Abstract
Animal models of disease have become increasingly critical in the field of gene therapy to examine not only proof of concept, but also to focus in on the more detailed and specific phenotypes and surgical delivery methods that may need to be addressed in the human patients. As we correct the most lethal phenotypes in severe genetic disease, new unknown phenotypes emerge that were not previously discovered because the animal models or the patients did not previously survive to that point, or the phenotypes were masked by the more severe symptoms.
Animal models, particularly large animal models, are critical to identifying these phenotypes and addressing treatments to them as appropriate so that we can create the best quality of life in patients of severe genetic diseases. In this issue of the journal, we can see the diversity of animal models required to address the unique challenges of each disease process. Here we highlight some of the specific challenges that we can see in these examples and how they relate to the field of gene therapy more broadly.
For Pompe disease, Han et al. addressed an on-going concern that gene therapy may be less effective in female mice, and how that might need to be addressed if human clinical trials indicate a similar deficiency in effectiveness in genetically female patients. 1 This highlights not only the usefulness of animal models, but also a potential pitfall. Are we addressing a solution to a problem that will only be seen in the mouse model, or is this a critically important difference in treating female patients that the mouse model will help offer a solution for?
Both transgenic murine models where the androgen issue can be addressed and naturally occurring large animal models (including cats, dogs, sheep, and quail) that recapitulate some of the complex neurological phenotypes are necessary to further hone therapies for this complex disease. As we move forward with addressing the need to treat female patients, these large animal models may shed light on the need to further address gender differences in therapeutic efficacy and inform what will need to be applied to female patients.
The preclinical testing of adeno-associated virus (AAV) gene therapy for Krabbe disease outlines the progression from mouse model to large animal model in the Hordeaux et al. study published in this issue. In their article, they describe the clinical therapeutic approach to Krabbe disease in the mouse model (Twitcher mice) as well as a canine Krabbe model. 2 Krabbe disease, like Pompe disease, is a lysosomal storage disease that affects the central and peripheral nervous system.
Because these lysosomal storage diseases can have a complex phenotype with multiple downstream effects, the large animal models can offer more clinically relevant assessments and endpoints than murine models can. As an example, for Krabbe disease, it is possible in the canine model to perform much more detailed neurological examinations of both the central and peripheral nervous system when compared with what is possible in the mouse model for the same disease. This can allow a better assessment of what can be corrected effectively with each new gene therapy design and delivery route.
The study by Bikou et al. in this issue highlights a lack of previous large animal models as one of the roadblocks to developing effective gene therapies for pulmonary hypertension, with <3% of new therapeutic targets translating to clinical trials. 3 Rodent respiratory systems have significantly different anatomy and pathophysiology of pulmonary hypertension, limiting their use as models for translating gene therapies from animals to human clinical trials. This is particularly true when trying to model delivery modalities in small rodents versus animals with a respiratory system similarly sized to a human patient.
For example, it is difficult to target only the lower respiratory track, as was done in this porcine study, to not lose the gene therapy agent (vectors, oligonucleotides, small molecules, etc.) in the upper respiratory track (nasal passages, pharynx, and trachea), or even have a portion swallowed if the mice are not sufficiently anesthetized. In this study, they were able to show that endobronchial delivery was superior in vector uptake and improved delivery to the lung when compared with intratracheal delivery. Although this specific outcome may vary depending on the disease model, it is critical that we have animal models that as closely recapitulate the anatomy as well as the disease state to test therapeutics in a meaningful way.
This was highlighted when the authors pointed to studies taking place in the 1990s and early 2000s when initial adenoviral and AAV gene therapies were first used to target the lung for cystic fibrosis gene therapy. From that period, we have made halting progress at best when it comes to pulmonary tissue targeting in gene therapy. This has to do, in large part, to physical and immune barriers of the lungs that are critical to protect from infection but hinder therapeutic entry, the cellular complexity of the lung, and the issues of constant cellular turnover of most of those cell types. Those issues cannot be sufficiently addressed and overcome without nonmurine large animal models.
In addition to neurological and cardiopulmonary diseases, this issue also highlights the need and use of animal models for developing therapeutics for musculoskeletal conditions. When it comes to the musculoskeletal system, large animals have provided naturally occurring models for therapeutic testing from early on. Naturally occurring dog, equine, rat, and rabbit models are just a few of the large animal species that have been critical to targeting musculoskeletal conditions. The studies by Senter et al. and Seol et al. both highlight the use of induced larger animal models of traumatic osteoarthritis in testing gene therapy strategies. 3,4
In summary, we can see highlighted in this single issue of Human Gene Therapy a wide range of diseases and animal models that are necessary for advancing preclinical gene therapy in a way that mice alone do not allow. As the gene therapy field becomes more comfortable with designing and interpreting these efficacy studies, we hope that they may also satisfy FDA pharmacology and toxicology study requirements and translatability to patients, hopefully reducing or eliminating the need for additional safety and toxicology studies in normal nonhuman primates. This could greatly reduce the time, cost, and animal use in the preclinical path to investigational new drug approval through the FDA. As well to hone the use of these models, the goal should be to expedite our path to clinical trials while using the fewest number of animals overall.