Proceedings of a Workshop
Microphysiological Systems: Bridging Human and Animal Research
Proceedings of a Workshop—in Brief
The National Academies of Sciences, Engineering, and Medicine (the National Academies) appointed a planning committee to convene a workshop on advances in microphysiological systems (MPS). MPS are complex, multi-cellular in vitro systems that commonly include three-dimensional (3D) aspects, fluid flow, changing pressure or stretch, and multi-organ interactions.1 According to the U.S. Food and Drug Administration (FDA), MPS model “functional features of a specific tissue or organ of human or animal origin by exposing cells to a microenvironment that mimics the physiological aspects important for their function or pathophysiological condition.”2 These systems are being developed to better mimic some aspects of specific organ systems or combinations of organ systems to improve upon standard two-dimensional (2D) cell systems, with the goal of eventually replacing animal models being used for hazard identification, risk assessment, and disease modeling, among other uses. The purpose of this workshop was to discuss current progress in developing MPS that realistically model in vivo animal and human physiology and to strategize about the potential to establish sustainable human and animal MPS banks.3 Speakers discussed how MPS fit within the portfolio of tools used in their fields of expertise, the limitations and areas of needed improvement for MPS, and how MPS may be used in the future as the technology develops. Presentations covered the following topics:
- MPS applications in drug discovery and development, environmental hazard assessment, and basic biological research;
- strategies for integrating MPS into regulatory decision making;
- opportunities for bridging human and animal studies using MPS;
- approaches for using in vivo and in silico model systems to optimize or augment MPS;
- proposals for standardizing and sharing MPS data more broadly and across sectors;
- methods for leveraging MPS to address zoonoses and fight viral pandemics; and
- state-of-the-art approaches to create multi-organ MPS to recapitulate the complexity of live human and animal biology in vitro.
The following summaries of the discussion represent individual perspectives and information and are not representative of all presenters’ views or the National Academies.
SESSION 1: WELCOME REMARKS
A panel of representatives from FDA, the National Institutes of Health’s (NIH’s) National Center for Advancing Translational Sciences (NCATS), the U.S. Environmental Protection Agency (EPA), and the biopharmaceutical industry provided introductory remarks about the potential uses of MPS and their role in various scientific fields. Topics included regulatory definitions and approvals of MPS and alternative methods, protections for animal welfare, the importance of MPS contexts
of use, the potential ability of MPS to improve the predictivity of translational assays and safety of pharmaceuticals, and the ethically and environmentally conscious aim of reducing drug testing in animals. Rear Admiral Denise Hinton from FDA stressed that researchers who plan to use cutting-edge technologies such as MPS in preparing regulatory submissions should engage FDA early in the process (e.g., via the FDA Alternative Methods Working Group) to ensure alignment with regulatory standards and expectations. Rear Admiral Estella Jones, also from FDA, detailed the potential role of MPS in FDA’s efforts—in connection with the One Health Initiative—to develop medical countermeasures such as biologics, drugs, vaccines, and personal protective equipment to counter environmental and zoonotic threats to humans, such as the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, as well as to protect animal health. The ability for pathogens to “jump” from wildlife to livestock to humans places pressure on the global scientific community to develop MPS banks for species susceptible to zoonotic organisms, which can be used for proactively testing and predicting the transmissibility of an organism from species to species. Such MPS banks would be subject to all federal animal welfare rules, including FDA regulatory standards for tissue banks.
Dr. Ekaterina Breous-Nystrom from Roche highlighted the potential for MPS to improve preclinical safety testing of experimental therapies, which may reduce the proportion of failed clinical trials, thus enabling drug developers and regulators to focus more of their resources and attention on the most promising therapies. Especially for complex biologics, differences between human and animal biology may limit the predictivity of nonhuman animal models. Companies such as Roche envision producing personalized MPS for individual patients, thus applying the principles of personalized medicine to preclinical toxicology testing.
Expanding on this theme, Dr. Christopher Austin from NCATS stressed the potential for MPS to accelerate the successful translation of advances in biomedical science into real-world improvements in clinical outcomes for patients. A major hurdle to improve the efficiency or effectiveness of translation to humans is the limited capacity of traditional testing systems to predict adverse or beneficial effects of chemicals and other agents to which humans are exposed, either as pharmaceuticals or by environmental exposure. Dr. Austin highlighted various NCATS programs that focus on solving this challenge through innovative testing platforms, such as the Toxicology in the 21st Century Program, induced pluripotent stem cell (iPSC)derived assays, organoid/spheroid systems, and microfluidic MPS. He noted that expanded development and the use of nonhuman MPS could potentially help to replace, reduce, and refine animal use in research; evaluate environmental chemical effects on wildlife species; advance veterinary research; and bridge animal and human data, which could improve the translational, regulatory, and drug development applications of MPS data. As Dr. Russell Thomas from EPA explained, the development and application of in vitro approaches play key roles in EPA’s strategy to eliminate mammalian testing by 2035, while generating informative data on environmental toxicants and chemical hazards (e.g., produced by industrial activity) that could have implications for human health.
