Abstract
Sharon Terry (ST): I am Sharon Terry, and today we are discussing drug development for rare and neglected disorders. We have been working together for almost two years: a group of us in a large coalition made up of advocacy, industry, academia, and government looking at the issues surrounding drug development, and particularly for rare and neglected disorders.
And while it is very hard for all diseases, rare diseases and neglected diseases present some additional challenges. And what is interesting and important about what we are trying to do is we are looking at this in a systematic way, thinking about what could be better in the system, what paradigms could change that would allow more rapid drug development for these diseases.
We are going to hear from four individuals to allow us to look at these issues from a couple of different perspectives. During the time that we have been working with each other on these issues, we have identified what we are calling walls, some places that are particularly difficult and are in need of gateways. The GRANDRx initiative, which is Gateway to Rare and Neglected Disorders Therapeutics, looks at how we build these gateways together. How do we find ways to collaborate and how do we build better systems?
I am going to ask three of our participants to speak for five min on each of the different walls and our fourth participant to give some comments on the issues surrounding the Food and Drug Administration (FDA) and the challenge of getting approval for these drugs. I will ask each of you to briefly introduce yourself and then to talk about the wall that you are particularly addressing. Jim Inglese, if you could please begin.
James Inglese (JI): Sure Sharon, I’m the, Deputy Director of the National Institutes of Health Chemical Genomics Center (NCGC). I will focus on a discussion of the early stages in discovery when strategies to develop and align assays for high-throughput screening (HTS) and follow-up progression of lead compounds are considered.
The topic of assay development is becoming fairly mainstream these days, especially as more and more academic investigators become involved in translational research. This trend is accelerating with increased university commitment to enabling the interdisciplinary research needs of drug discovery. Those involved realize that is a need for mechanisms to bridge basic biological findings with the ability to identify novel chemical substances; these compounds might be used either as a chemical probe to understand and explore the biology under investigation in more detail, or as a starting point for a therapeutic.
Most drugs, therapeutics, are small molecules, which are chemical substances. So it makes sense to identify new chemical matter to transform into chemical probes and then beyond into therapeutics.
The definition of an assay that is probably appropriate here—the way I like to think of it—is an efficiently designed experiment to measure the effect of a substance on a biological process of interest. Design and format are particularly important when one is developing assays for HTS, given the need to ensure compatiblity with automated systems, as I will elaborate on momentarily. High-throughput screening in academia is becoming routine, as I mentioned before, so opportunities to target a wider range of diseases, for example rare disorders, are now becoming possible at an unprecedented level. Today’s scientist is coming face-to-face with a technologically enabling field of applied science that creates an interface between chemical libraries and biological assays, and offers great opportunities to the rare and neglected disease community. As HTS assays are a specific type and format of biological assay, it is vital these assays be designed with knowledge of both the advantages and limitations of current technology, and as part of the larger development “critical path.” HTS assays have to be extremely efficient in their design and they have to be compatible with the automation that is often a part of the HTS process.
This is where we can run into some problems developing assays for HTS. While essentially rather uncomplicated in protocol (or number of steps), HTS assays are often complex in design. This sophistication stems from the fact that they have to embody the fidelity of the biological process in a rather straightforward, simple format that a robotic platform can manipulate and from which a signal can be readily detected.
At the NCGC, we have worked with hundreds of investigators over the past close to six years to develop assays for HTS. We have identified a number of issues that we are beginning to address; for example, how best to deal with assays designed to mimic low-affinity protein–protein interactions or the optimal utilization of specific reporter proteins in assays targeting signaling pathways. As we work with the scientific community to develop the HTS and follow-up assays, one thing is becoming clear: we need to continue to educate investigators about this process, the types of biological assays compatible with HTS, how to develop them, and most critically, the procedures to verify the biological relevance of compounds with apparent activity.
