WO2007091066A1 - Applications of non-integrating lentiviral vectors - Google Patents

Abstract

This invention provides a method of gene therapy comprising administering to a subject an integration-deficient lentiviral vector containing a nucleic acid that provides benefit.

Description

Applications of non-integrating lentiviral vectors

The invention relates to uses of lentiviral vectors, particularly for gene therapy.

Introduction

With the increase in knowledge of the genome of humans and other species, intervention in the expression of genes for therapeutic or other purposes has become a common goal in human and animal medicine. Various diseases are known or postulated to be caused by deleterious mutations in specific genes. Other diseases result from over-expression or other aberrant expression of particular genes. Treatment of such diseases by correcting, complementing or otherwise affecting gene expression frequently involves delivering DNA into a target cell in vivo. Various approaches have been proposed for in vivo DNA delivery.

Existing methods for delivery of DNA fall into two general categories: viral vectors or non- viral vectors. Non- viral vectors are generally compositions comprising DNA and particular surfactants, Viral vectors are viruses that have been manipulated to include the DNA of interest. Non-viral vectors have the advantage that they do not carry risks of aberrant viral activity in the body of the recipient. They are, however, generally less effective than viral vectors. Viral vectors are generally based on viruses that have in some way been inactivated or rendered non-infective. There remains a need for further, improved approaches to DNA delivery.

One class of viral vectors that have been used for in vitro gene transfer are lentiviral vectors.

Lentiviruses include human immunodeficiency virus (HIV) and feline immunodeficiency virus (FIV). They have been shown to mediate highly efficient transgene expression in a variety of animal tissues, However, the process of vector integration into host cell chromosomes carries with it a finite chance of causing insertional mutagenesis ' . This risk has been highlighted by the reproducible induction of malignancy in murine models, and development of lymphoproliferative disease in three severe combined immunodeficiency (SCID)-Xl patients treated with a gammaretroviral vector Insertional mutagenesis can be avoided by the use of gene therapy vectors that do not integrate.

Therefore, a key challenge for clinical therapies based on retroviral vectors is to achieve stable transgene expression whilst minimizing insertional mutagenesis. Retroviruses and retroviral vectors can be rendered integration-defective by mutations in the integrase coding sequence. Class II integrase mutations have pleiotropic effects, but class I mutations result in normal DNA synthesis, integration failure and accumulation in the cell nucleus as double- stranded DNA circles⁷. Two circular forms are produced from the unintegrated linear double- stranded DNA, presumed to be the result of intramolecular homologous recombination o n between the LTRs (1-LTR circle) or intramolecular end-joining (2-LTR circle) ' . These circles lack replication signals and in cultured cells dilute as a consequence of cell division¹⁰'¹¹.

Class I integration-deficient lentiviral vectors have been reported to be inefficient for transduction of dividing cells in vitroⁿ'ⁿ, though prolonged gene expression has been demonstrated in vitro in cultured non-dividing primary rat neurons and growth-arrested fibroblasts, and in SV40 T antigen-expressing proliferating cells through the incorporation of episomal replication signals. Recent in vitro studies have shown that integration-deficient lentiviral vectors can mediate stable transduction through persistence of episomal double- stranded DNA circles⁴'⁵'⁶ In contrast, direct administration of non-integrating vectors based on human immunodeficiency virus (HIV) or feline immunodeficiency virus (FIV) has not previously resulted in significant expression in v/vo^14"16. Naldini et al ^H demonstrated efficient transfer, integration and long-term expression of a lentiviral vector that was integration-proficient. In a control experiment, an integration-deficient vector was used (with D64V mutation in the integrase). That integration-deficient vector resulted in transduction in vitro to less than 2%

of that of the wild type. A concentrated stock of the vector with the mutant enzyme and β-gal

reporter gene was produced and it was injected into the brain of rats. In all injected sites, examined after 2 or 6 weeks, from none to a couple of rare cells could be detected expressing

β-gal. The authors concluded that expression of the transgene in vivo depends on its

integration.

Similarly, Loewen et at. ¹⁵ carried out experiments using feline immunodeficiency virus (FIV)

vectors. The vectors carried the β-gal reporter gene and they were injected in rats' eyes. The

wildtype integrase FIV vectors were compared with vectors with a class I mutation in the integrase. It was found that, in contrast to the wild-type vector, the integrase-mutant vector brought about transgene expression in no or only very rare cells.

Park et ø/.¹⁶ investigated the potential use of lentiviral vectors for the in vivo production of coagulation factors in liver cells. It was found that mice that received lentiviral particles with a functional integrase gene and a human Factor IX transgene produced therapeutic levels of Factor IX. However, no Factor IX was detected in mice that received integration deficient vector. It was thus concluded that lentiviral integration was essential for Factor IX expression. In contrast to the findings by Naldini, Loewen and Park, the present inventors have now surprisingly demonstrated that efficient and sustained transgene expression can be achieved in vivo in rodent ocular and brain tissues using integration-deficient lentiviral vectors. The inventors have demonstrated substantial rescue of clinically relevant rodent models of retinal degeneration by gene transfer. At the time of filing it is not clear why Naldini, Loewen and Park failed to achieve transduction.

Summary of the invention The invention provides a method of gene therapy comprising administering to a subject an integration-deficient lentiviral vector containing a nucleic acid that provides benefit.

In particular the invention provides a method of gene therapy comprising administering to a subject an integration-deficient lentiviral vector containing a nucleic acid, wherein the subject has a deficiency arising from aberrant gene expression or a defective gene and the nucleic acid complements or corrects the deficiency.

The invention further provides a method of gene therapy comprising administering to a subject an integration-deficient lentiviral vector containing a transgene wherein the subject has a deficiency or disease that is treatable by expression of a transgene. In particular, the invention may provide a method of treating a deficiency or disease where the transgene encodes a protein that, when expressed in the appropriate location corrects the deficiency or disease. The invention further provides a method of gene therapy comprising administering to a subject an integration-deficient lentiviral vector containing a nucleic acid, wherein the nucleic acid provides a prophylactic benefit to the subject. In particular, the invention may provide a method of gene therapy comprising administering to a subject an integration-deficient lentiviral vector containing a nucleic acid, wherein the nucleic acid encodes a protein that can vaccinate the subject against a particular disease.

The invention further provides a method of gene therapy comprising administering to a subject an integration-deficient lentiviral vector containing a nucleic acid, wherein the nucleic acid provides a genetic enhancement to said mammal.

The invention also provides a method for in vivo transfer of nucleic acid of interest into a target cell comprising bringing an integration-deficient lentiviral vector containing the nucleic acid of interest into contact with the target cell. In particular, the invention provides a method for in vivo expression of a transgene of interest in a target cell comprising bringing an integration-deficient lentiviral vector containing the transgene of interest into contact with the target cell.

The invention further provides a plasmid that mimics the structure of a circular lentiviral genome.

The invention enables high efficiency of nucleic acid transfer and gene expression mediated by lentiviruses to be harnessed in vivo without a requirement for vector integration. For therapeutic application to post-mitotic tissues, this system substantially reduces the risk of insertional mutagenesis. The nucleic acid of interest may include a gene which it is desirable to express in the target cell, for example because the copy of the gene in question carries a mutation. In that case, the nucleic acid of interest preferably includes various elements necessary to bring about efficient gene expression in the target cell.

Alternatively, the nucleic acid of interest may be transferred into the target cell for other reasons. For example, the nucleic acid of interest may be nucleic acid that can interact with or integrate into the target cell's genomic DNA or RNA in a specific fashion. For example, the nucleic acid may have a sequence that enables it to recombine with the genomic DNA in a targeted fashion and correct or introduce a mutation in a gene or its control sequences; or create a silencing mutation in a gene or its control sequences; or integrate in a genomic region permissive for transgene expression and with no risk of leading to tumourigenesis.

Brief description of the Figures

Figure 1 is a schematic representation of linear, integrated and circular forms of a generic lentiviral vector.

Figure 2 is a schematic representation of HIV-I vectors used in the Examples. Figure 3 shows that integration-deficient lentiviral vectors mediate transient transduction in cultured proliferating cells:

Figure 3a shows a time-course of eGFP expression after transduction of HeLa cells with integration-deficient and -proficient L-cSegfpW vector at the indicated MOI.

Figure 3b shows a schematic representation of linear, integrated and circular forms of

L-cSegfpW. Figure 3 c shows Southern blot analyses of vector forms and integration. Figure 4 shows In vivo eGFP fluorescence at various time points after subretinal injection with the indicated vectors.

