In his opening remarks for the session on translation, moderator Daniel Burch, global medical officer at PPD Biotech, echoed earlier remarks about the excitement and fast-moving nature of the gene-targeted therapy field, and characterized working in this area during the past few years as “drinking from a fire hose.”
R. Jude Samulski showed videos of two boys with Duchenne muscular dystrophy (DMD) who participated in clinical trials of a gene therapy treatment. DMD is a rare, fatal, X-linked genetic disorder caused by mutations in the dystrophin gene that prevent production of dystrophin, a protein essential for normal muscle function. Samulski said that within months of receiving the gene therapy by IV infusion, the boys progressed from not being able to climb steps to playing Little League baseball. The videos he showed illustrate what can happen when a technology developed in the laboratory is successfully translated to humans and what some of the challenges are. For example, asked Beverly Davidson, what happens when new vectors are developed that appear to be 10 times more efficient than the old vectors? Should trials of the old vectors be discontinued and replaced by trials of the new vectors, or can they be phased in gradually? What kind of bridging studies will be required to ensure safety?
ENABLING TRANSLATION WITH PRECLINICAL MODELS
Chapter 3 described Akshay Vaishnaw’s early work to develop conjugated small interfering RNAs as a therapeutic modality for central nervous system (CNS) disorders. To optimize the pharmacology of the compounds in development, he and his colleagues have initiated studies to understand the correct dose and frequency in animal models, with a goal of achieving
therapeutic levels with dosing every 6 months or annually. Other ongoing Investigational New Drug (IND)-enabling studies are exploring the distribution of the conjugates and gene knockdown in particular cell types as well as identifying novel ligands that exert knockdown only in specific cell types within the CNS. While cell models can be useful translational tools, animal models are especially important for CNS disorders, said Anastasia Khvorova. Particularly for psychiatric and neurodegenerative disorders, most cell models are non-predictive because they do not adequately model the complex infrastructure and environment in which cells function in the brain, she said.
In contrast with many other disease areas, a major translational bottleneck for psychiatric disorders is the limited availability of appropriate predictive models for both efficacy and toxicity, said Steven Hyman. Indeed, the successful gene-targeting approaches described in Chapter 2 were enabled in part by the availability of large animal models, both naturally occurring (inherited retinal disease) and genetically engineered (spinal muscular atrophy, or SMA) models. Samulski described how the golden retriever muscular dystrophy model—a model derived from dogs identified in the early 1980s with spontaneous dystrophinopathies and an X-linked pattern of inheritance (Kornegay, 2017)—was used to test an adeno-associated virus (AAV)-mediated gene therapy approach to a deficiency in dystrophin protein. For example, he showed photographs of one of these dogs who, after receiving a sufficient dose of the vector, lived a normal dog’s life and actually had increased muscle mass. More importantly, said Samulski, the dog displayed a good safety profile. Frank Bennett noted that in developing gene-targeted therapies for SMA, non-human primates were used only to study biodistribution and safety because humans are the only species to have the alternate SMN2 gene. Large animal models also allow preclinical studies to represent the effects of aging on uptake and distribution of vectors or antisense oligonucleotides, noted Lamya Shihabuddin.
While primate models may be most predictive because they are evolutionarily closer to humans, Jeffrey Kordower noted that the immune response of primates or their response to immunosuppressive drugs is highly variable. In addition, he said, little is known about the receptors for AAV vectors in different species.
CLINICAL TRIAL DESIGN CHALLENGES FOR GENE-TARGETED THERAPIES
In designing clinical trials, sponsors must make decisions regarding inclusion and exclusion criteria, endpoints, controls, and balancing risks and benefits. Bennett added that sponsors must also consider the natural
progression of the disease, the patient population available for a clinical trial, and the anticipated effect size of the drug. Kathleen Reape agreed, noting that the trial design will differ for a fatal debilitating condition compared to a slowly progressive disease. Chris Henderson added that one treatment may not be fully efficacious for all patients.
Michael Panzara, chief medical officer at Wave Life Sciences, Ltd., described the clinical program for survodirsen, a potential treatment for boys with DMD. Survodirsen is designed to restore functional dystrophin through the method of exon skipping, which enables production of a shorter but functional protein (Kole and Krieg, 2015). Based on a favorable safety and tolerability profile established in a recently completed Phase 1 single ascending dose study enrolling 40 boys, a Phase 2/3 study is now being planned, he said. In addition, boys who completed the Phase 1 study are eligible for an open-label extension study at a dose expected to lead to exon skipping, said Panzara. Data from this study are expected later in 2019, including assessment of dystrophin expression in muscle biopsy, and are intended to comprise an important component of the company’s submission for accelerated approval, he said.
