Retroviral vectors - Nature

Gene Therapy (2004) 11, S3–S9 & 2004 Nature Publishing Group All rights reserved 0969-7128/04 $30.00 www.nature.com/gt

REVIEW

Retroviral vectors: new applications for an old tool J Barquinero, H Eixarch and M Pe´rez-Melgosa Unitat de Diagno`stic i Tera`pia Molecular, Centre de Transfusio´ i Banc de Teixits, Psg. Vall d’Hebron 119–129, 08035 Barcelona, Spain

Retroviral vectors (RVs) have been used for stable gene transfer into mammalian cells for more than 20 years. The most popular RVs are those derived from the Moloney murine leukaemia virus (MoMLV). One of their main limitations is their inability to transduce noncycling cells. However, they have a relatively simple genome and structure, are easy to use, and are relatively safe for in vivo applications. For the last two decades, the artificial evolution of RVs has paralleled evolution in their applications, which now include those as diverse as the generation of transgenic animals, the stable delivery of small interfering RNA (siRNA) and gene therapy clinical trials. Recent reports of two successful gene therapy clinical trials in patients with severe immunodeficiency disease in France and Italy, and the

development of T-cell acute leukaemia in two of 10 children participating in one of these clinical trials, demonstrate the great potential of RVs, but also some potential risks which may be intrinsically associated with their use. Basic aspects of RVs and vector production were reviewed in detail in a previous supplement of this journal. This article will first summarize some general aspects of retroviruses and RVs. Thereafter, recent developments in gene therapy using RVs, novel applications such as stable RNA interference and some other recent issues related to retroviral integration, including clonality studies after haematopoietic stem cell transplantation, retroviral tagging and insertional oncogenesis will be discussed. Gene Therapy (2004) 11, S3–S9. doi:10.1038/sj.gt.3302363

Keywords: retroviral vectors; retroviruses; applications; insertional mutagenesis; insertional oncogenesis

Retrovirus and retroviral vectors: an overview For the last two decades, retroviral vectors (RVs) have been major players in the fields of gene transfer and gene therapy.1 In the early 1980s, they were the first genetic vectors to permit an efficient and stable gene transfer into mammalian cells.2 In 1990, RVs were the first vectors used in a gene therapy clinical trial (for adenosine deaminase (ADA) deficiency).3 In 2000, after a decade of hopes and relative frustration, RVs were used in the first successful protocol that actually cured a genetic disease, demonstrating proof of concept for gene therapy.4 In all these years, vectors based on the Moloney murine leukaemia virus (MoMLV) have been pivotal in thousands of experiments, and continue to constitute the best tool available for stable gene transfer into a number of cell types and applications. Keys to their enormous success include the relative simplicity of their genomes, ease of use and their ability to integrate into the cell genome, permitting long-term transgene expression in the transduced cells or their progeny. These characteristics render them ideal vectors for a stable correction of genetic defects. In this regard, stem cells in general, and haematopoietic stem cells (HSCs) in particular, constitute optimal targets for RVmediated gene transfer, since transgenes can be expressed long term in vivo and may give rise, throughout a continuous amplification process, to a large progeny of gene-modified mature cells. In fact, it was not a surprise Correspondence: J Barquinero, Unitat de Diagno`stic i Tera`pia Molecular, Centre de Transfusio´ i Banc de Teixits, Psg. Vall d’Hebron 119–129, 08035 Barcelona, Spain

that the first two successful gene therapy clinical trials, on X-linked severe immunodeficiency disease (SCID-X1) and ADA deficiency, were based on HSCs and used RVs.4,5 Subsequent reports of development of T-cell leukaemia in two of 10 children treated with this experimental therapy in France6 were totally unexpected and will be discussed in more detail later in this article. RVs are derived from retroviruses, lipid-enveloped particles containing two identical copies of a linear single-stranded RNA genome of around 7–11 kb. They usually require binding to a specific membrane-bound receptor for viral entry. Cells not expressing the appropriate receptor are resistant to infection by a specific retrovirus. In cytoplasm, viral reverse transcriptase retrotranscribes the viral genome into double-stranded DNA (dsDNA), which binds to cellular proteins to form a nucleoprotein preintegration complex (PIC), which contains karyophilic elements that facilitate its migration to the nucleus. Nuclear membrane is a physical barrier for these PICs. The size and ability of these PICs to migrate to the nucleus determine their capacity to reach the cellular genome and transduce quiescent cells. In fact, for most retroviruses, such as MoMLV, PIC cannot cross the nuclear membrane, which renders these viruses and vectors incapable of infecting cells unless they undergo a mitotic cycle which actually disrupts the nuclear membrane. HIV-1 does not have this limitation, and is capable of transducing quiescent cells, although transduction efficiency by lentiviral vectors (LVs) is much higher in dividing cells than in nondividing cells. Retroviruses infect birds and mammals, where they tend to establish chronic infections, which are usually latent and well tolerated for long periods of time and