SESSION 2: MPS FOR TOXICOLOGY TESTING
Among the most promising applications of MPS is their use for toxicology testing. Dr. Amy Avila from FDA’s Center for Drug Evaluation and Research (CDER) noted that FDA considers MPS to be a new approach methodology (NAM), which refers to a broad range of methods (e.g., in vitro, in chemico, in silico) that potentially improve predictivity of preclinical toxicology testing and contribute to the replacement, reduction, and refinement of animal use in development programs. Although MPS have not yet been used in any regulatory submissions for human pharmaceuticals, CDER has created a distinctive pathway for investigators to submit NAM-related information for review and comment and is actively exploring ways to incorporate MPS data into FDA’s regulatory decision-making framework as a tool for toxicology testing of human pharmaceuticals.4
Dr. David Strauss from CDER noted that full characterization and qualification of MPS for specific contexts of use may enable researchers and regulators to integrate MPS into a drug development paradigm that relies less on animal testing but improves the preclinical predictivity of drug safety, clinical pharmacology, and efficacy. CDER’s Division of Applied Regulatory Science plans to extend currently applied research on liver and heart MPS to other organ systems, with a focus on the performance of MPS compared to other culture systems, reproducibility of MPS within and between laboratories, and development of quality control criteria to ensure proper assembly and preparation of functional systems. Dr. Ivan Rusyn from Texas A&M University, a leader of the public–private partnership characterizing MPS called the TEX-VAL Consortium, highlighted the Consortium’s multi-tiered system for developing, validating, and defining contexts of use for MPS. The consortium aims to foster broader understanding and adoption of these systems, facilitate their integration into the regulatory decision-making process, and determine how and when MPS could complement or replace, to the extent possible, animal model usage.
The applications of MPS to toxicology testing extend beyond the field of drug development. Dr. Raja Settivari from Corteva Agriscience described how regulators worldwide require that agrochemical active ingredients and final products undergo extensive toxicology testing and risk assessment, including for long-term exposure. Recent advances in fields such as MPS suggest the potential to replace in vivo animal models with inanimate systems for agrochemical toxicology testing. Proof-of-concept studies have, for example, shown that MPS may aid in enabling early-stage screening of compounds, potentially substituting in vivo studies for some endpoints while still allowing product developers to select the best compounds from a group of chemical analogs. MPS have also demonstrated applicability to hazard assessments (e.g., to evaluate the time course of toxicity response and recovery across various tissue systems, and to improve the understanding of specific biological targets relevant to observed toxicities) and toxicokinetic assessments.
Biopharmaceutical companies have also used in vitro assays to assess patients’ risk of developing conditions involving electrophysiologic toxicity, such as Torsades de Pointes (TdP), a potentially fatal arrhythmia resulting from an irregular or delayed pattern of repolarization in heartbeat-controlling neurons. Dr. Gary Gintant from AbbVie presented the Comprehensive in vitro Proarrhythmia Assay (CiPA) initiative, which conducts in vitro studies in human iPSC-derived myocytes to corroborate in silico modeling of cellular electrophysiologic response, and thus to predict patients’ risk of developing arrhythmias such as TdP. Dr. Gintant highlighted several lessons learned from CiPA that are relevant for work with MPS, including the value of maintaining controlled culture conditions to reduce biological variability and promote reproducibility and the value of standardizing protocols to minimize variability, enhance statistical power, increase reproducibility, and promote assay use and adoption.
Several key themes emerged from discussions among Session 2 panelists, including the futility of establishing a standardized framework for the use of MPS in toxicology assessment, given that such frameworks depend on the specific context of use (COU) and the availability of data for specific endpoints of interest (e.g., using MPS to evaluate endothelial cell dysfunction may involve considerations that differ from those involved in evaluating vasodilation and vasoconstriction). Some panelists emphasized that different research questions and tissue systems may require varied testing approaches to gain regulatory approval for compounds of interest, and FDA representatives noted that it must be demonstrated that NAMs are at least as valid and reliable as the existing standard methods or assays before they are approved. Investigators who seek incorporation of specific NAMs into FDA’s regulatory decision making are generally encouraged to proactively engage with FDA (e.g., via pre-IND [investigational new drug] meetings) and to include data from MPS or other NAMs of interest within existing data submission packages (e.g., INDs or new drug applications) to demonstrate value.
SESSION 3: GOING FROM IN VITRO TO IN SILICO—DATA AND DEVELOPMENTAL TOOLS
Researchers are using combinations of in vitro models, including MPS, and in silico approaches, such as quantitative systems pharmacology (QSP), to model disease progression and predict drugs or drug combinations with a high probability of treating various diseases, including some that exhibit extreme heterogeneity, such as metabolic dysfunction–associated fatty liver disease (MAFLD). Dr. Lansing Taylor from the University of Pittsburgh developed an approach to studying MAFLD that combines human liver cell–based MPS—which can be experimentally manipulated to simulate stages of MAFLD progression—and QSP. This mixed approach enables investigators to model MAFLD heterogeneity and predict drugs and drug combinations that target mechanisms underlying MAFLD subtypes, and consequently halt or reverse disease progression. Specifically, researchers apply QSP computational analyses to patient liver biopsy samples, from which information is derived on MAFLD’s heterogeneity, mechanisms, and progression. Researchers then use MPS to model MAFLD subtypes experimentally, enabling in vitro testing of promising drugs. This iterative, in silico, in vitro, target-centric approach to drug discovery and validation holds promise for identifying effective drug candidates against MAFLD and a variety of other heterogeneous diseases that cannot be adequately modeled by current in vitro and animal model systems. Next-generation human biomimetic liver MPS will be based on patient-derived organoids and bioprinted (2D or 3D ink-jet printing of fluids containing cells into desired cell-containing shapes) iPSCs and will further expand the combined power of in silico, in vitro, and MPS testing platforms to advance the aim of replacing, reducing, and refining the use of living animal models.