Figure 5 shows eGFP fluorescence in cryosections of mouse retina and cornea six months after vector administration. Figure 6 shows GFP expression from integration-deficient vectors in rat RPE. Figure 7 shows transduction of human RPE with integration-deficient vectors. Figure 8 shows dose-response of in vivo eGFP expression after subretinal vector injection. Figure 9 shows efficient eGFP expression from integration-deficient vectors in mouse brain:

Figure 9a is a low power micrograph showing eGFP fluorescence after injection into the corpus striatum (st).

Figure 9b is a higher power image showing densely packed fluorescent cell bodies in the striatum.

Figure 9c shows morphological identification of most of the cell bodies expressing eGFP in the striatum as neurons. Figure 9d shows a blood vessel in the striatum, approximately 1 mm from the injection site.

Figure 9e shows the spread of the vector in the plane of white matter.

Figure 9f shows transduced neurons near an injection site in the hippocampus. Figure 10 relates to functional rescue of ocular function in Rpe65^{x rdl2} mice: Figure 10a shows subretinal delivery of L-cSrpe65W results in efficient RPE65 expression in mouse RPE.

Figure 10b shows sample scotopic ERG traces at 10, 100 and 1000 mcds/m² flash intensities from Rpe65'^{d uMn} eyes (3 weeks post-injection).

Figure 10c shows mean ERG b-wave amplitudes of the same paired groups as in Figure 10b. Figure 1Od shows improvements in ERG waveform and b-wave amplitude in Rpe65^{rά uMn} mice injected with integration-deficient or -proficient L-cSrpe65W (8 weeks post-injection).

Figure 11 shows functional rescue of ocular function in RCS rats: Figure 1 1a shows Sample scotopic ERG traces at 10, 100 and 1000 mcds/m² flash intensity from RCS eyes

Figure 1 Ib shows mean ERG b-wave amplitudes of the same paired groups. Figure 12 relates to analyses of HIV vector integration in vivo:

Figure 12a shows qPCR quantitation of 2-dLTR junction-containing molecules in transduced eyecups.

Figure 12b shows the strategy for detection of vector-host DNA junctions by LAM- PCR.

Detailed description The invention provides methods for cell transduction and gene expression using integration- deficient lentiviral vectors. Lentiviral vectors have particular advantages because of their efficiency of cell transduction, including post-mitotic cells, and paucity of immune recognition¹²'²⁶. However, previous studies^14"16 have concluded that integration-deficient lentiviral vectors are not effective in bringing about transgene expression in vivo. The invention is based on the inventors' surprising finding that, under particular conditions, integration-deficient HIV-I -based vectors are able to mediate sustained gene expression in vivo, and have the capacity to rescue representative models of clinical disease. Compared to other non-integrating vectors, perhaps the main potential advantages of integration-deficient lentiviral vectors are their greater transgene capacity (with respect to recombinant adeno- associated virus (rAAV) vectors) and their low immunogenicity. The method of the invention uses integration-deficient lentiviral vectors. Lentiviral vectors include the immunodeficiency viruses derived from human (HIV-I³² and HIV-2³³), feline

(FIV^" ), simian (SIV ^" ) or bovine (BIV^jδ) origins, the caprine arthritis encephalopathy virus (CAEV^j7). the equine infectious anemia virus (EIAV³⁸'³⁹ ) and the Jembrana disease virus

(JDV⁴⁰) of bovine origin. A full taxonomy of lentiviruses can be found at http://www.ncbi.nlm.nih. gov/ICTVdb/Ictv/index.htm . Lentiviruses are retroviruses so they exist as RNA but generate DNA transcripts by reverse transcription in a host cell. With the appropriate control elements, that DNA can be transcribed and translated to generate a gene product in the host cell.

Lentiviral vectors may be rendered integration-deficient by mutation of the viral integrase. Vectors with a Class I integrase mutation fail to integrate but result in normal DNA synthesis, and accumulate in the cell nucleus as double-stranded DNA circles⁷. Two circular forms are produced from the unintegrated linear double-stranded DNA, presumed to be the result of intramolecular homologous recombination between the long terminal repeats (LTRs) (1-LTR circle) or intramolecular end-joining (2-LTR circle)⁸'⁹. Fig. 1 shows a schematic representation of linear, integrated and circular forms of an example vector. Solid thick lines are vector DNA₅ discontinuous lines are genomic DNA, solid thin lines are restriction fragments, white boxes are dLTRs and the black box represents the Southern blot probe (not drawn to scale). For example, in the case of a vector based on HIV-I, the vector may have a mutant D64V integrase. Using pseudotyped HIV-I, such mutant integrase leads to 10000- fold reduced integration frequencies in comparison with wild-type integrase. Vectors with a class II integrase mutations also fail to integrate but analysis of such mutants is complicated, since class II mutations result in pleiotropic effects on viral function such as assembly and/or reverse transcription defects, as well as integration deficiency.

The vectors used in the methods of the invention may comprise various components. If the vector is to be used for expression in the transduced cell, it contains an expression cassette. An expression cassette typically comprises the gene of interest to be expressed and a suitable promoter. Additional elements that may be present include a termination sequence, a post- transcriptional regulatory element, an enhancer, a splice site and a poly-adenylation sequence. The promoter may be functional in a broad range of cells, or it may be cell- or tissue-type specific, ie it initiates transcription only in limited cell types. Suitable promoters for particular applications are known to the person skilled in the art. Examples of suitable promoters which are functional in a broad range of cells include the cytomegalovirus early promoter (CMV), the spleen focus-forming virus (SFFV) 3' LTR promoter, the thymidine kinase promoter, the SRalpha promoter and the human elongation factor 1 alpha (EFl alpha). Examples of suitable tissue-specific promoters include the neuron-specific enolase (rNSE) and the glial fibrillary acidic protein (GFAP) promoter for expression in the brain, the aldolase A promoter for expression in muscle and the factor IX promoter for expression in the liver. Such tissue-specific promoters would provide an additional safety feature designed to reduce the unintended consequences of viral therapy.

An expression cassette may further comprise various posttranscriptional regulatory elements. For example, it may comprise the Woodchuck posttranscriptional regulatory element (WPRE). In addition to an expression cassette, a vector for use in a method of the invention may comprise elements that assist in nuclear import of the virus. For example, a central polypurine tract (cPPT) and central termination sequence (CTS) may be included.

In its integrated form, a lentiviral vector comprises a long terminal repeat (LTR) sequence at each end as shown in Figure 1. Inside a host cell, an integration deficient vector is either degraded by host cell nuclease enzymes or else it forms circular molecules. Such circular molecules may comprise one or two LTRs.

For reasons of patient safety, it is preferred for a vector for use in gene therapy to comprise as little viral DNA/RNA as possible. A minimum lentivirus vector for use in the invention thus comprises (in its linear form) 2 flanking LTRs (one at the 5' end, one at the 3' end), a nucleic acid molecule of interest, an encapsidation sequence and an export enhancer. Optional further constituents are a cPPT/CTS and/or an mRNA stability element. In one embodiment a viral vector preferably comprises no further lentiviral genomic sequences.

In the case of a vector for use in expressing a gene of interest, the nucleic acid of interest comprises an expression cassette. In the case of a vector for use in bringing about a mutation or correction in the host cell, the nucleic acid of interest comprises a sequence with appropriate complementarity to the desired section of the host genome. It comprises a nucleic acid that has a sequence that enables it to recombine with the subject's genomic DNA in a targeted fashion. Optional additional elements in a vector for use in the invention include the elements described above, ie a central polypurine tract and/or a central termination sequence. Preferably, the vector contains as little further viral RNA/DNA as possible. A vector for use in the methods of the invention may be constructed using conventional molecular biological methods. Suitable methods have been described in references 12 and 18.

The invention further provides a plasmid mimicking the structure of a lentiviral genome, including the following nucleic acid sequence elements: i. One or two LTRs ii. RNA export enhancer iii. nucleic acid of interest or cloning site for insertion of a nucleic acid of interest

Such a plasmid may further comprise the following lentiviral-derived nucleic acid sequence elements'. iv. mRNA stability element

Such a plasmid may further comprise the following lentiviral-derived nucleic acid sequence elements: v. cPPT/CTS vi. encapsidation sequence

A plasmid of the invention finds use, for example as a transfection vector. Such transfection vectors are particularly useful for transgene expression, for siRNA or antisense RNA delivery or for the delivery of substrates for homologous/site specific recombination. The present inventors have demonstrated that integration-deficient lentiviral vectors are effective in bringing about long-term expression of a transgene in a mammal in vivo bringing about correction of a genetic defect.