New applications of retroviral vectors J Barquinero et al

S4

eventually result in the development of immunodeficiency or malignancy. The retroviridae family consists of seven genera: alpharetrovirus (whose type species is the Avian leucosis virus), betaretrovirus (Mouse mammary tumour virus), gammaretrovirus (Murine leukaemia virus), deltaretrovirus (Bovine leukaemia virus), epsilonretrovirus (Walleye dermal sarcoma virus), lentivirus (Human immunodeficiency virus 1) and spumavirus (Human spumavirus).7 The first five genera were formerly classified as oncoretrovirus. Although, strictly speaking, vectors based on lentivirus and spumavirus or foamy virus (FVs) are also RVs, this term is often used to refer to vectors based on the MoMLV or the former oncoretrovirus. Although the inspiration to use viral vectors for human gene therapy probably originated in the early 1970s,8,9 the knowledge and tools to engineer viral genomes did not sufficiently develop until some years later. The idea underlying the generation of RVs was conceived independently by three researchers in the early 1980s, Edward M Scolnick at the NIH, Howard Temin at the University of Wisconsin at Madison and Robert Weinberg at the Massachusetts Institute of Technology, who observed that retroviruses could pick up normal cellular genes from animal cells during infection. The large-scale production of RVs for clinical applications still faces several challenges such as high titre vector production and the potential generation of replication-competent particles. Improving vector designs, finding strategies for cell targeting, understanding stem cell biology, achieving regulated and tissue-specific expression, modulating factors determining transgene expression and silencing phenomena or controlling immune responses to the transgene products constitute important issues that must be resolved if RVs are to be widely applied to treat human disease. In recent years, significant improvements in vectorology resulted in the generation of novel RVs with enhanced abilities and a greater potential for gene therapy applications. One of these advances consisted of engineering the regulatory regions of RV to enhance transgene expression or reduce transcriptional silencing in specific target cells. This led to the generation of novel families of vectors such as the FMEV, which contains hybrid regulatory sequences,10 or MND, which contains an engineered LTR derived from the myeloproliferative sarcoma virus.11 These vectors provide higher levels of transgene expression in haematopoietic progenitors than conventional RVs. The use of heterologous envelopes (env) from other viruses in RV production (pseudotyping) was a further major advance. RVs pseudotyped with the env of the gibbon ape leukaemia virus (GALV) or the feline endogenous retrovirus (RD114) transduced HSCs more efficiently than amphotropic RVs.12,13 The G protein from the vesicular stomatitis virus (VSV-G) can also be used to pseudotype RVs. Since VSV-G is a fusogenic protein that interacts with membrane phospholipids to facilitate viral entry, RVs pseudotyped with VSV-G do not require a cellular receptor for cell entry, which permits a very broad spectrum of infectivity.14 As an additional advantage, RVs pseudotyped with VSV-G or RD114 are physically more stable and can be concentrated by ultracentrifugation. A third improvement in RV technology was the development of self-inactivating (SIN) vectors.15 SIN

Gene Therapy

vectors contain a deletion of 299 bp in the 30 LTR, which blunts the functionality of the enhancer and the promoter. During reverse transcription, the deletion is transferred to the 50 LTR, resulting in the transcriptional inactivation of the provirus in the infected cell. SIN vectors require an internal promoter to drive transgene expression, which permits the use of inducible or tissuespecific promoters to regulate gene expression. Many RVs and most LVs currently used are SIN. Additionally, the absence of enhancer and promoter sequences in both LTRs of the integrated provirus minimizes the risk of insertional oncogenesis, thus providing a safer alternative for human gene therapy.

Novel applications of RVs Gene therapy for immunodeficiencies: the first successes, at last The ultimate challenge for RVs is gene therapy. Since the beginning of the first gene therapy clinical trials in the early 1990s, MoMLV-based vectors have been the preferred tool for the permanent correction of genetic defects. A total of 254 gene therapy clinical trials, which account for 28% of the total clinical trials approved worldwide, have relied on this type of vector (http:// www.wiley.co.uk/genmed/clinical). Primary immunodeficiencies are ideal candidates for stem cell gene therapy for a number of reasons: the potential selective advantage for the gene-corrected cells and the theoretical lack of immune responses to vector system components or transgene products. These are good reasons to explain why they were the first diseases in which gene therapy succeeded. In some respect, gene therapy clinical trials for SCID can be considered the optimum test bench for gene therapy for other diseases. The first milestone in clinical gene therapy was reported by Cavazzana-Calvo et al (2000)4 in children with SCID-X1. The disease is characterized by a defect in the common gamma chain (gc) gene, which encodes a subunit of the receptors for interleukins 2, 4, 7, 9, 15 and 21. SCID-X1 accounts for approximately half of all SCID cases. Children with the disease have an almost complete absence of functional B, T and NK cells, which leads to severe and recurrent infections that are generally fatal during the early years of life. The gene therapy clinical trial conducted by Fischer and co-workers involved retrovirally mediated transfer of the gc gene into the SCID-X1 patients’ own marrow cells, which were subsequently reinfused back into the patients in the absence of any myelosuppressive treatment. A few months after infusion of the gene-corrected marrow cells, T-cell counts were restored to almost normal levels and immune function was significantly improved in nine of 10 children treated. These results were confirmed independently by researchers in the UK using a similar protocol in six additional children. The key to the success of this landmark achievement is probably the fact that gene-modified cells, able to respond to appropriate growth factors, have the selective advantage of growing and repopulating an empty compartment (T cells), whereas most noncorrected cells undergo apoptosis. The final result is an almost complete repopulation by functional immune cells and normalization of most


