Gene Therapy for Rare Diseases: Summary of a National Institutes of Health Workshop, September 13, 2012

Abstract

Gene therapy has shown clinical efficacy for several rare diseases, using different approaches and vectors. The Gene Therapy for Rare Diseases workshop, sponsored by the National Institutes of Health (NIH) Office of Biotechnology Activities and Office of Rare Diseases Research, brought together investigators from different disciplines to discuss the challenges and opportunities for advancing the field including means for enhancing data sharing for preclinical and clinical studies, development and utilization of available NIH resources, and interactions with the U.S. Food and Drug Administration.

Introduction

Over the past 5 years, clinical efficacy has been demonstrated in a number of gene therapy trials for rare diseases. These include trials for congenital eye diseases (Cideciyan, 2010) and a number of hereditary immune system disorders (Aiuti et al., 2009; Booth et al., 2011; Hacein-Bey-Abina et al., 2010). In addition, a gene therapy approach for lipoprotein lipase (LPL) deficiency (Stroes et al., 2008) has become the first licensed gene therapy product in Europe. These clinical successes are to be celebrated, and they also raise the hope for the development of gene transfer approaches for other rare diseases for which patients have few satisfactory therapeutic options.

Rare diseases are a heterogeneous group of diseases. According to the Orphan Drug Act [Public Law 97-414], which provides certain incentives for development of therapeutics for “orphan diseases,” an orphan disease is one that affects less than 200,000 individuals in the United States.¹ Some rare diseases affect far fewer than 200,000 individuals, such as hemophilia, which affects approximately 20,000 individuals in the United States (http://www.cdc.gov/ncbddd/hemophilia/facts.html), and there are other rare diseases for which the affected population may only be measured in the hundreds. Developing a new therapeutic for any disease presents challenges, both from a scientific and financial perspective, but these challenges may be even greater when the population is very small. An important question is whether it is possible to leverage the knowledge gained in these initial gene therapy trials for rare diseases to expand the technology more efficiently to other rare diseases that may share similar pathology, although not necessarily the identical target genes.

To try to answer this question, a workshop on “Gene Transfer for Rare Diseases” was organized and sponsored by the National Institutes of Health (NIH) as a joint effort of the NIH Office of Biotechnology Activities in the Office of Science Policy, Office of the Director, and the NIH Office of Rare Diseases Research (ORDR) in the National Center for Advancing Translational Sciences (NCATS). The overall goal was to discuss how clinical gene therapy for rare diseases can be advanced by increasing the efficiency of (1) the preclinical development processes through enhanced data sharing of preclinical and clinical studies, (2) development and utilization of available NIH resources, and (3) interactions with the U.S. Food and Drug Administration (FDA). The workshop was attended by investigators from the academic community, representatives from biotechnology companies, leaders from various NIH offices concerned with gene therapy and rare diseases, officials from FDA, and members of the NIH Recombinant DNA Advisory Committee. This article is a summary of the discussion and represents the views of the individual authors, not those of the NIH or FDA.

The Clinical Experiences to Date

The meeting included a summary of selected clinical experiences with gene therapy for hemophilia, Leber congenital amaurosis (LCA), LPL deficiency, and the severe combined immunodeficiency disorders (SCIDs) such as adenosine deaminase deficiency (ADA-SCID) and SCID-X1. In addition to reviewing the clinical successes, these talks also highlighted some of the ongoing challenges.

Dr. Katherine High presented her experience developing gene therapy for hemophilia B. Hemophilia B, an X-linked bleeding disorder caused by the absence of functional coagulation factor IX (FIX), has been an attractive target for gene transfer for several reasons including that efficacy requires only modest increases in FIX levels, and gene expression does not require precise regulation or a specific target tissue. The small size of the gene allows replacement using many vector systems. There are also appropriate small and large animal models for this disease. Early studies demonstrated that adeno-associated virus (AAV)-mediated gene transfer could transduce human liver cells and generate short-term expression of therapeutic levels of FIX. Of interest, the major challenges faced were not predicted by the preclinical studies, including the detection of vector sequences in the semen and an unexpected immune response to the AAV capsid that resulted in a transient transaminitis and loss of transgene expression (Manno et al., 2006). Data indicated that the loss of transgene expression was due to the destruction of the transduced liver cells by AAV capsid–specific CD8⁺ T cells (Mingozzi et al., 2007). In subjects with pre-existing neutralizing antibodies to AAV capsid, modest titers blocked transduction. The trials were amended in response to these events, in the first case through ongoing monitoring and barrier contraception, and in the second case, through the institution of immunosuppression protocols designed to blunt the immune system response to the AAV capsid so that transgene expression could continue.