The class I integration-deficient vectors used in the study carry a mutant D64V integrase which decreases integration of pseudotyped HIV-I 10,000-fold¹⁷. Paired integration- proficient and -deficient HIV-I vectors were produced by transient transfection using self- inactivating systems as described in further detail below and as shown in Fig 1 and Fig 2.

In vitro, integration-deficient eGFP-encoding vectors led to transient green fluorescence in a permissive cell line (Fig. 3). Southern blot assays demonstrated the disappearance of integration-deficient vector over time (Fig. 3) as the vector became diluted by cell division. It was thus confirmed that expression from the vectors was effective and that it decreased with time in dividing cells.

The ocular transduction patterns of paired integration-proficient and -deficient vectors in mice using GFP vectors with various promoters and regulatory elements were investigated (Fig. 4 and Fig. 5). All vector types produced identical spatial and temporal transduction patterns, regardless of integration proficiency or vector configuration. The transduction pattern was consistent with previous studies using integration-proficient lentiviral vectors ^" . In vivo fluorescence imaging of murine fundi and post-mortem sections demonstrated eGFP expression in the retinal pigment epithelium (RPE) after subretinal injection (Fig. 4 and Fig. 5). Injections in the anterior chamber led to expression predominantly in corneal endothelial cells (Fig, 5). eGFP expression was highly efficient by 7 days and remained so for the duration of the experiments (up to 9 months; Fig. 4). Similar results were obtained in rats (Fig. 6). These data clearly demonstrate that integration-deficient HIV-I vectors can mediate long-term, high-level trans gene expression in vivo. Additional in vitro experiments showed that integration-deficient HIV vectors are also proficient at transducing human RPE (Fig. 7).

To compare the transduction efficiency of integration-proficient and -deficient eGFP vectors dose-response experiments were performed in vivo. Mice were injected subretinally with non-saturating amounts of integration-proficient or -deficient vectors expressing eGFP. 11 days post-injection eyes were harvested and the amount of eGFP measured using a quantitative ELISA (Fig. 8). The differences between integration-proficient and -deficient vectors injected at equivalent doses were not statistically significant. Thus, it is concluded that efficient eGFP expression can be achieved in vivo with integration-deficient vectors.

To test the ability of integration-deficient vectors to transduce other tissues, stereotactic injections of L-cCegfp into the corpus striatum and the hippocampus in mice were performed. Brains were analyzed 7 and 30 days post-injection. eGFP fluorescence was more widespread at the latter time point. There was no obvious difference in transduction efficiency or in spatial transduction patterns between the two vectors (not shown). Many cells of the corpus striatum showed bright eGFP fluorescence (Fig. 9a-e). Neurons were more commonly transduced than glia, and there were no signs of retrograde neuronal transduction. Most of the transduced striatal cells were medium spiny projection neurons (Fig. 9b, c) and large aspiny interne urons. Endothelial and glial cells were more efficiently transduced in the white matter (Fig. 9d,e). In the hippocampus, fluorescent neurons were detected in the hilus of the dentate gyrus and the CA3 region (Fig. 9f). There were also strongly fluorescent glia in the fimbria (not shown). The ability of the methods of the invention to bring about reporter gene expression has thus been demonstrated. To demonstrate the utility of the invention in gene therapy, further experiments were carried out:

The ability of integration-deficient lentiviral vectors to correct genetic diseases using rodent models of retinal dystrophies was examined. Mice homozygous for the rdl2 mutation in the Rpe65 gene (Rpe65* ' ^{r n}) have a similar phenotype to patients with Leber Congenital Amaurosis (LCA), a severe form of early-onset autosomal recessive retinitis pigmentosa (RP) due to mutations in RPE65²². Rescue of mouse Rpe65 deficiency has been previously reported²³. The efficacy of subretinal delivery of integration-deficient L-cSrpe65W, which carries a functional human RPE65 transgene, in restoring visual function in these mice was evaluated. Expression of HRPE65 in transduced RPE was demonstrated by immunohistochemistry (Fig. IQd). Functional rescue was assessed by electroretinography (ERG), which measures the electrical response of the entire retina to a light stimulus²⁴. Substantial improvements in ERG waveform (Fig. IQh) and significant improvements in b- wave amplitude (Fig, 10c) were observed three weeks after injection of integration-deficient L-cSrpe65W. The b-wave response at 100 mcds/m² was almost 6-fold greater in L-cSrpe65W- treated eyes than in untreated contralateral eyes, indicating functional rescue (Fig. 10c). Furthermore, similar improvements in electrical response were observed in eyes treated with integrase-deficient or -proficient L-cSrpe65W (Fig. 1OZ? and Fig. 10c). These improvements were maintained for at least 8 weeks post-injection (latest time point studied; Fig. IQd). The mean b-wave amplitude measured in integration-deficient L-cSrpe65W-treated mice was approximately 50% of wild-type values (not shown). In contrast, no significant improvement was observed after injection of control vector L-cShrgfpW (Fig. \Qb and Fig. 10c). Integration-deficient vectors encoding Mertk were similarly efficient at rescuing the Royal College of Surgeons (RCS) rat model of RP (Figure 11). These results demonstrate that integration-deficient vectors can produce significant rescue of clinically relevant disease models.

Evidence has also been provided that shows that in vivo expression of reporter and therapeutic genes in eyes receiving subretinal injections of integration-deficient vectors originated from unintegrated vector templates. Using a specific qPCR (quantitative real-time PCR), it was determined that the 2-dLTR/total vector DNA ratio in mouse and rat eyecup samples combined was on average 8-fold higher with integration-deficient vectors than with their integrating counterparts, and that this was stable over time (Fig. 12α). These data indicate that 2-dLTR vector forms persist in eyes injected with integration-deficient vectors, but do not exclude the possibility that they could have integrated with that configuration. Given that genomic integration of circular vector forms could occur through DNA breaks located anywhere in the vector circle, a linear amplification mediated (LAM)-PCR²⁵ was used to scan three regions of the vector DNA (approximately 30-50% of the vector genome depending on vector size) for vector-host DNA junctions (Fig. \2b). LTR-host DNA junctions were easily detected in eyes injected with integration-proficient vectors. In contrast, an extensive search (815 LAM-PCR amplicons sequenced and mapped) for integration junctions in eyes injected with integration-deficient vectors revealed a single detectable integration event (Fig. 126). The sequence at this unique integration junction was that expected from an LTR-mediated integration event. Therefore, the extensive ocular expression of reporter and therapeutic genes that was demonstrated with integration-deficient vectors is most likely to derive from transcriptionally proficient, unintegrated vector episomes. The methods of the invention find particular application in long-term transgene expression in quiescent cells. In the absence of replication or integration of the viral vector, the vector will become diluted in a population of dividing cells. In time, that dilution leads to episomes ceasing to persist in the population of cells. The methods of the invention are thus particularly suitable for use for the long-term expression of a transgene in a quiescent target cell population. However, many tissues including muscle, liver, brain, and retina are essentially post-mitotic, particularly in adults, and are targets for the development of diverse gene therapy strategies. Even in tissues where transformation of target cells is unlikely, the risks associated with inadvertent transduction of by-stander cells or other cells (including germ cells) due to vector spread during in vivo administration, are dramatically reduced by use of an integration-deficient vector.

The methods also find use in the in vivo transduction of dividing cells in various circumstances. For example, if only short-term expression of a transgene is desired, the methods of the invention find application. Short-term expression may be desired in order to express a factor that influences the developmental progression of a cell. The presence of a particular factor at a particular stage of a cell's development may be used to cause the cell to differentiate into a particular desired cell type. Once the desired differentiation has taken place, the non-persistence of the vector is of no consequence.

Short-term presence of a vector in a cell may also be sufficient for a vector to bring about recombination with a target portion of the host genome (or other episome or plasmid present in the cell) to bring about a targeted correction or mutation in the host genome (or other episome or plasmid present in the cell). The methods of the invention have particular application in gene therapy. Efficient in vivo transduction of animal cells with integrase-deficient HIV-I vectors has been demonstrated by the present inventors as discussed in further detail below. That demonstration points to the application of the methods of the invention in human gene therapy, where an effective vector with highly reduced risk of genomic integration has clear advantages for biosafety. Combining the high efficiency of gene transfer mediated by lentiviruses with a stable non- integrating vector system is highly attractive for safe clinical application. The vectors of the invention may also have utility in vaccination protocols where resistance or immunity to infectious pathogens (for example HIV, hepatitis C virus, hepatitis B virus, herpes virus) or parasitic pathogens (for example tuberculosis or Leishmaniasis) or the elimination or induced quiescence of aberrant cells (for example cancer cells) is considered beneficial.

