3

Discovery Research

Drawing on the Global Health 2035 report, Adel Mahmoud, Professor in Molecular Biology and Public Policy at Princeton University, said that achieving convergence in global health involves rapid scale-up of existing and new life-saving interventions, and building the health care structure to deliver them.¹ However, many of the drugs, vaccines, diagnostics, and other devices needed to address global health challenges do not yet exist. Participants discussed current scientific tools, technologies, and capacities that enable the rapid discovery, development, and evaluation of medical products and highlighted gaps and needs for advancing research. Highlights and main points are summarized in the box below.

__________________

¹ For more information about the Global Health 2035 initiative see http://globalhealth2035.org (accessed October 30, 2015).

ADVANCING THE SCIENCE OF DISCOVERY

Platform Approaches to Vaccine and Therapeutic Research

Trevor Mundel, President of the Global Health Division of The Bill & Melinda Gates Foundation (BMGF), reviewed some of the platforms or approaches that could be used to address unanticipated pathogens for which there are no tools available, and for which traditional methods are unlikely to be successful. New vaccine platforms in development include nucleic acid–based vaccines (RNA-based vaccines, DNA plasmid vaccines) and live viral-vectored vaccines (the three main Ebola vaccine candidates are live viral-vectored vaccines). AAV vector technology is being studied not only as a vaccine approach to deliver immunogenic antigen, but also as a way to confer passive immunity through vector production of antibody. The potential for these platforms is the ability to switch in a nucleic acid “cassette” for the desired expression product. In a very short period of time, one could theoretically go from having the genetic sequence of a pathogen, to inserting a gene construct into an approved platform, to having a vaccine candidate(s) that can be taken into clinical trials. A key regulatory question, Mundel noted, is what set of evidence will be required for approval of a platform, so that preclinical studies do not need to be replicated?

Another platform is a convalescent plasma/fractionated plasma approach to therapy. This is based on the principle that patients who mount an immune response will produce effective neutralizing antibodies

against novel epitopes, and that plasma from those patients could potentially confer passive immunity to other patients. Mundel said that, during the Ebola outbreak, BMGF had bloodmobiles equipped with plasmapheresis and blood-cleansing capacities airlifted into West Africa to facilitate a convalescent plasma study (results are pending). New techniques are also being used to compare antibody responses to a given pathogen by survivors versus those who succumbed to fatal disease (including HIV, malaria, and Ebola). Mundel noted that there were challenges moving samples between countries and exporting samples out of West Africa for fractionation, and a meeting of the health ministers from the three countries was ultimately convened to secure permission for sample movement. New high-yield production systems have made the prospect of using neutralizing antibodies as therapeutics plausible. As noted above, there is also the potential to deploy vectored antibodies (e.g., AAV vectored) for sustained production of antibody in patients in an emergency circumstance. A key challenge for this platform is ensuring adequate process control to guarantee product quality, Mundel said.

As mentioned by Bilbe, one platform for rapid drug discovery is repurposing existing compounds. Companies have extensive compound libraries of well-characterized compounds, some of which have been through preclinical or clinical testing. Starting with a compound from one of these libraries can substantially reduce development timelines, often from years to months. One of the challenges from a regulatory point of view, Mundel noted, is scrutinizing and prioritizing the vast sea of compounds proposed for clinical testing, given the very limited capacity to conduct clinical trials.

Yamada raised the question of whether an AAV antibody platform could be preapproved, in similar fashion to the platform used for production of seasonal influenza vaccine. Could the gene coding for a protective antibody isolated from a survivor of an outbreak be inserted into the vector, and approved for passive immunization without the need for additional clinical trials? Hamburg responded that it might be possible to issue an emergency use authorization (EUA), depending on the foundation of data, the experience with the platform, and the understanding of the nature of the particular outbreak. The seasonal influenza vaccine experience does provide a model for the concept of having a flexible platform in which a new product can be rapidly approved based on an existing set of safety and efficacy data. However, it becomes more complicated, and there is less confidence, when the platform would be dealing with potential unknown factors including the nature of the disease and the level of risk people are willing to take in the context of a true public health emergency.