Combined in silico and in vitro MPS approaches—such as synthetic microsystems, artificial intelligence (AI), and artificial life models—are also being used to supplement in vivo animal models to study prenatal hazardous drug and chemical exposures during pregnancy, with the goal of replacing in vivo animal models. Dr. Thomas Knudson from EPA and his collaborators at the University of Michigan have used in vitro MPS to model cellular dynamics of human epiblasts during gastrulation and to study the impacts of various chemicals or growth factors (e.g., BMP4) on these dynamics, and then develop in silico models to simulate properties governing cell fate and behavior. These in silico AI models, in turn, translate specific genetic variants or biomolecular lesions into probabilistic renderings of adverse developmental outcomes. Similar in silico platforms are used to model the spatial dynamics of mesodermal formation during gastrulation (e.g., the formation of the “primitive streak,” which is a morphological hallmark of normal gastrulation), generating predictions that can then be
tested with in vitro MPS that use human-induced pluripotent stem cells (hPSCs). These approaches are being used to study human gastrulation and developmental processes and toxicities—without the use of in vivo animal models. Dr. Knudson stressed that in silico toxicodynamic models enable researchers to translate data on chemical effects from hPSC studies (e.g., from EPA’s Toxicity Forecaster [ToxCast]) into testable hypotheses about toxicological processes that occur within the intact human embryo.
Computational modeling of MPS data is also supported by the Microphysiology Systems Database (MPS-Db), a unique resource that supports the development and widespread implementation of MPS in basic biological research, drug discovery, and safety and efficacy testing. Dr. Mark Schurdak from the University of Pittsburgh explained that the MPS-Db allows aggregation, analyses, and computational modeling of experimental MPS data and serves as a centralized resource for sharing that data. Researchers use the MPS-Db to aggregate data from a wide variety of MPS models (e.g., animal and human cell–based models, individual organ models, multi-organ microfluidic models, and static microplate models), visualize and statistically analyze MPS study data, and assess MPS models’ characteristics. Users can also apply computational tools to probe the MPS datasets, allowing them to study mechanisms of disease and compound toxicities and to predict safety, efficacy, and pharmacokinetics behavior of specific compounds. The database further includes manually curated clinical and preclinical data that can be incorporated into analyses. The nonprofit version of the MPS-Db is being adapted into a private version to incorporate company firewalls and security standards. The MPS-Db helps ensure that MPS are fully leveraged (including the recording of data from safety studies with MPS) to advance the study of human disease and treatment and the safety and pharmacokinetics of compounds and drugs, while replacing, reducing, and refining the use of living animal models.
Within the Session 3 question and answer period, the use and advantages of iPSCs within MPS were discussed. As MPS continue to improve, researchers may be able to create MPS from fully matured iPSCs derived from individual patients. This approach would lead to MPS that reflect individual factors such as genetic variants and microbiome characteristics of individual patients, which may personalize the study of disease and treatment. Several panelists also emphasized the importance of developing MPS derived from animal tissues to interpret existing animal model data, to assess the ability of tissue chips to recapitulate intact in vivo organ function, and to study animal health and disease.
SESSION 4: PANEL DISCUSSION ON TOPICS OF PUBLIC HEALTH IMPORTANCE (SARS-CoV-2)
Amid the SARS-CoV-2 pandemic, zoonotic diseases have become an especially pressing area of interest, and a global community of scientists studying zoonoses has developed to further collaborations and knowledge sharing. Dr. Simon Funnell from Public Health England noted that the World Health Organization Research and Development Blueprint team helped to facilitate such collaborations, which together developed and refined MPS biological infection models to enable the development and assessment of drugs and vaccines—including drug repurposing. In this context, MPS may provide key advantages over traditional cell culture and live animal model systems, including better simulation of human in vivo infection. MPS thus represent a key opportunity to improve understanding of zoonoses, develop treatments and vaccines to combat them (and hence limit the impact of viruses with pandemic potential), and replace, reduce, and refine the use of in vivo animal models. Dr. Funnell explained that although producing highly differentiated tissues in the laboratory could previously only be accomplished on a small scale, new advances in MPS such as MucilAir™ and the Human Emulation System® have enabled these kinds of complex applications (e.g., development of MPS for pseudostratified ciliated columnar epithelial tissue) to be significantly scaled up. MPS of human airway and epithelial tissue have also been used, respectively, to model SARS-CoV-2 aerosol transmission and viremia and immune response. Eventually, MPS could potentially replace certain kinds of in vivo animal model studies currently considered indispensable in translational research pipelines.
Dr. Diane Bimczok from Montana State University presented on how bat and human gastrointestinal organoids are effective in vitro models to define pathogenic and protective responses to infection with SARS-CoV-2. Based on the observation that SARS-CoV-2 infection of bats’ gastrointestinal tracts does not result in pathology or disease (as it does in humans), Dr. Bimczok and colleagues hypothesized that antiviral immune mechanisms in the bat’s gastrointestinal tract may lead to a novel treatment approach to prevent SARS-CoV-2’s cytopathic effects in human respiratory and gastrointestinal epithelial cells. The lack of an appropriate in vitro system to test this hypothesis led Dr. Bimczok’s team to use cell lines extracted from the guts of Jamaican fruit bats to establish and characterize organoids that enable modeling of SARS-CoV-2 infection dynamics, including cell entry via ACE2 receptors and viral evolution in vivo. After these organoid MPS are fully established, the team will compare the responses of bat and human gut organoids to SARS-CoV-2 in order to understand the cellular mechanisms that allow bats to harbor coronaviruses without developing clinical diseases. This project shows how MPS can assist in the study of basic biological mechanisms underpinning health and disease in both humans and animals and can also support the investigation of novel treatment approaches for various diseases, including SARS-CoV-2 and other zoonotic viruses.
