Fundamental Approaches to the Study of Zebrafish Nutrition

Abstract

The value of the zebrafish model has been well established. However, culture variability within and among laboratories remains a concern, particularly as it relates to nutrition. Investigators using rodent models addressed this concern several decades ago and have developed strict nutritional regimes to which their models adhere. These investigators decreased the variability associated with nutrition in most studies by developing standardized reference and open formulation diets. Zebrafish investigators have not embraced this approach. In this article, we address the problems associated with the lack of nutritional information and standardization in the zebrafish research community. Based on the knowledge gained from studies of other animals, including traditional research models, other fish species, domesticated and companion animals, and humans, we have proposed an approach that seeks to standardize nutrition research in zebrafish. We have identified a number of factors for consideration in zebrafish nutrition studies and have suggested a number of proposed outcomes. The long term-goal of nutrition research will be to identify the daily nutritional requirements of the zebrafish and to develop appropriate standardized reference and open formulation diets.

Introduction

The term nutrition is commonly found in the scientific literature and is used (or misused) frequently in everyday language. Animal nutrition represents a basic concept, yet the approach to the study of nutrition varies widely among disciplines and researchers. Human nutrition is one of the most actively discussed but least understood areas of nutrition research. In this regard, much of what we know about human nutrition is implicated from epidemiological surveys or from studies of other mammalian species. For this reason, nutrition of traditional research animals, primarily rats and mice, has received much more attention in recent years, particularly in response to the discovery of chemical compounds contained in specific foodstuffs that may promote unwanted or undesired effects in studies of disease. Nutrition of domesticated agricultural animals is perhaps the most well studied area of nutrition, and many experimental protocols have been developed for the study of these commercially important species. More recently, studies of aquatic organisms, primarily finfish and shrimp, have increased in response to increased interest in aquaculture. Studies such as these have established a framework by which we should approach the study of nutrition of zebrafish held in culture.

To better develop a standardized approach to evaluating zebrafish nutrition, we must define the basic terminology associated with nutrition research. To more precisely evaluate nutrition in zebrafish, we will define a nutrient as any ingested or absorbed molecule, organic or inorganic, required for normal physiological function. Consequently, we can define nutrition as a process by which living organisms acquire and assimilate nutrients for survival, growth, reproduction, and replacement of tissues. To this end, daily dietary nutrient requirements should be determined in our model species. Nutrition research should be structured to show which nutrients are important and how they are obtained and assimilated. It is also important to understand how these nutrients possibly interact with other nutrients or non-nutrient compounds.

Despite the widespread use of zebrafish (Danio rerio) as a high-throughput development, toxicologic, and biomedical model, the daily nutritional requirements for this species have not been determined. The failure to estimate daily dietary nutrient requirements originally resides in the interests of those researchers who use the zebrafish as a tool. For example, live diets are a mainstay in first feed strategies of most zebrafish research labs. As zebrafish age, adequate formulated diets are available commercially. In combination, both provide reasonable growth and fecundity for zebrafish culture. However, the contribution of specific nutrients in any of these live or commercial diets is unknown, and the health of the fish consuming these diets cannot largely be determined. For some applications, use of undefined nutrient sources may not be of significant consequence, and we would minimally hope that the dietary management strategy is clearly outlined in publications about such applications. However, when zebrafish are used as a model of human disease, consideration of both quantity and quality of specific nutrients is essential. Zebrafish researchers must avoid repeating the mistakes made in the use of nondefined diets in mammalian models (primarily rats and mice) in previous decades. Consequently, it is essential to have the fish consume a nutritionally complete, chemically defined diet in those studies where disease onset may be affected or may have a contributory nutrient, dietary interaction, or antinutritional compound (ingested compound eliciting an aberrant physiological response) affecting disease onset and/or progression.

Considering the lack of published nutritional information for zebrafish culture, all of the information published to date is useful and can provide insight into nutrient requirements. Furthermore, a wealth of information on other fish species can provide relevant information. However, additional scrutiny and standardization are necessary to accurately quantify nutrient requirements (and nutrient toxicities) for zebrafish. This includes an adequate description of the environmental and nutritional history of the experimental animal and its parents. Because of the lack of dietary nutrient data, the goal of this paper is not to provide an in-depth review of the current state of nutrition research in zebrafish; our goal is to identify and summarize those criteria and methods that should be considered in conducting nutrition research with zebrafish.

Nutrition versus Dietetics

The procedures by which nutrition studies are conducted vary widely among researchers. Historically, researchers have engaged in studies that should be commonly referred to as dietetics, defined as the “practical application of diet in relation to health and disease” (American Heritage Dictionary 2007). Simply stated, these studies have evaluated common dietary items, usually natural live diets including Paramecium, brine shrimp nauplii, rotifers, or dried feeds prepared from these natural organisms. These studies are valuable in that they provide important information about feeding strategies and basic requirements in zebrafish and can provide direction on future nutrition research. However, evaluation of the effect of natural diets does not provide information adequate to define the daily quantitative nutrient requirement for an animal because the nutrient content of these natural diets is unknown and cannot be completely duplicated. Furthermore, these studies fall short of being repeatable due to qualitative differences in the dietary food items that occur seasonally or geographically. Also, these natural foods vary in physical and chemical characteristics, including size, shape, hardness, and basic chemistry, and may contain varying levels of attractants, stimulants, and deterrents. Despite their limitations, use of these diets represents a good foundation that can lead to well-characterized nutritional studies.