immune parameters. This is generally accepted as the first genetic disease actually cured using gene therapy. The second major success in gene therapy was reported in 2002 in children with deficiency of the enzyme ADA.5 ADA deficiency represents approximately 10–20% of all SCIDs. The enzyme catalyses the deamination of deoxyadenosine to adenosine, and its deficiency results in an accumulation of metabolites highly toxic for T cells, and it is clinically manifested as a SCID. As occurs in SCID-X1, bone marrow transplant from a matched sibling is the best treatment, if available. If not, patients can be treated with pegylated bovine ADA (PEG-ADA), which partially corrects the immune deficiency in the majority of patients. As mentioned above, the first human gene therapy clinical trial was carried out in children with ADA deficiency in 1990. The therapy consisted of using an MoMLV-based vector to insert a functional version of the ADA gene into the patient’s own T cells, which were subsequently reinfused back into the patients. The two children treated, who were maintained on PEG-ADA therapy, still have marked T cells in their peripheral blood, albeit at low levels. For many years, this clinical trial was considered unsuccessful, probably because ADA replacement therapy was blunting the selective advantage for the genemodified cells. However, detailed analysis of the two patients who received the therapy, almost 14 years ago, revealed that in one patient, ADA activity was approximately 25% of the normal level, with more than 15% of his peripheral blood mononuclear cells (PBMCs) still carrying the therapeutic gene. In the second patient, the results were far less satisfactory, with under 5% of the normal ADA activity and less than 0.1% of the gene marking level in PBMCs.16 A second gene therapy strategy, using autologous CD34+ cells from umbilical cord blood, instead of mature T cells, of prenatally diagnosed children also resulted in suboptimal levels of gene marking (1–10% of the circulating T cells) and little clinical benefit.17 These patients were also maintained on replacement therapy with PEG-ADA, which might have accounted for the limited success. A clear indication that gene-corrected cells have a selective advantage in this disease, and that PEG-ADA treatment actually blunted this selection, is that gradual discontinuation of the therapy was associated with an increase in the percentage of marked peripheral blood T cells from a stable level of 1–3% to virtually 100%, with a parallel increase in absolute T-cell counts and their functionality. The relative disappointment of these early studies resulted, however, in the generation of highly useful information for subsequent gene therapy clinical trials. In 2002, researchers in Milan reported the results of their clinical trial on ADA deficiency. In this trial, autologous CD34+ cells were transduced with an RV similar to that used in previous studies. However, a sublethal dose of busulphan was given to the patients as a partially myeloablative treatment prior to the infusion of the gene-corrected cells. Moreover, no PEG-ADA was administered after cell infusion. After several months, a significant improvement in immune parameters was observed in virtually all patients treated thus far, except in one who still has subtherapeutic reconstitution levels. Strikingly, a significant marking was also observed in the myeloid compartment, with up to 20% of marking in granulocytes, monocytes and megakaryocytes, which

also suggests a selective advantage for ADA-expressing myeloid cells.18 These first successful achievements in human gene therapy bring new hope to the field. The fact that two forms of SCID were the first genetic diseases to be cured by gene therapy was not a surprise. The strong selective advantage for gene-corrected cells to proliferate or escape apoptosis is essential to explain this success. Unfortunately, this is not going to be the case for the majority of candidate diseases that can potentially be corrected at stem cell level. Several other forms of SCID, now in the pipeline for gene therapy clinical trials at different stages of preclinical development, include Wiskott–Aldrich syndrome,19 JAK3 deficiency,20 RAG-1 and RAG-2 deficiencies21 and Artemis deficiency.