A recent trial conducted by investigators at St. Jude Children's Research Hospital and the University College of London evaluated the use of similar doses of an AAV serotype 8, codon-optimized, self-complementary vector (Nathwani et al., 2011). The majority of the six subjects dosed had increased FIX serum levels in the range of 2–6% of normal with decreased or abolished need for protein replacement therapy. A key to success was treating early rises in liver transaminase levels with short courses of steroids to control the anti-capsid T-cell response and preserve hepatic production of vector-encoded FIX. However, challenges remain in using this approach in patients with pre-existing antibodies for AAV, which may also limit redosing, and in enrolling subjects who may not be able to tolerate the immunosuppression needed to inhibit an immune response. Future efforts in the field will focus on improving the quality of the clinical AAV vector preparations by removing empty capsid-containing particles from the product and by developing AAV vectors expressing factor VIII for hemophilia A.

AAV-based gene therapy has also been used successfully to treat one of the molecular subtypes of LCA, a form of congenital blindness affecting the retinal pigment epithelium (RPE). LCA is caused by a mutation in 1 of 15 genes identified so far. Clinical trials to date have mainly focused on research participants with mutations in the RPE65 gene, which encodes the critical isomerase of the retinoid (visual) cycle. Dr. Samuel G. Jacobson described a Phase I dose-escalation study involving subretinal injections of AAV2 vectors expressing RPE65 in which all 15 research participants showed varying degrees of improved visual function and there were no serious adverse events (Jacobson et al., 2012). The effects have persisted at least 3 years. The development of this approach was detailed, including how the availability of both mouse and dog models facilitated preclinical development. In contrast to the hemophilia studies, targeted injection in the eye, a relatively immune-privileged site, has not yet resulted in a notable immune response to the vector. However, the surgical approach used is complex and not without potential complications. It remains to be determined whether intravitreal administration will be feasible with a similar vector. In addition, while LCA is a set of phenotypically similar diseases, not all the molecular subtypes may be amenable to the approach used for RPE65 mutations. In the short term, the LCA subtypes most likely to become candidates for gene transfer will be those in which there is dissociation between retinal function and structure (i.e., serious visual disturbance but sufficient photoreceptors and RPE to warrant treatment), and there are faithful animal models for proof-of-concept research.

Dr. Donald Kohn reviewed the experience with gene transfer for genetic diseases of blood cells in which some of the first and longest-followed successes have been reported. In the SCID trials, as well as other clinical applications for hematologic disorders, the approach has been to transduce hematopoietic stem cells using an integrating retroviral vector in order to achieve long-term gene correction. Again, impressive clinical successes have been achieved over time in trials for SCID-X1 and ADA-SCID that often compare favorably with standard stem cell transplants. However, in some trials immune reconstitution was balanced by challenges. In trials for SCID-X1 and Wiskott-Aldrich Syndrome (WAS), a number of the research participants, who had successful engraftment of the gene-modified cells, developed T-cell leukemias, and myelodysplasia developed in research participants in a trial for chronic granulomatosis disease (CGD), due to vector-mediated insertional mutagenesis (Aiuti et al., 2012; Boztug et al., 2010; Hacein-Bey-Abina et al., 2003; Stein et al., 2010). This genotoxicity was not predicted by the animal models. It is also not understood why it is not seen in similar approaches for other diseases, such as ADA-SCID. Thus, developing models that can be predictive of whether a novel integrating vector can result in similar toxicity remains a need and a challenge. Preclinical models that are more predictive of human experience are also being developed.

Other challenges include that the target for gene transfer in these protocols is stem cells, which were described by Dr. Kohn, as “rare, quiescent, and fragile” when manipulated ex vivo and more effective if administered fresh, not previously frozen. These characteristics lead to challenges for transduction and clinical administration. Efficient transduction of these cells must be balanced against the risk of genome perturbation with increased vector copies per cell, which may increase the risk of genotoxicity.