Historical versus Contemporary Approach

The approach used to study zebrafish nutrition can be adapted from studies with other organisms. Human nutrition has received the most attention of any organism. However, the study of human nutrition does not easily lend itself to empirical evaluation. Recent studies are largely corollary or epidemiological. Other animals that have received widespread attention are domesticated animals that are traditionally farmed for consumption, such as cattle, sheep, pigs, chickens, and other finfish. In fact, there is probably more information available about chicken nutrition than about nutrition of any other species. Additional studies with experimental laboratory animals, including mice and rats, suggest that there is much to learn about nutrition for all organisms. For example, soybean products are a common ingredient (protein or lipid source) in many mammalian diets. However, soy products may contain specific chemicals (isoflavones) that influence disease and animal health (Barnes et al. 1998), including antinutritional factors that can affect fish feeds. These studies indicate that an appropriate approach must consider the chemical and nutrient content of all feed ingredients consumed in the diet and that this approach must be organized to evaluate the appropriate outcome, whether that outcome is survival, growth, reproduction, aging, disease resistance, or another response. They have also demonstrated that standard, predictable, and valid methods must be used, which include adequate physical experimental conditions with diets having the desired physical characteristics and use of proper feed management strategies.

Field observations of holotypical populations of zebrafish provide some of the most valuable information on dietary requirements. Spence and colleagues (2007) provided an excellent review of the basic life history of zebrafish populations in south Asia. General observations and gut content analyses indicated that zebrafish consume a wide variety of animal and plant matter, including zooplankton and insects, phytoplankton, filamentous algae and vascular plant material, spores and invertebrate eggs, fish scales, arachnids, detritus, sand, and mud. A significant constituent of the gut contents was dipteran larvae. This diversity in foodstuffs is characteristic of an omnivorous species that is a generalist and can absorb and assimilate nutrients from a wide range of sources (not uncommon among cyprinids). Of particular interest is the fact that the diversity of foodstuffs indicates that material is ingested from different strata, including the surface, the water column, and the benthos. As such, the ability of zebrafish to feed in various strata will be important in providing diets in a physical form that is readily consumed, whether from the surface (floating feed) or from the water column or benthos (sinking feed).

Early interest in zebrafish nutrition was stimulated by the ornamental tropical fish industry. General fish diets (derived from commonly cultured fish) were fed to large populations of zebrafish often held in ponds or raceways. In many cases, natural productivity was stimulated in these culture systems, and natural food items (zooplankton and biofilm assemblages) were consumed, supplementing the formulated diets. Consequently, many of the dietary requirements were most likely met, and zebrafish culture was successful.

Adaptation to the laboratory was relatively easy (Creaser 1934). Early feeding strategies of zebrafish in the laboratory were basically adopted from the ornamental fish industry. It was quickly determined that zebrafish could be reared on tropical fish diets purchased from commercial sources. Little interest was shown in the quality of the diet as long as it proved successful. However, early concerns arose in the concept of “first feed.” At the time of completion of gut development (approximately 5 days postfertilization), anecdotal and published information suggested that survival and growth on formulated diets were relatively poor (Goolish et al. 1999). Batch-to-batch variability in survivorship was high. As was the case in many fish species, larval zebrafish performed best when fed a diet that modeled their natural diet of zooplankton (Spence et al. 2007). Larval zebrafish were often fed diets of Paramecium, rotifers, and brine shrimp nauplii because these feeds increased survival and early growth. Today, these diets are still used successfully in many laboratories around the world and are discussed in several handbooks (Westerfield 2007). After a period of early development (from several days to several weeks), formulated diets were introduced to developing zebrafish. The primary desire for the use of formulated diets was ease of feed delivery, storage, and preparation. Additionally, many of the smaller live diets (rotifers) could not be routinely cultured to produce the required biomass for large zebrafish populations in most laboratories. Success of a feed and/or feeding strategy was measured primarily as survivorship and reproduction of adults. If a diet produced sufficient survival of larvae and juveniles and good fecundity in adults, then that dietary combination was adopted by the lab or facility as “standard.” Unfortunately, many of these practices termed as “standard” have not been validated in scientific studies and are not comparable. This lack of standardization will have a significant impact on the success of zebrafish as a biological model organism.