The methods of the invention involve bringing a vector into contact with target cells. In vivo contact may, for example, be achieved by injecting a composition comprising the vector into target cell tissue or into tissue proximate to the target cells. Cell targeting elements may be included in a composition comprising the vector. For example, the vector may be bound to an antibody or other factor with binding specificity for a particular cell-type.

Ribonucleic acid (RNA), like DNA, is a nucleic acid polymer consisting of covalently bound nucleotides. In contrast to DNA, RNA nucleotides are based on a backbone of ribose rather than deoxyribose. There are many different types of RNA within the cell and RNA molecules play a major role in protein expression within the cell. For example, messenger RNA (mRNA) is transcribed from DNA by an enzyme called RNA polymerase. The mRNA then serves as the template for translation by ribosomes to produce proteins. During translation, transfer of amino acids to the ribosome is mediated by another type of RNA molecule known as transfer RNA (tRNA).

The methods of the invention may also alter the expression of a target gene via antisense RNA activity. Antisense RNA activity is achieved by exposing the cellular messenger RNA

(mRNA) molecules produced by a target gene to antisense RNA molecules that contain sequences complementary to a short portion of the mRNA molecule. The antisense RNA and the mRNA hybridise to form a double-stranded RNA molecule (dsRNA). dsRNA is cleaved inside the cell by nucleases, thus resulting in a reduction in the expression levels of the target gene.

The invention thus provides a method wherein the nucleic acid of interest comprises a sequence that, when reverse transcribed and transcribed, produces an RNA molecule that is capable of suppressing the expression of a target gene via antisense RNA interaction. The invention thus provides a method for altering gene expression comprising administering an appropriate vector.

The methods of the invention may also alter the expression of a target gene via RNA interference (RNAi). RNAi is achieved by exposing the cellular messenger RNA (mRNA) molecules produced by the target gene to double-stranded RNA (dsRNA) molecules that contain sequences complementary to a short portion of the mRNA molecule. Inside the cell, the double-stranded RNA molecules are cleaved to produce short (21-23 nucleotides long) single and double-stranded fragments which can bind to the target mRNA molecules. Such binding leads to the cleavage of the target mRNA by nucleases, thus resulting in a reduction in the expression levels of the target gene. In the present invention, the dsRNA is encoded by nucleic acid of interest that is inserted in the integration-deficient lentiviral vector and thus the present invention provides a method whereby effective levels of dsRNA can be stably expressed within the cell leading to RNAi of the target gene. The invention thus provides a method whereby nucleic acid of interest comprises a sequence that, when reverse transcribed and transcribed, produces an RNA molecule that is capable of suppressing the expression of a target gene via RNA interference. The invention provides a method for altering gene expression comprising administering an appropriate vector.

The invention may also provide a method where the nucleic acid of interest comprises a sequence that can act as a marker for transduced cells, whereby the marker sequence can be identifiable by a low-copy number nucleic acid detection method. Such methods include nucleic acid amplification methods, including PCR. Such a method would be useful in identifying cells which have been successfully transduced and would also allow identification of those cells which maintain the transduced DNA.

The invention may further provide a method of identifying a marker sequence in transduced cells using PCR, whereby the marker sequence has been introduced into the cells by an integration-deficient lentiviral vector.

The method of the invention may be carried out using a single vector comprising the elements described. Alternatively a method of the invention may use two or more vectors that each provide elements to the target cell whereby the elements together provide the desired effects in the target cell. For example the two or more vectors may each provide an expression element that provides a subunit protein and the subunit proteins together form a functional final protein. Alternatively, one vector may provide a nucleic acid unit (for example an nucleic acid encoding an expression cassette, an siRNA, antisense RNA or a recombination element) and a second vector may provide an element that enables regulation of the nucleic acid unit. Regulation may involve provision of suitable factors or it may involve recombination between the nucleic acid unit of the first vector and the regulation element of the second vector.

Examples

Example 1 In vitro titration analyses of integration-deficient lentiviral vectors a) Materials

Plasmids: HIV-I plasmids pD64VintCMVDR9 (mutant integrase), pCMVdR8.74, pMDLg/pRRE, pRSV.REV and pMD2VSV.G were obtained from Luigi Naldini and have been previously described¹²'¹⁸. Integrase-deficient, second and third generation plasmids pCMVdR8.74intD64V and pMDLg/pRREintD64V were made by cloning an Af[QJBcR fragment spanning the integrase mutation from pD64VintCMVDR9 into AfllllBcll-άigQsiQά pCMVdR8.74 and pMDLg/pRRE, respectively. Lentiviral transfer plasmids pHR'SIN-cPPT- SEW, pHR'SIN-CE, pHR'SIN-cPPT-CE and LNT.SFFV.Mertk have been previously described²⁷'²⁸ and were used to make vectors L-cSegfpW, L-Cegfp, L-cCegfp and L- cSmertkW respectively. Transfer plasmids pHR'SIN-cPPT-ShrgfpW (carrying the hrGFP gene) and pHR'SIN-cPPT-Srpe65W (containing a human RPE65 trans gene, a kind gift from Dr Susan Wilkie) were used to prepare vectors L-cShrgfpW and L-cSrpe65W, respectively. Figure 2 shows a schematic representation of HIV-I vectors used in this study. In this nomenclature "L" stands for lentiviral vector, "c" for the central polypurine tract/central termination sequence (cPPT/CTS), "S" and "C" for the spleen focus-forming virus (SFFV) 3' LTR promoter and the cytomegalovirus (CMV) immediate early promoter, respectively; this is followed by the transgene gene symbol in full and "W" for the woodchuck post- transcriptional regulatory element (WPRE), if pertaining. Shown are the double-stranded DNA vector forms after completion of reverse transcription. Relevant features are named on L-cSegfpW, and only those differing are named on the other vectors. dLTR is the 5'-deleted LTR present in these self-inactivating vectors. Not drawn to scale.

Cells: HeLa and 293T cells were cultured in DMEM with standard supplements.

b) Methods

Vector production, purification and titration. HIV vectors were produced by transient transfection and pseudotyped with the VSV-G glycoprotein envelope as described²⁷. p24 titrations were carried out using Beckman Coulter kits PN6604535 and PN626391 and following manufacturers instructions. For eGFP and qPCR titrations HeLa cells were

transduced with serial dilutions of vector stock in the presence of 8 μg/ml polybrene. eGFP-

positive cells were scored through FACS analysis. qPCR titrations of total vector DNA were done according to a published protocol for late reverse transcriptase amplicon quantification²⁹ but, to compare integration-proficient and -deficient vectors, cells were harvested 24 h post- transduction. DNA was purified using the DNAeasy kit (Qiagen, Crawley, UK). qPCR values were normalized by quantitating actin gene copies in the DNA extracts and correcting for HeLa cell aneuploidy.

c) Results In vitro titration analyses showed that paired integration-deficient and -proficient HIV-I vectors had similar titres (Table 1). The titres of integration-proficient and -deficient vectors produced in parallel were similar according to qPCR, p24 content and GFP gene expression measured in a permissive cell line (Table 1). Vectors encoding eGFP were initially tested for integration proficiency by titration in HeLa cells. For both integration-proficient and - deficient vectors peak expression occurred at 3 days. The percentage of eGFP-expressing cells transduced by integration-deficient vectors progressively declined over time, consistent with loss of transgene during cell division. The data are shown in Fig. 3a in which (a) shows the time-course of eGFP expression after transduction of HeLa cells with integration- deficient and -proficient L-cSegfpW vector at the indicated MOI. A representative experiment out of four performed is shown. Int- and int+ stand for integration-deficient and -proficient vectors, respectively. A 50 to 100-fold reduction in the percentage of cells expressing eGFP between 3 and 10 days was observed with every stock of integration-deficient vector tested. In contrast, the level of transduction mediated by integration-proficient vectors remained stable (Fig. Za and Table 1).

Table 1. Characterization of paired integration-deficient and -proficient HIV-I vector stocks.

