for 1 week before recovery of cell pellets. Genomic DNA was extracted using a commercial kit (Qiagen) and quantified by Nanodrop. 10μg of gDNA was digested with Xbal restriction endonuclease, which does not target any sequence within the provirus, but exists within the human genome at an incidence of approximately 279.3 sites per megabase pair. The digested gDNA was then column purified before ligation to circularise any released lentiviral backbone fragments, followed by transformation of electro-competent Escherichia, coli. Individual bacterial colonies were selected and grown in preparation for sequence analysis; these would only grow if they received a re- circularised copy of the vector backbone containing AmpR and ori. Recovered DNA was sequenced using the primers targetting the RRE region (which should be omitted from the final provirus), the bacterial origin of replication and the ampicillin resistance gene.

2. Results

2.1. Optimisation of the genomic RNA transcription cassette

The production of infectious lentiviral virions requires co-assembly of the viral genomic RNA with viral Gag, Gag-Pol and envelope proteins. Previous reports of codon-optimised Gag-Pol expression plasmids improved the intracellular expression of these proteins but did not result in higher titres (21;31). It is likely that expression of the viral RNA genome is generally the limiting factor for viral titre in many vector preparation protocols. Therefore, in order to maximise the titre obtainable with the LTR1 vector experiments were undertaken to optimise the elements used to drive viral RNA genome transcription.

Efficient polyadenylation is associated with higher steady-state levels of RNA polymerase II (Pol II) transcripts, and it has been shown that the simian virus 40 (SV40) late polyadenylation (polyA) signal acts as a stronger polyadenylation signal than the SV40 early polyA signal (32). The inventors therefore generated two LTR1 configurations: LTRl.O/PGK-eGFP-WPRE (LTR1.0/PEW) uses the SV40 early polyA signal present in the parental RRL backbone while LTR1.5/PEW uses an SV40 late polyA signal derived from pCI (Fig. 5). Lentiviral vector was prepared in parallel using these two constructs and a conventional RRL/PEW construct and titred on 293T cells by flow cytometry. The vector titres were calculated as LTR1.0 PEW 2.3 x 10 ⁴ transducing units/ml (TU/ml), LTR1.5 PEW 4.5 x 10 ⁴ TU/ml, RRL PEW 6.9 x 10 ⁷ TU/ml. As the SV40 late polyA signal yielded a higher titre, this element was used in all subsequent LTR1 constructs. Continued eGFP expression 14 days after transduction and following at least four passages of dividing cells suggests that the LTR1 vectors are competent for integrase-mediated chromosomal integration.

High steady-state levels of Pol II transcripts can also be achieved through the use of a strong promoter, so constructs were generated in which the Rous Sarcoma Virus (RSV) promoter derived from the parental RRL backbone was replaced by a human cytomegalovirus immediate early (CMV) promoter derived from a CCL plasmid. In order to optimise the alignment of the CMV promoter and the LTR1 5' PBS, three constructs were generated in which the PBS was positioned precisely on the reported transcription start site for the CMV promoter (33) (LTR1.7.672/PEW), 1 bp upstream (LTR1.7.671/PEW) or lbp downstream (LTR1.7.673/PEW). Lentiviral vector was prepared in parallel from these constructs, LTR1.5/PEW and RRL/PEW and titred on 293T cells by flow cytometry. The replacement of the RSV promoter with the CMV promoter resulted in a 25-fold increase in LTR1 vector titre (LTR1.5 PEW 2.3 x 10 ⁴ TU/ml, LTR1.7.671 PEW 5.6 x 10 ⁵ TU/ml, LTR1.7.672 PEW 5.4 x 10 ⁵ TU/ml, LTR1.7.673 PEW 4.7 x 10 ⁵ TU/ml, RRL PEW 5.8 x 10 ⁷ TU/ml). As the LTR1.7.671 CMV configuration yielded the highest titre, this promoter and alignment were used in subsequent LTR1 constructs. 2.2. Effect of translation of LTR1 transfer vector RNA