Similarly, Dr. Vivek Thacker from the École Polytechnique Fédérale de Lausanne described his laboratory’s work in developing novel, vascularized lung-on-a-chip MPS to study SARS-CoV-2 and tuberculosis infections. This MPS approach supports the study of disease characteristics that have not previously been modeled by traditional in vitro or in vivo model systems, but that are hypothesized to significantly impact the course of SARS-CoV-2 and tuberculosis infections. MPS studies of SARS-CoV-2, for example, have revealed cell-specific, endothelial inflammatory and antiviral activity, which suggests a key role for cell–cell communication in early-stage SARS-CoV-2 infection that non-MPS models have not been able to recreate. Similarly, MPS studies of tuberculosis have shown that heightened pulmonary surfactant levels inhibit bacterial growth in epithelial cells and macrophages. These results could not have been achieved using traditional monocultures, which cannot adequately recapitulate vascular physiology, or in traditional animal models, yet they indicate possible avenues to intervene early in SARS-CoV-2 and tuberculosis lung infections.
Dr. Don Ingber from the Wyss Institute for Biologically Inspired Engineering at Harvard University (Wyss Institute) presented on how human lung airway chips (microfluidic culture devices) have been used to test prophylactics and therapeutics against SARS-CoV-2 and other viruses (e.g., novel strains of influenza) with pandemic potential. These MPS support robust modeling of virus entry, replication, strain-dependent virulence, and host cytokine production, as well as the recruitment of circulating immune cells in response to infection by respiratory viruses. They have been used to screen drug combinations with strong scientific rationales against SARS-CoV-2 and to potentially discover new pathways of protein expression that result in the broad inhibition of viruses such as SARS-CoV-2, Middle East respiratory syndrome coronavirus, and influenza A. These results provide evidence for MPS for this specific COU as a more physiologically relevant preclinical platform for drug repurposing and discovery than more traditional in vitro models and indicate that MPS may become the preferred systems for preclinical study of human and animal organ-level biology, as well as of disease processes, prevention, and treatments. To rapidly address future viral pandemics, MPS will likely play a fundamental role in repurposing combinations of existing drugs, just as they have indicated a potential for the antiviral amodiaquine to combat lung infections of SARS-CoV-2.
Several participants noted the utility of designing MPS for specific COUs, and Dr. Ingber stated that tissue chips could be designed, for example, to study the long-term impacts of SARS-CoV-2 on lung epithelial tissue. Dr. Funnell also highlighted the potential to save time and resources by constructing MPS from specific organs obtained from model animals. This approach could enable researchers to study the effects of SARS-CoV-2 infection on specific organs without conducting many expensive and laborious animal model studies. Researchers could also use MPS of different animal species—perhaps derived from species-specific iPSCs—to gauge whether different animal tissues select for different variants of a pathogen of interest. MPS may also potentially contribute to replacing, reducing, and refining the use of living animal models. Such animal iPSC-derived MPS could also be used to support breakthroughs in animal health in connection with the One Health Initiative. One challenge in the use of iPSCs is that they sometimes fail to fully differentiate; MPS, by recreating a cell’s native microenvironment, can help to promote differentiation and thus may provide an ideal complement to develop more versatile animal and human iPSC models. Another avenue for further research is the potential for MPS to improve in vitro representation of the genetic diversity of patient populations.
SESSION 5: COMMERCIALIZATION, ENGINEERING, AND COLLABORATIONS
Pioneering MPS research has been conducted aboard the International Space Station (ISS). Tissue Chips in Space, a collaborative initiative between NCATS and the Center for the Advancement of Science in Space in partnership with the National Aeronautics and Space Administration, has delivered human kidney tissue chips to the ISS in order to study the impact of microgravity on human biology, with the ultimate goals of both improving human health on Earth and ensuring the health of astronauts on long-distance space missions (e.g., to Mars). Dr. Thomas Neumann from Nortis presented on a series of 2019 projects that shipped Nortis-manufactured human tissue chips to the ISS to study molecular pathways underlying critical kidney conditions that typically occur with aging and are accelerated by space travel. These projects studied the role of proteinuria in the progression of chronic kidney disease, the effects of reduced vitamin D bioactivation on bone loss in space, and the formation of kidney stones. The tissue chip systems achieved their functionality by replicating kidney tubule polarity and function in vitro (e.g., via specialized renal transporters). Tissue Chips in Space has emphasized the potential value of plug-and-play MPS that can be quickly and easily established for use in diverse conditions, especially with limited personnel, equipment, and time. It has also highlighted the need for robust performance, user friendliness, standardization, and stable quality in designing and developing compact, automated, and high-performing, next-generation MPS platforms for use in pharmaceutical drug screening and studies of human and animal disease. Dr. Neumann stressed that MPS developers should strive for consistent performance and highly reproducible structure, function, and output.
Use of MPS has also expanded into the contract research organization industry. Dr. Clive Roper from Charles River Laboratories presented on applications of 3D tissue models for respiratory toxicology and efficacy testing. Airway models
present opportunities for toxicology testing because many drugs, environmental pollutants, and pathogens enter the body via airways. Because most preclinical toxicology and efficacy tests have traditionally used rat models, transitioning to MPS may require rigorous comparison of the ability of human MPS and in vivo rat models to predict human toxicology and efficacy. Also important is the assessment of the ability of rat-based MPS to predict in vivo rat toxicology and efficacy. These results may help shed light on the possibility that analogous human tissue MPS may predict both safety and efficacy in human patients, which would be useful in the context of regulatory toxicology and efficacy testing. Charles River Laboratories is currently conducting such studies in MPS airway models, comparing a rat EpiAirway™ model to healthy MucilAir™ and diseased OncoCilAir™ human models. Dr. Roper noted that validation studies comparing human in vitro and animal in vivo data often fail and suggested that bridging these platforms with an animal in vitro comparator could be useful. He also noted that MPS have been used to screen out toxicants before assessment of toxicology in vivo, identify potential candidates for drug development following efficacy testing, generate human-equivalent concentrations in studies of occupational toxicology, and identify no-effect levels. For compounds that undergo in vivo testing, MPS can also potentially be used to inform dose range finding.
