To the Editor:

Gene therapy has promised much and delivered little for patients over many years, but a range of clinical projects are now showing clear signs of efficacy, giving considerable reassurance that technology from this field will soon enter into mainstream medicine. The studies showing clinical promise cover several different diseases and use a variety of vectors, ranging from simple oligonucleotides to replicating lytic viruses. However, one crucial unifying feature is that they all focus on realistic and achievable objectives, particularly in terms of effective delivery of the therapeutic agent to target cells, but also using levels and durations of transgene expression commensurate with the desired treatment outcome. Careful selection of disease targets on the basis of the vector systems available is an essential prerequisite to success. Here we highlight some of the more notable recent advances and try to put them into the context of the underpinning technological improvements; details of selected recent clinical studies are summarized in Table 1. These and many other areas of scientific and clinical progress were recently updated at a collaborative congress held in Brighton, UK, between the European Society of Gene & Cell Therapy (ESGCT) and the British Society for Gene Therapy (BSGT)1 (

Table 1 Some recent advances in clinical gene therapy

Technological landscape

Development of better genetic vectors over the last two decades has produced a number of therapeutic reagents that have now successfully transferred from laboratory study to clinical application. Most clinical studies now use viral-based vector systems that benefit from high efficiency but may be compromised by immunogenicity and toxicities specific to the type of vector employed. For example, integrating retroviral vectors mediate long-term expression of transgenes in replicating and expanding cell populations, but in some circumstances have produced toxicity from insertional mutagenesis. Whereas retroviral vectors are often used to mediate long-term stable gene expression, other viral vectors, such as those based on adeno-associated viruses (AAV), are being widely explored to mediate extended transgene expression in non-replicating cells, where they are thought to persist as non-integrated non-replicating episomes. However, all viral vectors are limited by their packaging capacity—in other words, by the size of the transgenic cassette, which may restrict capacity for incorporation of large genes or complex regulatory elements. Furthermore, very large-scale bio-production remains challenging for some viral vectors. Non-viral technologies based on different chemistries, targeting strategies and genetic constructs are being developed to remove some of these problems, and although efficiency has remained problematic, for some applications there are signs of genuine efficacy. Emerging successful strategies are therefore a 'best fit' based on knowledge of the target disease, the desired regulatory pattern of gene expression and the bioactivity of different vector systems.

Gene therapies for genetic disease

Congenital retinal blinding conditions affect approximately 1 in 2,000 people worldwide, yet are without effective treatments. They are clinically heterogeneous, with mutations in over 165 different genes identified as causing disease largely through dysfunction of photoreceptor, bipolar or retinal pigment epithelium cells. Although conventional medicine has little to offer, the eye presents one of the most accessible targets for localized delivery and therefore for application of gene (and cellular) therapies. Furthermore, there may be advantages in terms of immunological privilege and ability to monitor local outcomes and potential toxicities. Several studies have recently reported remarkable clinical improvements after treatment of patients with Leber's congenital amaurosis (LCA), for which alternative therapies are unavailable. Highly efficient delivery of AAV to retinal pigment epithelial cells was achieved by precise surgical subretinal injection. In one study, AAV serotype 2 (AAV2) expressing RPE65 (the retinal pigment protein that is lost in LCA) under transcriptional control of its physiological promoter was administered to three patients (Table 1)2. Although none of the recipients showed any change in retinal responses measured by electroretinography in this trial, one patient had significant improvement in visual function measured by microperimetry and dark-adapted perimetry, and also showed improvement in a subjective test of visual mobility. In another study, each patient showed a modest improvement in measures of retinal function in subjective tests of visual acuity, although normal vision was not achieved3. Finally, a dose-escalation study of AAV2 expressing RPE65 under control of the chicken β-actin promoter produced sustained improvements in subjective and objective measurements of vision, with at least a 2-log-unit increase in pupillary light responses in all 12 patients. The greatest improvement was noted in children, all of whom gained ambulatory vision3,4. The fact that maximal benefit occurred in younger recipients indicates that regenerative therapies for this and many other disorders, whether through gene or stem cell approaches, should be targeted as early as possible, before functional recovery becomes irretrievable. As a corollary to this, clinical trials should not necessarily dwell on extended study in patients for whom clinical benefit is unlikely, as these populations may not be predictive of efficacy or even safety. In many diseases, carefully structured studies in children are therefore of paramount importance.