Lenti viral RNA genomes acquire a 5' 7-methylguanylate cap (m7G) cap during transcription by PolII and hence are competent for translation by producer cell ribosomes. In the scanning model of translation initiation in mammalian cells, the ribosome loads onto the 5' cap and scans in a 5 '-3' direction until it reaches the 5 '-most ATG codon, at which point translation can initiate. In RRL and CCL RNA genomes the 5 '-most ATG occurs at the start of the truncated Gag element and is followed by a 21 codon open reading frame. In the LTR1 vector configuration, the 5 '-most ATG will be found within the transgene expression cassette, probably at a cryptic ATG within the internal promoter.

For example, in the LTR1.7.671 PEW vector the 5'-most ATG occurs close to the middle of the PGK promoter and is in frame with the downstream eGFP coding sequence with no intervening stop codons. Since LTR1.7.671 PEW was found to yield vector titres around 10 ² -fold lower than an RRL PEW vector it was hypothesised that this cryptic translation product might be interfering with virion production in producer cells.

To test this hypothesis the PGK promoter was replaced with the human GAPDH promoter which lacks ATG codons entirely. In LTR1.7.671 GAPDH-eGFP-wPRE (LTR1.7.671 GEW) the 5 '-most ATG is at the start of the eGFP open reading frame, so translation of full length vector RNA should produce eGFP only. Lentiviral vector was prepared from RRL PEW, RRL GEW, LTR1.7.671 PEW and LTR1.7.671 GEW and titred on 293T cells by flow cytometry. The resulting titres were RRL PEW 6.4 x 10 ⁷ TU/ml, RRL GEW 1.6 x 10 ⁸ TU/ml, LTR1.7.671 PEW 7.1 x 10 ⁵ TU/ml and LTR1.7.671 GEW 7.6 x 10 ⁵ TU/ml. Cryptic translation initiation does not appear to have been the limiting factor in this LTR1 vector preparation, but the GAPDH promoter was used in subsequent constructs as the mean fluorescence intensity indicates that it is a stronger promoter than the PGK promoter so is a better reporter of successful transduction. 2.3. Preventing cleavage and polyadenylation within the internal LTR

The SIN LTR located near the midpoint in the LTR1 vector configuration contains the HIV-1 polyadenylation signal. Although a previous report in which the HIV-1 LTR was relocated to the midpoint of a vector reported efficient full length transcription (8), it was decided to investigate the possibility that polyadenylation within the SIN LTR of an LTR1 vector might be preventing production of full length RNA genomes and hence causing the observed reduction in titre relative to third generation lentiviral configurations. In order to prevent cleavage and polyadenylation at the LTR polyA signal, the AATAAA motif which binds cleavage and polyadenylation specificity factor (CPSF) was mutated to AACAAA in construct LTRl.11.0 GEW. This mutation was previously reported to abolish cleavage and polyadenylation at the HIV-1 polyA signal (34). As a functional polyA signal is required for reporter gene expression in target cells, a fragment of the HIV-1 genome covering the region from the 5' R to the major splice donor (MSD) stem loop was introduced at the 5' end of this construct and the 3' PBS was deleted. This vector is predicted to initiate minus strand synthesis at the 5' PBS and undergo minus strand transfer and subsequent stages of reverse transcription in the same way as a third generation lentiviral vector so that the 3' mutated polyA signal is replaced by the functional 5' polyA signal. Construct LTRl.11.1 GEW is the same as LTRl.11.0 except that it retains a functional AATAAA motif.

