Recommendations to Improve Collaboration and Expedite Acceptance of Animal Models

The contents of this commentary are the sole responsibility of the authors and do not necessarily represent the official views of the Biomedical Advanced Research and Development Authority, the Office of the Assistant Secretary for Preparedness and Response, or the US Department of Health and Human Services.

In our review of animal models for diseases caused by Francisella tularensis, Burkholderia mallei, or Burkholderia pseudomallei, we identified several scientific gaps that need to be filled before any one model will stand up to regulatory scrutiny (this issue). These range from a lack of well-characterized challenge material to the absence of data on appropriate treatment triggers, both of which are necessary to support regulatory qualification of a model in accordance with the US Food and Drug Administration’s (FDA’s) “Animal Rule” (21 CFR 314.610 and 21 CFR 601.91). The Animal Rule provides a regulatory approval pathway for candidate medical countermeasures (MCMs) when human efficacy data are impossible to obtain due to ethical reasons. This alternative regulatory pathway is especially critical for scientists and developers of MCMs for chemical, biologic, and radiologic or nuclear threats (CBRN) and emerging infectious diseases. To support the development of novel MCMs and meet the FDA’s Animal Rule requirements, the Biomedical Advanced Research and Development Authority (BARDA), a division within the US Department of Health and Human Services (HHS), established the Nonclinical Development Network (NDN) to support and accelerate animal model research and development. However, in the 2 years since the NDN was established, we and others have observed 3 areas that still need improvement—specifically, data sharing, data integration, and research coordination. Public and private entities must come together to resolve these challenges. In this commentary, we recommend several policies that could improve research coordination and enable resources to be leveraged across government agencies and their academic and industry partners.

Under Federal Acquisition Regulations, the US government (USG) retains unlimited rights to data that are generated under the contracts it supports. This unfettered access to data and information should be the basis for a similarly unrestricted flow of information across the USG. Sharing data and information will allow for the timely identification of scientific and regulatory gaps, reduce the probability of redundant USG funding of studies, and ultimately reduce the number of animals that are used for experimentation. Due to the relative immaturity of animal models for Burkholderia, the USG has an opportunity to effectively implement processes for data sharing prior to engaging in a significant body of work. USG agencies that anticipate supporting Burkholderia model development could introduce common provisions in their contracts to ensure that contractors and grantees understand that their data and methods can, and will be, exchanged across agencies. The barriers to implementing such data-sharing processes should be low since this type of research and development work would be conducted independent of product development and therefore would not be deemed confidential or proprietary. BARDA and other agencies within the USG have begun to incorporate these provisions into their nonclinical development contracts.

The USG has supported the development of MCMs for the prophylaxis and treatment of anthrax exposure/infection, including vaccines, antibiotics, and antitoxin therapies. Due to the diverse nature of the anthrax MCM portfolio, many animal efficacy experiments have been conducted. As one approach to addressing some of the regulatory uncertainties regarding the development of MCMs, BARDA engaged its contractors and requested the integration of data from control animals challenged with Bacillus anthracis to conduct a meta-analysis of the data. As a result, data from a larger number of animal studies were obtained and model parameters were effectively refined with greater statistical rigor. This allowed for the identification of reproducible markers of disease that have aided in developing and further characterizing an animal model of anthrax. For pathogens where there are limited animal efficacy data being generated in the context of products, it is imperative that policies and a framework be established that allow for a similar integration of experimental data. Optimally, data integration would extend to all USG agencies conducting animal model development efforts so that a minimum number of animals would be required for studies to establish the natural history of the disease, treatment triggers, and correlates of protection.