Michael Pfleiderer, Head of the Viral Vaccines Section at the Paul-Ehrlich-Institut, highlighted the need for scientific convergence on product development issues of global concerns. For example, what is the right pan-

demic influenza construct (e.g., booster versus primary vaccination, number of doses) and manufacturing approach (e.g., if the amount of antigen per dose needed can be reduced, such as through the use of adjuvant, more vaccine can be made and more people protected).

Mahmoud, Pfleiderer, and others highlighted the gap (often called “the valley of death”) between the discovery research that takes place in academic, government, and industry laboratories and the product development that necessarily takes place in pharmaceutical companies. Hamburg noted that many opportunities are being missed because studies being done in academic or other early research settings are not being guided toward full product development. It is essential to work across the silos to shape the research and development plans from the beginning with the end in mind, so that the right studies are done as effectively and efficiently as possible, and products reach the people that need them, at the scale and timing needed. As an example of this gap, Mahmoud noted that Ebola virus is among the top 20 viruses considered by the U.S. government to be a potential biothreat, yet several Ebola vaccine candidates had been tested in animals 10 to 15 years ago and then sat in freezers with no further development until recently. To help span the development gap, Mahmoud and others have proposed a Global Vaccine Development Fund that would fund activities spanning from the discovery of a candidate to the end of phase II clinical testing (Plotkin et al., 2015). Funding would come from governments, foundations, and industry donors and would be awarded through a competitive process. Yamada added that, because much of the basic scientific advancement comes from academia, one of the challenges is how funding is distributed in academia.

Rex reminded participants of a prior Institute of Medicine (IOM) workshop that discussed selecting one or two pathogens from the threat list each year and running the exercise of bringing a product from discovery research through to a designated stopping point (phase I or phase II of clinical development) or to completion (acknowledging that further clinical trials might not be possible in the absence of cases). Many issues could be solved in advance of a crisis (e.g., methods, IP), and the result of the exercise would then be a product ready to be developed further, along with the developed expertise to do it, if the threat does emerge (IOM, 2015a).

Devices: Diagnostics and Personal Protective Equipment

Charles Goldstein, Chief Scientific Officer for Greater Asia at Becton Dickinson, emphasized the need to develop products that are appropriate for the intended market. Devices and diagnostics need to fit with the way health care is practiced in the targeted countries, and the availability of infrastructure and training for health workers.

Pauwels observed that many of the diagnostic solutions used in the Ebola response (e.g., the mobile laboratories) were slow to deploy, the logistics were costly and complex, and sustainability remains uncertain. Costs include not only the cost of the devices themselves, but, for example, the costs of disposal of consumables, transport and maintenance of equipment, and transport or training of people to do the work. Pauwels added that the physical transfer of samples must also be addressed. The practical field realities for deploying Ebola diagnostics included blood samples that were a “pink slurry” upon arrival, samples without names or addresses, or all samples labeled with the same name.

Pauwels emphasized the importance of bringing diagnostic testing capability closer in time and space to where the patients are (i.e., at the point of care). A typical situation in Sierra Leone, he said, was for exposed or symptomatic individuals to stay in holding or treatment centers for 2 to 3 days while awaiting results of the molecular diagnostic test. This meant that the 70 percent of people who ultimately tested negative for Ebola virus were sharing the same space with the 30 percent who were positive for Ebola. Building more laboratories across these regions is not necessarily the solution, Pauwels said. Data show that the accuracy and reproducibility of laboratory testing is highly variable, and the number of diagnostic errors is substantial. This is likely due to the complexity of the technologies, the instability of reagents, and other factors.