Lentiviral vector prepared from LTRl.11.0 GEW and LTRl.11.1 GEW yielded titres of 2.5 x 10 ⁸ TU/ml and 1.7 x 10 ⁸ TU/ml respectively compared to a CCL GEW parallel control titre of 1.3 x 10 ⁹ TU/ml. This suggests that the midpoint polyA signal is not efficiently used in an LTR1 configuration. It is possible that the MSD located downstream of the LTR is able to block polyadenylation at this site as occurs at the 5' polyA signal in a conventional lentiviral vector (35). 2.4. Effects of minus strand transfer and a 5' major splice donor

The greatly improved titres observed with the LTR1.11 configuration suggested two possible explanations. Firstly, initiation of minus strand synthesis at the 5' PBS and/or minus strand transfer might improve the efficiency of reverse transcription. Secondly, the presence of the MSD at the 5' end of the vector might improve the efficiency of RNA genome transcription. In order to test these hypotheses construct LTR 1.13.0 GEW was generated which resembles LTR1.7.671 except that the 5' end of the RNA includes the full HIV-1 genome between the PBS and the MSD stem loop instead of only the 18bp PBS. In this construct initiation of minus strand DNA synthesis is predicted to take place at the midpoint of the vector. Construct LTR 1.13.1 GEW is identical to LTR 1.13.0 GEW except that the sequence between the 5' PBS and the MSD stem loop has been deleted.

Lentiviral vector was prepared from LTR1.13.0 GEW, LTR1.13.1 GEW and LTR1.7.671 GEW and titred by flow cytometry. These constructs yielded titres of 2.6 x 10 ^s TU/ml, 9.2 x 10 ⁷ TU/ml and 3.2 x 10 ⁷ TU/ml respectively compared to a CCL GEW parallel control titre of 1.3 x 10 ⁹ TU/ml. These results suggest that the elimination of minus strand transfer from reverse transcription has no negative impact on viral titre and that the large increase in titre observed with the LTR1.11 configuration was due to the presence of the full sequence between the 5' PBS and the MSD stem loop. It has been previously reported that splice sites close to the promoter activate transcription of mammalian genes (36) and previous attempts to mutate the major splice donor in lentiviral vectors resulted in reduced vector titres (20;21). Therefore it appears that the presence of splicing factors close to the 5' end of the vector RNA genome is required for the production of higher titre lentiviral vectors in both a conventional and LTR1 configuration.

In order to mimic the transcriptional activating effect of the MSD in LTR-1 vectors while reducing HIV sequence, it was replaced by a chimeric intron from the pCI expression plasmid, inserted between the PBS and cPPT, to produce pLTRl .20/GAPDH-eGFP- WPRE. As the wild-type 5' LTR has been removed from LTR-1 the HIV MSD no longer functions to regulate the production of subgenomic RNAs in this vector and it can be exchanged. The use of the strong heterologous intron from pCI offers advantages over the inclusion of the MSD given that no flanking splice enhancers are required and it also includes a splice acceptor which facilitates its removal from viral RNA in producer cells.

Addition of the pCI chimeric intron in the pLTR1.20/GAPDH-eGFP-WPRE vector led to a 3-fold improvement in titre, as calculated by FACS.

2.5 Full-length PCR and sequencing of pLTRl .7.671/GAPDH-eGFP- WPRE provirus To demonstrate the structure of reverse transcribed proviral DNA from the pLTR1.7.671/GAPDH-eGFP-WPRE vector, HEK293T cells were transduced at a multiplicity of infection of 10. Genomic DNA was extracted 1 week after transduction. The full-length provirus was amplified by PCR using the oligos designated below to give a 2.1kb amplicon, which was extracted following separation on a 1% agarose gel and TA- cloned (Lifetechnologies) prior to sequencing.

U5 forward: 5' GGTAACTAGAGATCCCTCAGACCC 3' (SEQ ID NO. 1)

U3 reverse: 5' CGTTGGGAGTGAATTAGCCCTTCC 3' (SEQ ID NO. 2)

Sequence analysis showed that the provirus contained the correct 5' and 3' LTRs following reverse transcription, displaying the expected sequence with the gag-RRE region removed, when the provirus was examined.