S5

Stable siRNA delivery RNA interference (RNAi) is the phenomenon of sequence-specific, post-transcriptional gene silencing, elicited by small double-stranded RNAs (dsRNA) that are homologous in sequence to the silenced gene.22 Many experts consider RNAi the most useful tool for reverse genetics in the postgenomic era. Administration of synthetic duplexes of 21–23 nucleotide RNAs was found to mediate gene-specific RNAi in mammalian cells in vitro, without eliciting antiviral defence mechanisms.22 The use of these small interfering RNAs (siRNAs) constitutes a powerful tool to dissect gene function. Administration of these molecules can inhibit target gene expression both in vitro and in vivo. However, the limited lifespan of siRNAs in vivo and the short life of this effect prompted several groups to develop vector-based systems for specific siRNA delivery into mammalian cells. Efficient delivery and stable expression in target cells and their progeny necessarily requires the use of integrative vectors. Both types of vectors, RVs and LVs, have been reported to mediate an efficient and stable siRNA expression.23,24 Typically, the vectors use a pol III promoter, such as the U6, H1 or tRNA promoters. Vector systems for siRNA expression can be classified in two main groups depending on whether the expressed RNAs are hairpin-type or tandem-type. The former transcribe hairpin RNAs, which are subsequently processed into siRNA duplexes by endogenous RNase III. The tandemtype siRNA-expression vectors include both sense and antisense sequences, which are transcribed and subsequently annealed to generate the siRNA duplexes. Using an MoMLV-based vector encoding an siRNA specific for the p53 gene and a truncated form of human CD4 as a reporter gene, p53 expression was dramatically reduced in cultured cells.25 In another report by Rubinson et al,26 LVs encoding an siRNA significantly reduced CD8 expression in vitro. In addition, those investigators also showed that siRNA was able to inhibit transgene expression in vivo in an EGFP transgenic mouse model. Clonality studies in the haematopoietic system Each single retroviral integration or insertion site (RIS) in the genome of HSC constitutes a unique marker that is also transmitted to their progeny, thus turning RISs into distinctive clonal tags within the haematopoietic system. In the fields of gene therapy and stem cell biology, analysis of these RISs has two major applications: stem cell tagging for clonality studies and investigation of the Gene Therapy


S6

vector insertion site preferences in the genomes of target cells. A number of molecular techniques have been applied or specifically developed to analyse these RISs. Southern blot analysis of restriction fragments containing vector sequences permitted clonal characterization of murine and human haematopoiesis after transplantation of retrovirally transduced marrow cells.27–29 However, the introduction of PCR-based techniques affords a more precise characterization and mapping of RISs, since the amplified products containing the sequences flanking the proviral vector can be sequenced and mapped by in silico alignment search in mouse or human genome databases. These techniques include inverse PCR30 and, more recently, ligation-mediated PCR (LM-PCR),31 two-step PCR32 and linear amplification-mediated PCR (LAMPCR),33 which is currently the most popular among stem cell researchers. RIS analyses can be carried out in individual haematopoietic colonies that are, by definition, clonal. However, the newer and more sensitive approaches such as LAM-PCR allow analysis of whole blood samples or enriched phenotypically defined cell subpopulations such as neutrophils, T cells or NK cells. Since the information they provide is at clonal level, these techniques are unravelling questions that could not be properly answered in the past. LAM-PCR was used to analyse the clonal makeup of haematopoietic progeny in non-human primates transplanted with retrovirally transduced HSCs. More than 80 different transduced clones were detected in peripheral blood cells of these animals. Results of kinetic analyses of peripheral blood samples indicated that individual clones were continuously contributing to generating specific haematopoietic lineages for several years.33 In another setting, this technique was used to analyse the clonal pattern of the transduced haematopoietic progeny of ADA-deficient SCID children undergoing stem cell gene therapy. In these patients, T-cell reconstitution by gene-modified HSCs was found to be more oligoclonal than that of the previous report. In one patient, a single vector integrant was clearly predominant in T cells. At 8 years posttransplantation, isolated T cells with this prevailing integrant showed multiple patterns of T-cell receptor (TCR) gene rearrangement, thereby indicating that a single progenitor cell can generate the majority of the gene-marked cells over very long periods of time.34 In another study, analysis of the clonal dynamics of retrovirally transduced human HSCs grafted into NOD/SCID mice, including transplantation into secondary recipients, revealed that the stem cell compartment is highly heterogeneous and contains different types of repopulating cells, some contributing to short-term repopulation and others to long-term repopulation.29

Inserting tags in cancer pathways: retroviral tagging and insertional oncogenesis Until recently, very little was known regarding retroviral integration selectivity. Recent reports on insertional mutagenesis have renewed interest in investigating vector insertions. RVs were believed to integrate randomly in the genomes of host cells. Today, it is widely accepted that integration of retroviruses and RVs is skewed. RIS mapping studies have clearly shown that retroviruses have a striking integration preference for Gene Therapy