Presently, the field is moving forward with lentiviral vectors and modified retroviral vectors. Some clinical improvements have been observed in trials for X-adrenoleukodystrophy (Cartier et al., 2009) and β thalassemia (Cavazzana-Calvo et al., 2010) using lentiviral vectors, and trials are being planned for several other diseases. In addition, as is the case for congenital diseases of the retina, there are at least 20 genetic causes of human severe combined immunodeficiency; therefore, despite the success in treating some of these diseases, the challenge remains how to translate this research most efficiently into these other related diseases.

Dr. Carlos Camozzi, formerly from uniQure, focused on the development of Glybera (alipogene tiparvovec), the first gene therapy product to be authorized for marketing in the European Union (EU). Glybera is an AAV vector expressing lipoprotein lipase, an enzyme normally synthesized in muscle and adipose tissue that breaks down triglycerides. LPL deficiency is a rare disorder in which the only existing treatment is reduction of dietary fat. In a clinical trial of Glybera, 17 subjects experienced significant decreases in pancreatitis and hospitalizations. Dr. Camozzi described the review process by the European Medicines Agency (EMA). This included including multiple evaluations by the Committee for Advanced Therapies and Committee for Medicinal Products for Human Use before receiving positive opinions from both for approval for marketing in the EU. The process involved extensive interactions with the EMA, academic investigators, and patient advocacy groups. Dr. Camozzi discussed the importance of the development of a reproducible manufacturing process for biologics; preclinical and clinical studies, including efficacy endpoints and long-term follow-up; the necessary differences in industrial versus academic trial development; and the developmental timeline associated with these studies.

Can a Platform Be Developed?

One of the questions posed to the investigators at this workshop was whether any of these vector technologies could be used as a platform to streamline the preclinical development of vectors for additional indications. In the panel discussion on therapeutic platforms, Drs. Jude Samulski, Brian Sorrentino, and Jeffrey Bartlett considered the common characteristics that may allow for platform development: single target tissue, route of administration, vector backbone including regulatory elements, and identical vector production. Using the example of AAV vectors for eye diseases, the discussion focused on factors to be considered in developing a platform for a field in which patients were eager to see the success of the RPE65 trials applied to other LCA or eye disorders. In all of these diseases, the eye will be the target tissue. The route of administration for the LCA protocols involves subretinal injection; however, for some diseases vitreal injection may be possible and allow for the use of a less invasive and simpler procedure that could be administered outside of highly skilled centers. As for the vector, AAV2 has been used successfully for ophthalmologic gene therapy and immune responses have not been an apparent limitation, even with repeat dosing. However, if the site of delivery was altered, there might be the need for modifications to the vector, including different AAV serotypes or tissue-specific promoters, to obtain adequate biodistribution across the intravitreal space. Another consideration is the variations in methods of vector production because there is no standardization yet in the field. Even comparing dosing across trials may be a challenge. For example, in the hemophilia trial that used a self-complementary AAV8 vector, the initial dose was discovered to be almost 10-fold higher than originally measured because the quantitative polymerase chain reaction assay did not accurately detect the copy number due to the secondary structure (self complementary) of the vector genomes (Fangone et al., 2012).

The use of lentiviral vectors for hematological disorders was discussed as another possible candidate for platform development. There are three major hematologic target tissues: stem cells, T cells, and dendritic cells. In this model, genodistribution and potential genotoxicity due to use of integrating vectors are perhaps more important than biodistribution. Even when similar vector backbones are used, the genodistribution and risk of oncogenesis may differ depending upon the regulatory elements (e.g., promoter/enhancer), envelope, transgene, and the disease phenotype. The differences in the risk of leukoproliferative complications and common insertion sites observed in trials for different diseases such as CGD, WAS, SCID-X1, and ADA-SCID are not yet fully understood. Animal models have proven more useful for predicting efficacy than safety in humans and disease-specific animal models are not available for all diseases. As with AAV vectors, vector production methods vary with both transient and stable packaging systems being used in different cell lines making comparisons across protocols difficult and likely limiting the ability to rely on previous studies.

Industry appears to have taken the first steps to explore the use of vector platforms. In 2010, GlaxoSmithKline (GSK) announced collaboration with the San Raffaele and Telethon Foundations to develop new gene therapy treatments for rare diseases. GSK stated that it would gain exclusive license to develop and commercialize an ex vivo therapy for ADA-SCID using a murine leukemia retroviral virus. In addition, GSK gained option rights for six additional applications using a new gene therapy technology currently developed by the San Raffaele Telethon Institute of Gene Therapy (http://us.gsk.com/html/media-news/pressreleases/2010/2010_pressrelease_10113.htm). The six additional applications propose to use a lentiviral based gene therapy product for ex vivo hematopoietic stem cell transduction (Luigi Naldini, personal communication).