This increased volume, exchange, integration, and analysis of data would be most effectively facilitated and managed through a robust system architecture and database so that data can be standardized, collected, stored, and accessed in a centralized fashion. As a paradigm, the biodefense community could use the successful database approaches that have been established for the collection and management of toxicology data from clinical trials. It is critical that standards be established for the type and format of data entered into the database as well as their transfer between vendors, sponsors, and regulatory agencies. Ensuring compliance with these requirements will necessarily fall to the users of this database. Guidelines for analyzing the data sets within the database to produce statistically rigorous results should similarly be established. We also believe it is important that data from each individual animal be entered independently into the database. Of course, there are challenges associated with establishing and managing such a database. Significant resources will be required to maintain and conduct quality control of any database to ensure continued utility and accuracy of the stored data, especially given the scope and scale of data to be collected and integrated. Furthermore, institutional frameworks will need to be established to ensure that all agencies involved in animal model development upload all relevant data into the database in a timely fashion; without this, the MCM development community cannot benefit. One of the first uses for this database should be a scientific gap analysis so that key studies that need to be conducted can be identified and resources coordinated to ensure that they are addressed in a timely fashion.

The most effective path for animal model development would be one in which government agencies coordinate resources and manpower to expedite animal model development and qualification. A model that has emerged in the USG is to establish working groups to coordinate and inform ongoing activities. The Filovirus Animal Model Non-Clinical Working Group (FANG) has been effective at establishing near and long-term objectives for animal model development for these threat agents. ¹ This group comprises representatives from the Department of Defense, HHS, the Department of Homeland Security, and multiple biosafety level 4 laboratories around the country. Specifically, the FANG has identified multiple research gaps and initiated efforts to address them. To facilitate these efforts, the FANG has established a consensus position regarding the Ebola strain (Kikwit R4368) and source of filovirus to be used in animal experiments. As an iterative improvement on this successful model, the Burkholderia Interagency Working Group (BIWG) was initiated and includes representatives from the Defense Threat Reduction Agency (DTRA), National Institute of Allergy and Infectious Diseases, FDA, and BARDA. The goal of this group is to exchange data and coordinate research efforts and resources between USG agencies supporting Burkholderia animal model research. In doing so, the BIWG will avoid duplication and promote the formation of a coherent and consistent strategy for the development of Burkholderia spp animal models. Furthermore, DTRA and BARDA will use the data collected by the BIWG to sponsor and submit a coordinated Letter of Intent (LOI) to the FDA’s Animal Model Qualification Program (AMQP) for the rhesus macaque as a model of melioidosis following inhaled B. pseudomallei. The LOI is the initial formal communication with the FDA identifying the sponsoring entity or consortium of the animal model, the agent to be investigated, the animal model to be developed with the agent, and the proposed use of the model (eg, postexposure efficacy testing of an antibiotic). In response to increased interest in the Animal Rule, the FDA established the AMQP to facilitate the review of LOIs and animal model data packages and provide a route of communication between the FDA and the sponsoring entity. Although USG coordination will enhance animal model development, timely execution of experiments will occur through the BIWG; partner agencies will remain autonomous so that experiments are conducted in a timely fashion and not trenched in unnecessary bureaucratic approval processes.

The iterative process of conducting experiments to elucidate a disease process in a specific animal model, reporting those results to the FDA, and returning to the laboratory to resolve data gaps within the qualification package requires multiple stakeholders. The ability of these stakeholders to have access to interpretable data, efficiently exchange them through a database, and draw upon each other’s resources via working groups will greatly improve communications with the FDA and accelerate the qualification process. Adopting these policies early in the animal model development process, as the FANG and BIWG are doing, will hopefully facilitate and expedite animal model and drug development efforts. One of the most concerning observations we have made over time and during the preparation of our recent review article is that a number of relevant studies have been completed but have not been published or otherwise publicly shared. As a consequence, the larger research community is not benefiting from this knowledge, product developers are left with an incomplete awareness of the technical maturity of a model, and experiments are needlessly repeated. The impact is that it will take longer to qualify an animal model, and the development and/or approval of much-needed MCMs will be delayed. The coordination of resources and technical activities at the tactical level, coupled with an integrating architecture for data and policies that encourage the unfettered exchange of data, would expedite animal model development in the near to midterm.