Executive Summary
Drug development is an expensive, high risk process that takes 10-15 years.1 In most cases, one in a thousand synthetized compounds progress to the clinic. Before reaching the clinic, new compounds must undergo a rigorous process involving in vitro and in vivo testing to establish their pharmacology and biochemistry. Failure of a new drug to reach the market can be very costly; however, the effective application of translational research measures to a development program in Phases I and II could lead to the identification of efficacies that can result in a potentially quicker route of drug approval.1 Although translational research is expected to facilitate the development of safe and efficacious therapies, the results have been mixed.2 This has been attributed to flaws in the process to translate basic research findings to clinical research. This white paper presents the areas in translational research that have contributed to the failure of moving therapies to the clinic and the actions taken by experts to correct such deficiencies.
Although HGPS children appear normal at birth, the clinical manifestations of the disorder, including partial to total hair loss, loss of subcutaneous fat, bone changes, nail degeneration, skin abnormalities over the abdomen and upper thighs, and delayed teething, surge during their first to third year of life.3 Later disease manifestations include hearing loss, dental crowding, lack of secondary tooth eruption and cardiovascular disease.
The etiology of HGPS was unknown in 2002, but technological advances proved to be useful at dissecting the molecular mechanisms involved in this genetic disease.3 Through the use of novel techniques, it was discovered that HGPS was caused by a point mutation on the laminin A gene C-terminal region.3 In vitro models showed that laminin A gets modified through the addition of a farnesyl group at the C-terminus, allowing laminin A to interact with the nuclear membrane and to be released from the nucleus after the cleavage of the farnesyl group by the enzyme ZMPSTE24.3 The point mutation of laminin A’s C-terminal region prevents the cleavage of the farnesyl group resulting in the accumulation of farnesylated laminin A in the cell nucleus.3
Genetically engineered mouse HGPS models were used to test the theory that the inhibition of farnesyl transfer to laminin A could ameliorate cellular abnormalities.3 Although the mouse models lacked all the abnormalities seen in HFPS children, they provided data supporting the use the farnesyltransferase inhibitor lonafarnib in HGPS patients. Later animal studies also provided evidence for the beneficial effects of statins and bisphosphonates for the treatment of cardiovascular disease in HGPS patients and the planning and implementation of new treatment strategies using rapamycin, a drug that prevents organ rejection.3 Thanks to translational research, the lives of HGPS children have been improved. From this success story it became clear that successful translational science involves a team effort based on tight communication and collaboration between basic scientists and clinicians.3
What is Translational Research?
Translational research is defined as the transfer of knowledge gained from basic science to develop new methods to diagnose, prevent and treat diseases. It is an iterative process wherein scientific discoveries are integrated into clinical applications, and, conversely, clinical observations are used to generate new basic research – “bench to bedside and back to bench”.4 The basis for translational research lies in sound scientific and clinical principles involving systematic approaches beginning from sample and data collection, lab investigations, data analysis, preclinical testing, clinical trials, treatment efficacy monitoring, and the evaluation of therapeutic results. This requires an enormous effort from clinicians, scientists, regulators, patient advocates, drug developers and others to overcome obstacles encountered during the path of drug approval.4 Translational medicine aims at providing guidelines to increase the effectiveness by which clinical testing can be applied in go/no go decisions and, therefore, integrates innovative pharmacologic tools, biomarker identification, testing and validation, and study designs to better understand the biology of diseases and to select drug targets with greater confidence.4
Key for the successful translation of basic research to human clinical trials is the interrogation of biological samples for the identification of biomarkers. The path of translational research involves various phases, including discovery, development and application, each of which is supported by biomarkers, either for the definition of objectives, proof of concept, or risk and feasibility.5 Biomarkers typically used in translational research are classified as: 1) target biomarkers measuring the interactions between a drug and its target; 2) mechanism biomarkers that measure downstream biological effects; and 3) outcome biomarkers that reflect safety and efficacy.5 The value of these biomarkers lies primarily in the prediction of biological activity for early decision-making, thereby limiting the costly and late attrition of clinical trials.5
Challenges in Translational Research
The process of translational research is, however, arduous and prone to errors. Despite the massive investment in preclinical studies to guide the development of targeted agents, not all success at the bench is translated to success at the bedside. New therapies or interventions shown to be effective in animal studies are often found to be less effective or ineffective in clinical trials and some have been proven to be harmful to humans. In fact, only 1 in 10 drugs that start the clinical phase make it to market, in part due to significant adverse events and unforeseen toxicities. This has been attributed to fundamental differences in the genetic makeup of humans compared to the controlled genetics within inbred animals, but the reasons for these shortcomings are multifactorial and also relate to failures at both the bench and the bedside.