Dr. Camozzi noted that UniQure intends to develop their AAV platform for other diseases. The platform will use the same production method because the only way to ensure the quality and the standardization of the vector from batch to batch is to have a validated, reliable platform for production that is approved by the regulatory agencies. In May 2012, Avalanche Biotech and Lonza Group, a contract manufacturer of cell- and virus-based therapeutics, announced a manufacturing collaboration focused on process development and scale-up efforts for the manufacturing collaboration to produce high-yield AAV vectors and make the technology available to third parties in exchange for a share of the revenues (Corporate News Release May 16, 2012, Avalanche Biotechnologies, Inc. and the Lonza Group Ltd. http://www.avalanchebiotech.com/pdf/press-release-05-16-2012.pdf).

The consensus of the panel discussion was that while development of platform technologies may be an ultimate goal, there are many challenges that must be addressed to develop a platform and several reasons it may yet be premature to lock into platforms. The choice among vector platforms and continued vector refinement may ultimately be determined by the availability of funding and resources. Despite some advantages in tissue trophism or lower immune responses in humans demonstrated with other AAV serotypes, AAV2 vectors are predominantly used in clinical studies because of the previously available preclinical and clinical data. However, one question is whether and how to use the experience with AAV2 vectors to evaluate new serotypes rather than committing to a specific platform before the optimal vector/approach is defined. There was general agreement that the platform development could begin with standardized methods such as the assays for titers, replication competent virus detection, and genotoxicity assays for retro- and lentiviral vectors. Development of uniform production methods is also a critical piece that needs to be addressed before moving toward platforms, otherwise preclinical studies done for one vector may not be comparable to the next study.

Improving the Efficiency of Translational Research

Developing a platform technology may not be the only way to facilitate the development of new clinical applications. A key question is whether there are safety and toxicology data from previous applications that can be used and built upon in developing subsequent clinical Investigational New Drug applications (INDs). The FDA recently published a draft Guidance for Industry: Preclinical Assessment of Investigational Cellular and Gene Therapy Products (http://www.fda.gov/downloads/BiologicsBloodVaccines /GuidanceComplianceRegulatoryInformation /Guidances/CellularandGeneTherapy/UCM329861.pdf). The FDA noted that investigators are able to cross-reference studies from other INDs provided there are no “significant” differences between the products, target tissue, route of administration, and disease. For example, one might be able to share biodistribution data, if there was no significant changes in the vector, including formulation changes, route of administration, dosing, or transgene.

However, even if preclinical data from a previous study is not sufficiently similar for cross-referencing in an IND, previous studies, including pharmacology/toxicology studies (pharm/tox) or efficacy studies in animal models, may be useful to guide the development and structure of new studies in phenotypically related diseases. Accessing such data can be difficult because it is often not published. Dr. Sorrentino related the difficulty that his group had in publishing in vitro and in vivo work that was done in support of their lentiviral trial for SCID-X1, data that are valuable as the field explores the use of lentiviral vectors instead of other retroviral vectors. Challenges in publishing negative data are likely to be even greater.

Several initiatives are underway to facilitate the sharing of data via publications or databases. Dr. James Wilson, Editor-in-Chief of Human Gene Therapy, announced a new journal, Human Gene Therapy Methods that will provide a venue for publishing data relevant to regulatory review, such as preclinical animal and in vitro studies. Both positive and negative studies will be published. Dr. Malcolm Brenner, Editor-in-Chief of Molecular Therapy stated that their planned methods journal will likewise provide a forum for such research to be published. Of course in both cases, the articles will be peer reviewed and therefore some research may not be published.

The ability to publish may foster new collaborations. For example, if one group has done significant preclinical work on a certain vector system, they may be more willing to share the vector if an agreement could be made that any publication of preclinical work done with this vector would be a joint publication, allowing both groups to share the credit. For industry, sharing of preclinical data is also contingent on whether the information could be critical to an intellectual property (IP) claim in the future. If so, there will be more restrictions on sharing. While certain toxicology and data on persistence of the vector and transgene expression may not be critical to IP claims, industry may need to take a more measured approach to publishing, only releasing data when they are confident their IP position will not be compromised. However, as in the case of the discussion of platforms, without standardized production such that there is good reproducibility, the ability to use these data may be limited due to the lack of comparability.