Pauwels described Idylla, a real-time polymerase chain reaction (PCR)-based molecular diagnostics system developed by Biocartis that is fully automated from clinical sample to result. It is a compact, cartridge-based, closed system (no dedicated PCR laboratory infrastructure is needed, and contamination is reduced); all reagents are on board (no cold chain needed); results are rapid; and there is less than 2 minutes of hands-on time by the operator. The system is fully integrated, scalable, and can test for up to 30 biomarkers in one sample simultaneously. Local people can be trained as Idylla users, and minimal infrastructure is needed (an autonomous, mobile diagnostic laboratory powered by batteries and an onboard generator are in development).

Pauwels added that Idylla can be used to create a real-time, sustainable diagnostic grid for surveillance. Idylla sentinels transmit test results to a central control room that monitors the data for trends. Pauwels said that feedback from local clinicians indicates that they need syndromic diagnostic panels (to test for, e.g., respiratory infections and central nervous system infections). There is also a need to assess viral loads and monitor evolving pathogenicity and emergence of drug resistance. Yamada suggested that a syndromic diagnostic panel including the top 10 priority diseases, for example, would be very informative and would help to identify emerging diseases. Mundel also noted the potential of diagnostic platforms as tools

for surveillance and added that machine learning tools are being deployed for automated monitoring of media reports in local languages to detect signals of emerging infections.

Craig Colton, Division Scientist at 3M Personal Safety Division, said that PPE (e.g., respirator, suits, gloves, and goggles) plays an important role in delivering health care during infectious disease outbreaks. PPE is regulated relative to both approval for marketing and conditions for use, as proper use is essential for protection of the worker. Colton pointed out that the way each infectious agent is transmitted (e.g., inhalation, touch, both, or other methods) affects how PPE will be used and drives the research and testing that need to be done for each product. Parameters include, for example, conditions of use, length of wearing time, or use in conjunction with other PPE. The product must not only be effective, but must be robust enough for the particular environment in which it will be used. For example, the Ebola outbreak posed new challenges for the development and use of respirators in that no skin was to be exposed, Colton said. This affects donning and doffing, including how double-gloving might impact the ability to manage the respirator strap. This drove research on new strap materials and new ways to grab hold of the strap. Another challenge was finding materials that would be better suited to the heat and humidity in the outbreak regions, and that might actually cool the workers. In some cases, there are surrogates for predicting performance (e.g., filter effectiveness in the face of various biological challenges). However, for other parameters (e.g., fluid resistance), the tests are not as developed. Colton added that some research and development of PPE (e.g., the effectiveness of a biocide as part of a product) could require working with infectious agents that the PPE industry does not generally have the facilities for.

Shanelle Hall, Director of the Supply Division at the United Nations Children’s Fund (UNICEF), underscored that products cannot be developed in isolation from how they will be used, and PPE use by untrained health workers in a community care center will be very different than that used by highly qualified people in an emergency treatment unit. Development is an iterative process, informed by how products are actually being used.

Regulatory Science Capacity

Developing new platforms and technologies is only the first step, Mundel said. Effectively deploying interventions requires partnership with regulators, including developing their competence and capacity in regulatory science, as well as addressing the broader community and political constructs. Pfleiderer observed that the regulatory field has changed drastically in recent years, and now the process is more of a partnership with

product developers. Hamburg agreed and added that the U.S. Food and Drug Administration (FDA) experience shows that the time for the research and development can be reduced when the regulator is involved early, and in a continuing way, in the research and development plan.

Hamburg highlighted the need for a new emphasis on building regulatory science capacity. Academic, industry, and government science need to come together to ensure that the knowledge and tools are available to be able to assess a promising candidate product for safety and efficacy; achieve an acceptable risk–benefit ratio; and ensure that the product can be manufactured consistently, with quality, and scaled up. Hamburg highlighted several critical areas of science that are ripe for further development including identification, characterization, and validation of biomarkers; the use of surrogate end points; the development of innovative clinical trial designs; the use of information technologies for disease detection and surveillance or post-market pharmacovigilance; the use of modeling and simulation; and predictive toxicology.

