Patients have benefited greatly from the steady stream of new medical devices approved each year by the US Food and Drug Administration and by regulatory agencies in countries around the world. But the number of such devices for tissue engineering and regenerative medicine is just a small fraction of these, perhaps 7 of the 194 devices (<5%) approved by the FDA since 2008 [1]. Few new tissue engineering and regenerative medicine devices have reached the clinic despite the proliferation of biomaterial matrices/scaffolds, cells, and regulators of cell function (e.g., growth factors); i.e., the three types of tool currently available for implantation or injection into a specific type of defect to facilitate regeneration (the paradigm of regenerative medicine), or for the formation of tissues and organs in vitro for subsequent implantation (the tissue engineering plan). While the number of new tools added to the tissue engineer's toolbox each year is continuing to grow dramatically, there are few tools being implemented for the production of new medical devices undergoing human trial (figure 1). What are the bottlenecks in getting the tools out of the toolbox and into clinical use to advance patient care?
Figure 1. The many ideas generated for biomaterials (B), cells (C), and regulators of cell function (R) as tools for tissue engineering and regenerative medicine generally first undergo testing in the laboratory to determine if their physical and biological properties are suitable for the anticipated clinical application. Many of those that successfully pass this first screening may go directly into the toolbox, while others undergo further testing in an in vivo model before being accepted for the toolbox (in this schematic two of the four failed the in vivo test). It is only relatively few tools in the toolbox that are then evaluated in a definitive animal model for the targeted clinical application. Of these, certain tools may fail when tested individually (as shown by the dashed blue line), but succeed when evaluated in judicious combination (dashed ellipse) with other tools.
The supposition of this editorial is that our toolbox already likely contains the solutions to many clinical problems, but there are two specific issues contributing to the holdup in the translation of tools to clinical use: (1) the necessity of using select tools in combination; and (2) the need for the systematic evaluation of such combinations of tools in a definitive animal model (figure 1), with the attendant challenge of funding. Dealing with issues to accelerate the implementation of tools for the production of medical devices for use in the clinic is, of course, of paramount importance to the patients with a myriad of problems waiting for medical devices to prolong their lives and/or improve the quality of the life that they have. Funding agencies need to know how to deal with these issues in order to judiciously allocate their financial resources, and companies/investors would like to know for obvious reasons. And journals, too, need to know the answers to questions related to the implementation of our tools in order to inform their scope and to select the most meaningful papers for publication, to do their part in disseminating important knowledge for the good of science and health.
Today's toolbox
What does the tissue engineer's toolbox look like? Quite a bit different from the little black bag which physicians in the 19th and first half of the 20th century used to carry in hand when they visited patients in their homes; that as small as it was, could contain all of the most essential tools available at the time. The tissue engineer's toolbox (also, the 'regeneration medic's bag') needs to have three quite different compartments to be able to accommodate: (1) the pre-formed biomaterial matrices and the liquid formulations of polymers that will undergo gelation once injected into the body; (2) the many cell types available for use; and (3) the regulatory molecules and apparatus to control the behavior of cells and tissues (figure 1). Over the past two decades numerous discoveries and technologies have enabled tissue engineering and regenerative medicine, and have filled our toolbox to its brim. The technologies include methods for the: synthesis of absorbable polymeric and mineral scaffolds with a wide range of properties; synthesis of limitless quantities of regulatory proteins using recombinant and other technologies; and isolation and proliferation in vitro of differentiated cell types, and stem and progenitor cells which can be driven along specific differentiation pathways. The development of these technologies underlying the generation of new tools has been driven by biomedical scientists and engineers dedicated to searching for solutions to medical problems and stimulated by the process of discovery. Moreover, professional interests and the financial rewards attendant to the commercial development of these agents for tissue regeneration, and the support of funding agencies and desire of journals for novelty, have motivated the generation of new tools.
But the bonanza in the types and numbers of tools resulting from these discoveries and technologies has not been followed by windfalls of new medical devices introduced into the clinic. Why haven't we yet been able to identify which existing tools to use clinically? One answer, and way forward, may relate to the necessity of evaluating the efficacy of tools in combination in clinically-relevant, controlled animal studies.