Since unpublished studies may still be relevant to other investigators, the panelists proposed that a mechanism be developed to facilitate the transfer of data submitted to these journals, both published and unpublished, to a detailed, readable, and sharable database. The NIH/National Heart, Lung and Blood Institute (NHLBI)-supported National Gene Vector Biorepository (NGVB) pharm/tox database provides such a mechanism by which investigators can share data. The NGVB accepts gene therapy pharm/tox studies, and the data are then available to other investigators, including to reference for their IND. NGVB will assist in arranging letters of cross-reference in support of an IND. To date there are 36 studies in the database; however, utilization could increase as the number of completed pharm/tox studies grows along with greater awareness of the benefits of use of the database (Table 1).

Table 1.

National Institutes of Health and U.S. Food and Drug Administration Resources to Facilitate Development of New Protocols

• Gene Therapy Resource Program (GTRP) (www.gtrp.org)

○ National Heart, Lung, and Blood Institute (NHLBI) supported program designed to facilitate translation of gene transfer applications into clinical interventions

○ Core facilities for preclinical and clinical Good Manufacturing Practices–grade vectors (adeno-associated virus, adenovirus, and lentivirus), pharmacology, toxicology, and immunology studies, and a clinical coordinating center.

○ Clinical trial performance and regulatory application support

• Bridging Interventional Development Gaps (BrIDGs) (http://www.ncats.nih.gov/research/reengineering/bridgs/bridgs.html)

○ National Center for Advancing Translational Sciences (NCATS) program providing resources and contract services to generate data for filing Investigational New Drug applications (INDs) for clinical candidate products

• Genetic Modification Clinical Research Information System (GeMCRIS) (http://www.gemcris.od.nih.gov/Contents/GC_HOME.asp)

○ Information resource and analytical tool that allows public users to access basic reports about gene transfer trials and develop specific queries (e.g., vector, transgene, medical condition)

• Rare Diseases Clinical Research Network (RDCRN) (http://rarediseases.info.nih.gov/ASP/resources/extr_res.asp)

○ Program coordinated by the Office of Rare Diseases Research, NCATS to support and facilitate research by creating consortia of research for related rare diseases with the goals of cost sharing, establishing uniform protocols, and conducting collaborative clinical research

• Clinical and Translational Science Award (CTSA) (http://www.ncats.nih.gov/research/cts/ctsa/about/about.html)

○ Program within the Division of Clinical Innovation, NCATS, provides funding to individual universities to increase the efficiency and speed of clinical and translational research across the country.

• National Gene Vector Biorepository (NGVB) (https://www.ngvbcc.org/)

○ Archiving services, insertional site analysis, pharmacology/toxicology (pharm/tox) resources, and reagent repository

• U.S. Food and Drug Administration (FDA) Orphan Products Grants Program (http://www.fda.gov/ForIndustry/DevelopingProductsforRareDiseasesConditions/WhomtoContactaboutOrphanProductDevelopment/default.htm)

○ Support the clinical development of products for use in rare diseases

Presently much of the data captured in the NGVB are Good Laboratory Practice (GLP) studies funded by NIH in support of an IND. However, there was agreement that beyond pharm/tox studies, there are other types of data that would be useful if shared, including data on gene expression, distribution, persistence of the vector, and long-term effects of vectors, such as oncogenic effects. Particularly useful for sharing would be data from long-term follow-up studies of large animals, such as the dogs and nonhuman primates because this information is valuable regarding the long-term safety of certain vectors. One proposal was to develop templates that could facilitate entry of data into the NGVB as well as targeted searches.

Another way to promote data sharing is by having it linked to funding. NIH has a data-sharing policy for its funded research (http://grants.nih.gov/grants/sharing.htm). If an applicant seeks over $500,000 in direct costs in any single year, it is expected that the application will include a plan for sharing data or an explanation as to why data sharing is not possible. The timeliness of that data sharing is also recognized as important and NIH's policy states that data sharing should have occurred no later than the acceptance for publication of the main findings from the final dataset. The policy recognizes that for large projects the data may be released in waves as they become available. The policy attempts to balance the need to provide initial and continuing benefits from the use of data but not to permit prolonged exclusive use. While the NIH expects that data from gene therapy studies will be shared whenever possible, there is no specific policy requiring gene therapy studies to deposit data with the NGVB database.