Herein lies the problem facing today’s zebrafish laboratory: there is no standardized diet for common application. There is no reference diet to which new diets can be compared. We do not know the basic ingredients in many of the commercial diets, nor the level of contribution of specific nutrients. No open-formulation diets are available for comparison. We do not know the best diet for first feed (known in other species to be a critical period affecting gut development). We do not know which nutrients are required for normal growth and development, and we do not know which diets promote optimal fecundity, fertilization success, and healthy embryos. We use commercial diets that can contain dyes, lakes, antinutritional factors, and preservatives with little thought as to their impact on genetics, genomics, transcriptomics, proteomics, or metabolomics, or even epigenetics. The zebrafish has current and potential value as a biomedical model and for toxicological assessments, yet we often unknowingly feed zebrafish commercial diets containing compounds that could easily compromise the results of many scientific studies. One could argue that dietary conditions in many laboratories are held constant and that reported results of many previously published studies have the appropriate controls. However, if the presumed nutritional control does not actually exist, scientific conclusions could be compromised as well. This concern would be particularly important for the use of zebrafish as models of disease wherein nutrition is hypothesized to have an important role in disease onset and progression. Quite simply, the daily dietary nutrient requirements for any of the various life stages (larval vs. juvenile vs. adult) of Danio rerio are unknown.

Nutritionally complete diets, highly desired for zebrafish culture, are being developed in several laboratories. However, not all biologists working with zebrafish are adequately trained in nutrition research, potentially resulting in inappropriate experimental design and experimental methods that commonly vary among laboratories. Consequently, what is an appropriate approach to the study of nutrition and diet development in zebrafish? Can we borrow approaches used in other species (fish and other vertebrates), or are we destined to make similar mistakes by ignoring factors with obvious (or not so obvious) impacts? What are the appropriate outcomes that we should measure? Survival, length, weight gain, condition factors, stress test responses, food conversion ratios, general health, disease susceptibility, and fecundity are some of the historical outcomes, but others may be more appropriate. Additional outcomes that demand consideration are gamete production, health and/or disease resistance, and organ development and function. Will there be differences in the nutritional requirements of larvae, juveniles, and adults? All these questions represent important considerations as we develop appropriate outcomes for zebrafish nutrition research.

Numerous commercial feeds for zebrafish have appeared on the market in the last decade, and several formulations have been published (Kaushik et al. 2011; Siccardi et al. 2009). Many of these feeds are made by pelleting or extrusion cooking procedures, with moisture contents of less than 15% for dry feeds. Many contain ingredients that are highly refined. They can be produced in large quantities and have a stable shelf life. Similar processes were used in the development of rodent diets. However, rodent diets experienced a renaissance when it was determined that nutrient (or antinutrient) content must be available to scientists. Many rodent diets are now open formulation (public access to all ingredients and nutrient contents) and are recommended by the American Institute of Nutrition, the National Academy of Sciences, and other agencies (Barnard et al. 2009). Standards for nutritional studies have been established and have increased the fidelity of many studies investigating these models as they relate to human health and disease. Development of similar approaches to dietary standards for zebrafish will be challenging. First and foremost, the nutritional requirements for any zebrafish genus or strain have not been adequately determined. Open-formulation diets have not been developed for zebrafish. Most diets are largely practical (having relatively unrefined sources of proximate constituents), and purified (chemically defined) diets are not available. Furthermore, zebrafish feed is obviously placed in aquatic environments; thus, proffered food must be consumed immediately or risk significant leaching of required nutrients. Consequently, feed management will have an increasingly important role in determining nutritional requirements.

In the remaining portion of this paper, we attempt to outline basic considerations for nutrition research in zebrafish. These considerations should provide investigators with a strong foundation to determine the nutritional needs and requirements for any strain of Danio rerio.

Basic Considerations in Zebrafish Nutrition Research

Knowledge of Zebrafish Biology is Essential

In nutrition as well as in other research areas, one cannot develop strategies to manage an organism without understanding the biology of that organism. A fundamental understanding of factors that affect feed acquisition and assimilation is critical for nutrition research. Fortunately, population analyses and behavioral observations from natural populations have been well reported for zebrafish (Engeszer et al. 2007; Spence 2011). These observations help promote diet development and feed management strategies that adapt to the natural biology of the organism.

In the wild, juvenile and adult zebrafish are generally found in shallow water habitats, including flood plains or with rice agriculture (Spence 2011). They also can be found in open water or in association with vegetation. Generally speaking, zebrafish are usually found in habitats with sandy or silty soils and are not considered riverine species. They avoid areas of hydrodynamic activity and prefer relatively still waters (currents, 0-0.1 m/sec) (Engeszer et al. 2007). As such, they are well suited for culture in aquaria of different sizes and shapes with limited water exchange or flow. These aquaria can be composed of glass, fiberglass, or high-quality plastics. Metal surfaces are not desirable because they release metal ions into water with high conductivity, the toxicity of which has not been well described for zebrafish held in culture. In culture, zebrafish can often be found grazing on the surface of their containers; therefore, excess food and the presence of biofilms will compromise nutrition research.