Necessity (and power) of tools in combination
The hope of a tool developer is that the agent may be the solution to a clinical problem. As new tools are developed they are generally tested individually in the laboratory to determine their properties relevant to select clinical applications, and the cellular responses that they elicit in vitro (figure 1). In vivo investigations may be conducted to evaluate certain features of the tissue response, and collectively assessments can be made regarding the safety and clinical 'promise' of the tool. But relatively few tools are then tested in combination with other tools to see if the desired effect is further enhanced. One of the challenges is that the knowledge and resources necessary for the handling and evaluation of the three types of tools can be quite different, necessitating a dedicated effort of the tool developer to venture out of his/her own area and to obtain collaborative assistance.
For the treatment of some medical problems a single tool can have a profoundly beneficial effect. But for most applications is it not more likely that a combination of tools will be necessary to achieve the desired outcome? Regeneration which occurs spontaneously in the body (e.g., bone regeneration) is due to an exquisite choreography of many extracellular matrix and regulatory molecules and cell types. One exogenous therapeutic agent may initiate a cascade of processes involving many endogenous processes, but it is more likely that several tools will have to be delivered in conjunction in some fashion to achieve a meaningful degree of regeneration or improved reparative response. One notable example is the FDA-approved medical device, Medtronic's Infuse Bone Graft ®, which is a combination of bone morphogenetic protein (BMP)-2, an osteoinductive regulator of cell function, in an absorbable collagen sponge (ACS). Either tool (i.e., BMP-2 or ACS) employed alone has little effectiveness in stimulating an osteogenic response, but in combination they are a potent product to stimulate bone formation, even at ectopic sites [2]. A second example is a scaffold implanted or injected into a cavitary defect in order to provide an exogenous matrix permissive of endogenous cell migration into the matrix-filled defect. Adding a chemoattractant to the matrix would likely recruit a greater number of the target cells into the defect. For example, incorporation of nanoparticles delivering stromal cell-derived factor (SDF)-1α, a known chemoattractant for endogenous neural stem and progenitor cells in the brain, was found to increase substantially the number of such cells migrating into matrices in vitro [3]. Many other examples relate to cell therapies, which normally involve a bolus injection of an aqueous suspension of cells. Incorporation of the cells into a pre-formed or injectable matrix which also delivers certain regulatory molecules could serve to retain the cells at the desired site and provide them with the necessary cues to regulate their function. For example, in a rat model, subretinal injection of retinal stem-progenitor cells in hyaluronic acid-methylcellulose [4] and in hyaluronic acid [5] matrices resulted in a more even distribution of the cells than could be achieved with injection of the cells in a saline solution, and was permissive of differentiation of the progenitor cells into photoreceptor cells [5].
Failure of a tool to live up to expectations may only signal that it requires certain adjunctive agents to achieve clinical effectiveness.
Systematic evaluation of tools in the most clinically-relevant animal model
How to definitively test the tools already in the toolbox, whether alone or in combination? One straightforward answer could be the implementation of a standardized animal model for a specific clinical problem for the systematic testing of various tools. While many of the tools in our toolbox have been tested in vivo , relatively few have been evaluated in an animal model that is recognized as being the most meaningful for a specific disorder. For example, there are rat models that have been employed for the pre-clinical testing of agents for the treatment of some of the most complex and complicated of problems, the central nervous system disorders of stroke [6, 7], spinal cord injury [8, 9], and retinal disease [10]. Agents that have yielded positive results in certain of these rat models have proceeded directly to human trial, reflecting the relevance of the models. Evaluation of various tools in the same animal model for comparison with relevant controls and with each other would: identify the products in our toolbox ready for human trial; guide the selection of other tools to test; and inform the development of new tools. What then are the obstacles to employing an animal model to test various combinations of tools in a systematic fashion?
One obstacle, of course, is that validated animal models do not yet exist for many medical problems. That the absence of a definitive animal model is the bottleneck for the translation of tools from the toolbox to the clinic, further underscores the importance of animal model development. But even when accepted animal models are available, an obstacle to the systematic evaluation of combinations of tools is that not all tool developers have the background to conduct the animal work, and the groups that routinely employ the animal model or those that have the capability to do so, may not be engaged in work dealing with the local delivery of therapeutic agents, and may not be in range of collaborators who know how to handle such tools. And even when the groups with the necessary skills get together, it takes time and effort for each to climb their respective learning curves, to validate the model in their hands, and to learn to work together.