2.6 Examination of integrated LTR-1 provirus sequences by 'plasmid rescue'

In order to stringently investigate the sequences of LTR-1 proviruses and confirm the absence of the deleted HIV-1 packaging sequences, a technique was employed in which the ampicillin resistance marker and bacterial origin of replication were removed from the pLTR1.20 plasmid backbone and inserted within the transgenic region, between the LTRs. This enables pLTR1.20/AmplR-Ori proviral DNA to be excised from transduced HEK 293T cells, recircularised and transformed into Escherichia. coli bacteria which would replicate and form colonies upon uptake of a vector genome containing an ampicillin resistance gene (AmpR) and bacterial origin of replication (ori) between the LTRs. Propagation of proviral sequences in bacteria allows the reverse transcribed sequence to be studied in detail. Sequencing of the proviral DNA rescued in bacteria reads through the LTR and into a region within, host cell chromosomal DNA which was identified by BLAT search. The internal sequence of the provirus confirmed the mechanism of reverse transcription did result in the expected structure as shown in Fig. 4. It was not possible to sequence the RRE element, suggesting that it was absent from the integrated provirus as predicted.

2.7. In vivo function of pLTR1.20/SFFV-eGFP-WPRE

To demonstrate LTR-1 vector function in vivo, 1 day old neonatal CD-I mice were injected intravenously with 4.5xl0 ⁵ vector particles of LTR1.20/SFFV-eGFP-WPRE. Mice were sacrificed and livers were dissected 1 week after vector administration and imaged revealing GFP -positive cells visible in the livers of injected animals demonstrating that LTR-1 is capable of gene deliver to cells in vivo. 3. Discussion

A number of different retroviruses have been developed as retroviral vectors for mammalian gene transfer (1-5). The largest change which takes place during vectorisation of a retrovirus is the separation of viral components into cis elements which must remain present on the nucleic acid being transferred into the target cell and trans elements which need only be provided in producer cells to enable production of infectious particles. Examples of retroviral cis elements include the HIV-1 RNA packaging signal (Ψ) and Rev Response Element (RRE) which must remain covalently attached to the transgene expression cassette in order for it to be packaged into virions, while trans elements include coding sequences for the Gag and Gag-Pol polyproteins and the viral envelope glycoprotein.

As with other retroviral vectors, the cis elements within HIV-1 -based lentiviral vectors are located between the viral long terminal repeats (LTRs) so are reverse transcribed and remain associated with the transgene expression cassette within the target cell. The presence of these cis elements within target cell proviruses creates a number of empirically observed and theoretically possible problems for the practical application of lentiviral vectors, particularly for gene therapy. The cis elements contain splice sites able to splice with and dysregulate host genes involved in the control of cell proliferation (14), CpG islands able to undergo DNA methylation (16), large untranscribed regions associated with transcriptional silencing of episomal DNA molecules (17), and RNA packaging signals able to mediate vector remobilisation (15). Lentiviral cis elements also occupy up to 2kb of the reverse transcript which may reduce the size of transgene expression cassettes which can be delivered with these vectors.

In this report, an approach to eliminating HIV-1 -derived cis elements from the DNA delivered to target cells is described. In the novel LTR1 configuration the cis elements are located downstream of the 3' LTR. These elements are therefore present in the viral RNA genome so that it can be packaged into virions, but are outside of the region of the genome that is reverse transcribed and so are not present in the vector DNA in the target cell. In the first configuration tested, transcription of the LTR1 RNA genome was driven by an RSV promoter and SV40 early polyA signal cassette. This configuration produced functional vector titres 3000-fold lower than a conventional RRL vector. A modified cassette using the CMV promoter and an SV40 late polyadenylation signal improved titres significantly to approximately 100-fold lower than a conventional RRL vector.

By relocating the HIV-l-derived 5' leader sequences the new configuration raised the possibility of aberrant translation products originating from within the transgene cassette. It was suggested that aberrant translation could be reducing the vector titres obtained with these constructs. The inventors addressed this possibility by replacing the PGK promoter which contains a cryptic start codon with a GAPDH promoter which lacks ATG codons. No difference in titre was observed.