Zebrafish are eurythermal, tolerating temperatures of 6-38^oC (Spence 2011), and many investigators rely on static renewal, recirculating, or flow-through systems. A range of temperatures can be used for successful culture (Lawrence 2007). Although zebrafish are tolerant of a wide range of temperatures, changes in temperature during nutritional trials will have consequences by influencing feeding activity, ingestion, digestion, and reproduction. Gamete production (quantitative) will most likely be impacted by temperature, and, as a consequence, energy allocation will also be affected. Several reference texts have indicated that 28^oC is a preferred temperature for zebrafish culture. If investigation studies were performed at a similar standard temperature, then nutrition studies would be comparable. If this is not possible, an adequate description of the temperature regime during the experiment should be provided. Metabolic demands (based on oxygen consumption) vary according to temperature and age, with Q₁₀ values ranging from two to five (Barrionuevo and Burggren 1999), indicating that energy use (nutritional demand) is highly sensitive to changes in water temperature. This information supports the plea for standardization of culture temperature when nutritional requirements are evaluated. We also hypothesize that temperature extremes alter proximate composition and metabolism, and that high temperatures could induce differences in gene expression and protein metabolism (stress response) (Vergauwen et al. 2010).

Inorganic ions are required for normal physiological activity in most teleosts. Unlike terrestrial organisms that consume ions (or other minerals) in the diet, fish absorb a significant percentage of their essential minerals from the water. However, the requirements for chloride and other individual ions have not been determined and zebrafish nutrient requirements can be affected by the mineral content of the water. Zebrafish are freshwater, basically stenohaline, fish and do not tolerate water with high ionic content (usually reported as salinity or conductivity) in culture (Boisen et al. 2003). They can tolerate water of low ion content, primarily supported by a series of efficient ion transporters (Craig et al. 2007). In fact, the average conductivity of water supporting zebrafish in wild populations ranges from 10 to 271 μS/cm (Engeszer et al. 2007). General culture manuals indicate that a salinity of less than 1 ppt is reasonable for normal survival and growth; this level is usually maintained by the addition of synthetic sea salts. Other macrominerals, such as calcium, can be obtained from the water, but few reports are available. Padgett-Vasquez and colleagues (2008) reported that dietary calcium requirements in the water are minimal, suggesting that zebrafish are very efficient at removing these important cations from the water. To standardize nutrition experiments, we highly recommend that purified water sources (using reverse osmosis or ion exchange resins) be used and that a standard synthetic sea salt be added to attain a defined conductivity or salinity (unless the requirements for specific ions are being investigated). When using synthetic sea salts, we recommend 0.8 ppt salinity (approximately 1600 μS/cm) as a reasonable salinity for conducting nutrition experiments. Addition of synthetic commercial salt preparations at this level will not only provide physiologically relevant ions, including sodium, potassium, and chloride, but will also provide quantities of the minerals calcium, magnesium, zinc, and phosphate. Because most of these ions are nutrients as well, it is important to provide the methodology of water preparation in any publication. Hyper- or hyposaline exposure will increase the energetic requirements of zebrafish, thereby affecting the nutritional requirements of the species.

Field observations indicate that zebrafish are generally rheophobic, residing in areas of low water flow (Engeszer et al. 2007; Jayaram 1999; McClure et al 2006; Sterba 1962; Talwar and Jhingran 1991), but they will orient to a current. Under this condition, energy expenditure due to maintenance of position in water currents would be minimal. There are no reports that water flow (hydrodynamic action) enhances or impedes feeding activity. For zebrafish held in culture, water flow (turnover) in culture systems for nutritional studies must be monitored and reported (e.g., volume exchange per hour or per day). Minimum rates will depend on aquaria shape and size, but rates of exchange should be adequate to maintain oxygen saturation greater than 6 mg/L (Matthews et al. 2002) and to maintain adequate water quality. Obviously, high rates of turnover during periods of feeding, particularly live diets, will result in a waste of feed. In nutritional studies, another consequence of water flow is the leaching of nutrients from formulated feeds not consumed immediately. Although a number of variables affect leaching (including particle size and binders), it becomes an issue when feed is not consumed immediately by the fish. We recommend that water flow be monitored and reported in nutrition research (Goolish et al. 1999), and the effect of water flow or hydrodynamic action should be considered in future studies.

Photoperiod influences reproduction, locomotion, and feeding in zebrafish, but its influence on feeding is not clear. Zebrafish exhibited higher locomotory activity during the day; however, under a 12:12 light-dark cycle, zebrafish show a nocturnal feeding pattern (del Pozo et al. 2011). Locomotion was also influenced by feeding time (Blanco-Vives and Sánchez-Vásquez 2009). Although photoperiod represents an important factor, a more important factor influencing feeding under controlled laboratory conditions may be consistency in feeding times relative to the photoperiod or another entrainment event. Scheduled feedings can result in feed anticipation activity, and it has been hypothesized that this type of behavior may improve feed acquisition and nutrient utilization by improving consumption at the time of feeding (Comperatore and Stephan 1987). Therefore, adherence to a regular feeding schedule could improve nutrient acquisition from feeds (reduced leaching time) as well as increase total consumption. In any case, photoperiod and feeding time and rate should be reported.

A number of abiotic and biotic factors must be considered or controlled during nutrition research. Although not comprehensive, Table 1 lists some of these factors important in nutrition research. In general, abiotic factors should be maintained as constant as possible during the experiment, and the prior history of the experimental animal should be reported in the methodology. As a general criterion, a survival that is at least 90% would be optimal in a control group (often fed live feeds) and used as a minimum criterion for any dietary nutritional experiment. Values significantly below this level suggest confounding interactions or a poor control population.

