Between Bench and Bedside—Which Research Manuscripts Are Truly Translational?
Article Outline
Recognizing the value of translational studies and the resources and time required to perform them, journals place a high priority on articles that make the transition from bench to bedside.
Over the past few years, there has been increased emphasis on the need to facilitate translational research. In 2004, the National Institutes of Health launched a roadmap to provide stronger infrastructure to accelerate the pace of translational research and many pages in academic journals have been devoted to debating the best ways to promote the translation of scientific discovery between bench and bedside—between the basic scientists who study disease at the cellular or molecular level and the clinical researchers who provide new insights from patient-oriented studies that can stimulate further basic investigations.
The goal of translational research is to improve health by rationally designing new therapies or diagnostic tools based on a prior understanding of the molecular and cellular mechanisms of disease pathogenesis. But what components make up the best translational papers and which manuscripts have truly made the transition from the bench to the bedside, or vice versa? It can take years to move from a basic to a clinical study, but the best translational research papers do indeed cover a broad range of this continuum.
Bedside
Definitions of translational research vary. A narrow definition encompasses studies that demonstrate the direct application of basic research to patients (with the end point of the study being investigated in the patient; eg, a clinical trial of a new therapy) or provide new insight into pathogenesis by characterizing some aspect of a disease process directly in patients. Although human studies are obviously an essential component of medical research, there are some important limitations. It can be difficult to prove the significance of proposed pathogenic mechanisms in humans; these studies are by nature often correlative and descriptive and the relevance of the observations is, therefore, uncertain. It can be a challenge to determine whether a particular abnormality is a cause of a pathogenic process or a consequence of the disease.
Given the limitations, what ingredients contribute to the strongest manuscripts reporting translational findings from human studies? Key criteria are the novelty and the importance of the question that was asked and whether there is interesting new biology presented that generates new ideas about pathogenesis. Of course, the studies should also include adequate statistical analyses, robust numbers of patients, and appropriate control groups. With respect to human studies that demonstrate the diagnostic or prognostic potential of a new biomarker for a particular disease, one should expect the robustness of the assay to be validated in an independent, blinded cohort with appropriate disease controls; authors should also demonstrate that their approach is better than existing techniques.
Clinical trials investigating a new therapeutic intervention offer more potential to directly investigate the physiologic relevance of a particular pathway. Although it is exciting to find that a particular therapy has worked, in terms of translational potential it is also important to provide some insight into the mechanism by harnessing new technology to study the cellular or molecular changes that result from the intervention. These new findings can then be used to refine the model of pathogenesis and produce new hypotheses for future testing.
The potential rewards from translational studies in humans can be enormous. Psoriasis vulgaris, for example, is a disease for which human studies have been enormously important in driving the development of new therapies.1 Aside from human psoriatic skin xenotransplanted to immunodeficient mice, there is no mouse model that faithfully recapitulates all aspects of psoriasis; psoriasis is an autoimmune skin disease and mice have a skin architecture that differs from that of humans. But studies of the cellular and molecular basis of psoriasis pathogenesis in humans have been made easier by the accessibility of the affected organ. Early histologic observations of epidermal hyperplasia in affected skin led to testing of reagents assumed to modulate keratinocyte proliferation. Advances in basic science, such as the development of monoclonal antibodies and immunohistochemical techniques, revealed an immune-mediated component to the disease—a hypothesis that was tested in clinical trials when the necessary biological reagents became available. As discussed elsewhere,1 animal models of psoriasis, particularly the xenotransplantation models, have proven helpful in supporting and generating hypotheses about pathogenic mechanisms of the disease. Such models are likely to continue to prove useful in the future; they allow potentially pathogenic processes to be manipulated in ways that might not be possible in humans. But the most useful of these animal studies have been and will be closely informed by an understanding of human pathophysiology. This is true for translational research for any human disease.
Bench
Based on the narrow definition that translational studies are conducted in patients, only a small proportion of papers published in the field of molecular medicine would be considered truly translational. Nonetheless, studies in preclinical animal models are an essential part of medical research and have provided substantial new insights into mechanisms of pathogenesis that have ultimately been translated into new therapies. Studies in cell culture systems alone have even revealed important new insights with ultimate translational potential. The successful development of several different HIV drugs that are adding decades to patients' lives, for example, has been a direct result of basic research on the molecular mechanisms of HIV replication in human cells. In vitro investigations of cellular host factors required for HIV replication are continuing to reveal potential new targets for antiviral drugs.2
Unfortunately, much of basic science does not directly translate to successful human therapy. The Th2 cytokines interleukin (IL)-4 and IL-5 are important drivers of airway inflammation and hyper-responsiveness in mice, yet biological reagents that target the IL-4 and IL-5 pathways have produced disappointing results in clinical trials for asthma.3 Similarly, hopes that a vaccine against HIV will be developed any time soon have diminished in the wake of failed efficacy trials of the most promising vaccine candidate to date4—these are just 2 of many examples of the difficulties of the bench-to-bedside translation.