Another critical characteristic of this research is that there are active trials in both the U.S. and Europe, and indeed many of the first clinical successes in SCID-X1 and ADA-SCID came in European trials. Harmonization and collaboration with European partners will be critical to maximize the efficient use of resources. This will also require harmonization of regulatory standards to foster trans-Atlantic clinical trials. The FDA and EMA are already engaged in harmonization activities and it will be critical that data-sharing policies extend across the Atlantic.

The Rare Diseases Clinical Research Network (RDCRN), which is led by ORDR in collaboration with eight institutes and centers in NIH, provides another mechanism to bring together researchers, both in the United States and abroad (Table 1). The RDCRN comprises 19 consortia that study multiple related diseases grouped in categories such as the North American Mitochondrial Diseases consortium or the Primary Immune Deficiency Treatment consortium. Each consortium includes multiple clinical sites that perform collaborative clinical research and involve patient advocacy groups as partners in the research effort. In each consortium, one study must be a longitudinal natural history study. The RDCRN currently has 225 clinical sites studying 200 diseases through 83 clinical studies with over 14,000 subjects enrolled. The RDCRN also has a Data Management and Reporting Center, which supports the network by providing tools and technologies to collect standardized clinical research data. There is a data-sharing policy that is mandated for all studies supported through this network, and a data access committee is being established. Clinical trial results are required to be shared after publication of the results or by 3 years after enrollment of the last participant. Longitudinal studies have a slightly different timeframe, either 1 year after publication or 5 years from the date the data were collected, whichever comes first. These data will be available to the members of the RDCRN and also to the public.

In summary there are a number of mechanisms to foster sharing of data. Some, such as the new journals are just beginning. Others, (Table 2) such as the NGVB may benefit from linking to these new initiatives or from funding policies that encourage data sharing through this central database. Specific data needs should be defined by the field and the standardization of production and reagents may facilitate these data being used to support preclinical data for new trials.

Table 2.

Data Sharing

The sharing of data generated from various preclinical studies (e.g., pharmacology, toxicology, biodistribution, genotoxicity, and immunology studies) may be possible.

Challenges

• Determination of what studies need to have in common to be useful: vector, serotype, promoters, transgene, animal models, route of administration?

• Limited accessibility to data

○ Preclinical studies and negative results infrequently published

○ Pharma: Restrictions on sharing data to protect intellectual property claims

○ Academia: concerns about authorship and grants

• Lack of incentive to share data

Possible solutions

• Inclusion of letters in the IND submitted to FDA that authorize cross-reference to other applicable regulatory files (e.g., Master Files, INDs)

• GeMCRIS queries can identify previous research using similar vectors, transgenes, diseases etc. facilitating sharing

• New journals being launched to publish such data

○ Human Gene Therapy Methods

○ Molecular Therapy Methods Journal (in development)

• Additional source or summary data could be submitted by principal investigators to searchable, readable database

○ Expand NGVB pharm/tox database

Funding of Clinical Trials

Funding for research is always a challenge, and gene therapy for rare diseases is no exception. Dr. Kohn reviewed the development timeline for one of his protocols for ADA-SCID and noted that in the preclinical realm, it is often difficult to fund preclinical studies with traditional R01 awards for toxicology, good manufacturing practices vector, and reagents. NIH programs designed to assist investigators in preparing the IND package include the Gene Transfer Resource Program (GTRP) from NHLBI, and the Bridging Interventional Development Gaps (BrIDGs), which is run by NCATS (Table 1). Both programs provide contract services for development of new protocols.

In the case of GTRP, the service is available at no cost to investigators who are working on NHLBI-funded diseases. GTRP can provide resources to investigators funded by other NIH Institutes if the funding Institute pays for the services through a transfer of funds to NHLBI. GTRP provides preclinical vector production services for AAV, adenoviral, and lentiviral vectors, including immunology testing, and can provide clinical grade AAV and lentiviral vectors. It also provides toxicology and biodistribution studies in various animal models, including nonhuman primates and prepares the final study reports for the IND. A limited amount of GTRP funds are available to support Phase I/II clinical trials for heart, lung, and blood diseases. Finally, it includes a clinical coordinating center that provides regulatory assistance to investigators and can manage clinical funding distribution and support clinical trial infrastructure. As of September 2012, 113 requests for service applications have been approved, with 81 of them focusing on preclinical vector production, and the remainder on clinical vector production (eight), immunology testing (seven), pharm/tox studies (five), regulatory support (nine), and clinical trial services (three).