Running in parallel, the Phase 2/3 study called DYSTANCES 51, set to begin in July 2019, was selected by the Food and Drug Administration (FDA) for its Complex Innovative Trials Designs pilot program,1 an initiative under the 21st Century Cures Act, said Panzara. Among the innovative trial design features planned for DYSTANCES 51 is the leveraging of historical control data to help augment the placebo arm. This approach is intended to reduce the number of participants required to deliver conclusive efficacy results, minimize the number of participants in the placebo treatment arm, and accelerate the study program, he said. A second key innovation is the use of Bayesian repeated measure modeling to adapt the trial based on interim dystrophin analyses, said Panzara. Simultaneously, they will develop a Bayesian disease progression model, which will incorporate historical control data and interim biopsy data to predict the probability of success and potentially to adjust enrollment in an ongoing fashion to improve the efficiency of the trial.
Panzara noted that the historical control data are being contributed by several companies that have conducted DMD clinical trials. The Critical Path Institute will assist in this process as a neutral convener, housing the datasets for those who would prefer that the data sit with a third party. The modeling work and placebo data collected in this trial will be shared with the field to leverage learnings and propel the field forward. Panzara
___________________
1 For more information about FDA’s Complex Innovative Trial Designs Pilot Program, go to https://www.fda.gov/drugs/development-resources/complex-innovative-trial-designs-pilot-program (accessed June 7, 2019).
added that he hoped this would serve as a model for future rare disease clinical trials.
The innovations in clinical trial design described by Panzara address the significant recruitment and retention challenges sponsors face when conducting clinical trials for life-threatening disease. These challenges are exacerbated for rare diseases, where access to patient populations is critical but may be limited, said Vaishnaw. Natural history studies may provide access to potential trial participants, and patient registries have been developed in several disease areas to improve the efficiency of recruitment. For example, in the Huntington’s disease (HD) space, the CHDI Foundation developed a platform called ENROLL-HD,2 which has been running for more than 6 years and has 20,000 participants, said Cristina Sampaio, chief medical officer at the CHDI Foundation and professor of clinical pharmacology at the University of Lisbon. Registries have helped fuel translation in other disease areas as well. Vaishnaw said that in DMD, the size of the population has allowed several registries to prosper, collect meaningful data, and provide an important resource for academia and industry. For micro-orphan diseases, however, he said that competing registries in both academia and industry can be a significant impediment to the drug development process.
The selection of clinically meaningful endpoints will differ depending on the condition and the population affected, said Reape. For example, in establishing what constituted a clinically meaningful change in the novel endpoint developed for the voretigene studies (discussed in Chapter 2), the investigators had to consider the real-world meaningfulness of restoring vision in children compared with adults who have been blind since birth, she said. Vaishnaw added that appropriate endpoints and biomarkers are particularly difficult to identify for rare diseases for which natural history studies are so difficult.
Balancing risks and benefits and establishing minimal effect sizes will also differ depending on other disease-specific factors, such as whether a condition is fatal and debilitating or slowly progressive, said Reape. Kordower agreed, noting that it may be appropriate to accept smaller benefits in certain subpopulations in which the disease is especially aggressive. In early-stage trials when there is a lot of uncertainty, risks are higher so the potential benefit to trial participants should also be high, said Petra Kaufmann. Kordower added that safety and tolerability studies are typically underpowered to answer efficacy questions and cautioned sponsors not to try to assess efficacy from a safety and tolerability study. However, Henderson suggested that efficacy data from a safety and tolerability study, if appropriately interpreted, can speed up the progress of a trial.
___________________
2 For more information about ENROLL-HD, go to https://www.enroll-hd.org (accessed June 17, 2019).
Reape noted that the RPE65 gene therapy trials included detailed responder analyses to try to determine if characteristics such as age influenced response. There were two participants who showed no change in performance (improvement or decline) on the multiluminance mobility test, she said, but they were unable to identify any single underlying characteristic that both had in common. However, they learned from the natural history study that age is a very loose indicator of retinal degeneration, with a high degree of individual variation, as would be expected with a progressive disease. Because the treatment requires the presence of viable retinal cells, surrogate markers such as retinal thickness and visual field were used to assess retinal viability, although she noted that those tests do not directly assess retinal function.