Another potential problem for vector titre is the presence of a functional polyadenylation signal within the LTR located near the midpoint of the vector. Termination of transcription at this site would prevent production of full length transfer vector RNA competent for packaging into virions. A point mutation was introduced to the AATAAA hexamer which was previously shown to abolish cleavage and polyadenylation at the HIV-1 polyA signal but observed no increase in titre. The inventors hypothesise that the HIV-1 major splice donor (MSD) located downstream of the LTR in the LTR1 configuration is able to block cleavage and polyadenylation at this site just as it does in wildtype HIV-1 and conventional lentiviral vectors (35).

The inventors then reintroduced sequences between the PBS and the MSD stem loop to the 5' end of the LTR1 genomic RNA and observed a significant increase in titre to within 5- fold of a CCL-based construct. The presence of splicing factors close to the 5' end of the vector genomic RNA appears to aid the production of high titre lentiviral vectors in both the conventional and LTR1 configurations but is not essential (20;21). The LTR1 vector configuration described in this report could potentially replace all third generation HIV-1 -based lentiviral vectors in the applications in which they are currently used. Furthermore, the mechanism of reverse transcription is highly conserved among retroviruses so an LTRl configuration could be applied to vectors based on retroviruses other than HIV-1.

References

(1) Suerth JD, Maetzig T, Galla M, Baum C, Schambach A. Journal of Virology 2010 Jul l;84(13):6626-35.

(2) Hildinger M, Abel KL, Ostertag W, Baum C. Journal of Virology 1999 May l;73(5):4083-9.

(3) Zufferey R, Dull T, Mandel RJ, Bukovsky A, Quiroz D, Naldini L, et al. Journal of Virology 1998 Dec l;72(12):9873-80.

(4) Jarraya B+, Boulet S, Scott Ralph G, Jan C, Bonvento G, Azzouz M, et al. Science Translational Medicine 2009 Oct 14;l(2):2ra4.

(5) Heinkelein M, Dressier M, Jirmy G, Rammling M, Imrich H, Thurow J, et al.

Journal of Virology 2002 Apr 15;76(8):3774-83.

(6) Burns JC, Friedmann T, Driever W, Burrascano M, Yee JK. Proceedings of the National Academy of Sciences 1993 Sep l;90(17):8033-7.

(7) Malim MH, Hauber J, Le SY, Maizel JV, Cullen BR. Nature 1989 Mar 16;338(6212):254-7.

(8) Cockrell A, van Praag H, Santistevan N, Ma H, Kafri T. Retrovirology 2011;8(1):51.

(9) Sirven A, Pflumio F, Zennou V, Titeux M, Vainchenker W, Coulombel L, et al.

Blood 2000 Dec 15;96(13):4103-10.

(10) Higashimoto T, Urbinati F, Perumbeti A, Jiang G, Zarzuela A, Chang LJ, et al.

Gene Ther 2007 Jun 28; 14(17): 1298-304.

(11) Cesana D, Sgualdino J, Rudilosso L, Merella S, Naldini L, Montini E. J Clin Invest 2012 May l;122(5): 1667-76.

(12) Moiani A, Paleari Y, Sartori D, Mezzadra R, Miccio A, Cattoglio C, et al. J Clin Invest 2012 May 1; 122(5): 1653-66.

(13) Almarza D, Bussadori G, Navarro M, Mavilio F, Larcher F, Murillas R. Gene Ther 2011 Jul;18(7):674-81.

(14) Cavazzana-Calvo M, Payen E, Negre O, Wang G, Hehir K, Fusil F, et al. Nature 2010 Sep 16;467(7313):318-22.

(15) Hanawa H, Persons DA, Nienhuis AW. Journal of Virology 2005 Jul 1;79(13):8410-21. (16) He J, Yang Q, Chang LJ. Journal of Virology 2005 Nov l;79(21): 13497-508.

(17) Lu J, Zhang F, Xu S, Fire AZ, Kay MA. Mol Ther 2012 Nov;20(l l):2111-9.

(18) Kumar M, Keller B, Makalou N, Sutton RE. Hum Gene Ther 2001 Oct 10;12(15):1893-905.

(19) Descours B, Cribier A, Chable-Bessia C, Ayinde D, Rice G, Crow Y, et al.