HTS assays often can take anywhere from months to several years to construct and validate. So, it is quite an effort in some cases, and that needs to be understood and acknowledged. Resources required for this development process should continue to be made available, and enhanced, as this process represents a critical link between moving from the basic research realm to therapeutic development. Combined with the ingenuity and creativity of our academic counterparts, the possibilities for advances in this field are limitless.
ST: Chris Austin, could you talk now about what happens next? Let’s imagine someone is successful and builds a robust assay that can be used for HTS. You then move to the molecular libraries process and to the next wall, which we are calling the preclinical wall.
Christopher Austin (CA): This is Christopher Austin. I am the Director of the NCGC and Senior Advisor of the Director for Translational Research at the National Human Genome Research Institute (NHGRI).
Jim gave a great introduction to the problem of moving basic genetic or biochemical or cellular discoveries into a format that is amenable for HTS. Once those assays work well in 96-well format, our collaborators frequently apply via the NIH peer-review process for formal access to the NCGC or one of the other NIH Molecular Libraries centers, where the assay is optimized for robotic screening, in our case in 1,536-well format.
At the NCGC, all primary screens are performed in dose–response format, at generally seven concentrations covering four orders of magnitude. This process, which we call quantitative HTS, produces very rich datasets on over 300,000 compounds in every screen. Informatics analysis defines the pharmacology of the actives and their structure–activity relationships, and provides immediately actionable data for follow-up biology and/or chemistry optimization. We work closely with our collaborators to confirm actives and test them in secondary assays that confirm that the biology found in the original screening assay is actually scientifically robust in more physiological testing systems.
In order to produce compounds that have the desired properties (eg, potency, selectivity, cellular permeability), we frequently perform medicinal chemistry optimization. The chemical probes produced are then used in cell-based or animal model systems to test the project’s biological or therapeutic hypothesis.
Useful as these probes are for elucidating biological mechanisms, they are just the start of drug development. The next stage, producing a compound suitable for human testing, is what we have called the “chemistry brick wall,” or “preclinical” wall, in the GRANDRx vernacular. This stage in the biopharmaceutical environment requires 3–5 years, during which 3,000–5,000 individual compounds will be made by medicinal chemists and tested in an increasingly long list of biological assays in the hopes of retaining the original efficacy and selectivity that was found in the chemical probe stage, but add all the pharmaceutical properties that are required for testing for safety and efficacy in animals and then in safety and efficacy in people.
This preclinical stage, sometimes called the “valley of death” because of its high failure rate and high cost, is particularly difficult to traverse for rare and neglected diseases, where anticipated return on investment is low. GRANDRx envisions a dedicated effort to build a gateway through this preclinical wall (see Fig. 1). Since the GRANDRx vision was first articulated, the National Institutes of Health (NIH) Therapeutics for Rare and Neglected Diseases (TRND) program was started and directly addresses it. TRND is a congressionally mandated program to build on the successes of the NIH chemical genomics programs to develop new drugs that can be tested in the clinic for rare and neglected diseases. TRND bridges the gap in time and resources that often exists between basic research and human testing of new drugs for these diseases, and focuses not only on individual diseases, but also on paradigm improvements that may make the preclinical development process more efficient. The NIH Office of Rare Diseases Research (ORDR) will handle oversight and governance of TRND. Researchers will perform TRND’s laboratory work in a new facility administered by the intramural program of the NHGRI.
Barriers to the progress of drug discovery for rare and neglected diseases identifi ed by GRANDRx. Source: Image courtesy of Darryl Leja.
ST: Great. Thanks very much Chris. David Meeker, we ask you to talk about the rubber hitting the road, when a clinical candidate is ready for being tested in real people.
David Meeker (DM): Thanks Sharon. My name is David Meeker. I am an Executive Vice President at Genzyme and have worked there for 14 years on a number of the drug development programs in rare genetic diseases. We have a fair amount of experience dealing with the clinical wall as it is described.