There are several challenges impacting the successful translation of outcomes from animal research to humans in a clinical setting. These include:
lack of solid understanding of human pathophysiology in complex, multifactorial diseases
the use of pre-clinical models that are too simplistic and poor predictors of effectiveness in humans
biological differences between species, strains and cell lines
inappropriate preclinical pharmacokinetic evaluation to guide dose selection in Phase 1 clinical trials
unrealistic approximations of drug exposure
lack of insight about the mechanism of action of compounds
poor or inadequate study designs and sample accrual, collection, storage and analysis
lack of pre-clinical study randomization and blinding
use of protocols that differ from clinical studies
insufficient reporting of details, animals, methods and materials
the exclusion of important negative or neutral data leading to publication bias
use of inappropriate statistical power and analysis to provide conclusive validation of potential biomarkers or
surrogate markers
inappropriate selection of patient populations for Phase I/II studies to successfully assess toxicity, identify optimal
biological dose, characterize kinetics and better predict the biological or clinical effectiveness of the treatment
unknown epidemiology and biologic relevance of selected biomarkers in the targeted population
inability to link clinical information from high-throughput technologies to quality sample collection
limited funding to cover the needs of biomedical researchers in the basic/clinical interface to identify clinically
useful surrogate markers and to assess the best methodology for clinical evaluation
conflict of interest
In light of these translational challenges, it is not surprising to see deviations between results obtained from clinical trials and animal studies. This has led experts to recommend the use of systematic reviews in 2002 not only as a solution to better understand how animal research data is relayed to humans, but also to: 1) expose biases and inadequate methodologies, 2) establish differences between study designs in animal and clinical studies, and 3) identify optimal animal models for evaluating treatments before testing in humans. In 2004 there was a call for a large-scale program of systematic reviews of animal studies to help improve the quality of evidence derived from animal data. In the same year, the chair of the National Institute for Clinical Excellence called for ‘‘detailed scrutiny of the totality of all available evidence’’ to assess the ‘‘real predictive power’’ of the preclinical biological studies used in the research and development process.
A systematic review of stroke research publications conducted in 2005 revealed that 36% of published studies reported randomization and only 29% of published studies were blinded.6 The studies also lacked adequate statistical analysis, and power analysis was only documented in 3% of published studies.6 These findings, combined with significant amounts of unpublished negative or neutral data, caused an overestimation of the efficacy of published positive data.
A historical overview of modern preclinical research projects citing the Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES) also provided significant insight into the basis for the lack of efficacy of drugs developed to treat neurodegenerative disease conditions, as well as many other conditions that affect the human population during the aging process.6 These findings prompted the recommendation of guidelines to improve translational research. One such guideline, the RIGOR report produced by the National Institute of Neurological Disorders and Stroke (NINDS),6 recommended that preclinical studies should incorporate the following experimental design characteristics:
sufficient rationale for the model selection and endpoint measurement
the justification of sample size, including complete power analysis calculations for primary endpoints
an adequate number of control groups using an appropriate route of administration and the incorporation of
timing of intervention delivery to reflect eventual use in the patient population (i.e.: oral, intravenous, acute,
chronic)
randomization and blinding
statistical analysis for the interpretation of results consistent with the study design
The stroke therapy academic industry roundtable (STAIR) also recommended:
efficacy in two or more laboratories
replication in a second species
consideration of sex differences
consideration of the route of administration
implementation of good laboratory practices (GLP) to:
o eliminate randomization and assessment bias
o define inclusion and exclusion criteria
o conduct full power analysis and sample size calculations o disclose potential conflict of interest
after initial evaluations in young, healthy male animals, further studies should be performed in female aged animals, and animals with co-morbidities that reproduce the pathophysiological state of patients