Table 1.

Abiotic and biotic factors presumed potentially important in nutrition research

Abiotic	Biotic
Temperature	Life stage
Salinity or conductivity	Size
Photoperiod	Sex
Light intensity	Disturbance
Oxygen	Dominance and aggression
Water flow	Strain or phenotype
Water quality (e.g., ammonia, nitrite, nitrate, pH)	Disease
Ions	Symbioses or specific pathogen free

Many of the factors in Table 1 have been considered in ecological evaluations but not in the context of nutrition. Undoubtedly, all would affect nutrient acquisition and production. It is possible to control most abiotic factors in nutrition studies, and studies evaluating the importance of each factor would be valuable contributions to our understanding of zebrafish biology and nutrition. Water quality will be one of the most important factors to evaluate. Water quality can be considered in terms of both waste production (ammonia, nitrite, nitrate, dissolved organic compounds from leachate) and those ions and the chemical environment considered necessary for optimal survival, growth, and osmoregulation (calcium, magnesium, trace minerals, alkalinity, pH, and so on). Biotic factors are sometimes difficult to control, but even anecdotal information should be reported. Of particular interest will be interactions related to sex. Nutritional investigations that use mixed-sex populations or sex-specific populations could affect results. In fact, the nutrients themselves could affect animal interactions (S.A. Watts, personal observation). Many biotic factors, such as animal interactions, dominance, and aggression, can be controlled by rearing cultured zebrafish individually in aquaria or other containers, but such an approach may not be practical. Of great interest will be the evaluation of nutrition in specific pathogen–free or transgenic lines of zebrafish. The role of specific symbioses (e.g., microsporidian infestations) or gene manipulations on energy demands and nutritional requirements needs to be investigated.

Another challenging aspect of zebrafish nutrition research is the pattern, or lack thereof, of ingestion (feed intake). Feed intake in most juvenile or adult zebrafish occurs rapidly and without apparent pattern, making direct observations of feed intake difficult. Formulated feeds for zebrafish are small in size and irregular in shape, unlike the large formulated pellets for large aquatic species, which can be observed to be rapidly gulped. Increased surface area relative to volume of these small pellets or flakes would promote the leaching of specific nutrients. Proffered food can remain in the water for some time prior to feed intake, again allowing significant leaching. It is possible that feeding attractants that enhance the feeding response in zebrafish could be identified (Lindsay and Vogt 2004), thereby allowing more defined measurement of feed intake. Further observations, perhaps with high-speed photography, are needed to evaluate the biologic (behavior) and physical characteristics of food detection, acquisition, and satiation. This will include evaluation of particle size selection or preference as well as particle color. Also, factors that affect buoyancy (whether the diet is floating or sinking and, if sinking, the speed and action of the descending pellet or flake) may affect feed intake (Goolish et al. 1999).

An additional consideration in zebrafish nutrition concerns the role of dissolved organic material. The contribution of dissolved organic material to total production is not known; however, even a small percent contribution would be important, particularly in terms of chemical attractants or signaling molecules. Uptake of dissolved non-nutrient organic toxicants should also be considered. Because of the lack of dissolved organic material in most culture systems, the importance of dissolved organic material may be minimal compared with the potential role of dissolved inorganic material (macro- and microminerals) discussed previously.

Culture System Designs

Nutrition research is usually accomplished in what is termed “blue water” conditions—that is, clean freshwater in highly controlled environments. However, conditions can vary depending on the desired outcomes of the study. Specific nutrient requirements should be determined under well-defined conditions of water quality and other abiotic and biotic factors. Accurate descriptions of all culture conditions in published reports will allow comparisons among studies.

Many different culture systems for zebrafish have been used, and they are often specific to a particular lab or institution. They can be as simple as a few glass aquaria or as elaborate as rack systems currently produced by several commercial vendors. Most culture systems are generally adapted from those used in the culture of other aquatic species. For our purposes, most culture systems will be defined in terms of water source and filter characteristics. These systems are generally referred to as flow-through, recirculating, a combination of the two (semirecirculating), or static renewal. Specific aquaria/tank/rack concepts or designs will not be addressed, but the size (volume) of the individual culture vessel will affect the number of individuals held (density), with potential density-dependent effects on zebrafish growth demographics.

Flow-through culture systems are the least common systems used by researchers, but they do have many attractive attributes that make them appropriate for conducting nutritional research. Municipal or well water (filtered or unfiltered) is pumped into a culture facility and returned to the environment, usually after remediation. Once established, flow-through water systems are reasonably cost effective to operate. Large quantities of water can be moved quickly though the culture system, allowing virtually unlimited supplies of fresh clean water. Waste material does not accumulate appreciably and is inexpensive to remove. Consistent water quality depends on the source of the water and whether that source changes over time (one or more wells or impoundments could be used seasonally in some municipalities). In addition, proximity of the water source intake to potential sources of pollutants or toxicants should be considered because stormwater runoff into freshwater aquifers or impoundments can be a source of toxicants that could be harmful over time. Temperature and conductivity may be difficult to maintain without costly inputs. Another important concern is the transient occurrence of potential pathogens (e.g., bacteria, viruses) that could have deleterious effects on growth and survival of zebrafish. Flow-through water can be filtered and treated before exposure to the zebrafish but with an obvious cost input. Source, availability, and cost to maintain will be the deciding factors in using these systems.