I would start this part of the discussion by dropping back one step into Chris’s world, which is the challenge of animal models. In many of these rare diseases, no animal model exists, and a potential therapeutic may move into the clinic without a strong proof of principle. Secondarily, if the model does exist, it may or may not be predictive or it might be predictive but we would have no idea to what extent it is predictive since clinical trials may never have been conducted in that specific rare disease. We would only learn that later. In short, we often start with a hope and a belief based on what may be fairly limited data.
The clinical challenges are significantly impacted by the concept of rarity. First and foremost is that often little is known about the disease. It may not have been studied previously. Often what drives the initial surge in information is the availability of a potential therapeutic and that then causes people to rally to the particular disease, and information may grow rapidly.
But at the start of the clinical program, very little may be known, and there may be very few disease experts with which you can work.
So one of the most important things that a community can do to help prepare and facilitate ultimate clinical development is to organize themselves and build an understanding of the natural history of the disease. If you have a therapeutic that is capable of reversing some aspect of the disease, a trial may be relatively straightforward in the sense that you can have a comparator group and show that your treatment impacts or reverses that endpoint.
Alternatively, if you are not able to reverse anything—and this is a not uncommon outcome—but you are able to stabilize or prevent the decline, then the challenge is showing against a comparison group that you are slowing the decline. That becomes a much longer potential exercise. In this case, the value of the natural history is, of course, extremely valuable.
In addition to organizing the community, defining the natural history of the disease, a key third element is the development of validated biomarkers. As noted, clinical outcome measures may change slowly over time. And in the absence of prior clinical trials in that disease, there may not be a good understanding of the amount of time that may be required to show a change. And so the use and development of biomarkers becomes extremely important.
Often, these are hypothesis-driven and absent from prior clinical trials, and so these biomarkers will not have been validated.
Finally, tissue banks are another important potential facilitating element whereby a systematic collection of tissue from affected individuals in advance of a therapeutic being developed may create a source of information that will allow trials to be designed in a more efficient way.
An additional confounding variable in performing clinical trials in rare disease populations is the heterogeneity of the population. Ideally, you would like to test your new therapy in a homogeneous population where the only variable is the drug. A small heterogeneous trial population may make interpretation of the clinical endpoints difficult. Secondly, if you do need to do multiple trials, meaning a phase I/II and then a phase II/III or whatever the sequence, the selection of a homogeneous population for the first trial may make recruitment of a similar homogeneous population for the pivotal trial more difficult simply because the patient population is so small. As we discussed earlier, a well-organized patient population often through the use of registries can facilitate the selection of the optimal population for the trial.
Finally, it is critical to match the manufacturing scale up of the treatment with the clinical development program to insure that the product that will ultimately be used post-approval has been used in the pivotal trial.
ST: Thanks very much. Pat Terry, would you comment on the issues for the FDA and for drug approval?
Patrick Terry (PT): My name is Patrick Terry. I am a business entrepreneur—in the biotechnology and self-organized disease advocacy community. I think all of the previous speakers clearly outlined the complexities and the challenges facing us. It is not only the science, but also the technology involved here.
Biology, in general, is highly complex, as indicated by David’s comment about heterogeneity of disease not only in the disease state, but also in how individual patients respond to a therapeutic intervention. With all of these complexities, there is the additional challenge of regulatory science and for high-quality manufacturing of these assays and compounds. There is a single yardstick or measure on which products are evaluated around safety and efficacy. The role and responsibility for what the U.S. FDA and the European agencies are responsible is statutorily and legislatively defined. These standards are not optimized for today’s realities and technologies.
Most of the regulatory processes or current legacy system was established by bad precedents: harms being done to individual patients and a haphazard construct of remedies. We have an obstacle course to run when taking a compound from discovery to preclinical animal studies, designing high-quality trials, and recruiting patients to adequately perform multicenter trials that satisfy these regulatory standards.