Example 2 - In vitro Southern blot analysis of vectors a) Methods

Southern blot. HeLa cells were transduced with integration-proficient or -deficient vectors (multiplicity of infection (MOI) 25 eGFP TU/cell or a similar p24 amount) in the presence of

8 μg/ml polybrene. The cells were harvested 1 day or 5 weeks post-transduction, included in

agarose plugs and subjected to Southern blotting following standard protocols. The probe was a 1-kb NotVXhol fragment spanning from the rev-response element to the cPPT/CTS.

b) Results

A Southern blot assay was used to distinguish the various vector forms in cultured cells. Southern blot analyses of vector forms and integration are shown in Figure3c. Total DNA from HeLa cells transduced with the indicated vectors was harvested 1 day (4/711 or EcoRY) or 5 weeks post-transduction (Aflll) and subjected to Southern blot. +: integration-proficient; -: integration-deficient; M: mock-transduced sample. The positions of molecular weight markers (in kbp) are indicated to the right of the gels and expected fragment sizes are indicated on the right and in Fig. 3b. For this HeLa cells were transduced with integration-deficient or proficient L-cSegfpW, L- cCegfp or L-cSmertkW. An EcoRΪ Southern blot of cells harvested 1 day post-transduction showed that linear DNA and circular forms were present with both integration-proficient and -deficient vectors, but the smear diagnostic of vector integration was only noticeable with integration-proficient vectors (Fig, 3c). To study the disappearance of integration-deficient vectors over time, HeLa cells were also harvested five weeks post-transduction. In this case an AfIU digest was used, which releases the same band from all vector forms irrespective of integration (Fig. 3b, c). Although similar amounts of integration-proficient and -deficient vector were observed after 24 h, no integration-deficient vector was detectable five weeks post-transduction indicating that it had been diluted out by cell division.

Example 3 In vitro transduction of human RPE by integration-deficient vectors

The capacity of integration-deficient lentiviral vectors to transduce human retina was evaluated. a) Methods

Firstly, the human RPE cell line ARPE- 19 was transduced with either integration-deficient L- cSegfpW or L-cCegfp vectors. Secondly, we transduced human retinal tissue excised from patients with age-related macular degeneration (AMD) undergoing RPE translocation surgery.

Culture and transduction of human RPE cells. Human ARPE- 19 cells (a kind gift from Dr Maria Balda) were cultured as described⁴¹ and transduced at an eGFP MOI of 20 with either integrase-deficient L-cSegfpW or L-cCegfp.

Ex vivo transduction of excised human retinal tissue. Human retina was obtained with written consent from patients undergoing surgery as part of a clinical trial. Excised retina was co-cultured with 20 μl of L-cCegfp (7.50E+08 TU/ml) in DMEM for 20 min, 40 min or

overnight at 37°C in 5% CO₂. Specimens were washed three times with DMEM+glutamax

(Gibco/Invitrogen, Paisley, UK), 1% antibiotic-antimycotic solution (Gibco/Invitrogen, Paisley, UK) and 3% foetal calf serum. This solution was used for subsequent tissue culture at 37°C and was changed every 48 hours. After 7 days specimens were fixed with 4%

paraformaldehyde, cryoembedded in OCT and serially sectioned at 12 μm for confocal

microscopy.

b) Results In the human ARPE-19 cell line experiment, ARPE-19 cells were found to express high levels of eGFP using either vector (Fig. 7). In Figure 7 the Upper left panels show efficient in vitro transduction of human ARPE-19 cells with integrase-deficient L-cSegfpW or L-cCegfp. In the Upper right, centre and bottom panels transduction of excised human retina by integration-deficient L-cCegfp is shown. Shown are confocal images from cryosections of cultured human retina. No eGFP expressing cells could be identified in DAPI-counterstained (blue) neurosensory retina in the upper right panel. In the centre panel, there is shown a high magnification, longitudinal section through an eGFP-expressm^' g RPE cell showing apical microvilli and an autofluorescent (red) deposit. Presumed intracellular lipofuscin granules in RPE show bright red autofluorescence, consistent with diseased RPE in patients with age- related macular degeneration. In the bottom left panel, it is seen that almost all the fluorescent cells were found at the edges of cultured tissue and had a hexagonal shape consistent with macular RPE morphology. In the bottom right panel, there are seen abundant autofluorescent

granules in the diseased RPE cells. Scale bars: 10 μm. Abbreviations as in Fig. 5. In the transduced human retinal tissue experiment, eGFP fluorescence was first detected after 3-4 days. Strong fluorescence was observed in approximately 10-20% of cells in the RPE sheets after 1 week (Fig. 7), Transduction efficiency was similar for cells exposed to vector for 20 min, 40 min or overnight (not shown). Almost all the eGFP-expressing cells were found at the edges of the cultured tissue, in areas most exposed to the vector suspension. Morphologically, these cells had a hexagonal structure and apical microvilli consistent with macular RPE (Fig. 7). Bright autofluorescence was detected from intracellular lipofuscin granules which, in the retinas of AMD patients, are found exclusively in the RPE (Fig. 7). eGFP was not detected in neurosensory retinal cells (Fig. 7). These results demonstrate that lentiviral vectors have the capacity to transduce diseased human RPE and that a brief period of exposure to vector suspension is sufficient to achieve efficient transgene expression.

Example 4 - Transduction patterns of integration-proficient and -deficient GFP vectors in mice a) Methods

Intraocular injections. All animals were treated with procedures approved by the UK Home Office and in compliance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. Subretinal injections were performed as previously described¹⁹.

Ocular GFP imaging in mice. Adult female C57B1/6J mice received eGFP vectors (Table 2). In vivo eGFP expression was visualized using a modified human ophthalmic slit lamp. Mice were imaged 1 week (all eGFP constructs) and between 1 and 6 months (L-cSegfpW and L- cCegfp constructs) post-injection. Some mice were kept until 9 months (L-cSegfpW). eGFP expression was also analyzed using fluorescence microscopy on cryosections between 3 days and 6 months post-injection.

Table 2. HIV-I -mediated reporter gene studies in rodents

b) Results

The ocular transduction patterns of paired integration-proficient and -deficient vectors in mice using GFP vectors with various promoters and regulatory elements are shown in Fig. 4. which shows in vivo eGFP fluorescence at various time points after subretinal injection with the indicated vectors (abbreviations: int+, integration-proficient; int-, integration-deficient). AU vector types produced identical spatial and temporal transduction patterns, regardless of integration proficiency or vector configuration. The transduction pattern was consistent with previous studies using integration-proficient lentiviral vectors^19"21. In vivo fluorescence imaging of murine fundi and post-mortem sections demonstrated eGFP expression in the retinal pigment epithelium (RPE) after subretinal injection (Fig. 4 and Fig. 5). In Figure 5, eGFP fluorescence in cryosections of mouse retina and cornea six months after vector administration are shown. The sections were counterstained with propidium iodide (red). eGFP fluorescence observed in the region of the neurosensory retina originates from detached RPE microvilli. Intracameral injection with L-cSegfpW vectors transduced corneal endothelium (right panels) and trabecular meshwork (not shown). Similar transduction patterns were obtained after intracameral delivery of L-cCegfp (not shown). GcI: ganglion cell layer; inl: inner nuclear layer; onl: outer nuclear cell layer; rpe: retinal pigment epithelium; cho: choroid; epi: corneal epithelium; str: stroma; end: corneal endothelium.

Thus it is seen that subretinal injections led to RPE expression, while injections in the anterior chamber led to expression predominantly in corneal endothelial cells (Fig. 5). eGFP expression was highly efficient by 7 days and remained so for the duration of the experiments (up to 9 months; Fig. 4). Similar results were obtained in rats (See example 5 below) (Fig. 6). These data clearly demonstrate that integration-deficient HIV-I vectors can mediate long-term, high-level transgene expression in vivo.

Example 5 Mediation of ocular GFP expression in rats by integration-deficient vectors a) Methods

Ocular GFP imaging in rats. 2 week-old RCS rats received hrGFP vectors (Table 2). In vivo hrGFP expression was visualized using a modified human ophthalmic slit lamp. Images were obtained between 2 and 13 weeks (latest time point) post-injection. hrGFP expression was also analyzed using fluorescence microscopy on cryosections between 6 and 8 weeks post-injection.

b) Results

GFP expression from integration-deficient vectors in rat RPE. Fluorescence imaging was performed after subretinal injection of rats with integration-deficient L-cShrgfpW. (Figure 6

Left) In viyo imaging of rat retina thirteen weeks post-injection showing bright hrGFP fluorescence confined to the injected area. (Figure 6 Right) Propidium iodide-counterstained

(red) retinal cryosection six weeks post-injection. ScI: sclera. Other abbreviations as in

Figure 5. It is seen that subretinal injection of integration-deficient L-cShrgfpW led to high expression of hrGFP in rat RPE cells at 2 weeks (earliest time point examined), and remained efficient at 13 weeks (Fig. 6). As in mice, histological analysis demonstrated exclusive expression in the RPE (Fig. 6).