INTELLECTUAL PROPERTY AND SHARING OF DATA AND REAGENTS

Researchers, public health institutions, and companies alike have concerns over the access to the building blocks of knowledge, key research tools, and technology platforms necessary for developing a diagnostic, drug, or vaccine, said Anthony So, Professor of the Practice of Public Policy and Global Health at Duke University. As an example, So described the case surrounding the “ownership” of the Middle East respiratory syndrome (MERS) coronavirus (see Box 3-1). At the core of the controversy is the Erasmus Medical Center Material Transfer Agreement (MTA). Although it was claimed that the Erasmus MTA was based on the Uniform Biological Materials Transfer Agreement (UBMTA),² the Erasmus MTA in fact gave Erasmus ownership rights over any inventions made by recipients that directly relate to the material, in perpetuity. Under the UBMTA, the agreement may be terminated if the material becomes generally available from third parties (e.g., a reagent catalog or public depository).Yamada, So, and others discussed that there is no definitive answer on the status of the ownership of a vaccine strain that has been derived from a patient. Some countries consider it to be a naturally occurring substance that would not be patentable; others assert that under the Convention on Biological Diversity countries have sovereign control over such samples.

__________________

² The UBMTA is intended to facilitate the transfer of biological materials between the institutions, while protecting the rights of the provider to commercialize the material, and allowing the recipient to publish research findings in a timely way.

So highlighted some of the challenges to sharing data and reagents. There are transaction costs of assembling knowledge, and institutional arrangements need to include efforts to search, cross-license, and keep in check undue royalty stacking. Scientists in competition for grant support might slow the dissemination of research materials or findings in their quest to publish first. Producing information or materials in response to requests also takes time and effort away from research. Sharing is also affected by whether the data or material have known commercial value. A company might be concerned about liability risk associated with the pursuit of a secondary indication (e.g., drug repurposing studies might find new adverse drug reactions, or result in deaths when tested in a sicker population than the original intended targeted population).

Patents, licensing, MTAs, and other IP arrangements can powerfully shape the innovation ecosystem, So said. Patent holdouts can occur when the original company refuses or delays the licensing of IP to others. Patent thickets result from overlapping IP rights of multiple patent owners, all of which must be dealt with before moving forward. There can also be a

temporal lag between the pace of emergent disease outbreaks and the prosecution of patents (i.e., it takes longer to resolve the patent claims than for the outbreak to subside).

There are many variations of collecting IP for medical research and development. Patents and licensing range from exclusive licenses of proprietary information, to patent pools of nonproprietary information. MTAs range from negotiated access to private compound libraries to public repositories of compounds. Similarly, data sharing spans negotiated access to proprietary data, to open-access databases. Publication of data ranges from traditional paid subscriber access to open-access journals.

Pooling arrangements have often relied on centralized institutions as facilitators. For example, the Biomarkers Consortium³ pools data on biomarkers that might be used in the regulatory process. Public- and private-sector contributors offer a limited, nonexclusive, royalty-free license to their IP to others in the pool. In the World Intellectual Property Organization (WIPO) Re:Search consortium,⁴ contributors agree to grant recipients royalty-free licenses to both data and other IP (e.g., reagents) for research and development of products addressing specific neglected diseases in the least developed countries. Pooling arrangements can also involve a proprietary compound library that is made available for screening to a product development partnership by a research group (e.g., the Tres Cantos Open Lab Foundation established by GlaxoSmithKline [GSK]). Hits identified during the compound screening process remain confidential, and the compound owner has the right of first refusal for development of the compound. The converse, So said, is essentially crowd-sourcing promising compounds for company-defined disease targets. Eli Lilly’s Open Innovation Drug Discovery Initiative,⁵ for example, provides compound design tools in the cloud for outside investigators to use. In silico screening is done on the neutral network, protecting information from access inside or outside of the company. Outside investigators can then decide whether to submit their compound designs to Eli Lilly or not. To be effective, these pooling arrangements must have strategic fit within a larger innovation ecosystem, So said. He cited the Pandemic Influenza Preparedness Framework as one such innovation ecosystem that brings together multiple stakeholders who, in exchange for access to virus samples, commit to donating pandemic vaccine or drug or reserving production capacity to supply the World Health Organization (WHO) stockpiles for low- and middle-income countries. There are lessons from other sectors as well, So said. The International

__________________

³ For more information see http://www.biomarkersconsortium.org (accessed October 30, 2015).