The Nuffield Council on Bioethics published a report in 2005 stressing the importance of more systematic reviews of animal studies to understand ethical and scientific issues in animal research.7 On the same year, the American Council on Science and Health called on US government agencies adopted new guidelines for toxicology studies to identify cancer- inducing substances, and the UK SABRE research group was established to promote the systematic review of animal studies to promote better health care and to protect patients and research volunteers from unsound research.7 In 2009, the National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs) surveyed the quality of reporting, experimental design and statistical analysis of 271 published animal studies and still found a number of deficiencies in experimental design, statistical analysis and reporting.7
Due to poor reporting and methodological inadequacies of animal studies, some systematic reviews were unable to provide reliable effect estimates to address the issue that the discordance between results obtained from animal and clinical studies resulted from biological differences. To address this issue, new guidelines were created to improve the reporting of animal data.7 Such guidelines, the Animal Research: Reporting In Vivo Experiments (ARRIVE guidelines) and the Gold Standard Publication Checklist (GSPC), were based on the Consolidated Standards of Reporting Trials (CONSORT) guidelines developed to improve the quality and transparency of clinical trial reports.6 The ARRIVE guidelines provide a checklist to guide authors preparing manuscripts for publication to optimize the information provided on the design, conduct, and analysis of the experiments.7
Recommendations for the Improvement of Translational Research
A review of 26 guidelines addressing the design and execution of preclinical animal efficacy studies to support translational research for neurological or cerebrovascular drug development was conducted by Henderson et al.8 These guidelines presented 55 experimental practices commonly recommended by preclinical researchers to evaluate the strength of preclinical data supporting clinical testing. The most recurrent recommendations were proposed as a starting point for developing preclinical guidelines for animal studies in other disease domains (Table 1).
Table 1. Recommendations for the coordination of experimental design practices across research programs.
Recommendation
Examples
Choice of sample size
Power calculation, larger sample sizes
Randomized allocation of animals to treatment
Various methods of randomization
Blinding of outcome assessment
Blinded measurement or analysis
Flow of animals through an experiment
Recording animals excluded from treatment through to analysis
Selection of appropriate control groups
Using negative, positive, concurrent, or vehicle control groups
Study of dose-response relationships
Testing above and below optimal therapeutic dose
Characterization of animal properties at baseline
Characterizing inclusion/exclusion criteria, disease severity, age, or sex
Matching model to human manifestation of the disease
Matching mechanism, chronicity, or symptoms
Treatment response along mechanistic pathway
Characterizing pathway in terms of molecular biology, histology, physiology, or behavior
Matching outcome measure to clinical setting
Using functional or non-surrogate outcome measures
Matching model to age of patients in clinical setting
Using aged or juvenile animals
Replication in different models of the same disease
Different transgenics, strains, or lesion techniques
Independent replication of studies
Different investigators or research groups
Replication in different species
Rodents and nonhuman primates
Inter-study standardization of experimental design
Coordination between independent research groups
Defining programmatic purpose of research
Study purpose is preclinical, proof of concept, or exploratory
In addition to the recommendations above, sharing data originating from preclinical animal studies has become a priority since a wealth of information can be lost when transferred from researchers to clinicians.3 In Europe, the sharing of data from vertebrate animals by companies, scientists and citizens is already mandatory and is facilitated through the European Chemical Agency (ECHA).7 In 2011, Figshare was created to allow researchers to publish research findings, including neutral results, while allowing them to maintain ownership over the data.7 In 2012, calls for opportunities to present negative data were also proposed by experts.6 Calls have been also made for the prospective registration of all animal studies. Although a registry system does not exist yet, initiatives have already begun.7 In 2012, the Dutch parliament defined a general database for the registry of animal studies to prevent unnecessary duplication and to reduce publication bias.7
Other Obstacles in Translational Research
Although the failure to get new therapies to market has been attributed primarily to flaws in the design, execution and reporting of animal studies, other barriers working against translational research have been identified. These are:
1. Complex etiology of some human diseases. During the 1st International Conference of Translational Medicine, experts suggested that hypotheses used in translational medicine should originate from the careful observation of
www.databean.com