When you look at what has happened and the infrastructure that resulted, the ecosystem around drug development and drug and biomarker development, along with patient accrual and biologic sample collection, it is a daunting challenge to design a truly efficient system of regulation. When you take a look at the system challenges that are exaggerated by rare and neglected diseases, it is tough to imagine a clear pathway forward for how you would take a systematic or industrialized approach to develop compounds in a systematic and efficient way. But I believe the GRANDRx approach can provide new opportunities and new efficiencies to redefine the pathway.
Currently, there are about 7,000 genetic diseases that have been described in humans. Most of these are loss of function status, so you have human knockout models of highly penetrant disease or clear disruption in normal biologic function. Each of them represents unique challenges as well as opportunities that currently do not really have a robust clinical and translational enterprise around these disorders because of the limited number of patients involved. But I think they can be extremely informative opportunities to understand basic disease pathobiology as well as therapeutic interaction that could extend to more common disease conditions.
The agencies are trying to think creatively, and I think industry is also trying to think creatively in this area. Certainly, NIH and Congress have been thinking creatively with their efforts around the TRND program and the recent emergence of the academic-based translational and applied science activities in rare disease. Industry is getting more and more interested in redefining the precompetitive space here and understanding how they can share best practices, information, and learnings from previous failures.
The FDA is trying to figure out how they leverage the repository of knowledge that they have accumulated when they review all of these compounds and failed applications and get to a place of best practices or a system in which community learning can be had so we can improve this process, especially in rare disease.
But ultimately when you look at the costs associated with going through this current regulatory regime, it is almost an unfathomable concept to imagine how you could take 7,000 rare diseases through this process and manufacture these highly expensive compounds for a patient community in any reasonable amount of time.
There is a need for a new kind of sensibility. I think the sociological and organizational framework of the process of assay and drug development is changing broadly because of the recent economic crisis in the United States, the extensive divestiture of the pharmaceutical industry from basic science and discovery work, along with the extensive outsourcing of this work. There are new market entrants. NIH is getting involved, as we see in the TRND program. There are a variety of other players, clinical research organizations, and others entering this space, such as the venture capital community, who are seeking models for taking a product through to a provisional approval in a phase IV type of registry so you could get commercial capacity and products to market under quicker time lines.
You could actually do small cohort trials in a really scientific, robust way, hand-in-hand with the regulators to create a new paradigm for regulatory science for these types of unmet needs. So there is a real willingness and flexibility in the regulatory agencies. Within Congress, and certainly in the folks on this call, individuals of like-mind thinking about new ways to create products and treatments in this arena of clear unmet need.
ST: Great. Thanks very much. So I want to spend a little bit of time speaking with each other about this. We have been working together for some time—and not only us certainly, but also a large number of government, industry, academia, and advocacy folks. I think all of us have been jazzed to do this work, not because we are going to find 1, 2, 3 treatments for rare or neglected diseases, but because we believe that by working with one another we will actually be changing the landscape and improving the possibility of more drugs being developed for these diseases.
I thought it would be interesting for us to talk about system change, which is certainly more compelling than just figuring out again one or two hits here or figuring out one or two wins. The systems change that will accelerate the entire transformation of this space from basic science to treatments is certainly formidable. We have seen some examples of it in other parts of biomedical science.
What do you anticipate or what do you hope for? You can speak to just one wall or you can speak to the whole and give us some idea of what we think might change as we work together.
PT: I would like to immediately jump in. Even though I reflected on the regulatory and process challenges, I think that from my own personal experience from working with rare genetic diseases from basic science to translation to animal models to clinical trial design, probably the most disruptive element of change here that is just emerging is the idea that patient communities can self-organize. They can, as David said, define the true disease status, the natural history of the disease, the various clinical phenotypes, and the appropriate clinical endpoints associated with disease mitigation and treatment response.
The community can get organized, drive research in a patient-centered way, can develop work-arounds to the unique challenges associated with rare disease drug development and have a sense of urgency to navigate the system. I think once the broader patient community, and certainly Genetic Alliance and other stakeholders in the GRANDRx initiative, conduct capacity training and build the wherewithal of the community to be truly effective, then that starts to ferment and gather momentum to develop best practices on how to engage drug development, how to organize clinical cohorts, and so forth. I think that is going to have a dramatic impact and positive impact on this challenge.