Dr. Michael Moore from Tulane University presented on another 3D in vitro tissue model system that he and his team developed to recapitulate rats’ sensory peripheral pain circuitry. Dr. Moore harvested rats’ dorsal root ganglion and spinal cord tissues and used them to grow separate spinal cord spheroids containing functioning synapses, which could potentially be used to test novel analgesics. Importantly, each analgesic tested perturbed the circuit’s firing patterns in unique dose-dependent fashion and produced quantifiable signatures. This MPS system recapitulates critical aspects of in vivo physiology and cytoarchitecture, such as the aspects involved in afferent nociceptive signaling from the periphery to the spinal cord, and may soon provide a viable alternative to behavioral testing in in vivo animal models for early-stage investigation of analgesics—at a lower cost and with higher throughput. Similar MPS are being built using human iPSCs to extend these capabilities for the study of candidate analgesics in human tissues in vitro. Expanding the use of these MPS for preclinical testing of candidate analgesics could potentially accelerate the replacement of opioids as a standard of care for treating chronic pain and could potentially thereby play a role in resolving the twin public health crises of chronic pain and opioid overuse and addiction.
Panelists noted that they obtain tissues for their MPS from various sources, depending in part on whether the MPS are based on animal or human tissues. Regardless of tissue source, many of the panelists called for rigorous and consistent quality control procedures. Some commercial MPS developers have sought to establish their own internal tissue banks (e.g., of human kidney tissues) that undergo rigorous characterization and standard preprocessing and fully characterized “immortalized” cell lines for use in MPS (e.g., organoids). Panelists acknowledged the desire of some researchers to establish universal cell banks that a wide range of investigators can access, but suggested that, because cells are highly dynamic entities, establishing a universal, fully characterized and fully quality controlled cell bank may present practical challenges. For example, one vendor unexpectedly discovered that it could not ship its rat and human epithelial cell MPS in the same manner, which had significant unforeseen consequences for a project being conducted by Charles River Laboratories; any seemingly minor practical challenge could become a similarly unforeseen yet consequential issue in establishing universal MPS banks or cell banks. Panelists also noted that commercializing animal tissue MPS for veterinary health applications could contribute to the One Health Initiative’s animal health objectives. Panelists also discussed the importance of considering the COU and the research question before selecting or developing a model suitable for answering that question—in some cases, research questions may not be appropriately addressed by any MPS.
SESSION 6: MULTI-ORGAN CHIPS AND EMERGING APPLICATIONS FOR BIOLOGICS STUDIES AND INTEGRATED MULTI-ORGAN SYSTEMS
MPS applications have extended beyond modeling single organs, as multi-organ chips have been applied to modeling both animal and human in vivo systems. However, industry adoption of MPS and associated assays has been challenged by these systems’ high level of complexity. Dr. Uwe Marx from TissUse® presented on how TissUse®’s HUMIMIC® microfluidic platform has modeled interactions among various human organ models in up to four-organ arrangements. Sixteen different single-organ equivalents have been established on this platform, and 12 organ combinations have been tested for their ability to model in vivo organ-to-organ crosstalk (e.g., thyroid–liver model). TissUse® seeks to translate a HUMIMIC®-based platform that combines liver, kidney, and intestine equivalents—and that contains an extra culture compartment for another organ model—into a commercially viable tool for absorption, distribution, metabolism, and excretion profiling. To support the long-term performance and industry acceptance of such complex MPS, substantial challenges regarding design criteria, tissue supply, and qualification standards would need to be overcome. Both single- and multi-organ MPS can provide qualified COU-specific assays for hazard identification and for safety and efficacy testing, although development of MPS to reach these goals will often be iterative and conditional to the COU.
When developing multi-organ tissue chips, researchers typically seek to balance the competing demands of establishing physiological communication between the different tissues while preserving their individual phenotypes. This challenge becomes especially acute when attempting to model whole-body physiology and systemic diseases. Dr. Gordana Vunjak-Novakovic from Columbia University explained that one approach to meeting each of these conflicting requirements is to establish a modular, configurable, multi-organ platform in which each human tissue is cultured in its own optimized environment and separated from the system’s recirculating flow by a selectively permeable endothelial barrier. Under these conditions, the tissues linked by vascular perfusion can maintain stable molecular, structural, and functional phenotypes for at least 4 weeks. Furthermore, because tissues and endothelial and circulating cells in these models can be derived from individual patients’ iPSCs, these MPS can potentially support the generation of individualized models. Modularization of these multi-organ MPS enables researchers to select the modules relevant to their research questions and adopt a plug-and-play approach. The resulting multi-tissue platforms can potentially serve as high-fidelity models for studies of development, regeneration, and organ-specific and systemic diseases. Key challenges include establishing and maintaining mature tissue phenotypes for longer duration studies and determining how simple or complex multi-tissue models must be to adequately mimic in vivo human tissue systems.
Multi-organ 3D microfluidic systems hold specific promise for modeling human chronic inflammatory diseases, which affect multiple organ systems and remain among the most difficult drug development targets, in part because they are challenging to mimic using in vivo animal models. Dr. Linda Griffith from the Massachusetts Institute of Technology presented on how she and her team have developed a multi-organ platform (i.e., a three-organ “interactome”) that is outfitted with tunable microfluidic pumps to model the human gut–liver axis in inflammatory bowel disease and the gut–liver–central nervous system axis in Parkinson’s disease. These MPS generate key multi-omic and systems biology data that currently complement animal study data, but that could ultimately surpass animal studies to become robust, in vitro efficacy models capable of testing complex biologics against chronic inflammatory diseases such as endometriosis and adenomyosis. Mouse molecular data have also been used to train computational models on known mouse phenotypes so that those computational models, when fed human molecular data, can potentially predict human phenotypes without the need for animal MPS.