gene-rich regions.35 In addition, they may also prefer some chromosomes to others. Laufs and co-workers used retrovirally transduced human PBSCs transplanted into immunodeficient mice to analyse RISs in the haematopoietic progeny generated in recipient mice 8 weeks after transplantation. The authors found that, although most of the chromosomes were targeted, RV insertions occurred relatively more often in chromosomes 17 and 19 and in specific regions of chromosomes 6, 13 and 16.31 Regarding lentiviral insertion preferences, a recent report by Schroder et al characterized 524 RISs, after a single round of HIV infection in a human lymphoid cell line. Almost 70% of the RISs analysed were found within genes.36 In a more recent report, 903 MoMLV and 379 HIV-1 RISs were analysed in human cells. The study showed that both RVs and LVs prefer to integrate close to genes; however, the authors found significant differences between them. MoMLV-based vectors were more prone to insert close to the regulatory regions and the start of transcriptional units (either upstream or downstream), whereas integrations of HIV-1-based vectors did not show this preference and occurred anywhere in the transcriptional unit but not upstream of the start site.35 Another integrative virus, the adeno-associated virus (AAV) serotype 2, was recently reported to also prefer active genes for integration.37 Although viral insertion sites have been used for many years to identify potential oncogenes and signalling pathways involved in cancer, recent advances, such as high throughput PCR-based insertion site cloning, the availability of genetically modified animals and completion of the mouse genome project, have enormously widened the potential of this approach.38,39 A number of investigators reported the identification of hundreds of common integration sites (CISs, defined as insertion sites found in more than one tumour) associated with newly identified cancer genes in MoMLV-induced murine haematopoietic malignancies.40–42 In general, the vast majority of insertions were found outside the coding regions. Less than 10% of the RISs actually disrupted genes, which can thereafter be considered putative tumour suppressor genes. Interestingly, approximately 17–18% of ISs targeted transcription factors. All this information is now freely available at the Retroviral Tagged Cancer Gene Database (RTCGD, online access at http://RTCGD.ncifcrf.gov).43 The database, which contains more than 3100 RISs and more than 230 CISs from near to 1000 retrovirally induced haematopoietic tumours, manages several high-throughput insertional mutagenesis screening projects and contains all the information available on RISs reported in murine leukaemias and lymphomas. By definition, all insertions of genetic material into a cellular genome are mutagenic. However, the risk of insertional oncogenesis after standard gene transfer procedures using integrative vectors in preclinical animal studies, although sporadic cases have been reported,44 was believed to be very low.45 For this reason, the finding that two of the 10 children treated in the French clinical trial for SCID-X1 developed T-cell leukaemia due to single vector insertions, in both cases near the regulatory region of the LMO2 gene, was extremely shocking and unexpected. In recent months, a host of new information has accumulated that may finally translate into a clear idea of what caused these


two severe adverse effects (SAEs). This is crucial for the future of gene therapy so that the risks associated with a specific therapy or procedure can be minimized. Common sense tells us that individual observations apparently contradicting a large body of knowledge are probably exceptions to a general rule. The occurrence of T-cell leukaemia in two children undergoing the same gene therapy procedure, and the discovery that the same oncogene was probably involved in both SAEs, argued against the notion that risks of insertional oncogenesis for these procedures were very low. However, similar SAEs had not previously been reported, either around 200 patients treated so far with analogous gene therapy protocols for other diseases using integrative vectors, including ADA deficiency, or, albeit anecdotally, in the thousands of experimental animals undergoing preclinical gene transfer studies conducted to date.45 This raised the crucial issue of whether gene therapy using integrative vectors was much more dangerous than previously thought, or whether these two cases were more related to context-dependent factors and this could not be extrapolated to other settings. The question of why a retroviral insertion near the same proto-oncogene was related to the two cases of leukaemia was especially intriguing. There are at least two hypotheses, mutually nonexclusive, to respond to this question. One postulates that RVs insert preferentially into specific hotspots, and the second supports the idea that there are no significant preferences for integration near specific genes, but a strong selection of integrants carrying critical or synergistic oncogenic insertions, as previously reported in mice.39 The available data clearly support the second hypothesis. LMO2, the gene targeted in the two cases of leukaemia, encodes a transcription factor which is considered as a central regulator of normal haematopoiesis. Abnormal expression of LMO2 is a frequent finding in childhood T-cell acute lymphocytic leukaemia, mostly related to chromosomal translocations. Transgenic mice overexpressing LMO2 are more prone to developing T-cell leukaemias. In humans, LMO2 is highly expressed in bone marrow CD34+ progenitor cells, moderately expressed in immature thymic cells (CD34+ CD1a) and low or absent in more mature thymic cells (CD34+ CD1a+).6 Among the potential factors that might have contributed to these SAEs, one must consider the high number of marrow cells grafted in the two children (both patients received the two largest grafts). It has been estimated that the odds of an insertion in a given gene are B1 105. This is considering insertions less than 10 kbp away from the transcriptional units. However, their potential for transactivation is probably higher, since the retroviral enhancer has been reported to transactivate genes over much longer distances (up to 100 kbp). If we assume that each patient received at least 106 transduced progenitor cells, the grafts must have included, on average, 1–10 progenitor cell clones carrying insertions near the LMO2 locus.6 This has significant repercussions in determining the optimal graft size for these patients, since larger grafts are likely to be associated with higher chances of containing cells with dangerous insertions than smaller ones. Another potential contributing factor could have been the young age of the children (1 and 3 months) at the time they received