As these highly encouraging trials progress into phase II and eventually phase III, trials are planned for other genetic forms of inherited blindness, as well as for use of the same or similar technologies to treat acquired disease such as vascular retinopathies and age-related macular degeneration. At the same time, the technologies are continually being refined to enhance efficacy and safety—for example, through the use of alternative AAV serotypes shown to transduce retinal cells more efficiently in vivo. The small size of the target organ being treated means that production of vector for large numbers of patients will be relatively straightforward. Furthermore, the prospect of a one-time treatment for what is essentially a lifetime illness with significant healthcare and social costs implies that this novel therapeutic approach will be extremely cost-effective. The potential for the development of gene therapies into mainstream medicine for ocular disorders appears remarkably good.

While localized delivery of gene therapeutics has clear applications for some disorders, preclinical studies have suggested that some AAV serotypes injected intravascularly will predominantly transduce hepatocytes, providing a useful platform for local or systemic expression of a wide range of therapeutic proteins from this tissue. Although attempts to use this strategy for treatment of metabolic disease and hemophilia have previously been disappointing as immunological responses to the viral proteins caused toxicity or prevented long-term expression, new data from an experimental treatment of hemophilia B has demonstrated that AAV serotype 8 (AAV8) can be administered intravenously to deliver a codon-optimized gene encoding Factor IX (FIX) into liver5. AAV8 was selected for this study partly because of its hepatic tropism, but also because most individuals have not been previously exposed to the virus, diminishing the likelihood of virus-specific memory T cells that might attack transduced hepatocytes, as noted in a previous study using AAV2. Eligible patients were selected on the basis of having only low levels of antibodies against AAV8. Transgene expression was regulated by core domains from the human apolipoprotein hepatic control region and the human α-1-antitrypsin gene promoter. The approach has now been explored in six patient volunteers, and all have shown signs of durable clinical benefit associated with recovery of 2–11% of normal levels of FIX. There were signs of transient inflammation at higher AAV doses, perhaps reflecting immune recognition of transduced hepatocytes, although this was managed effectively using glucocorticoids with only small effects on FIX levels.

Unlike inherited blinding disorders, hemophilias and some metabolic diseases can be treated by exogenous-factor or enzyme-replacement therapy. However, the prospect of a one-time treatment as opposed to lifelong drug administration is clearly attractive on a number of counts, not only from the quality-of-life perspective but also economically. It is also important to remember that a significant proportion of the global population with these diseases does not have access to what are regarded as conventional therapies. Other tissues may also provide tractable targets for a similar approach. For example, systemic injection of AAV serotype 9 has been shown to result in widespread transduction of neurones and is soon to be evaluated in patients with spinal muscular atrophy, among others. The highly promising early data from the hemophilia B study therefore raises the real possibility of successful long-term treatment by gene therapy of a large number of patients with inherited disease5,6.

Gene transfer to hematopoietic stem cells (HSCs), first shown to have major therapeutic effects in severe combined immunodeficiency (SCID) over 10 years ago, continues to show particular promise, with applications now broadening to other hematological (including Wiskott-Aldrich syndrome (WAS) and B thalassemia) and metabolic disorders (X-linked adrenoleukodystrophy and metachromatic leukodystrophy)7. The long-term data in over 90 patients with a number of inherited primary immunodeficiencies who have received gene therapy using conventional γ-retroviral vectors over the last decade show well over 90% overall survival, with the vast majority experiencing significant clinical benefit, despite insertional mutagenesis in around ten patients to date7. These striking results become more impressive when compared with published outcomes for alternative therapies, bearing in mind that the patients selected for these trials were usually without human leukocyte antigen (HLA)–matched hematopoietic stem cell donors. The observed toxicities in these trials are now known to share a common mechanism, namely upregulated expression of proto-oncogenes induced by powerful enhancer sequences within the long terminal repeats (LTRs) of the early-generation γ-retroviral vectors that were used. During the last decade, conventional γ-retroviral vectors have been largely replaced by so-called 'self-inactivating' γ-retroviral vector systems and particularly lentiviral vector systems, which can tolerate larger transgene sequences. These are designed to limit the risks of insertional mutagenesis through use of regulatory sequences that are much less likely to interfere with expression of host-cell genes adjacent to the vector integration site. Some benefits may accrue from changing the class of vector from gammaretrovirus to lentivirus, as the intrinsic integration strategy of the latter is less prone to targeting regulatory regions of the genome. Lentiviral vectors may also be more efficient for transduction of HSCs, as they do not require breakdown of the nuclear membrane during mitosis to access the chromosomes and can therefore transduce HSCs through minimal ex vivo culture times, thereby preserving engraftment capability. One of the first studies using a lentivirus to transduce HSCs expressed the ABCD1 gene and showed successful protection from demyelination in two boys with adrenoleukodystrophy8, establishing the impressive potential of this approach. As examples of the ongoing process of technological refinement, new trials for X-linked SCID, WAS and chronic granulomatous disease have recently entered the clinical arena. However, large-scale production of these vectors to Good Manufacturing Practice–grade remains a biotechnological challenge that will need to be addressed before widespread application to large numbers of patients can be contemplated.