Recirculating systems are used primarily by investigators who conduct research where flow-through water is not available or where the supply and/or quality of water may be marginal. In recirculating systems, the investigator has the ability to control water quality and even eliminate undesired factors. Water can be treated by a variety of filter systems, and specific ions can be added to the water. In our laboratory, potable water is filtered with sediment and charcoal filters, reverse osmosis, and ion-exchange columns to produce a highly purified fresh water source. Then synthetic sea salt is added to achieve an appropriate conductivity to maintain fish health.

Despite offering some apparent advantages, recirculating systems also have significant shortcomings that must be considered, particularly the control and monitoring of water quality in these systems. Water quality can be controlled by the use of mechanical filters and clarifiers in combination with biological filters, ultraviolet sterilization, protein skimming, and ozonation (or combinations thereof). These systems, particularly biological filtration, often require an adjustment period, whereby appropriate levels of autotrophic nitrifying bacterial populations need to be established. Filter efficiencies will fluctuate over time if water parameters such as alkalinity, hardness, and pH are not monitored and corrected as needed. These systems will commonly require inputs of specific macro- and microminerals, the required levels of which have yet to be established in zebrafish studies. Costly initial infrastructure investments as well as recurring maintenance costs, particularly for larger, more complex systems, will be necessary. Use of these systems requires a reasonable knowledge of culture engineering technologies and water management. However, if properly monitored, these systems can be effectively used in nutrition studies.

Static renewal systems can be inexpensive and effective under small-scale conditions. For these systems, water is replaced incrementally in small tanks or raceways, at levels ranging from 5% to 50% of water volume exchange per day. Water exchange is not without consequence because water change can disturb zebrafish in culture. Water quality is somewhat inconsistent in these systems, pulsing in association with the proportioned quantity and frequency of exchange to decrease soluble and solid waste material. These systems are not recommended for use with large populations of zebrafish and are least desirable for use because of an erratic rather than consistent control of water in the culture units.

For any type of culture system (particularly new systems), it is prudent to clean and rinse the aquaria before introduction of fish. New tanks often have been exposed to various cleaning agents during the manufacturing process, and these can contaminate a culture system for some time. Adequate rinsing prior to the introduction of fish is required. It may be wise to let water stand in the culture unit to collect any leachate. In general, we rinse a new system with several flushes of clean water during the week prior to the introduction of fish, and a weak acid treatment might facilitate the process of removing alkali. Furthermore, in systems where a biological filter is used, we introduce a population of bacteria to the filter and then “feed” the filter with 2 mg/L each of ammonium chloride and ammonium nitrite (for at least a week or more) so that the bacteria in the biological filter are established prior to the introduction of fish. Monitoring the rate of conversion of these compounds to nitrate will determine the stage of development of the filter and its efficacy.

Recently, two major concerns have been discussed concerning zebrafish culture systems. For those recirculating systems that use biological filtration, the source of the biota must be considered. Mixed cultures of autotrophic nitrifying bacteria are required to establish the functional biofilms (nitrification: converting ammonia to nitrate) in biological filters. Usually these cultures are obtained from other “feeder” tanks with established biota, or, given enough time, biota will establish themselves in the systems. Some systems can have undesirable bacteria (e.g., mycobacteria) that could compromise the health of the fish in the system and interfere with response variables of the fish. Ultraviolet or ozone purification downstream of the biological filter can help to minimize the problems associated with these undesirable bacteria, but testing may be required to monitor their populations. A second issue concerns the ultimate disposal of the water and/or solid waste material from these systems. Concern about the release of potential microbes, viruses, or even fish from culture systems into municipal wastewater systems has increased. Several laboratories are considering disinfection of the wastewater prior to disposal; additional measures of disinfection may be required in the future.

Diet Development

Purified, Semipurified, and Practical Ingredients

A full understanding of the nutritional requirements of zebrafish and of those factors that affect nutritional requirements requires empirical studies of diets that contain chemically defined nutrients and ingredients. We recommend using a step-wise evaluation of the daily dietary nutritional requirements and the nutritional value of specific ingredients for diet development. Because of numerous nutrient interactions (known and unknown), this process requires a continuous reevaluation of specific nutrients as requirements for other nutrients are determined. Diet development will allow a biological evaluation of those nutrients that affect growth, reproduction, and other physiological and molecular functions (particularly relating to disease onset and progression).

Diets are classified based on ingredient and nutrient composition (Table 2). Some diets (research diets) are used to determine specific nutrient requirements under defined experimental conditions. Others (practical diets) have commercial applications and are designed for large-scale production and use, based on least-cost formulations to determine ingredient composition. This means that commercial producers can change ingredients without notice. Because most zebrafish diets are used in research laboratories, lowest cost should not be a determinant for use.

Table 2.