⁴ For more information see http://www.wipo.int/research/en (accessed October 30, 2015).

⁵ For more information see https://openinnovation.lilly.com/dd (accessed October 30, 2015).

Treaty on Plant Genetic Resources for Food and Agriculture is noteworthy for having codified a compensatory liability regime that allows researchers to freely take the materials for any research purpose, without the need for any permission to use, and requires payment of equitable compensation if the research leads to a commercial application and gain.

Shaping the innovation ecosystem is the responsibility of governments, So said, and demands a multilevel framework that creates an enabling environment for sharing of data and reagents, and enabling medical research and development to meet emerging needs.

Stoffels suggested that large companies secure patents to preserve the freedom to operate. A company can decide to give patented IP away, but if it does not patent, it ends up paying royalties to third parties to use that IP. So suggested that one approach could be to create a research semi-commons, whereby patent offices would waive the patent filing and maintenance fees for IP that is “parked” for the main purpose of preserving the freedom to operate. This IP would then be available more readily for research.

Reid Adler, founder of Practical Innovation Strategy, noted that the ecosystem for transfers of infectious materials and related data involves many different control points (e.g., national laws; international treaties; public health organizations; the rules and procedures of research institutions, funders, companies, publishers, biological materials depositories, and others; and interpersonal relationships). Numerous factors affect or impede the sharing of biological material and data, including

• regulations,
• expectations of affordable access to products developed through use of shared materials,
• time and administrative burden of dealing with legal provisions in agreements,
• cost and logistics of sharing materials,
• concerns about opportunity costs,
• competition among researchers/institutions,
• diversity of materials being shared,
• misperceptions or frustrations about IP, and
• expectations about access to training or technology (poorly characterized, processed, or labeled materials).

Adler mentioned several successful models and frameworks that address sharing, including the National Institutes of Health (NIH) grants policy, the UBMTA, the WHO Pandemic Influenza Preparedness Framework, and the BMGF AIDS research agreement. Examples of data-sharing efforts discussed by Marks and others included Data Sphere (which facilitates sharing of de-identified patient-level data from cancer clinical trials), GA4GH

(the Global Alliance for Global Health, facilitating responsible sharing of genomic data), and TransCelerate BioPharma (creating standards for clinical data transparency that preserve patient privacy). Dzau referred to a recent IOM report that made specific recommendations for the timing of the release of clinical trial data (IOM, 2015b).

Adler emphasized the need for frank and open communication, relationship building, and establishing trust, all of which are more difficult to do in the midst of an emergency. Trusted rules or agreements should be negotiated and agreed to in advance of when needed, he said. Ask researchers and institutions what they need to be able and willing to share data and materials, he suggested, and learn from experiences where sharing of materials was not optimal. Providing practical training to researchers and institutions about material sharing and related agreements (e.g., patents, MTAs, informed consent, and benefit sharing) could increase understanding about sharing and make the system work more smoothly.

Michelle Mulder, Manager of Technology Transfer, Grants, and HIV Program at the South African Medical Research Council (SAMRC)⁶ said that SAMRC has a dual role in both funding and conducting health research, and supports IP protection for three main reasons: to ensure control over how it is used, to leverage partnerships, and to incentivize private investment. The approach to IP and data sharing seeks to balance the global access principles with profit motive. The primary objective is to get products to those who need them most, at affordable prices. SAMRC is bound by the South African Intellectual Property Rights Act of 2008, which is aimed at ensuring that benefits from publicly funded research accrue to South Africa. Mulder noted that the Act allows SAMRC to award free licensees to IP for research. Together with international partners, SAMRC has also developed a socially responsible licensing guide. SAMRC also promotes public and precompetitive access to data (generally after protection/publication). For example, the Regional Prospective Observational Research for Tuberculosis South Africa is a network of institutions and investigators who will collaborate on established clinical studies, using a standardized set of definitions, standardized protocols, and well-characterized populations. All collaborating institutions will be required to place certain data and specimens in a central repository for availability to all collaborating institutions within the network. The broader tuberculosis research community may apply for access under certain conditions. Similar data-sharing provisions are part of the Malaria Drug Discovery Program, the H3 Africa consortium (Human Heredity and Health in Africa), and the HIV Reagent and Data repositories.

