Workshop Report: Zebrafish and Other Fish Models—Description of Extrinsic Environmental Factors for Rigorous Experiments and Reproducible Results

Abstract

Current standards for husbandry and maintenance of zebrafish and other aquatic species in the laboratory are diverse, and are subject to laboratory performance, engineering, and practice standards (the Guide), institutional interpretation, national animal welfare laws, and cultural differences. Consequently, it is difficult, and probably not advantageous, to establish a single standard in view of the hardy nature of zebrafish and the diversity of research requirements it is used to address. Based on their natural habitat, zebrafish can thrive in a variety of environmental conditions, which is a specific advantage for working with this laboratory organism. However, it also makes reporting and reproducibility difficult, because variations in the husbandry and environmental conditions, including the environmental conditions before and during experiments, are often underreported in the scientific literature. This lack of consistency presents a potential problem for research reproducibility. To begin addressing this emerging scientific gap, the National Institutes of Health's (NIH) Office of Research Infrastructure Programs (ORIP), Division of Construction and Instruments (DCI), hosted a workshop in late 2017, entitled “Zebrafish and Other Aquatic Models: Reporting of Environmental Husbandry Conditions for Rigorous Experiments and Reproducible Results,” that was attended by ∼60 participants. The objectives of the workshop were to bring together a diverse group of stakeholders—researchers, facility managers, veterinarians, journal editors, commercial vendors, and others to (1) review current husbandry and environmental management practices for the care of zebrafish and other aquatic organisms in the laboratory and to (2) propose a process for the development of a minimal set of environmental parameters that should be reported in publications to ensure rigor and robustness of experiments and reproducible outcomes. The participants also discussed how these recommendations, as an initial step, might be collected, disseminated, implemented, and improved upon after future iteration.

Introduction

Current husbandry standards for zebrafish (Danio rerio) are diverse and based on documents such as the Guide for the Care and Use of Laboratory Animals¹ (the Guide) or Guidance on the Housing and Care of Zebrafish² (Royal Society for Prevention of Cruelty to Animals). Another example is a joint working group report by members of the Federation of European Laboratory Animal Science Association and the European Society for Fish Models in Biology and Medicine, which is currently in preparation for publication (under review) for Laboratory Animals (SAGE Publications). The joint Working Group Report lists and defines among others a set of ranges for commonly used housing, husbandry, and environmental conditions to serve as standards for the European zebrafish community³ (P. Aleström, pers. comm.). Such standards are often equally subject to local interpretation by researchers and animal welfare committees or officials overseeing the humane use of animals in teaching and research. However, the broad tolerance of the species to variable environmental conditions has also, in part, driven its growth in the laboratory, and this presents a specific advantage for working with this species. Today, this flexibility also presents challenges as the scale and complexity of zebrafish and aquatic species in biomedical research expand. One challenge involves the overall lack of reporting standards for the husbandry and management of the fish in the laboratory. Although this trait has afforded researchers with diversity as to how they maintain animals before, during, and after experiments, it has also contributed to a situation wherein the importance of environmental conditions is frequently overlooked. This problem is reflected in the literature, wherein the environmental conditions of animals are either inadequately described in material and methods sections or simply not included at all.

This dynamic is taking place within a larger backdrop where the integrity of science is being questioned. Growing concerns of a “reproducibility crisis” in research have both scientists and the general public expressing doubts about the reliability of research and publication practices.⁴ This is due to failure on the part of researchers to adequately describe research methods, a situation that is particularly pronounced in animal studies.^5,6 The rapid rise of the zebrafish—with its diversity of standards for husbandry and environmental conditions—places the model squarely in the middle of this conversation. How strong are all of the new data being generated with fish if basic methods are not adequately reported to ensure reproducible testing?