the gene-modified cells. Both were the youngest patients treated in the French trial. A younger age might be associated with a different gene expression profile in HSCs, a different proportion of HSCs subsets or an increased predisposition to oncogenic transformation. Another potential factor that may have contributed to development of the leukaemia is the high number of actively proliferating lymphoid progenitors that were collected from these patients and subsequently transduced and transplanted. This situation, which is probably specific for this particular disease, may have increased the number of progenitor cells potentially predisposed to dangerous vector insertions. Additionally, the tremendous selective pressure of the genecorrected progenitor cells to repopulate a virtually empty compartment in vivo (1000–10 000-fold expansion) may also have played a role.46 A second major question was whether disease- and/or protocol-specific factors might have played a role in the development of the leukaemias. In this regard, one possibility is the unregulated expression of the gc transgene itself, which is driven by the retroviral LTR. gc is a subunit of receptors such as IL2R which, upon activation, promotes T-cell proliferation and inhibits apoptosis. Whether expression of the other components of these receptors can limit or modulate the expression of the functional receptor is not yet known. However, it is noteworthy that Hacein-Bey-Abina et al6 were not able to detect any abnormalities of the gc-signalling pathway in the patients’ leukaemic cells. Involvement of gc overexpression and signalling in the pathogenicity of these leukaemias would be in agreement with the hypothesis of the multiple hit for cancer development. Also of interest was the finding that leukaemic cells of the SCIDX1 patients with insertional activation of LMO2 were CD3+ with fully rearranged TCR,6 whereas T-cell clones observed in children with ALL and LMO2 translocations are usually CD3, indicating that transformation occurs before TCR rearrangement. This suggests that a different mechanism of transformation probably operates in both types of LMO2-associated leukaemia. Insertional oncogenesis was also reported in experimental animals receiving bone marrow cells transduced with a recombinant retrovirus encoding truncated nerve growth factor receptor (NGFR). A total of 10 mice receiving a marrow graft of the same origin developed leukaemia. Analysis of the leukaemic cells showed that the RV was inserted within the Evi 1 gene, which was deregulated, and that the truncated and theoretically defective NGFR was in fact able to transduce a growth signal upon stimulation by its ligand.44 Retroviral insertional activation of the Evi 1 gene is one of the most common events associated with transformation in murine myeloid leukaemia. The role of the truncated NGFR in the leukaemic transformation remains controversial, as recently observed by Bonini et al. These authors were not able to detect any transformation events in a series of experiments using this transgene.47 Early this year, Dave´ et al, at the National Cancer Institute (USA), unveiled a key information that could finally explain the development of leukaemia. By searching their above-mentioned Mouse Retroviral Tagged Cancer Gene Database, they found two murine leukaemias associated with insertions at the LMO2 gene locus, and two with insertions at the IL2RG gene locus.

S7

Gene Therapy


S8

One of them actually contained insertions at both loci.48 The chances of these two unlikely events occurring in the same cell are extremely remote. This new finding further supports the notion that the leukaemic transformations were context dependent in these two children. This could be very good news for the field and may finally help to restore the image of RVs and confidence in gene therapy.49 The two cases of leukaemia reported by the French researchers triggered rapid unintentional responses by the media and society, and scepticism by the scientific community. Several regulatory agencies in many countries banned gene therapy protocols based on integrative vectors, and some important companies even reoriented their lines of research away from RV. Altogether, these unfounded reactions became a pessimistic wave of opinion at several levels, and led to the paradox that children with SCID-X1 were not allowed to receive a therapy that could save their lives. It is worth mentioning that unrelated bone marrow transplantation, a therapy widely accepted by society and regulatory agencies and which constitutes the only alternative therapy for these children, has a mortality rate of around 30%, a much higher risk than that of gene therapyinduced leukaemia in this disease (two of 16 cases treated worldwide). In addition, survivors of this procedure often have to face other medical complications such as graft-versus-host disease or an increased risk of lymphoma. This posed a great ethical dilemma, since gene therapy for SCID was still the best therapeutic option for these children. The facts are that current gene therapy protocols for SCID-X1 in Europe have saved 15 of 16 patients treated (10 in France and six in the UK), and that both children developing leukaemia subsequently underwent an unrelated marrow transplant and are currently in remission. At present, many countries have reopened gene therapy clinical trials, or permit them on a case-by-case basis. However, the image of gene therapy has been damaged, and the idea that ‘gene therapy causes cancer’ will remain in the minds of many people. Even if this finally clears, it will take a long time and a lot of effort for society to regain the confidence lost in gene therapy.

Future prospects for RVs RVs compete with LVs in many applications and for several target cells, including HSCs and other stem cells. LVs have already shown their huge potential in preclinical gene therapy studies. LVs can transduce quiescent cells,50 including terminally differentiated cells such as neurons,51 whereas RVs cannot. The use of LVs may reduce prolonged cell manipulation required for an efficient transduction using RVs, which would maintain their engrafting and long-term repopulation abilities. However, at the same time, they might be associated with higher risks of insertional oncogenesis since they have been reported to produce a high copy number per cell.52 At present, at least one clinical trial has already been approved for AIDS using LVs (http:// www.virxsys.com/news_and_events.html). Other potential competitors for RV as integrating vectors include foamy viruses, which are able to transduce quiescent cells, and transposable elements such as the sleeping beauty (SB) transposon,53 recently engineered as a nonviral vector. SB transposons permit an efficient cutGene Therapy

and-paste transposition in mammalian cells in vitro and in somatic tissues or the germ line of the mouse in vivo, and have already proved to be able to mediate gene transfer at clinically significant levels.54 As for RVs, they may still have a lot to say in the future. At present, no other vectors can substitute them for most gene therapy applications. However, most researchers agree they must be improved. There is much room for improvement, including better vector designs, systems for conditional expression, use of selectable genes, use of insulators to eliminate surrounding promoter effects of the retroviral enhancer or, alternatively, use of SIN vectors in conjunction with safer tissuespecific promoters. Regarding clinical trials, each vector system will have to be carefully validated on a case-bycase basis. It will take some time for gene therapy to be better understood and for tools to be improved; however, there is no doubt that, currently, it constitutes the best weapon and the only hope to cure genetic diseases.