Although there are many reasons for the failure of promising preclinical results to translate to humans, a major stumbling block seems to be how well the preclinical animal models replicate the human disease. Mouse physiology is different than human physiology, although the extent of these differences varies between tissues. Mouse models of bone loss have been relatively successful in identifying pathogenic processes and new therapeutic targets that have subsequently translated well to humans. A notable success story is the identification, in mice, of the necessary role of the RANK ligand (RANK-L) in osteoclast development and function. Mouse studies showed proof-of-principle that RANK-L is a therapeutic target for osteoporosis5 and a monoclonal antibody to RANK-L, denosumab, has recently completed Phase III clinical trials.
A second reason for the failure of some approaches to translate to humans is that the large degree of genetic diversity among human populations does not exist for most animal models. A particular intervention that seems to be effective in 1 strain of mouse might not work so well in a different strain, let alone the vastly heterogeneous human population.
It is important to emphasize that animal models can provide a great deal of important information about molecular mechanisms that regulate particular biological processes. They are of great use in investigating new concepts that could have broad relevance, for example, to identify functions of novel T-cell subsets such as regulatory T cells and T-helper 17 cells.6 These sorts of studies generate new ideas about biological processes that can then be examined in human disease. However, when animal models are used specifically to gain insight into pathologic mechanisms that underlie 1 particular disease, their relevance to humans must be more carefully scrutinized.
In such cases, it is important to ensure that the best animal model available is used to study a particular disease. For example, those in the sepsis field consider cecal ligation and puncture a better model of human sepsis than injection of lipopolysaccharide. When authors describe the therapeutic efficacy of blocking a particular pathway, it is useful to show that the approach is effective in more than 1 animal model of disease—this is particularly true in studies of colitis, for which several different mouse models exist. Relevance to humans should always be a consideration when modeling a specific disease; the relevance of important findings made in mouse models should be supported by parallel studies on human tissue.
When knockout or overexpression models are used to test the function of a gene or pathway that is associated with a particular disease in humans, it is important to be certain that the phenotype observed in the mouse is relevant to the human disease. If new observations are made in the mouse model, they should be taken back to the patient to evaluate whether they also hold true for the human disease. If a particular mutation is associated with a disease phenotype in humans, it is not always sufficient to investigate the effect of knocking out that gene in mice—researchers should also provide insight into the functional consequences of that particular mutation in human cells. Finally, before using an animal model to investigate a pathogenic mechanism, researchers should have solid evidence that that pathway is indeed dysregulated in human disease.
A Two-Way Street
Translational research is an iterative or cyclical process—the bedside-to-bench component is just as important as bench-to-bedside. Research from the clinic should be available to the scientists that develop animal models, so that they can improve the relevance of their model to the human condition (or at least better understand its limitations). Translational research requires expertise in many different areas, which means teamwork is required. Collaborations should be forged at the earliest stages of planning a project so that basic scientists working at the bench can plan the most relevant experiments and clinicians can harness the best and newest techniques to maximize the information they collect from their patients.
The challenge to journal editors is to publish papers that provide important new mechanistic insights that will ultimately translate to improvements in the diagnosis or treatment of human diseases. The more difficult challenge to those who care about translational research is to figure out how best to harness the expertise of basic researchers and clinicians to allow them to work in synergy.
References
- . Psoriasis: evolution of pathogenic concepts and new therapies through phases of translational research. Br J Dermatol. 2007;157:1103–1115
- . Knockdown screens to knock out HIV. Cell. 2008;135:417–420
- . The mouse trap: it still yields few answers in asthma. Am J Respir Crit Care Med. 2006;174:1171–1173
- Nonhuman primate models and the failure of the Merck HIV-1 vaccine in humans. Nat Med. 2008;14:617–621
- Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell. 1998;93:165–176
- . A brief history of Th17, the first major revision in the Th1/Th2 hypothesis of T cell-mediated tissue damage. Nat Med. 2007;13:139–145
Conflicts of interest The author discloses no conflicts.
PII: S0016-5085(09)00439-9
doi:10.1053/j.gastro.2009.03.031
© 2009 AGA Institute. Published by Elsevier Inc. All rights reserved.