Classification of feeds

Type	Application	Content	Relative cost
Purified	Experimental feeds, used to evaluate  macro- and micronutrient requirements	Purified, chemically defined ingredients, usually  with defined lot numbers and specifications	High
Semipurified	Experimental feeds, used to evaluate  macro- and some micronutrient  requirements	Both purified and practical ingredients,  not completely defined	Moderate
Practical	Commercial production feeds, produced  in mass quantities for aquaculture  operations, cannot be used to evaluate  nutrient requirements	Practical ingredients, not  chemically defined	Low

Purified, semipurified, and practical diets are composed of ingredients obtained from commercial sources. Purified ingredients are usually identified by specific ingredient and lot numbers, are chemically defined, and can be compared chemically with similar ingredients from various vendors. Practical ingredients are usually crude foodstuffs that may be relatively unprocessed, dried, and ground to an appropriate particle size. These ingredients may vary significantly in composition (based on source), vendor, and season of collection. Practical ingredients are used to provide essential quantities of proximate constituents such as proteins, lipids, and carbohydrates but can also contribute significant levels of micronutrients. These ingredients are combined with ingredients that have binding properties (e.g., wheat gluten, carboxymethyl cellulose, alginate with or without sodium hexametaphosphate). The binder aids in the formation of pellets that have low moisture content (<15%) using heat and pressure. The nature of the binder also affects stability and texture of feeds. Semimoist diets (usually 20-40% moisture) can be formulated and used in nutrition studies, but they generally have reduced shelf life and have not received widespread use in zebrafish culture.

In many cases, combinations of specific ingredients are formulated to produce purified, semipurified, or practical diets. Combinations of ingredients (e.g., fish meal, soy protein, gluten, casein as protein sources) are often preferred to single-ingredient nutrient sources, particularly in semipurified and practical diets, because they can effectively provide a suite of compounds suggested to be important nutrients. Single-ingredient nutrient sources (e.g., casein as a sole protein source) are often used in purified diets to better determine the nutritional requirement for a specific nutrient or nutrient source.

Survival and growth rates using practical diets are higher than those achieved when using semipurified diets, which in turn are higher than those achieved using purified diets (Carvalho et al. 2006). Although practical diets are not appropriate for nutrition studies in zebrafish, it is possible to gain insight into the quantities or qualities of specific nutrient combinations that promote animal health. For example, corn meal can be used in practical diets; however, such diets contribute carbohydrates (starch), protein (gluten), lipids (n6 fatty acids), and other nutrients that may or may not be identified but can confer an advantage for animal growth. Corn meal is a good ingredient in many practical feeds, but the nutrients derived from corn that are actually important cannot be determined without fractionation and purification. It is very difficult to determine dietary nutrient requirements using practical diets because controlling to achieve a very low limiting nutrient level is difficult and it is impossible to vary only one dietary nutrient. Regardless of the type of feed, an adequate diet for the determination of dietary nutrient requirements requires a survival of greater than 80% of animals and minimal growth rates that are 75% of the growth rate for animals living in natural environments under comparable conditions. Ideally only the nutrient being studied should vary in the diets, and only purified or semipurified diets can be used in such an investigation. Some dietary nutrient requirements such as microminerals (Table 3) cannot be easily evaluated with semipurified diets and usually require the use of a purified diet.

Table 3.

Classes of nutrients and sources suggested to be important in zebrafi sh nutrition

Nutrient Class	Description
Proximate nutrients	Protein (nitrogen); essential and nonessential amino acids, dipeptides, tripeptides,  hydrolysates, intact protein
	Carbohydrate (digestible); monosaccharides, oligosaccharides, polysaccharides
	Lipids; essential fatty acids, polyunsaturated fatty acids, long chain polyunsaturated fatty acids,  n3 and n6, phospholipids, cholesterol
	Ash (non-nutrient, e.g., diatomaceous earth)
	Fiber (insoluble and soluble, nondigestible)
Macrominerals	Calcium, magnesium, phosphate
Microminerals	Iron, zinc, manganese, copper, selenium, cobalt, sodium, chloride, potassium, boron
Vitamins	Ascorbic acid, tocopherols, calciferols, naphthoquinones, retinols, thiamine, ribofl avin,  pyridoxine, niacin, pantothenic acid, biotin, folic acid, choline, inositol, etc.
Carotenoids	Carotene and xanthophylls