I think if we focus completely on drugs for rare diseases, we also lose sight of the value in actual improvement of the quality of life, alleviation of burden of disease in social networks that create services. These disease communities have expertise, and perhaps they cannot advance actual therapies, but certainly they can provide disease interventions or processes to ameliorate the severity of disease in this long and arduous road to develop drugs.
ST: Great. Other comments.
DM: I’d like to echo a couple of things there. Drug development and the job of the regulators are all about defining safety and efficacy. The rare disease package may be a much smaller number of trials and patients putting the regulators in the more difficult decision of having to make a risk/benefit analysis without the full benefit of the large package.
Factors that will strengthen the regulatory submission and give the regulators more confidence that they are approving a safe and effective therapy include information on the natural history and importantly on the biology of the disease. In other words, the better we understand the disease, the stronger the argument that the therapeutic makes biologic sense. The combination of supporting clinical data with a strong biologic rationale may allow a product to be approved with additional post-marketing clinical trial commitments.
CA: What I think we have seen in the last five years is a recognition by many in the patient/advocacy, academic, and biopharma communities that the problem we are all trying to solve is so difficult and so multidimensional that it will not be solved by any of these groups alone.
That, in my mind, is a sea change from the go-it-alone ethos that has traditionally existed in all three communities. The challenge—and the opportunity—over the next few years will be to establish new mechanisms by which these three groups will work together to achieve their common goal of new medicines for patients with rare and neglected diseases. This is a time of disruptive innovation, and traditional roles are changing rapidly. We have academia getting into areas that were previously the domain of biopharma. We have companies like Genzyme who are working in diseases that are highly unconventional and less common than have been worked on in the traditional biopharma model. And we have disease advocacy groups that are taking a much more involved role managerially and/or scientifically in the science and/or therapeutic development efforts that their organizations undertake and/or fund.
This realignment raises a lot of very important questions about business models, intellectual property, credit, publishing, and go/no-go decisions. Discussions of these questions led to the creation of the GRANDRx model we have been discussing.
JI: I’d like to comment briefly about the physical infrastructure that is appearing on the landscape sparking the relay race or the football analogy. This in part is due to the efforts of the federal government, individual institutes, or academic centers installing the laboratories and hiring the experts to do screening and more medicinal-like chemistry, as opposed to the type of methods development that was traditionally the primary focus of chemistry in an academic setting. Also, populating these new labs in the NIH’s network and throughout academia are the types of instrumentation and robotics (eg, reagent and compound handling devices and sensitive detectors) that we need to configure and make measurements from these low-volume specialized assay formats described earlier. Much of this technology was developed by funding from the pharmaceutical industry in prior decades to enable their drug discovery paradigm, and now all of us benefit from that investment; it is often not appreciated, and certainly, the industry never gets credit for it.
Today, we see some of those pharma investments as the technological infrastructure that will be key to allowing, for example, an academic investigator advancing our understanding in the biology of neuronal development to consider now becoming directly involved in the development of assays for HTS, and subsequent development of a therapeutic for rare peripheral neuropathies, such as Charcot-Marie-Tooth disease. In the past, they might not have been motivated because “selling” their idea to pharma or biotech was one of the only ways of initiating a drug discovery program, and that would be virtually impossible for a rare or neglected disease target.
Tapping into the intellectual capital of academia, for example, could be realized with a distributed model for assay development. Within a university, there may be not necessarily a large-scale screening center, but there may likely be components that could act as an assay development center: one can think of this as a facility that focuses primarily on developing and fine-tuning assays for HTS, which then could be imported to the NIH’s Molecular Libraries Initiative or other such programs. These assay development centers may also function to handle follow-up activities and in-depth testing of molecules that are optimized in medicinal chemistry labs local or elsewhere in close collaboration with the university faculty.