Researchers have also used multi-tissue microfluidic MPS to model human reproductive diseases, such as polycystic ovary syndrome (PCOS), a complex endocrine disorder that presents with a broad range of phenotypes and often involves metabolic dysfunction. Dr. Joanna Burdette from the University of Illinois Chicago noted that complex multi-organ diseases such as PCOS are difficult to study in patients and animal models. She presented on how MPS are now being used with reproductive organ cultures (on a platform called Lattice) that contain a range of both reproductive and nonreproductive tissues, including the ovary, fallopian tube, pancreas, and uterine endometrium. This work has established, for example, that rat ovary explants, when combined with human pancreatic islets, retain their structure and function in culture for up to 21 days. Dr. Burdette’s team has also developed a method of generating scaffold-free endometrial organoids that contain both primary epithelial and stromal cells and that express androgen, estrogen, and progesterone receptors. These MPS and tissue culture models are being used to study the biological pathways and mechanisms underpinning conditions such as PCOS (e.g., by introducing synthetic hormones or studying the effects of altered gene expression on reproductive function). Lattice is also being expanded to support eight cross-talking organ systems in a microfluidic environment, which will help researchers to study how endocrine-disrupting compounds might impact reproductive function.
In addition to screening small molecule compounds, MPS can potentially play a role in predicting the in vivo performance of regenerative medicine advanced therapies, including cell-based products that rely on either iPSCs or multipotent stromal cells. Dr. Kyung Sung from FDA’s Center for Biologics Evaluation and Research emphasized that the successful clinical translation of such cell-based products requires the evaluation of large numbers of complex parameters. In turn, such evaluation involves the development of high-throughput assays to assess cellular function and heterogeneity and to enhance the throughput, reliability, and clinical relevance of cell-based screening platforms. MPS that recapitulate in vivo physiological conditions could potentially aid the development of improved methods for product development, testing, and characterization and the identification of product attributes that are predictive of safety, potency, and effectiveness. One area in which MPS have been used to evaluate cellular function is the study of myocardial infarction. Dr. Kevin Kit Parker from the Wyss Institute and his team have engineered human laminar cardiac tissue to develop an ischemia-reperfusion injury (IPI) model on a chip in order to study how contents of endothelial extracellular vesicles appear to induce cardioprotective effects among cardiomyocytes during IPI. By leveraging an instrumented chip to continuously obtain functional readouts, these studies generated substantially more granular data than would have been obtainable from an in vivo study and produced evidence to support the possibility of developing cell-based cardiotherapies based on the cardioprotective abilities of endothelial exosomes.
Dr. Thomas Hartung from Johns Hopkins University stated that advances in bioengineering organ architecture and functionality via the use of multi-tissue MPS have greatly increased the complexity of in vitro modeling, and that this
increased complexity—especially of stem cell–derived systems—requires greater efforts to standardize, qualify, and improve quality control for these systems. The complexity of MPS is compounded by the introduction of infectious or therapeutic agents, the induction of cancers and immune responses, and the manipulation of genetic and environmental factors (e.g., oxygen or growth factors), as well as environmental toxins or toxicants. Commercial providers may standardize MPS to ensure broad availability and comparability of test systems and to ensure FDA’s and other regulators’ ability to confidently and meaningfully assess data generated on these platforms. A Good Cell Culture Practices (GCCP 2.0) standard is being finalized by a task force organized by the European Centre for the Validation of Alternative Methods of the European Commission, while good in vitro reporting standards are being developed. As MPS become more prevalent, it may become necessary for investigators to rethink methods for validating these platforms (e.g., by transcending traditional validation methods that use multi-laboratory cross-validations, perhaps in favor of mechanistic validation methods) so that regulators possess the evidence they need to determine whether they can trust a new or alternative method. In developing novel MPS to study various physiological or pathophysiological processes, researchers could heed Albert Einstein’s maxim and design them to be “as simple as possible and no simpler.”
Beyond the potential capacity for MPS to accelerate preclinical compound screening, several panelists expressed optimism for the potential for MPS to substantially improve clinical trial design. Specifically, if investigators can harvest iPSCs from specific patients with a particular disease and use those iPSCs to develop an individualized tissue chip, then they could test the experimental compound on that tissue chip and use the results to decide whether to enroll that patient into a Phase III clinical trial. This potential would be especially significant in clinical trials for therapies against rare and ultra-rare diseases.
SESSION 7: PERSPECTIVES AND STRATEGIES ON THE NEED FOR ANIMAL CELL AND CHIP BANKS
As MPS developers design systems to address specific scientific and regulatory questions, these questions can extend beyond human health to also address animal health. Dr. Kevin Greenlees from FDA’s Center for Veterinary Medicine, which regulates animal drugs, foods, and devices, presented on how animals are enrolled in “target animal studies,” which are analogous to human clinical trials. He noted that target animal studies have some of the same challenges that human clinical trials do, and that MPS have the potential to likewise be used to address some of these challenges. Yet, the development of MPS to study animal drugs also presents unique challenges. For example, most veterinary drugs are developed for multiple species and for many breeds within those species. Moreover, researchers consider the safety of not only the animals, but also the humans who are handling or eating food derived from these animals. MPS and organoids are among the multiple tools being used for safety evaluations, with the promise to respond to these multi-parameter research questions by improving in vitro modeling for multiple COUs. Given the complexity of these questions, researchers may need to be cautious about overextending the use of a given MPS beyond the COU for which it is specifically designed. Dr. Bernadette Dunham, advisor to One Health for FDA, stressed the potential importance of MPS in supporting the One Health Initiative’s focus on the interconnections between human and animal health and the environment, and on replacing, reducing, and refining the use of animals in studying these interconnections. Other challenges to address before MPS can be fully integrated into the regulatory decision-making process include scaling, validation, reproducibility, translation, perfusion, and automation. Such integration would increase the impact of MPS across the pharmaceutical, food, cosmetic, chemical, defense, and other industries.