In response to this, the National Institutes of Health's (NIH) Office of Research Infrastructure Programs (ORIP), Division of Construction and Instruments (DCI), organized a workshop on the subject that took place during September 11–12, 2017. The workshop, which convened ∼60 experts from academia, industry, publishing, and animal health, was entitled “Zebrafish and Other Aquatic Models: Reporting of Environmental Husbandry Conditions for Rigorous Experiments and Reproducible Results.” A detailed log of the meeting is available.^7,8

The sessions had two overarching goals. The first was to bring together stakeholders from various sectors of the field to (1) review the breadth of current husbandry and environmental management practices for zebrafish and other aquatic organisms used in research and (2) examine how these practices might affect research results. The second goal—taking the previous into account—was to try to come up with a process to develop a recommended minimal set of environmental parameters that should be reported in publications to ensure rigor and robustness of experiments and reproducible outcomes. Included in this part of the discussion were concepts around the collection and eventual dissemination of these data.

Results

Plenary presentations focused on physical and chemical parameters, pathogens, nutrition, welfare and behavior, and data collection/dissemination issues in aquatic facilities. Examples included how light wavelength,⁹ type of feed,¹⁰ variations of temperature,¹¹ water chemistry and quality,¹² disease,¹³ and other environmental factors affect gene expression, physiology, and behavior. These examples further illustrated the need to account for environmental metadata to ensure the robustness of experimental setups and accurate descriptions of experimental protocols.

The information gathered during the plenary sessions provided material for discussion aims for breakout sessions on the second day. These discussions addressed husbandry and environmental issues in more detail and considered strengths, weaknesses, opportunities, and threats of the current practices. The breakout groups formulated a number of proposals for which husbandry and environmental data should be collected and reported, and how it should be shared either in publications or when fish lines are exchanged between facilities.

As with any complex issue with many different stakeholders, broad consensus was developed for few issues raised during the session. However, a number of areas of common agreement were identified. First, and probably foremost, it was widely agreed that various husbandry conditions of the animals may be important for particular research outcomes. Indeed, there was little debate that the various environmental factors can exert widespread and potentially pervasive effects on the fish. The “environment” encompasses a range of conditions, including nutrition, water quality, pathogen status, social, and physical parameters, including light and temperature. The primary concern that many researchers expressed was production-related to the extent that these factors impact fish growth and reproduction. The concept of husbandry conditions being an important “nonprotocol induced variable” was less readily accepted and demonstrates one of the major challenges associated with this initiative; namely, determining which conditions are important enough to be reported for specific experiments.

Another area of broad consensus was for the need to support funding for basic research on the effects of various environmental parameters on the fish and research outcomes that depend on their physiology. This includes, but is not limited to, nutritional requirements, water chemistry, and physical parameter preferences, as well as social/behavioral impacts and pathogen status.

The final area of agreement was that technology will improve definition of key environmental conditions and the collection/dissemination of metadata. Examples of the former are tools for feeding fish defined amounts of feed and measuring various water quality parameters. For the latter, there was discussion that centered around the tools for collecting data and assimilating it into a common format. It was acknowledged that many existing platforms for data collection exist but there are (1) no straight-forward ways to share these data in its raw form and (2) there is little to no crosstalk between different platforms that exist in the field. Both of these factors limit the impact and scope of these data to improve reproducibility.

A key recommendation resulting from the workshop was on the process for next steps. First, recommendations included broad outreach to the community to seek critical input on the creation of community guidelines on minimal husbandry and environmental parameters to be reported and metadata formats for reporting them. After the workshop, participants compiled a set of reportable husbandry and environmental parameters, collected based on discussions during the workshop's breakout sessions. This preliminary data set is openly available,^7,8 requires community feedback, and is summarized in Supplementary Table 1 (Supplementary Data are available online at www.liebertpub.com/zeb). Implementing and adopting such guidelines would require appropriate planning and tools for data collection, management, and sharing to make these processes efficient and straightforward for facility staff and researchers. Such tools would also aid the communication within the community and help with data sharing.

Disclosure Statement

No competing financial interests exist.