REGULATORY PATHWAYS
Gene therapies fall into the regulatory category of advanced therapy medicinal products, which also includes cell therapies and xenotransplantation, said Peter Marks, director of the Center for Biologics Evaluation and Research (CBER) at FDA. He added that gene therapies for serious conditions are eligible for several of FDA’s expedited development programs, including Fast Track, Priority Review, Accelerated Approval, Breakthrough Therapy, and Regenerative Medicine Advanced Therapy (RMAT). RMAT became law as part of the 21st Century Cures Act at the end of 2016 to expedite gene and cell therapies, tissue engineering products, and certain combination approaches, he said. To get this designation, the product must address a serious or life-threatening disease or condition and there must be preliminary evidence of its potential to address an unmet medical need. RMAT-designated products may also be eligible for priority review and accelerated approval, Marks noted. In the past 2-plus years, Marks said 33 products—mostly cellular or cell-based gene therapy products—have been granted this designation.
To help advance the development of gene and cell therapies, Marks said FDA has also issued several guidance documents, taken steps to reduce the administrative burden of regulatory approval, established several clinical development and manufacturing initiatives, and helped develop standards. He also mentioned the INitial Targeted Engagement for Regulatory Advice on CBER producTs (INTERACT) program,3 which enables sponsors to meet with FDA for a non-binding, relatively informal pre-IND meeting to discuss preclinical, manufacturing, and clinical issues related to therapies in early stages of development.
___________________
3 For more information about the INTERACT program, see https://www.fda.gov/vaccinesblood-biologics/industry-biologics/interact-meetings-initial-targeted-engagement-regulatory-advice-cber-products (accessed June 8, 2019).
The European Medicines Agency (EMA), though organized quite differently from FDA, functions in a similar manner, according to Rune Kjeken, scientific director for advanced therapies at the Norwegian Medicine Agency and a member of EMA’s Committee for Advanced Therapy and the Scientific Advice Working Party. Structured as a network of regulatory agencies from the 28 European Union member states plus Iceland and Norway, the work of EMA is conducted by scientific committees and working parties, said Kjeken.
Like FDA, EMA is responsible for assessment and decision making at all steps of the regulatory pathway, including final marketing approval, said Kjeken. Also like FDA, EMA writes guidelines, including disease-specific and modality-specific (e.g., gene therapy) guidelines. EMA also has a program similar to INTERACT, which is called the Innovation Task Force.4 Approval of clinical trials, however, remains with the competent authorities of individual nations, he said.
EMA also has an early access mechanism called PRIME (PRIority MEdicines), similar to what FDA calls “breakthrough designation,” to foster development of medicines with high public health potential, said Kjeken. Since December 2018, he said, about 50 products have been accepted into the PRIME program, about 20 of which are gene therapy products.
Kjeken said EMA has a procedure for parallel scientific advice with FDA. It also works in parallel with health technology assessment bodies to ensure that consideration is given to the potential value of a new drug and how it will perform in the real world. Kjeken predicted this will become increasingly important in coming years as the more drugs developed for rare diseases exert an ever-greater impact on overall health care costs.
Matching Modalities and Regulatory Pathways to Specific Disorders
Regulators evaluate gene-targeted therapies differently from more traditional pharmacological therapies for several reasons, including the invasiveness of the interventions, the durability of effect, and issues related to placebo controls and participant recruitment, according to Sampaio. For the treatment of CNS disorders, the invasiveness of gene therapy approaches is somewhat more acceptable than for systemic disorders, she said, because of the 20-year history of using deep brain stimulation as a treatment approach for Parkinson’s disease (PD); however, intravenous delivery would represent a major step forward over intrathecal administration.
For complex polygenic disorders where the pathophysiology of the disease is not fully understood, drug development is even more challeng-
___________________
4 For more information about EMA’s Innovation Task Force, see https://www.ema.europa.eu/en/documents/leaflet/innovation-task-force_en.pdf (accessed June 8, 2019).
ing, said Marks. Sampaio agreed, adding that while combination therapy is widely believed to be necessary to treat neurodegenerative diseases, regulators have been reluctant to embrace these approaches. Nonetheless, she said, there is a combination gene therapy product called ProSavin that is currently being tested in clinical trials for the treatment of PD (Palfi et al., 2018). Using a lentiviral vector, ProSavin delivers genes for three enzymes involved in the biosynthesis of dopamine.