Acknowledgements We thank Ana Limo´n, Robert Tjin and Christine O’Hara for critical review of the manuscript. The group is supported by grants from the Ministerio de Ciencia y Technologia FIS, the Spanish Collaborative Network on Transplantation and the Vth Framework Programme of the European Commission (http://www.inherinet.org).

References 1 Pages JC, Bru T. Toolbox for retrovectorologists. J Gene Med 2004; 6: S67–S82. 2 Cone RD, Mulligan RC. High-efficiency gene transfer into mammalian cells: generation of helper-free recombinant retrovirus with broad mammalian host range. Proc Natl Acad Sci USA 1984; 81: 6349–6353. 3 Anderson WF. September 14, 1990: the beginning [editorial]. Hum Gene Ther 1990; 1: 371–372. 4 Cavazzana-Calvo M et al. Gene therapy of human severe combined immunodeficiency (SCID)-X1 disease. Science 2000; 288: 669–672. 5 Aiuti A et al. Correction of ADA-SCID by stem cell gene therapy combined with nonmyeloablative conditioning. Science 2002; 296: 2410–2413. 6 Hacein-Bey-Abina S et al. LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science 2003; 302: 415–419. 7 Pringle CR. Virus taxonomy – 1999. The universal system of virus taxonomy, updated to include the new proposals ratified by the International Committee on Taxonomy of Viruses during 1998. Arch Virol 1999; 144: 421–429. 8 Friedmann T, Roblin R. Gene therapy for human genetic disease? Science 1972; 175: 949–955. 9 Anderson WF, Fletcher JC. Sounding boards. Gene therapy in human beings: when is it ethical to begin? N Engl J Med 1980; 303: 1293–1297. 10 Baum C et al. Novel retroviral vectors for efficient expression of the multidrug resistance (mdr-1) gene in early hematopoietic cells. J Virol 1995; 69: 7541–7547. 11 Challita PM et al. Multiple modifications in cis elements of the long terminal repeat of retroviral vectors lead to increased expression and decreased DNA methylation in embryonic carcinoma cells. J Virol 1995; 69: 748–755.


S9 12 van Hennik PB et al. Highly efficient transduction of the green fluorescent protein gene in human umbilical cord blood stem cells capable of cobblestone formation in long-term cultures and multilineage engraftment of immunodeficient mice. Blood 1998; 92: 4013–4022. 13 Van Der Loo JC et al. Optimization of gene transfer into primitive human hematopoietic cells of granulocyte-colony stimulating factor-mobilized peripheral blood using low-dose cytokines and comparison of a gibbon ape leukemia virus versus an RD114pseudotyped retroviral vector. Hum Gene Ther 2002; 13: 1317–1330. 14 Yee JK, Friedmann T, Burns JC. Generation of high-titer pseudotyped retroviral vectors with very broad host range. Methods Cell Biol 1994; 43: 99–112. 15 Yu SF et al. Self-inactivating retroviral vectors designed for transfer of whole genes into mammalian cells. Proc Natl Acad Sci USA 1986; 83: 3194–3198. 16 Muul LM et al. Persistence and expression of the adenosine deaminase gene for 12 years and immune reaction to gene transfer components: long-term results of the first clinical gene therapy trial. Blood 2003; 101: 2563–2569. 17 Engel BC, Kohn DB. Gene therapy for inborn and acquired immune deficiency disorders. Acta Haematol 2003; 110: 60–70. 18 Aiuti A et al. Gene therapy for adenosine deaminase deficiency. Curr Opin Allergy Clin Immunol 2003; 3: 461–466. 19 Strom TS et al. Defects in T-cell-mediated immunity to influenza virus in murine Wiskott–Aldrich syndrome are corrected by oncoretroviral vector-mediated gene transfer into repopulating hematopoietic cells. Blood 2003; 102: 3108–3116. 20 Bunting KD, Lu T, Kelly PF, Sorrentino BP. Self-selection by genetically modified committed lymphocyte precursors reverses the phenotype of JAK3-deficient mice without myeloablation. Hum Gene Ther 2000; 11: 2353–2364. 21 Yates F et al. Gene therapy of RAG-2/ mice: sustained correction of the immunodeficiency. Blood 2002; 100: 3942–3949. 22 Elbashir SM et al. Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. Nature 2001; 411: 494–498. 23 Yu JY, DeRuiter SL, Turner DL. RNA interference by expression of short-interfering RNAs and hairpin RNAs in mammalian cells. Proc Natl Acad Sci USA 2002; 99: 6047–6052. 24 McManus MT, Sharp PA. Gene silencing in mammals by small interfering RNAs. Nat Rev Genet 2002; 3: 737–747. 25 Barton GM, Medzhitov R. Retroviral delivery of small interfering RNA into primary cells. Proc Natl Acad Sci USA 2002; 99: 14943–14945. 26 Rubinson DA et al. A lentivirus-based system to functionally silence genes in primary mammalian cells, stem cells and transgenic mice by RNA interference. Nat Genet 2003; 33: 401–406. 27 Lemischka IR, Raulet DH, Mulligan RC. Developmental potential and dynamic behavior of hematopoietic stem cells. Cell 1986; 45: 917–927. 28 Barquinero J et al. Efficient transduction of human hematopoietic repopulating cells generating stable engraftment of transgeneexpressing cells in NOD/SCID mice. Blood 2000; 95: 3085–3093. 29 Guenechea G, Gan OI, Dorrell C, Dick JE. Distinct classes of human stem cells that differ in proliferative and self-renewal potential. Nat Immunol 2001; 2: 75–82. 30 Nolta JA et al. Transduction of pluripotent human hematopoietic stem cells demonstrated by clonal analysis after engraftment in immune-deficient mice. Proc Natl Acad Sci USA 1996; 93: 2414–2419.