BrIDGs facilitates collaborations between NIH staff with expertise in preclinical drug development and extramural and intramural labs with candidate small molecule or biologic agents, including gene vectors. The program assists investigators with the development of IND data packages in support of clinical candidates for which there is preclinical proof of concept. BrIDGs is not a grant program. Similar to the GTRP, it provides preclinical contract services for production/bulk supply, good manufacturing practices manufacturing, pharmacokinetic testing, animal toxicology, manufacture of clinical trial supplies, and advice on IND filing. In some cases, BrIDGs will use contract resources offered by the GTRP to advance gene therapy projects. BrIDGs (formerly the NIH-RAID program) has funded 34 projects since 2005. Of these, three gene therapy projects have been approved to develop AAV vectors for Parkinson's disease, L amino acid decarboxylase deficiency, and osteoarthritis. Applications for the development of therapies for all diseases or disorders are accepted.

Another program administered by NCATS, the Therapeutics for Rare and Neglected Diseases (TRND) program, focuses only on rare diseases. Like BrIDGs, TRND also facilitates collaborations between NIH staff and extramural and intramural labs. Projects enter the program at various stages of development, including lead selection, and may be advanced far enough to attract licensure for future development by an external organization. Unlike BrIDGs, limited clinical trial support and regulatory affairs assistance are available. Also in contrast with BrIDGs, in which intellectual property is retained by the owner, the NIH may generate intellectual property under the TRND program. Currently, TRND does not accept gene therapy applications.

Once an investigator succeeds in developing a successful IND package, funding for the clinical trial must be obtained. A number of the gene therapy trials have required costly and medical resource–intensive procedures such as apheresis and stem cell harvest for immunodeficiency diseases, eye surgery, or neurosurgery, and different hospitals may have very different charges for some of these services. With multisite trials, negotiating for similar rates across sites can be a challenge. A possible resource may be the Clinical and Translational Science Award (CTSA) consortium (Table 1). Also located in NCATS, the CTSA program funds a consortium of about 60 medical research institutions with a focus on providing specialized infrastructure support to NIH funded scientists, engaging community partners, and training the new generation of clinical and translational scientists. In addition to local support for investigators, CTSAs can also facilitate collaboration with investigators at other institutions, particularly those with CTSAs. CTSA-based resources may be particularly valuable for clinical trials with small target populations.

The National Institute of Neurological Disorders and Stroke has a U01 grant mechanism for translational research, which is milestone-driven and has been used by gene therapy investigators (http://www.ninds.nih.gov/research/translational/index.htm). The goal of this grant mechanism is to focus on identification, optimization, and preclinical testing of candidate therapeutics. A funding mechanism that spans a longer timeframe and uses milestones, was seen as one possible mechanism to facilitate full project funding while lowering the risk of committing significant funding to a project that would not succeed. Another example of such a funding mechanism comes from the California Institute for Regenerative Medicine. Known as Disease Team Research Awards, they are described as providing funding for actively managed multidisciplinary teams engaged in milestone-driven translational research for the development of stem cell–based therapy. The mission of these teams is to support the necessary research and regulatory activities to prepare and file a complete, well-supported IND. Dr. Kohn noted that under this program there is a very detailed timeline for the whole project and progress towards milestones are reviewed quarterly. NHLBI staff noted that they are also moving toward milestone-driven contract awards.

Finally, participants noted the important role that foundations provide in filling the funding gaps. This includes funding for both preclinical studies and for clinical trials. However, it is not clear that foundation funding alone can continue to meet current needs.

Conclusions

Gene therapy for rare diseases appears to hold great promise as demonstrated by clinical success in several diseases to date. The ability to apply this technology to other diseases will depend upon a number of factors. Diseases such as hemophilia in which the severe clinical manifestations can be ameliorated with low levels of protein replacement, sometimes just 5%–10% of wild-type, are feasible candidates for gene therapy. For those diseases in which long-term gene correction is needed, the fact that the target cells—hematologic stem cells—were accessible and could be manipulated ex vivo also facilitated gene therapy approaches. Not all rare diseases will have characteristics that will allow the approaches used to date to be successfully translated, even for phenotypically similar diseases. For example, Dr. High noted that for hemophilia, there are important differences in the diseases that may be amenable to gene transfer. Hemophilia A is due to deficiencies in factor VIII while hemophilia B is due to deficiencies in FIX. In both cases replacement of the missing factor at levels just 5%–10% of normal can lead to significant clinical improvement. However the frequencies with which individuals develop antibodies to factor VIII is much higher than for FIX, and there may be reasons for this that could impact gene therapy–mediated replacement. In addition, the factor VIII gene, even truncated, may exceed the packaging limit for an AAV vector, leading investigators to evaluate other vector systems for this form of hemophilia.