Clinical trials for gene-targeted therapies are often constrained by the small number of potential participants, even in non-rare diseases, because not everyone will be a candidate for gene therapy, particularly if there is invasive administration, said Sampaio. Small numbers of participants and the severity of the condition being treated may also lead sponsors to propose alternatives to placebo-controlled trials. However, Sampaio argued that while placebo controls in very small trials may not produce the statistical power to demonstrate efficacy, they can nevertheless ensure blinding. This is important to avoid safety misreporting and to facilitate a balanced interpretation of biomarkers, keeping in mind that even biochemical markers can change with placebo. The need for using a sham intervention as a placebo may also add logistical constraints, she said.
Identifying appropriate and clinically relevant outcomes may also be challenging, particularly if the clinical readout is long or there are no reasonably like surrogates, said Sampaio. She argued, however, that the potential for very large effect sizes from gene therapy—even possible cures—may mitigate some of these problems, particularly those that result from the necessity of conducting trials with small numbers of participants. Although the potential of having a durable effect increases the appeal of gene therapy, it can be difficult to prove, she said.
Transitioning from First to Second Generation Vectors
In January 2019, a statement from Peter Marks and Scott Gottlieb, then FDA Commissioner, predicted that by 2020, FDA would be receiving more than 200 IND applications per year and that by 2025, they would be approving 10 to 20 cell and gene therapy products per year.5 Samulski expressed concern that with this “tsunami of therapeutics” coming forward at a time when the technology is in the midst of a shift from first to second generation technologies, drug developers might be unwilling or
___________________
5 Statement from FDA Commissioner Scott Gottlieb, M.D., and Peter Marks, M.D., Ph.D., director of the Center for Biologics Evaluation and Research on new policies to advance development of safe and effective gene therapies. See https://www.fda.gov/news-events/press-announcements/statement-fda-commissioner-scott-gottlieb-md-and-peter-marks-md-phd-director-center-biologics (accessed June 5, 2019).
unable to take advantage of superior vectors because only the inferior first generation vectors have been fully evaluated, endorsed, or approved by regulators. He suggested that developers and regulators will need to work together to craft strategies for bridging studies that will enable an efficient shift from one formulation to another even as that change in formulation may affect the specificity of targeting, the immune response to the capsid, payload size, transduction efficiency, and need for repeat dosing. Vaishnaw added that regulatory pathways are also needed that would enable quick bridging studies from one disorder to another, given that the platforms for developing therapies hold much in common even when designed for different conditions.
MANUFACTURING CAPACITY
The best vectors in the world are useless if they cannot be manufactured in sufficient quantities, said Robert Kotin. Thus, a major challenge for gene therapy drug development is chemistry, manufacturing, and control, he said, noting that vector production for both non-clinical and clinical vectors is difficult and expensive. Moreover, while relatively small doses are needed for subretinal injections to treat ocular indications, other CNS gene therapies may require much larger doses depending on the delivery method and indication. For example, systemic dosing for the treatment of diseases such as DMD require relatively large doses, said Kotin.
Over the past two decades, methods of producing AAV vectors have evolved to enable large-scale production of vectors at good manufacturing practice (GMP) facilities, said Kotin. Collaboration between academic researchers and industry has been critical to this evolution, he said. For example, a partnership between the University of Massachusetts Medical School (UMMS) and industry partners brought together the expertise of virologists and vectorologists at UMMS with the industry’s engineering expertise to generate the large quantities of good laboratory practice vectors needed to support large-animal, dose-escalation studies. These processes may then be transferred to a GMP facility for large-scale manufacturing.
Developing and delivering gene therapies to the many patients with rare diseases will only be possible if manufacturing costs can be driven down to a sustainable level for common diseases with a big market potential, added Marks. Thus, he said, FDA’s efforts to streamline manufacturing are especially important. A commonality among all advanced therapy medicinal products is that product quality, safety, and efficacy are inextricably linked, he said. Thus, he noted that for these products a controlled manufacturing process and an understanding of critical quality attributes is essential. Manufacturing AAV vector products, for example, requires dealing with multiple manufacturing challenges, including empty capsids, purification,
and contaminating nucleic acids, he said, adding that only a few companies have mastered this process to date. He suggested that the translation of scientific advances made in academic laboratories to commercially manufactured products that help patients could be advanced by developing a set of non-proprietary AAV vectors and a “cookbook” of how to engineer vectors that could transfer easily into proprietary systems.