Aside from viral vectors, antisense oligonucleotides based on relatively simple chemistries have recently been used to promote exon skipping in some patients with Duchenne muscular dystrophy (DMD). By allowing the spliceosome to 'skip' a frameshift mutation in exon 51 of the dystrophin gene (which characterizes a major subset of patients with DMD), this approach has successfully induced expression of a short, though 'in frame' and partly functional dystrophin protein9. A real step forward here is the observation that oligonucleotide treatments may be given systemically to access dystrophic muscle systems throughout the body, thereby obviating the very significant difficulties associated with multiple local deliveries. In principle, exon skipping could be useful to induce production of in-frame though truncated proteins for other genetic diseases such as the dysferlinopathies10 as well as for disrupting aberrant expression of proteins in cancer. However, the successful tissue uptake of antisense oligonucleotides in DMD is thought to reflect unusually high membrane permeability of dystrophic muscle cells, so widespread use of this technology may require continued focus on delivery mechanisms, both viral and non-viral.

Regenerative gene therapy

Outside the context of Mendelian genetic disorders, very similar approaches to those described above are being evaluated in a number of 'degenerative' conditions. After observations that chronic heart failure is associated with low levels of sarcoplasmic calcium, the coronary arteries of patients with advanced heart failure were injected with AAV vectors (serotype 1, AAV1) encoding the enzyme responsible for reloading the sarcoplasmic reticulum with calcium during relaxation, the sarcoplasmic reticulum Ca2+-ATPase (SERCA2a), under transcriptional control of the cytomegalovirus immediate early promoter and enhancer. Patients with high levels of AAV1-specific antibodies were excluded from the study, given that the route of delivery was intravenous, and a total of 39 patients were randomized into the placebo and three treatment groups11. At the highest dose (n = 9 for test and n = 14 for placebo), the treatment resulted in a positive outcome for several predefined functional parameters: high-dose treatment, compared with placebo, led to substantial increases in time to clinical events, lower frequency of cardiovascular events observed at 12 months, and shorter mean duration of cardiovascular hospitalizations (0.4 versus 4.5 days). Similarly, lentiviruses engineered to express dopamine-biosynthetic enzymes are being injected stereotactically directly into the corpus striatum of patients with Parkinson's disease. Encouraging and objective signs of efficacy have been recorded using the Unified Parkinson's Disease Rating Scale and quality-of-life measures, with some patients now approaching 3 years since treatment. These very encouraging results make a strong case for conducting larger trials.

Infectious disease and cancer

Recombinant viruses provide a powerful vaccine platform for infectious diseases, allowing expression of proteins and epitopes from target pathogens in the context of strong background immune stimulation by the viral vector. A range of vaccine strategies are being explored and extensive trials conducted for prophylaxis of malaria, tuberculosis and HIV12. Particularly exciting are new innovations using gene-based approaches as a treatment to alleviate disease in people who are already infected. Among the most advanced are clinical studies using 'locked nucleic acid' oligonucleotides targeting microRNA 122 as anti–hepatitis C agents. There are several anti-infective approaches on the horizon, including strategies based on transfer of viral restriction factors. As an example, long-term cellular protection against HIV infection might be achieved through modification of hematopoietic stem cells to express a TRIM5a–cyclophilin A fusion protein, designed to bind incoming HIV virus and target it for proteasomal degradation before reverse transcription can occur13.

Targeted endonucleases are being developed to facilitate repair of DNA mutations through homologous recombination and have already shown some efficacy in preclinical models including hemophilia14. Although promising, this approach may not yet be efficient enough to enter the clinical arena, and some questions remain regarding selectivity and off-target activity. However, endonucleases also provide the possibility of selective disruption of specific genes through induction of double-strand DNA breakages and host-cell repair by nonhomologous end-joining. Following the observation that HSCs lacking expression of the CCR5 coreceptor for HIV produce CD4+ T cells that are HIV resistant, early trials in patients with HIV are showing signs of progress after infusion of autologous CD4+ T cells subjected to genome editing ex vivo with zinc-finger nucleases to prevent CCR5 expression15. Although this is one example of current clinical application, endonuclease-based technologies are extremely versatile, and their use to modify the cellular genome may eventually find a broad array of applications in many different inherited and acquired diseases.