The specific needs of the pharmaceutical industry make both human and animal cell and chip banks important for drug discovery and development. Dr. Lorna Ewart from Emulate, Inc. described the ability of animal-based MPS to bridge the translational gap between preclinical compound screening and Phase I clinical trial research. Drug developers often obtain inconsistent safety signals when testing drug candidates in in vitro, small animal, and large animal models, and it can be difficult to determine which safety signal will be most predictive of human response. Dr. Ewart presented on how Emulate has used rat and dog in vivo data to optimize rat and dog liver chips, and then applied the resulting insights to develop human liver chips that predict drug-induced liver injury in human patients. As these MPS approaches advance, the pharmaceutical industry may coordinate its efforts with regulators and other stakeholders, as well as each other, to accelerate progress. Dr. Patrick Devine from Novartis Institutes for BioMedical Research relayed the activities of the precompetitive International Consortium for Innovation & Quality in Pharmaceutical Development’s (IQ’s) MPS affiliate, which is working to leverage its large amount of in vivo animal model data to optimize MPS that consist of both human and animal cell and tissue cultures. For example, the IQ MPS affiliate is working to cross-reference animal cell–based MPS data with in vivo animal data in order to accelerate the development and validation of novel MPS technologies and applications. The IQ MPS affiliate is also helping to broker key collaborations within the pharmaceutical industry and with funders such as NCATS and regulators such as FDA. The IQ MPS affiliate has also been critical in helping to establish desired standards for organotypic cultures and by encouraging MPS data sharing, all with the aim of improving the pharmaceutical industry’s success rate in drug discovery and development.
The greater availability of differentiated cell types for use in organotypic cultures and MPS also present opportunities in areas outside of drug development, such as environmental hazard identification and safety assessment. Dr. Sidney Hunter from EPA and his team, for example, have used MPS to study the influences of chemical toxicants on early-stage human embryology. Dr. Hunter, at the beginning of MPS studies, called for researchers to think carefully about key decisions, such as what species to use, what endpoints to measure, what biological processes to model, what sources to use for cells of interest, what criteria to use in selecting cells for the model, what level of cellular complexity to include in the platform (i.e., which cell types could be included in the platform versus which cell types must be included to address the research question), and whether to create an intact model or create conditions for progenitor cells to self-organize and assemble. In their study of chemical exposures during early-stage embryonic development, Dr. Hunter and his team have successfully created both static and dynamic 3D models of the embryo–fetal neurovascular unit. This model demonstrates the ability to recapitulate blood–brain barrier function and thus to enable the testing of how specific chemical exposures may impact early neural development.
As MPS become more important to various industries, academic researchers, and regulators, the field will strive to establish shared resources such as cell and tissue banks for sourcing raw materials, and to develop best practices (e.g., benefits and drawbacks of using fresh versus frozen tissues) to help guide and standardize research processes and materials. Panelists noted that large companies that conduct many animal studies may be prime candidates to begin establishing cell and tissue banks and that broad and collaborative efforts would benefit from external funding. Other promising areas for further MPS development are the characterization and cross-species comparisons of markers such as drug transporters (e.g., multi-drug resistance proteins, breast cancer–related protein, P-glycoprotein), membrane (e.g., aryl hydrocarbon receptor, growth factor receptors) and nuclear receptors (e.g., pregnane X, estrogen, androgen, and thyroid receptors), and densities to improve translational validity across MPS of different species (including humans), as well as human MPS for applications for which no suitable animal models are available (e.g., human microbiome cultures). Several panelists considered it important to raise awareness within their professional associations, and among external stakeholders and the wider public, about the promise of MPS so that the field receives support to drive progress and replace, reduce, and refine the use of animal models and the areas for consideration and improvement.
CONCLUDING THOUGHTS AND FUTURE DIRECTIONS
Many workshop participants representing government agencies (both funding and regulatory) and the industry and nonprofit sectors supported the development and use of MPS as a NAM. Panelists highlighted various ways in which MPS, in line with the One Health Initiative, can contribute to improving human and animal health and environmental protection, while helping to replace, reduce, and refine the use of animals in research. No MPS data have yet been submitted to FDA. However, FDA welcomes such data as part of normal submission packages and encourages investigators interested in using MPS to engage with FDA early so that standards and best practices can be discussed and agreed to before regulatory decisions are made. Whether animal chip data are needed for qualification of human cell or tissue MPS will depend on the specific COU for those data. Several panelists commented that appropriate COU is necessary for regulatory acceptance of MPS data, emphasizing that researchers must choose their models thoughtfully in light of their specific research question.