Example 6 - Comparison of transduction efficiency of integration-proficient and - deficient eGFP vectors in mice a) Methods eGFP ELISA. Mouse eyes

per group) were injected subretinally with 1 or 5 ng p24 of integration-proficient or -deficient vectors encoding eGFP. Eyes were harvested 11 d later. Whole eye extracts were prepared following standard techniques and clarified supernatant used for analysis. Enzyme-linked immunoadsorbent assays were carried out following standard procedures. eGFP was captured using a monoclonal antibody (abl218-100, 1:10,000 dilution; Abeam, Cambridge, UK). Absolute quantification of eGFP was done using recombinant eGFP (Clontech, Cowley, UK) as a standard.

Statistical analyses. ELISA measurements of eGFP levels were analyzed for statistical significance (PO.05) using the one-way non-parametric Kruskal-Wallis test with Dunn's correction.

b) Results In Figure 8 the dose-response of in vivo eGFP expression after subretinal vector injection is shown. Error bars represent standard errors of the mean. Int- and int+ stand for integration- deficient and -proficient vectors, respectively. The differences between integration-deficient and -proficient vector samples at the same doses were not statistically significant. We conclude that efficient eGFP expression can be achieved in vivo with integration-deficient vectors.

Example 7 - Ability of integration-deficient vectors to transduce brain tissue a) Method

Brain injections and histological analysis. Adult female C57B1 mice were injected with 0.8

μl (6.50E+08 eGFP transducing units (TU)/ml) of integration-proficient or -deficient L- cCegfp vectors over 25 minutes into the striatum (n-5) or the hippocampal formation (n-2).

Brains were retrieved at 7 and 30 days post injection. 40 μm-thick coronal cryosections were

mounted in DABCO and examined with a fluorescence microscope. All procedures were approved by the UCL Ethical Committee and the UK Home Office.

b) Results

In Figure 9 there is shown eGFP fluorescence in coronal sections of mouse brain, one month after injection of integration-deficient L-cCegfp vector, (a-e) Corpus striatum, (f) Hippocampus, (a) Low power micrograph showing eGFP fluorescence after injection into the corpus striatum (st). The cerebral cortex (ct) shows little fluorescence indicating that corticostriate projection neurons were not retrogradely transduced. Scale bar: 40 μm. (b) Higher power image showing densely packed fluorescent cell bodies in the striatum. Scale bar: 20 μm, also applies to (d-f). (c) Morphological identification of most of the cell bodies expressing eGFP in the striatum as neurons. Scale bar: 10 μm. (d) Blood vessel in the striatum, approximately 1 mm from the injection site. The vector spread along the outside of blood vessels, transducing perivascular astrocytes and some endothelial cells, (e) The vector also spread in the plane of white matter as shown by the transduced glia in the subcortical white matter separating cerebral cortex and striatum, (f) Transduced neurons near an injection site in the hippocampus.

In summary, eGFP fluorescence was more widespread at the 30 day time point than at the 7 day time point. There was no obvious difference in transduction efficiency or in spatial transduction patterns between integration-deficient and integration-proficient vectors (not shown). Many cells of the corpus striatum showed bright eGFP fluorescence (Fig. 9a-e). Neurons were more commonly transduced than glia, and there were no signs of retrograde neuronal transduction. Most of the transduced striatal cells were medium spiny projection neurons (Fig. 9b,c) and large aspiny intemeurons. Endothelial and glial cells were more efficiently transduced in the white matter (Fig. 9d,e). In the hippocampus, fluorescent neurons were detected in the hilus of the dentate gyrus and the CA3 region (Fig. 9f). There were also strongly fluorescent glia in the fimbria (not shown).

Example 8 - Correction oiRpe65 deficiency in mice a) Method

Rescue of Rpe65 deficiency in mice. 5 week-old Rpe65^τάl2Mn mice received unilateral subretinal injections of either integration-deficient L-cSrpe65W (46.5 ng/μl p24; n=β) or L- cShrgfpW (45.0 ng/μl p24; n=3); contralateral eyes were untreated. Additional mice («=4) received unilateral injections of integration-deficient L-cSrpe65W; contralateral eyes received integration-proficient L-cSrpe65W (70 ng/μl p24). ERG analysis was performed three or eight weeks later as previously described³ . This study design controls for the inter-animal and test- retest variability inherent in rodent ERGs. Some animals were sacrificed three weeks post- injection for RPE65 immunohistochemistry as previously described (D. A. Thompson, N. Hemati, D. L. Allen, K. L. Feathers, K. J. Biederman, A. J. Mears, A. Swaroop & T. J. Carlson. Association for Research in Vision and Ophthalmology 2003, poster 402/B377; the anti-RPE65 antibody was a kind gift from Dr Debra Thomson). Statistical analysis: For ERG measurements, the b-wave amplitudes were compared with those from the contralateral eyes. As ERG amplitude data do not follow a normal distribution, a non-parametric test for paired samples (Wilcoxon matched pairs test) was used to evaluate significance (PO.05). Statistical analyses were performed with GraphPad Prism software.

b) Results Expression of HRPE65 in transduced RPE was demonstrated by immunohistochemistry (Fig. IQa). Retinal cryosections were immunostained for RPE65 (green) and counterstained with propidium iodide (red). We observed RPE65-positive immunohistochemical labelling in the RPE layer of wild-type mice (left) and in Rpe65^{rdmd 12} mice three weeks after subretinal injection of integrase-deficient L-cSrpe65W (centre left). RPE65 is not detected in sections from L-cSrpe65W-injected Rpe65^{x 12/rd l} mice incubated without primary antibody (centre right) nor in uninjected Rpe65^rdnMl2 mice (right). Some background fluorescence from the secondary antibody is visible in the choroid and sclera in all sections. Abbreviations as in Fig 5.

Functional rescue was assessed by electroretinography (ERG), which measures the electrical response of the entire retina to a light stimulus²⁴. In Figure 10b, there are shown sample scotopic ERG traces at 10, 100 and 1000 mcds/m² flash intensities from Rpe65^rdmdn eyes injected with integration-deficient L-cSrpe65W and untreated contralateral eyes, integration- deficient and -proficient L-cSrpe65W, and integration-deficient L-cShrgfpW and untreated contralateral eyes, three weeks after subretinal vector administration. Abbreviations: ref, reference value; a, a-wave trough; b, b-wave peak. Arrow indicates b-wave. Substantial improvements in ERG waveform were observed three weeks after injection of integration- deficient L-cSrpe65W.

In Figure 10c there are shown mean ERG b-wave amplitudes of the same paired groups. Error bars represent standard deviations. P values are indicated for the corresponding pair-wise comparisons. Asterisks indicate statistically significant differences. Significant improvements in b-wave amplitude (Fig. 10c) were observed three weeks after injection of integration- deficient L-cSrpe65W. The b-wave response at 100 mcds/m² was almost 6-fold greater in L- cSrpe65W-treated eyes than in untreated contralateral eyes, indicating functional rescue.

Furthermore, similar improvements in electrical response in eyes treated with integrase- deficient or -proficient L-cSrpe65W were observed. The mean b-wave amplitude measured in integration-deficient L-cSrpe65W-treated mice was approximately 50% of wild-type values (not shown). In contrast, no significant improvement was observed after injection of control vector L-cShrgfpW.

In Figure IOd it is shown that the improvements in ERG waveform and b-wave amplitude in Rpe65^{rά UM l2} mice injected with integration-deficient or -proficient L-cSrpe65W were maintained for at least 8 weeks post-injection. Integration-deficient vectors encoding Mertk were similarly efficient at rescuing the Royal College of Surgeons (RCS) rat model of RP (See Example 9 below). These results demonstrate that integration-deficient vectors can produce significant rescue of clinically relevant disease models.

Example 9 - Correction of Mertk deficiency in the RCS rat model a) Methods

Rescue of Mertk deficiency in rats. 14 day-old RCS rats received unilateral injections of either integrase-deficient L-cSmertkW (92.1 ng/μl p24; n=9) or L-cShrgfpW (45.0 ng/μl p24;

received injections of integration-deficient L-cSmertkW in one eye and integrase-proficient L-cSmertkW (62.9 ng/μl p24) in the contralateral eye. ERG analysis was performed 6 weeks later as previously described²⁴. This study design controls for the inter-animal and test-retest variability inherent in rodent ERGs. b) Results:

In Figure 11a, there are shown sample scotopic ERG traces at 10, 100 and 1000 mcds/m² flash intensity from RCS eyes injected with integration-deficient L-cSmertkW and untreated contralateral eyes, integration-deficient and -proficient L-cSmertkW, and integration-deficient L-cShrgfpW and untreated contralateral eyes, six weeks after subretinal vector administration. In Figure 1 Ib there are shown the mean ERG b-wave amplitudes of the same paired groups. Abbreviations, error bars, P values and asterisks as described in Fig. lθb,c.