For diseases related to those for which gene therapy has shown clinical promise, there are several potential strategies that may help accelerate the efficient development of gene therapy applications (Table 3). First, while it may be premature to develop a single or even two different platform vectors for gene therapy for rare diseases, it is not too early for the field to begin to think about standardizing production and reagents for those vectors most commonly used in these trials. This may be the first step toward developing platforms for which a streamlined regulatory approach might be possible in the future. Indeed, industry seems to be moving in that direction. Short of developing a standard platform, the sharing of data, not only pharm/tox studies but also studies on genodistribution as well as genotoxicity, and data from long-term follow-up in large animal models, may promote more efficient development of new approaches. The commitment of both Human Gene Therapy and Molecular Therapy to establish dedicated journals to publish preclinical studies is an important step in facilitating and incentivizing the sharing of information. However, as with all journals, not all of the data submitted will be published; creating a mechanism to link both published and unpublished data to the NGVB database may provide further opportunities for data sharing. Funding agencies for such research, such as NIH, may want to encourage the data-sharing plans for such grants to include depositing the studies with the NGVB. Finally, having more formal systematic reviews of certain vectors for which there is considerable experience may assist investigators as they take these vectors into new applications.

Table 3.

Suggestions for Next Steps

• Encourage data sharing through funding mechanisms

• Launch of new journals to increase publication of preclinical studies

• NGVB to explore expanding pharm/tox database into repository

○ Preclinical, long-term follow-up of animal and clinical data

• Development of additional reference standards, assays

• Multiphase, milestone-driven funding mechanisms

Clinical trial funding is a challenge for many different fields. In gene therapy, the investigators noted the challenges of funding the various development stages, including vector production, through traditional R01 mechanisms. A single funding mechanism is not often available to bring the project from proof of concept though Phase I clinical trial; therefore, a project may be derailed by delay or failure to obtain the next grant. Programs such the GTRP are seen as very helpful to move through the preclinical development stage but of note the majority of projects to date have been for vector development, which may be indicative of an area of need for the field. Milestone-based funding that might provide more seamless funding across stages of development was put forth as an attractive mechanism, as was having access to rare diseases translational programs such as TRNDS. There was also support for the RDCRN, which facilitates collaborations between investigators in the United States and abroad and provides critical natural history data that are invaluable for many rare diseases.

Finally, the clinical successes to date have been in a number of different diseases with different approaches and vectors. This conference brought together investigators from different disciplines to speak generally about the challenges and opportunities to move the field forward. There was consensus that more focused meetings on particular areas, for example, congenital eye diseases, may lead to new strategies for promoting specific areas of research.

Acknowledgments

The authors thank all of the speakers and panelists for their contributions. The views expressed in this paper are those of the individual authors and do not represent the views of the NIH.

Author Disclosure Statement

JMW is a consultant to ReGenX Holdings, and is a founder of, holds equity in, and receives a grant from affiliates of ReGenX Holdings. In addition, he is an inventor on patents licensed to various biopharmaceutical companies, including affiliates of ReGenX Holdings. BJB is a founder and holds minor equity in Applied Genetic Technologies Corporation (AGTC). In addition, he is an inventor on patents owned by Johns Hopkins University and the University of Florida. KC is founder and holds equity in Rimedion Inc. but is not employed by the company. JB is an employee of Calimmune, Inc. KH holds patents in the area of AAV-FIX, and has waived all financial interest in these. I am also an inventor on patents related to AAV manufacture and purification, and lentiviral vector manufacture. I hold equity in bluebird bio, Inc, which has clinical programs in X-linked adrenoleukodystrophy and thalassemia. I have served as a consultant to BioMarin Pharmaceuticals, Genzyme, Novo Nordisk, Nordic Biotech Advisors, and Shire regarding gene therapy for genetic disease. SGJ has no financial interests in the subject matter or materials discussed in the manuscript. YX, DBK, JB, PJB, RGC, RGS, and SIS have no conflicts to disclose.