In the field of cancer, cytotoxic T cells modified to recognize specific tumor targets using engineered T-cell receptors, or chimeric antigen receptors (CARs) formed using single-chain antibodies (as the extracellular domain) fused to internal signaling and additional costimulatory domains, are showing increasing promise as the technology matures16. Tumor-killing 'oncolytic' viruses, where lytic viruses replicate selectively in cancer cells and lyse them before spreading to adjacent cells, are also showing promise. Virus-mediated cytotoxicity is particularly attractive for cancer treatment as it may be able to overcome deficiencies in cellular apoptosis mechanisms that characterize many drug-resistant tumors. Wild-type reovirus has been studied in over 20 clinical trials treating a range of tumor types and has shown signs of therapeutic activity both as a single agent17 and in combination with chemotherapy18. Several trials are ongoing, and the agent is now in a phase III trial with paclitaxel (Abraxane) and carboplatin (Paraplatin) for treatment of squamous-cell carcinoma of the head and neck. Conditionally replicating vaccinia virus modified to express granulocyte-macrophage colony-stimulating factor (GM-CSF) has shown a good toxicology profile and signs of activity in a range of phase I and phase II trials, including extended survival after direct injection of the virus into non-resectable liver cancer19. A conditionally replicating herpes virus, also expressing GM-CSF and designed to stimulate an anticancer immune response after direct injection into tumor nodules, has likewise performed well in early-phase trials. It is now in phase III trials for treatment of melanoma and seems to be another promising candidate for product licensure20. Oncolytic viruses can spread from cell to cell, providing the possibility of increased penetration into tumor nodules after the initial infection; nevertheless, it remains essential that the initial delivery can reach enough tumor cells to allow productive infection and adequate virus replication. For viruses with poor stability in human blood, the greatest success to date has been achieved after direct injection into tumor nodules21, although recently a great deal of attention has been given to exploring the feasibility of using even these oncolytic viruses systemically to treat metastatic disease22.


The clinical benefits seen from a variety of gene therapy approaches, coupled with growing enthusiasm for applying genetic technologies in medicine, strongly indicate that gene therapy has now passed beyond the proof-of-principle stage and may soon provide realistic approaches for several previously intractable medical disorders. Strategies to enable transgene expression in a sufficient proportion of target cells are central to progress, together with achieving an appropriate duration of transgene expression without unwanted immune response. This may present significant challenges, particularly where very high-level or tightly regulated gene expression is necessary, although technology is evolving quickly and the regulatory climate may be changing to facilitate testing of mechanism and molecular performance as well as toxicology in early-phase clinical trials. Some of the best candidate diseases for gene supplementation therapy are those in which the transgenic cells are endowed with specific growth and survival advantages owing to functional correction, allowing physiological factors to regulate their repopulation of the diseased environment. However, in other settings a degree of host manipulation may be used to ensure sufficient engraftment or survival of transduced cells (e.g., through chemotherapeutic or immunosuppressive conditioning of the bone marrow or host immunity).

No real consensus has yet emerged on whether transgene expression must be controlled by cognate promoters to allow physiological regulation of activity. Indeed, for many disorders the profile of gene expression required to achieve significant therapeutic effect is achievable with relatively simple expression systems. For example, the frequency of spontaneous hemorrhage is greatly reduced in hemophilia B with a small fraction of normal circulating levels of FIX. Similarly, a fraction of functional circulating phagocytes is sufficient to offer substantial protection against infection in some inherited immunodeficiencies. Diseases most amenable to effective treatment with current gene therapy approaches may therefore be those where the therapeutic window for functional protein expression is relatively broad, and where a large number of target cells are accessible to transduction.

Previous expectations that gene therapy would produce a 'cure-all' solution for intractable diseases were unrealistic. The agents in question are a diverse spectrum of nucleic-acid based medicines that are formulated in very different ways, yet can be used to prevent, alleviate and provide long-lasting treatments for a wide variety of diseases both inherited and acquired. In some cases they can now realistically provide physicians and patients with new therapeutic options where more conventional approaches have failed, a testament to the huge amount of scientific research in the field over the last 10–20 years. By decreasing the requirement for frequent repeated interventions, many gene therapy approaches can lead to substantial savings in the costs of lifetime medication. However, it is also noteworthy that without clinical trials, the field would not have progressed nearly so rapidly. The fact that some gene therapy strategies are finally beginning to deliver on their potential may well herald a raft of new and imaginative interventional approaches designed to exploit recent insights into cell biology and disease processes.