Retrovirology 2012 Oct 23;9:87. doi: 10.1186/1742-4690-9-87.:87-9.

(20) Cui Y, Iwakuma T, Chang LJ. Contributions of Viral Splice Sites and cis- Regulatory Elements to Lentivirus Vector Function. Journal of Virology 1999 Jul l;73(7):6171-6.

(21) Kotsopoulou E, Kim VN, Kingsman AJ, Kingsman SM, Mitrophanous KA. Journal of Virology 2000 May 15;74(10):4839-52.

(22) Koldej R, Anson D. BMC Biotechnology 2009;9(1):86.

(23) Delviks KA, Pathak VK. Journal of Virology 1999 Oct l;73(10):7923-32.

(24) Srinivasakumar N. Peer J 2013 Jun 4;l:e84. doi: 10.7717/peerj.84.

Print;%2013.:e84.

(25) Torne-Celer C, Moreau K, Faure C, Verdier G, Ronfort C. Intervirology 2008;51(6):447-52.

(26) Torne-Celer C, Moreau K, Faure C, Verdier G+, Ronfort C. Arch Virol 2008;153(12):2233-43.

(27) Luche RM, Enssle J, Kiem HP. Sci Rep 2012 Jun 1;2.

(28) Fang Y, Gong X, Xu M, Zeng F, Zhang J. J Gene Med 2013 Feb 1;15(2): 102-12.

(29) Voronin YA, Pathak VK. Journal of Virology 2004 May 15;78(10):5402-13.

(30) Aki T, Yanagisawa S, Akanuma H. Journal of Biochemistry 1997 Aug l;122(2):271-8.

(31) Fuller M, Anson DS. Hum Gene Ther 2001 Nov;%20;12(17):2081-93.

(32) Carswell S, Alwine JC. Molecular and Cellular Biology 1989 Oct l;9(10):4248-58.

(33) Isomura H, Stinski MF, Kudoh A, Nakayama S, Murata T, Sato Y, et al. Journal of Virology 2008 Jan 15;82(2):849-58.

(34) Perkins KJ, Lusic M, Mitar I, Giacca M, Proudfoot NJ. Molecular Cell 2008 Jan 18;29(l):56-68.

(35) Ashe MP, Pearson LH, Proudfoot NJ. EMBO J 1997 Sep 15;16(18):5752-63. Furger A, Binnie JMO, Lee BA, Proudfoot NJ. Genes & Development 2002 Nov l;16(21):2792-9.

Appendix

Construct sequences

The constructs have the sequences according to the sequence identifier below. The actual sequence of each construct is provided in the accompanying sequence listing. Features of these constructs are also shown in Figure 5. a) pRRL/PGK-eGFP-WPRE - SEQ ID NO. 3.

b) pRRL/GAPDH-eGFP-WPRE - SEQ ID NO. 4.

c) pCCL/GAPDH-eGFP-WPRE - SEQ ID NO. 5.

d) pLTRl.O/PGK-eGFP-WPRE - SEQ ID NO. 6.

e) pLTR1.5/PGK-eGFP-WPRE - SEQ ID NO. 7.

f) pLTR1.7.671/PGK-eGFP-WPRE - SEQ ID NO. 8.

g) pLTR1.7.672/PGK-eGFP-WPRE - SEQ ID NO. 9.

h) pLTR1.7.673/PGK-eGFP-WPRE - SEQ ID NO. 10.

i) pLTR1.7.671/GAPDH-eGFP-WPRE - SEQ ID NO. 11.

j) pLTRl .11.0/GAPDH-eGFP-wPRE - SEQ ID NO. 12.

k) pLTRl.l l. l/GAPDH-eGFP-WPRE - SEQ ID NO. 13.

1) pLTR1.13.0/GAPDH-eGFP-WPRE - SEQ ID NO. 14.

m) pLTR1.13.1/GAPDH-eGFP-WPRE - SEQ ID NO. 15.

n) pLTR1.20/GAPDH-eGFP-WPRE - SEQ ID NO. 16.