I believe that this type of network, the components of which have been growing over the last six or so years, is poised to make it physically possible for us to accomplish all of these advances about which we have been discussing.
ST: Fabulous. In addition to the excitement of the convergence of the right kind of science to make this pathway more robust, we are also hearing from each of you about culture change. As a result, things will be different, starting with Pat’s concepts around the roles of advocacy communities and into certainly the points made by Chris, David, and Jim. You all pointed to different ways that the system works.
It sounds like we are seeing a sea change, and that is really a fabulous thing. I am also wondering about this change in terms of the community, in terms of the people, in terms of the infrastructure and the capacity of the facilities, what change will we also see in the science? And in this regard, I am wondering if you could speak to cross disease research, more pathways science, more bundles, or whatever term we might use, for molecules rather than single entities? What do you see changing in the science as we start to put the culture together in new ways?
CA: I think any of us could speak to that. That question is a very, very important question, and one need look no farther than the name of the center that Jim and I are in to indicate where we think this field or this area is going.
One of the ways that we use the term “chemical genomics” is to describe using small molecules to draw connections between parts of the genome and parts of the functional infrastructure of the cell and the organism that were not previously appreciated. And we see this kind of thing all the time anecdotally in effects of pharmaceuticals that are unexpected, whether on the good side in something like Viagra for pulmonary hypertension, or on the bad side where we see something like progressive multifocal leukoencephalopathy (PML) from Tysabri. Compounds virtually always have more than one effect, whether due to interaction with the desired target in an unintended way, or a different target (ie, an “off-target” effect). However, because academic investigators, companies, foundations, and advocacy organizations all tend to have specific diseases they are interested in, those commonalities are frequently ignored and the opportunity for making new disease connections is lost. This despite the fact that it has repeatedly been the case, particularly in rare or genetic diseases, that critical insights have come from the studies of other diseases that turn out to be related.
Our approach, within both NCGC and TRND, is to take advantage of this connectedness of science, taking an explicitly cross-cutting, cross-disease approach, and capturing data on every project in a comprehensive way that can be used to generate insights into other targets and diseases. This approach is not only most efficient, but also the only strategy appropriate to the scale of the overall problem of the 7,000 rare and neglected diseases that Pat referred to. We are never going to address any appreciable number of those 7,000 diseases with the current one disease at a time paradigm. It has to be done by making connections between diseases that are not currently appreciated.
DM: I think that is exactly right. If you look at this from an industry perspective and think about what industry is trying to achieve, although we would love to decrease the cost of a specific aspect of drug development, what we are really interested in doing is decreasing the cost of failure. I think that for us right now is our biggest challenge.
I do think what Chris said happens, which is a program that is actively pursued and then fails may go on the shelf. It is unlikely given the portfolio management process inside the company that there would be a willingness to endlessly pursue a therapeutic or a pathway without some confidence that the probability of success was higher, for example, than the path that had just failed.
That is where, I think, a common mechanism like what is currently ongoing or being developed at the NIH, where information on failed compounds is made accessible has real potential. We would love to find a way to be more efficient, of course. We would love to find arrangements where our work was not lost and could be built on by somebody else with the hopes that we could come in later and pick up at a point where we can bring the kind of value that we as industry can bring to the equation.
ST: Terrific. I want to thank everyone for participating in this roundtable. From where I sit, which is working with the 1,000 disease-specific advocacy organizations and thousands of other academic and industry groups, as well as our government friends, this is a new age. We heard it from each of you: there are new and exciting ways that organizations and entities are interrelating that will bring about change that we are very, very hopeful for as we continue to work with the many people who suffer from rare and neglected diseases, the millions and in some cases more than millions of people, especially in developing nations who look forward to the kinds of treatments that we will bring forward.
So I thank you all very much not only for your time today, but also of course for your work as we continue to push forward. Thank you.