Another obstacle to the systematic evaluation of tools, likely to be encountered, is funding. More than ever, most federal agencies prize novelty of the project as the principal criterion for funding, based on the understandable (but as we see here, uncertain) supposition that if a solution to a clinical problem had already been found (i.e., if a useful tool already existed) it would have been implemented; therefore, the need to conceive and investigate new tools. This obstacle to funding also generally applies to foundations even though they may be heavily supported by the patients who likely would be inclined to explore options currently in the toolbox in parallel to the continued search for new tools. And this approach of funding the testing of existing tools is not likely to appeal to commercial concerns whose basis for the evaluation of a tool is principally and understandably whether it is their intellectual property.
Finally, still another obstacle to the comparative testing of the tools already in our toolbox is finding a journal to publish such studies unless there is a notable success, which may take many trials to achieve. The challenge here, of course, is to formulate the study at the outset with working hypotheses or research questions and outcome metrics that are meaningful even if the principal goal of a successful treatment is not achieved. But there is no denying that that reports of the systematic evaluation of existing agents for the treatment of a clinical problem, no matter how profound, may be perceived as uninspired by reviewers.
Getting tools out of the toolbox and into patients
So how do we answer the patient who grabs our toolbox and entreats: isn't there already something in here that can help me? Shouldn't the simple answer be: we are currently conducting such tests. We do need more tools, and they are sure to come with the new knowledge being acquired from ongoing scientific investigations, but we also need to test the tools that we already have, in standardized animal models which have some of the features of the human condition to be treated. There is generally enough knowledge about how the tools will function and about past failures to make informed and rational decisions as to which combinations of tools to test. Perhaps an increasing awareness of the current situation by funding agencies and pressure from patient-advocacy groups will elevate this work to a higher priority. As regards publication of results of such testing, this journal has always appreciated the importance of in vivo evaluation of biomaterial matrices, alone or in combination with the other tools of tissue engineering, and will continue to see this work as compelling for dissemination to the scientific and clinical communities.
Perhaps this assessment of the current state of affairs will make the case for a 'tissue engineer's toolbox manifesto ': the systematic evaluation of combinations of existing tools in the definitive animal model by a consortium of laboratories? Don't we owe it to the patients who are counting on us to overcome these obstacles to execute this manifesto?
Acknowledgments
The author gratefully acknowledges the important contributions of Jonathan M Spector, MD, and Teck Chuan Lim, BS.
References
[1] www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/DeviceApprovalsandClearances/Recently-ApprovedDevices/default.htm
[2] Hsu H P, Zanella J M, Peckham S M and Spector M 2006 Comparing ectopic bone growth induced by rhBMP-2 on an absorbable collagen sponge in rat and rabbit models J. Orthopaedic Res.24 1660–9
[3] Lim T C, Rokkappanavar S, Toh W S, Wang L S, Kurisawa M and Spector M 2013 Chemotactic recruitment of adult neural progenitor cells into multifunctional hydrogels providing sustained SDF-1alpha release and compatible structural support Faseb J.27 1023–33
[4] Ballios B G, Cooke M J, van der Kooy D and Shoichet M S 2010 A hydrogel-based stem cell delivery system to treat retinal degenerative diseases Biomaterials31 2555–64
[5] Liu Y, Wang R, Zarembinski T I, Doty N, Jiang C, Regatieri C, Zhang X and Young M J 2013 The application of hyaluronic acid hydrogels to retinal progenitor cell transplantation Tissue Eng. A19 135–42
[6] Popp A, Jaenisch N, Witte O W and Frahm C 2009 Identification of ischemic regions in a rat model of stroke PLoS One4 e4764
[7] Otero L, Zurita M, Bonilla C, Rico M A, Aguayo C, Rodriguez A and Vaquero J 2012 Endogenous neurogenesis after intracerebral hemorrhage Histol. Histopathol.27 303–15
[8] Filli L and Schwab M E 2012 The rocky road to translation in spinal cord repair Ann. Neurol.72 491–501
[9] Metz G A, Curt A, van de Meent H, Klusman I, Schwab M E and Dietz V 2000 Validation of the weight-drop contusion model in rats: a comparative study of human spinal cord injury J. Neurotrauma17 1–17
[10] Lu B, Malcuit C, Wang S, Girman S, Francis P, Lemieux L, Lanza R and Lund R 2009 Long-term safety and function of RPE from human embryonic stem cells in preclinical models of macular degeneration Stem Cells27 2126–35