events relevant to human diseases based on skillful studies in human subjects, at relevant time points, and where the pathophysiological process and its perturbation is evolving (from bedside-to-bench-to-bedside).9 This suggestion was based on the assumptions that: 1) there is lack the understanding of the pathophysiology of complex and multifactorial diseases; 2) scientist hypothesize on human diseases based on their comfort zone; and 3) pre-clinical models are poor predictors of effectiveness in humans because they often lack salient aspects of pathophysiology. To better understand and target human diseases, scientists should consider an array of factors such as genetics, epigenetics (environmental exposures that impact genetics positively or negatively), predisposition, co-morbidities, length of the disease, treatment history, and responses to therapies.9
2. Lack of translational research education and training. Multidisciplinary teams need to work in concert to allow the bidirectional exchange of knowledge. A main issue debated in translational research is the lack of didactic training, experiential learning opportunities, and limited opportunities of basic scientists to interact with clinicians and patient populations. To foster scientific collaborations, basic scientists should be exposed to the clinical environment while clinicians are exposed to bench research.10 Both, basic scientists and clinicians should also receive training in the areas of clinical trial design and conduct as well as in medical statistics. Institutions should incorporate didactic courses in pathobiology and pathophysiology into graduate education, provide case-based learning opportunities, and organize seminars that integrate clinical issues into the curricula of basic science programs to provide the foundations for understanding human health and disease processes.10 Institutions should also facilitate the interactions between basic scientists and clinicians by providing them access to the necessary research resources and developing funding mechanisms and strategies that encourage interdisciplinary partnerships.10
3. Need to increase the number of biobanks for human specimens. The collection of human materials has been hindered by several barriers of practical and ethical nature.4 To identify relevant biomarkers that can predict the outcome of new therapies, there is the need for unique patient samples for scientists to measure specific endpoints. Well-preserved tissue specimens from biobanks could allow researchers to link information obtained from innovative technologies, including molecular techniques, with clinical information. This could speed the discovery and development of therapies. A considerable effort is, however, needed for the standardization of protocols used for the collection of human specimens and the validation of predictive biomarkers of disease that could provide a therapeutic advantage with no preconceived bias.
Summary
The failure of investigational drugs to reach the market is expensive and could pose potential harm to clinical trial participants. This failure has been attributed to flaws in preclinical research as evidenced by the difficulty to replicate preclinical studies, publication bias, and unstandardized methodological practices. The need for more rigorous assessments of animal studies has led to the systematic review of animal data and the development of guidelines for the effective conduct of preclinical studies. Among the most common recommendations for the proper conduct of preclinical studies are the selection of proper animal models and controls, the use of adequate numbers of animals, and study randomization and blinding. Preclinical research should be able to be replicated in different animal models and species, and by different scientific groups. This requires data sharing, including negative and neutral results. A significant obstacle in translational research is impaired communication between basic researchers and clinicians. Thus, achieving success will depend upon ways to improve communication between clinicians and the animal researchers who conduct the pre- clinical work.
Although translational research faces many challenges, there are ways to optimize its outcomes. Imperative in translational research is the rigorous review of pre-clinical and clinical data to identify biomarkers and techniques for testing in appropriate human populations. This requires the ability to use strong bioinformatics systems to integrate and trace data, ensure proper data analysis, and implement procedures for effective decision-making. Databean’s experts incorporate bioinformatics, biostatistics and clinical informatics to identify scientific challenges and to develop predictive and preventive tools to be used in the design and conduct of Phase I and II clinical studies. Such predictive analysis permits the prompt anticipation of road blocks and the efficient mitigation of risk. Combined with adaptive design and data monitoring tools such as TrialPointTM, Databean identifies optimal treatments without compromising the validity and integrity of clinical studies.
Databean, LLC is a clinical research organization providing end-to-end clinical trial management services to biotechnology, biopharmaceutical and medical device companies with targeted therapeutics in the areas of immunomodulation, solid organ transplantation and class III medical devices. By combining our expert scientific knowledge and cutting edge technologies, Databean provides efficient, affordable and scalable solutions to guide life science companies through the complex regulatory and clinical processes necessary for product approval.
by Carmen Perrone, VP Scientific Affairs, Databean, LLC