The ERG analyses performed six weeks after subretinal vector administration demonstrated more regular ERG waveforms and 3 -fold greater b-wave amplitudes in integrase deficient L- cSmertkW-injected eyes than in untreated eyes. Efficacy in delaying functional loss was comparable between integration-proficient and -deficient L-cSmertkW. The average b-wave amplitude scored using integration-deficient L-cSmertkW was approximately 30% of wild- type levels (not shown), similar to that previously demonstrated using integrating vector. No significant rescue was observed after injection of control vector L-cShrgfpW.

Example 10 -PCR evaluation of eyecup DNA a) Method Molecular analyses of vector forms in injected eyes. qPCR and LAM-PCR analyses were performed on eyecup DNA prepared using the DNAeasy kit (Qiagen, Crawley, UK). A 2- dLTR junction qPCR assay for self-inactivating HIV vectors was performed using oligos 2- dLTR forward 5 '-AACTAGAGATCCCTCAGACCCTTTT-S', 2-dLTR reverse 5'- CTTGTCTTCGTTGGGAGTGAATT-S' and a Taqman MGB probe 5 ' -FAM- CTAGAGATTTTCCACACTGAC-S'. LAM-PCR was performed as previously described²⁵ with the following modifications: Tsp509l , HpyCH4lY or enzyme mix HpallFspllStul were applied for DNA fragmentation, and we used three primer sets to study different regions of the vector molecule. For LAM-PCR primer set 1, linear PCR was performed using oligonucleotidess LTRlbio 5'-GAGCTCTCTGGCTAACTAGG-S ' and LTR2bio: 5'- GAACCCACTGCTTAAGCCTCA-3'; the first exponential PCR was done with oligonucleotides LTR3bio 5'-AGCTTGCCTTGAGTGCTTCA-S ' and LCI. This was followed by an additional magnetic capture step of the first exponential PCR product, which served as template for the second exponential PCR with oligonucleotides LTR4bio 5'- AGTAGTGTGTGCCCGTCTGT-3' and LCII. For LAM-PCRs with primer set 2 we used LTR5bio 5'GCTAGAGATTTTCCACACTG-S' and LTRόbio

5'CTAAAAGGGTCTGAGGGATC-3'for the linear PCR, LTR7bio 5'- CTAGTTACC AGAGTCACACA-S' and LCI for the first exponential amplification, and LTR8 5'-TGAGGCTTAAGCAGTGGGTTC-S' and LCII for the second exponential amplification. For primer set 3 the linear PCR was performed with Ll bio 5'GATAGTAGGAGGCTTGGTAG-S ' and L2bio 5 'AGTGAATAG AGTTAGGC AGG-3', the first exponential PCR with L3bio 5'-CAGGGATATTCACCATTATCG-S' and LCI, and the second exponential PCR with L4 5'-CACCATTATCGTTTCAGACCC-S' and LCII. LAM-PCR products were visualized on a high-resolution Spreadex gel (Elchrom Scientific, Cham, Switzerland). LAM-PCR products were also purified, shotgun cloned into the TOPO TA vector (Invitrogen, Carlsbad, CA) and sequenced (GATC, Konstanz, Germany). Sequences were aligned to the human (assembly May 2004), mouse (assembly March 2005) or rat (assembly June 2003) genome database using University of California, Santa Cruz BLAT search tools (http://genome.ucsc.edu/).

b) Results Mouse eyes were injected with L-cSegfpW or L-cCegfp, rat eyes were injected with L- cSmertkW. Figure 12α shows qPCR quantitation of 2-dLTR junction-containing molecules in the transduced eyecups. A normalized ratio 2-dLTR molecules/total vector is shown. The average ratio for integration-proficient vectors has been arbitrarily set to 1 (solid line), while the average ratio for integration-deficient vectors is indicated by a dashed line. It is seen that the 2-dLTR/total vector DNA ratio in mouse and rat eyecup samples combined was on average 8-fold higher with integration-deficient vectors than with their integrating counterparts, and that this was stable over time. This indicates the persistence of 2-dLTR species in eyes injected with integration-deficient vectors, but does not exclude the possibility that these could have integrated with that configuration.

Given that genomic integration of circular vector forms could occur through DNA breaks located anywhere in the vector circle, we used linear amplification mediated (LAM)-PCR²⁵ to scan three regions of the vector DNA (approximately 30-50% of the vector genome depending on vector size) for vector-host DNA junctions. The strategy for detection of vector-host DNA junctions by LAM-PCR is shown in Figure 126. Three different LAM-PCR reactions were designed to scan vector regions downstream of the 5'-dLTR (reaction 1), upstream of the 3'-dLTR (reaction 2) and downstream of residual env sequences (a region including the cPPT/CTS and the internal promoter, reaction 3) for integration junctions. The vector region scanned with each LAM-PCR reaction spans the distance between the corresponding forward oligonucleotide of the second exponential PCR and the first downstream restriction site for the enzyme used in the assay. In practice, we estimate (from the length of sequenced LAM amplicons) that the size of the scanned region can be between 600 and 1,000-bp if there is no intervening restriction site. Shown are schematic representations of integrated (through normal LTR-mediated integration) and circular forms of a generic lentiviral vector. Solid lines are vector DNA, discontinuous lines are genomic DNA and white boxes are dLTRs. Also indicated is the forward oligonucleotide of the second exponential PCR for each of the LAM-PCR assays (numbered arrows), the region potentially scanned with each assay (dotted lines) and relevant 7!5^509I sites (solid lines interrupting some reactions with primer set 1). The bottom panel describes the frequency of unique integration events detected with the various LAM-PCR assays in eyecup samples injected with integration-proficient or -deficient vectors. Abbreviations: Int, integrase proficiency; Site, distance from indicated restriction site.

LTR-host DNA junctions were easily detected in eyes injected with integration-proficient vectors. In contrast, an extensive search (815 LAM-PCR amplicons sequenced and mapped) for integration junctions in eyes injected with integration-deficient vectors revealed a single detectable integration event. The sequence at this unique integration junction was that expected from an LTR-mediated integration event. Therefore, the extensive ocular expression of reporter and therapeutic genes that we have demonstrated with integration-deficient vectors is most likely to derive from transcriptionally proficient, unintegrated vector episomes.

REFERENCES

1. Baum, C. et al. Side effects of retroviral gene transfer into hematopoietic stem cells. Blood 101, 2099-114 (2003).

2. Li, Z. et al. Murine leukemia induced by retroviral gene marking. Science 296, 497 (2002).

3. Hacein-Bey-Abina, S. et al. LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-Xl. Science 302, 415-9 (2003). ^'4. Saenz, D.T. et al. Unintegrated lentivirus DNA persistence and accessibility to expression in nondividing cells: analysis with class I integrase mutants. J Virol 78, 2906-20 (2004).

5. Vargas, J., Jr., Gusella, G.L., Najfeld, V., Klotman, M.E. & Cara, A. Novel integrase- defective lentiviral episomal vectors for gene transfer. Hum Gene Ther 15, 361-72 (2004).

6. Lu, R. et al. Simian virus 40-based replication of catalytically inactive human immunodeficiency virus type 1 integrase mutants in nonpermissive T cells and monocyte-derived macrophages. J Virol 78, 658-68 (2004). 7. Engelman, A. In vivo analysis of retroviral integrase structure and function. Adv Virus Res 52, 411-26 (1999).

8. Shank, P.R. et al. Mapping unintegrated avian sarcoma virus DNA: termini of linear DNA bear 300 nucleotides present once or twice in two species of circular DNA. Cell 15, 1383-95 (1978).

9. Hsu, T. W., Sabran, J.L., Mark, G.E., Guntaka, R.V. & Taylor, J.M. Analysis of unintegrated avian RNA tumor virus double-stranded DNA intermediates. J Virol 28, 810-8 (1978).

10. Pierson, T.C. et al. Intrinsic stability of episomal circles formed during human immunodeficiency virus type 1 replication. J Virol 76, 4138-44 (2002).

11. Butler, S. L., Johnson, E.P. & Bushman, F.D. Human immunodeficiency vims cDNA metabolism: notable stability of two-long terminal repeat circles. J Virol 76, 3739-47 (2002).