Panelists also described the breadth of potential MPS applications, including prioritizing drug candidates; conducting pivotal studies to assess efficacy, safety, and potency endpoints for drugs and biologics; addressing emerging crises such as the SARS-CoV-2 pandemic and opioid crisis; and studying zoonotic transmissions. In addition to preclinical safety assessment, other applications for MPS include modeling human and animal diseases, personalized medicine, and screening patients for clinical trials, especially for cancer therapeutics, pediatric disorders, or rare diseases. Several panelists stressed that in vitro models, including MPS, can be combined with machine learning and computational modeling approaches to improve their ability to predict in vivo and clinical outcomes. To fully maximize the potential of MPS, some panelists expressed a desire to establish a standardized and reproducible source of cells and tissues that are both fully characterized and renewable (e.g., fully matured iPSCs). Establishing such a resource may require decision making about which species’ cells and tissues to include and developing or optimizing cell and/or MPS storage and transport procedures. Several panelists mentioned the European Collection of Authenticated Cell Cultures as a possible institution to spearhead the establishment of an MPS cell bank.
DISCLAIMER: This Proceedings of a Workshop—in Brief was prepared by Rose Li and Associates, Inc., as a factual summary of what occurred at the workshop. The statements recorded here are those of the individual workshop participants and do not necessarily represent the views of all participants; the planning committee; the Institute for Laboratory Animal Research; or the National Academies of Sciences, Engineering, and Medicine. Preparation of earlier versions of this proceedings by the following individuals is gratefully acknowledged: Dana Carluccio, Lucas Smalldon, and Nancy Tuvesson. References to proprietary products or services do not imply the adoption or endorsement of any product or service by the National Academies or any participant in the workshop.
REVIEWERS: To ensure that it meets institutional standards for quality and objectivity, this Proceedings of a Workshop—in Brief was reviewed by Suzanne C. Fitzpatrick, U.S. Food and Drug Administration; George W. Lathrop, Jr., U.S. Department of Veterans Affairs; Donna L. Mendrick, U.S. Food and Drug Administration; and Nicolette Petervary, U.S. National Institutes of Health.
PLANNING COMMITTEE: Patrick Devine (Co-Chair), Novartis Institutes for BioMedical Research; Danilo A. Tagle (Co-Chair), National Center for Advancing Translational Sciences, National Institutes of Health; Ashutosh Agarwal, University of Miami; Szczepan Baran, Novartis Institutes for BioMedical Research; Suzanne Fitzpatrick, U.S. Food and Drug Administration; Sean Gehen, Corteva Agriscience; David M. Kurtz, National Institute of Environmental Health Sciences, National Institutes of Health; Milica Radisic, University of Toronto; John Rogers, U.S. Environmental Protection Agency.
ABOUT THE ROUNDTABLE ON SCIENCE AND WELFARE IN LABORATORY ANIMAL USE
This roundtable was created to promote the responsible use of animals in science, provide a balanced and civil forum to stimulate dialogue and collaboration, and help build trust and transparency among stakeholders. Roundtable members comprise entities with strong interests in the use of laboratory animals in research, testing, and education, including government agencies, leading pharmaceutical and consumer product companies, contract research organizations, animal advocacy groups, professional societies, and prominent academic institutions.
ROUNDTABLE MEMBERS: Jill Ascher (Chair), Office of Research Services, Division of Veterinary Resources, National Institutes of Health (NIH); Bonnie V. Beaver (Vice Chair), Texas A&M University; Szczepan Baran, Novartis Institutes for BioMedical Research; Saverio (Buddy) Capuano III, Wisconsin National Primate Research Center; Carol L. Clarke, Animal and Plant Health Inspection Service, U.S. Department of Agriculture; Robert C. Dysko (Past Chair), University of Michigan; Michael Fallon, U.S. Department of Veterans Affairs; James G. Fox, Massachusetts Institute of Technology; Gloria J. Gaito, Pfizer Worldwide Research and Development; Alema Galijatovic-Idrizbegovic, Merck & Co., Inc.; Jim Gnadt, NIH Blueprint for Neuroscience Research; Gail C. Golab, American Veterinary Medical Association; Debra L. Hickman, Indiana University School of Medicine; LaWanda Holland, Janssen Pharmaceutical Companies of Johnson & Johnson; Michael Huerkamp, Emory University; Rich Krauzlis, National Eye Institute, NIH; David M. Kurtz, National Institute of Environmental Health Sciences, NIH; Margaret S. Landi, GlaxoSmithKline; Malcolm Martin, Viral Pathogenesis and Vaccine Section, National Institute of Allergy and Infectious Diseases, NIH; Joseph T. Newsome, University of Pittsburgh; Lori S. Palley, Massachusetts General Hospital; Patricia Preisig, Yale University; Barry Richmond, National Institute of Mental Health, NIH; Brianna L. Skinner, U.S. Food and Drug Administration; Edda (Floh) Thiels, National Science Foundation; Sally Thompson-Iritani, University of Washington; Joseph Thulin, Medical College of Wisconsin; Patricia V. Turner, Charles River Laboratories; Axel Wolff, Office of Extramural Research, Office of Laboratory Animal Welfare, NIH; and Robert H. Wurtz, National Academy of Sciences and National Academy of Medicine.
SPONSORS: The workshop was sponsored by NIH National Center for Advancing Translational Sciences; U.S. Department of Veterans Affairs Office of Research and Development; U.S. Environmental Protection Agency; U.S. Food and Drug Administration; and USDA Animal and Plant Health Inspection Service, Animal Care.
For additional information regarding the workshop, visit https://www.nationalacademies.org/event/01-19-2021/microphysiological-systems-bridging-human-and-animal-research-a-workshop.
Suggested citation: National Academies of Sciences, Engineering, and Medicine. 2021. Microphysiological Systems: Bridging Human and Animal Research: Proceedings of a Workshop—in Brief. Washington, DC: The National Academies Press. http://doi.org/10.17226/26124.
Division on Earth and Life Studies
Copyright 2021 by the National Academy of Sciences. All rights reserved.