Marks pointed out that when discussing the potential costs of transpar-

__________________

⁶ Further information about SAMRC is available at http://www.mrc.ac.za (accessed October 30, 2015).

ency and data sharing, it is important to recognize that the alternative—if researchers have to keep reproducing the same work—will cost even more. He said that GSK is emphasizing a transparency model, placing both positive and negative data in the public domain, and the Tres Cantos Open Lab Foundation is just one model of how to manage IP and sharing of information and reagents. Marks observed that there are ever fewer players in the infectious disease space, and showing them examples of successful ways to manage transparency issues, data sharing, and IP could remove a barrier to entry. Yamada observed that companies have repeatedly shown they are willing to work together on an issue in times of crisis, but they are prohibited by antitrust laws from discussing pricing.⁷ Pricing is a critical issue for the distribution of a vaccine, and the inability to discuss price leads to incomplete discussion and sharing of the science.

A question was asked about compulsory licensing in pandemic situations, noting that it applies only to patents and not to materials.⁸ So responded that compulsory licensing is currently done at the national level, but there has been some discussion of the issues of collective action, such as an economic bloc acting as a group to implement a joint compulsory license. So suggested that a compulsory license is not very valuable if the market is for a single, small country, especially a developing country. Mulder said that South Africa has compulsory licensing and there have only been three applications thus far, and all have been denied. She noted that even though compulsory licensing applies only to patents, there are requirements that the materials needed to enable the patent must be deposited in a publicly available repository.

Sharing Negative Data

The importance of sharing negative data was emphasized. A participant observed that during the Ebola outbreak there were negative data that were not publicly available, and treatment decisions were informed only by the publicly available data. It was noted that industry is now releasing both positive and negative data, but negative data from academic and government laboratories are often not published or released. Marks noted that most journals are not interested in publishing negative results.

__________________

⁷ Though not presented at the workshop, discussion and proposals about price, particularly that aim to ensure that a product is affordable, can be undertaken with relevant stakeholders, such as the case of the MenAfriVAC vaccine under the Meningitis Vaccine Project. For more information see http://www.meningvax.org/faq.php (accessed November 13, 2015).

⁸ Briefly, a compulsory license is granted by a government to a generic drug-maker for the production of a generic version of a patented product without the patent owner’s consent. Grounds for granting compulsory licenses are determined by the individual country.

LIABILITY

Participants also briefly discussed liability concerns, which can be a disincentive for engagement in research. So said that a review by WHO of international no-fault systems for vaccines found that 19 countries have vaccine compensation systems to address liability: 13 in Europe, and none in developing countries. Typically, national governments are involved, funding may be derived from a manufacturer’s tax, the products that are covered vary, and eligibility and compensation decisions might depend on a standard of proof. In the United States, for example, systems include the National Vaccine Injury Compensation Program, and the countermeasures injury compensation provisions of the Public Readiness and Emergency Preparedness (PREP) Act. Venkayya added that, before passage of the PREP Act, potential liability was the primary impediment to companies engaging in pandemic influenza preparedness, and biodefense preparedness after the terrorist attacks. After the PREP Act was implemented, this liability was no longer a barrier to engagement. This legislation, he said, was critical to creating an environment in which companies were willing to make investments to address unknown threats, or known threats with an unknown time frame for emergence. Venkayya called for globalization of this concept, before the next crisis, and he suggested that an independent entity propose legislative language for consideration by countries.