12. Naldini, L. et al. In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector. Science 272, 263-7 (1996).

13. Case, S. S. et al. Stable transduction of quiescent CD34(+)CD38(-) human hematopoietic cells by HIV-1-based lentiviral vectors. Proc Natl Acad Sci U S A 96, 2988-93 (1999).

14. Naldini, L., Blomer, U., Gage, F.H., Trono, D. & Verma, LM. Efficient transfer, integration, and sustained long-term expression of the transgene in adult rat brains injected with a lentiviral vector. Proc Natl Acad Sci USA 93, 11382-8 (1996).

15. Loewen, N. et al. Comparison of wild-type and class I integrase mutant-FIV vectors in retina demonstrates sustained expression of integrated transgenes in retinal pigment epithelium. J Gene Med 5, 1009-17 (2003). 16. Park, F., Ohashi, K. & Kay, M. A. Therapeutic levels of human factor VIII and IX using HIV-1-based lentiviral vectors in mouse liver. Blood 96, 1173-6 (2000). 17. Leavitt, A. D., Robles, G., Alesandro, N. & Varmus, H.E. Human immunodeficiency virus type 1 integrase mutants retain in vitro integrase activity yet fail to integrate viral

DNA efficiently during infection. J Virol 70, 721-8 (1996). 18. Dull, T. et al. A third-generation lentivirus vector with a conditional packaging system. J Virol 72, 8463-71 (1998). 19. Bainbridge, J. W. et al. In vivo gene transfer to the mouse eye using an HIV-based lentiviral vector; efficient long-term transduction of corneal endothelium and retinal pigment epithelium. Gene Ther 8, 1665-8 (2001). 20. Bemelmans, A. P. et al. Retinal cell type expression specificity of HIV-I -derived gene transfer vectors upon subretinal injection in the adult rat: influence of pseudotyping and promoter. J Gene Med 7, 1367-74 (2005).

21. Gruter, O. et al. Lentiviral vector-mediated gene transfer in adult mouse photoreceptors is impaired by the presence of a physical barrier. Gene Ther 12, 942-7 (2005).

22. Pang, JJ. et al. Retinal degeneration 12 (rdl2): a new, spontaneously arising mouse model for human Leber congenital amaurosis (LCA). MoI Vis 11, 152-62 (2005).

23. Dejneka, N. S. et al. In utero gene therapy rescues vision in a murine model of congenital blindness. MoI Ther 9, 182-8 (2004). 24. Smith, AJ. et al. AAV-Mediated gene transfer slows photoreceptor loss in the RCS rat model of retinitis pigmentosa. MoI Ther 8, 188-95 (2003).

25. Schmidt, M. et al. Polyclonal long-term repopulating stem cell clones in a primate model. Blood 100, 2737-43 (2002). 26. Abordo-Adesida, E. et al. Stability of lentiviral vector-mediated transgene expression in the brain in the presence of systemic antivector immune responses. Hum Gene Ther 16, 741-51 (2005).

27. Demaison, C. et al. High-level transduction and gene expression in hematopoietic repopulating cells using a human imunodeficiency virus type 1 -based lentiviral vector containing an internal spleen focus forming virus promoter. Hum Gene Ther 13, 803- 13 (2002).

28. Tschernutter, M. et al. Long-term preservation of retinal function in the RCS rat model of retinitis pigmentosa following lentivirus-mediated gene therapy. Gene Ther 12, 694-701 (2005).

29. Butler, S.L., Hansen, M.S. & Bushman, F.D. A quantitative assay for HIV DNA integration in vivo. Nat Med 7, 631 -4 (2001 ).

30. AIi, R.R. et al. Restoration of photoreceptor ultrastructure and function in retinal degeneration slow mice by gene therapy. Nat Genet 25, 306-10 (2000). 31. Hlavaty J, et al.₄ Effect of posttranscriptional regulatory elements on transgene expression and virus production in the context of retrovirus vectors. Virology. 2005

Oct lO;341(l):l-l l 32. Naldini L. et al. In vivo gene delivery and stable tranduction of non-dividing cells by a lentiviral vector. Science 272: 263-267 (1996). 33. Poeschla E. et al. Identification of a human immunodeficiency virus type 2 (HIV-2) encapsidation determinant and transduction of non-dividing human cells by HIV-2- based lentivirus vectors. J. Virol. 72: 6527-6536 (1998). 34. Poeschla E. et al. Efficient transduction of non-dividing human cells by feline immunodeficiency virus lentiviral vectors. Nat. Med. 4: 354-357 (1998). 35. Mangeot P.E. et al. Development of minimal lentivirus vectors derived from simian immunodeficiency virus (SIVmac251) and their use for gene transfer into human dendritic cells. J. Virol. 74: 8307-8315 (2000). 36. Berkowitz R. et al. Construction and molecular analysis of gene transfer systems derived from bovine immunodeficiency virus. J. Virol. 75: 3371-3382 (2001). 37. Mselli-Lakhal L. et al., Defective RNA packaging is responsible for low transduction efficiency of CAEV-based vectors. Arch. Virol. 143: 681-695 (1998).

38. Olsen J.C., Gene transfer vectors derived from equine infectious anemia virus. Gene Ther. 5: 1481-1487 (1998).

39. Mitrophanous K. et al. , Stable gene transfer to the nervous system using a non-primate lentiviral vector. Gene Ther. 6: 1808-1818 (1999).

40. Metharom P. et al, Novel bovine lentiviral vectors based on Jembrana disease virus. J. Gene Med., 2: 176-185.

41. Dunn et al. , Exp Eye Res 62, 155-69, 1996

Claims

Claims

1. A method of gene therapy comprising administering to a subject an integration- deficient lentiviral vector containing a nucleic acid that provides benefit.

2. A method as claimed in claim 1 wherein the subject has a deficiency arising from aberrant gene expression or a defective gene and the nucleic acid complements or corrects the deficiency.

3. A method as claimed in claim 1 wherein the subject has a deficiency or disease that is treatable by expression of a transgene.

4. A method as claimed in claim 1 wherein the nucleic acid provides a prophylactic benefit to the subject.

5. A method as claimed in claim 1 wherein the subject is a healthy mammal and the nucleic acid provides a genetic enhancement to said mammal.

6. A method as claimed in claim 1 in which the vector is an HIV-I vector.

7. A method as claimed in claim 6 in which the vector is class I integration deficient.

8. A method as claimed in claim 1 in which the vector comprises an expression cassette comprising a functional copy of the gene that is aberrantly expressed or defective in the patient or a gene that provides therapeutic benefit.

9. A method as claimed in claim 1 in which the vector further comprises central polypurine tract/central termination sequence.

10. A method for in vivo transfer of a nucleic acid of interest into a target cell comprising bringing an integration-deficient lentiviral vector containing the nucleic acid of interest into contact with the target cell.

11. A method as claimed in claim 10 where the nucleic acid of interest comprises a sequence that, when reverse transcribed and transcribed, produces an RNA molecule that is capable of suppressing the expression of a target gene via RNA interference.

12. A method as claimed in claim 10 wherein the nucleic acid of interest comprises a sequence that, when reverse transcribed and transcribed, produces an RNA molecule that is capable of suppressing the expression of a target gene via antisense RNA interaction.

13. A method as claimed in claim 10 where the nucleic acid of interest comprises a sequence that, when reverse transcribed produces a DNA molecule that is capable of expressing a gene of interest.

14. A method as claimed in claim 10 where the nucleic acid of interest comprises a sequence that can act as a marker for transduced cells, whereby the marker sequence is identifiable by a low-copy number nucleic acid detection method.

15. A method as claimed in claim 14 where the marker sequence is identifiable by PCR.

16. A method for in vivo expression of a transgene of interest in a target cell comprising bringing an integration-deficient lentiviral vector containing the transgene of interest into contact with the target cell.

17. A plasmid mimicking the structure of a lentiviral genome, including the following nucleic acid sequence elements: i. One or two LTRs ii. RNA export enhancer iii. nucleic acid of interest or cloning site for insertion of a nucleic acid of interest

18. A plasmid as claimed in claim 17 further comprising the following lentiviral-derived nucleic acid sequence elements: iv. mRNA stability element

19. A plasmid as claimed in claim 17 or 18 further comprising the following lentiviral- derived nucleic acid sequence elements: v. cPPT/CTS vi. encapsidation sequence

20. A method of identifying a marker sequence in transduced cells using PCR, whereby the marker sequence has been introduced into the cells by an integration-deficient lentiviral vector.