31 Laufs S et al. Retroviral vector integration occurs in preferred genomic targets of human bone marrow-repopulating cells. Blood 2003; 101: 2191–2198. 32 Gentner B et al. Rapid detection of retroviral vector integration sites in colony-forming human peripheral blood progenitor cells using PCR with arbitrary primers. Gene Therapy 2003; 10: 789–794. 33 Schmidt M et al. Polyclonal long-term repopulating stem cell clones in a primate model. Blood 2002; 100: 2737–2743. 34 Schmidt M et al. Clonality analysis after retroviral-mediated gene transfer to CD34(+) cells from the cord blood of ADAdeficient SCID neonates. Nat Med 2003; 9: 463–468. 35 Wu X, Li Y, Crise B, Burgess SM. Transcription start regions in the human genome are favored targets for MLV integration. Science 2003; 300: 1749–1751. 36 Schroder AR et al. HIV-1 integration in the human genome favors active genes and local hotspots. Cell 2002; 110: 521–529. 37 Nakai H et al. AAV serotype 2 vectors preferentially integrate into active genes in mice. Nat Genet 2003; 34: 297–302. 38 Dudley JP. Tag, you’re hit: retroviral insertions identify genes involved in cancer. Trends Mol Med 2003; 9: 43–45. 39 Mikkers H, Berns A. Retroviral insertional mutagenesis: tagging cancer pathways. Adv Cancer Res 2003; 88: 53–99. 40 Suzuki T et al. New genes involved in cancer identified by retroviral tagging. Nat Genet 2002; 32: 166–174. 41 Lund AH et al. Genome-wide retroviral insertional tagging of genes involved in cancer in Cdkn2a-deficient mice. Nat Genet 2002; 32: 160–165. 42 Mikkers H et al. High-throughput retroviral tagging to identify components of specific signaling pathways in cancer. Nat Genet 2002; 32: 153–159. 43 Akagi K et al. RTCGD: retroviral tagged cancer gene database. Nucleic Acids Res 2004; 32: Database issue D523–D527. 44 Li Z et al. Murine leukemia induced by retroviral gene marking. Science 2002; 296: 497. 45 Kohn DB et al. American Society of Gene Therapy (ASGT) ad hoc subcommittee on retroviral-mediated gene transfer to hematopoietic stem cells. Mol Ther 2003; 8: 180–187. 46 Baum C et al. Chance or necessity? Insertional mutagenesis in gene therapy and its consequences. Mol Ther 2004; 9: 5–13. 47 Bonini C et al. Safety of retroviral gene marking with a truncated NGF receptor. Nat Med 2003; 9: 367–369. 48 Dave UP, Jenkins NA, Copeland NG. Gene therapy insertional mutagenesis insights. Science 2004; 303: 333. 49 Berns A. Good news for gene therapy. N Engl J Med 2004; 350: 1679–1680. 50 Uchida N et al. HIV, but not murine leukemia virus, vectors mediate high efficiency gene transfer into freshly isolated G0/ G1 human hematopoietic stem cells. Proc Natl Acad Sci USA 1998; 95: 11939–11944. 51 Blomer U et al. Highly efficient and sustained gene transfer in adult neurons with a lentivirus vector. J Virol 1997; 71: 6641–6649. 52 Woods NB et al. Lentiviral vector transduction of NOD/SCID repopulating cells results in multiple vector integrations per transduced cell: risk of insertional mutagenesis. Blood 2003; 101: 1284–1289. 53 Ivics Z et al. The Sleeping Beauty transposable element: evolution, regulation and genetic applications. Curr Issues Mol Biol 2004; 6: 43–55. 54 Yant SR et al. Somatic integration and long-term transgene expression in normal and haemophilic mice using a DNA transposon system. Nat Genet 2000; 25: 35–41.

Gene Therapy