Physical and Chemical Characteristics of Feed and Storage

The nutrient content of a diet is the most important determinant for optimal outcome assessment in nutrition studies. However, ingestion of a diet must occur before nutrient content can be assessed and related to biological outcomes. To this end, several physical and chemical characteristics of a diet must be considered in the development of a diet that will be suitably consumed. Physical characteristics include size, shape, texture, hardness, stability, and handling of a feed pellet or flake. Rounded or cylinder-like pellets are the most common shape produced by commercial extrusion or pelleting processes. However, commercially produced pellets are generally too large for all but adult zebrafish to ingest (400 μm diameter is generally at the lower end of the size range for most extruders). For this reason, flake diets or pelleted “crumbles” are the most common commercial sources of zebrafish diets. For the typical nutrition experimental laboratory, extruders are not available, and diets are prepared without heat (cold processing) and ground to small particle sizes using small mills or grinders. After grinding, coarse-powdered diets of various size classes can be uniformly separated into the desired size class by passing them through a series of metal sieves of particular mesh sizes. Powdered diets as small as 30 μm diameter can be produced and are adequate for newly feeding larval zebrafish (Kaushik et al. 2011). Empirical studies that compare responses to the physical properties of diets are lacking (Goolish et al. 1999). Based largely on anecdotal observations with natural foodstuffs, chemically based characteristics such as attractability and palatability are suggested to be important in increasing the rate of feed intake (Lindsay and Vogt 2004), but evidence that supports specific classifications of these compounds for zebrafish is lacking. Additional studies are needed to further determine which dietary characteristics enhance feed intake.

Storage of diets can be a challenge for any study of nutrition. Once a diet has been prepared, it must be either fed directly or stored for some period of time before use. Wet or moist diets greater than 15% moisture must be used immediately or stored under refrigerated conditions. These diets have a reduced shelf life, usually less than a week, because they are susceptible to fungal and bacterial degradation, even in the presence of specific antifungal or antibacterial inhibitors. Dry diets (moisture content <15%) are not as susceptible to fungal or bacterial degradation. Storage in cool, dry, and dark (some nutrients photo-oxidize) conditions will lengthen their shelf life. Storage in frostless freezers or under carbon dioxide or nitrogen gas also greatly lengthens shelf life of both moist and dry diets. Commercial feeds are often packaged in airtight, lined containers and may be stable for up to a year or more after production. For these diets, oxidation of specific nutrients is a greater risk to the nutritional value of the diet. Proteins and lipids are particularly vulnerable to oxidative degradation. To minimize oxidation of these nutrients, diets may contain strong chemical antioxidants, which themselves could affect dietary requirements of nutrient antioxidants. Rancidity tests can be applied to determine the levels of oxidation for freshness determination of the diet.

Practical feeds sometimes contain specific grains or cereals, a relatively inexpensive source of many essential nutrients. However, during the production or storage of these grains, specific fungi may grow, particularly under conditions of increased water content and high temperatures, and produce aflatoxins. Aflatoxins have been shown to be very toxic to many aquatic and terrestrial species. The full response of zebrafish to dietary aflatoxins is not known, but Troxel and colleagues (1997) have shown that zebrafish can metabolize aflatoxin B1 postinjection, exhibiting some resistance to hepatocarcinogenicity. In addition, Spitsbergen and Kent (2003) found that all life stages of wild Florida zebrafish were remarkably resistant to aflatoxin B1, but little is known about inbred research strains held in long-term culture. Doubt about the quality of a diet should always result in its discard.

Basic Nutrient Requirements

Many organisms share common metabolic pathways and, therefore, will have similar nutrient requirements. Table 3 presents a synopsis of those nutrients that will be required in the diet. The quality and quantity of classes of proximate nutrients (e.g., proteins, carbohydrates, neutral and polar lipids) will differ with source (e.g., menhaden, soybean, casein, starch). Most of these potential nutrients have not been evaluated in zebrafish, but it is likely there will be a dietary requirement for many of these nutrients, and it is also likely that excessive levels of some of these nutrients will be toxic and either reduce growth rates significantly or cause mortality.

Few studies have specifically evaluated protein requirements in zebrafish (Zhu et al. 2009). Most information on protein requirements is for other cyprinids, with suggested levels of 30-53% protein (Ulloa et al. 2011) and with minimal reporting as to the quality of the protein (Carvalho et al. 1997; Ostaszewska et al. 2008). Protein (amino acid) requirements should be an important area of future research. Carbohydrate requirements have not been determined; however, Robison and colleagues (2008) reported that a lack of dietary carbohydrate decreased growth rates and affected the transcriptome. They also stated that growth rates of zebrafish were unaffected by high levels of dietary carbohydrates (up to 35% starch), but body composition did change. Lipid requirements have been minimally evaluated in zebrafish. Meinelt and colleagues (1999; 2000) stated that both n3 and n6 fatty acids are required for growth, egg production, and fertilization success. Jaya-Ram and colleagues (2008) reported that fatty acid profiles of various organs changed in response to changes in the source of dietary lipids. They also suggested that a balanced ratio of n3 to n6 sources of lipid is necessary for optimal egg production and hatching rate. Stoletov and colleagues (2009) found that excessive dietary cholesterol negatively impacts physiological function. Hypercholesterolemia was induced in zebrafish fed a commercial diet enriched with cholesterol, and high weight gain, high triglycerides, and plaque formation in vascular tissues were also observed. Fiber and ash requirements (or tolerances) have not been evaluated; however, Robison and colleagues (2008) varied levels of dietary alpha-cellulose and silica (each up to 17.5%) and observed minimal effects on growth. These results are valuable to the preparation of dietary formulations for research because these indigestible nutrients (when used as bulking agents) can be used to manipulate quantities of digestible nutrients.

Mineral and vitamin requirements are not fully characterized but should be similar to those found in related aquatic species, particularly other cyprinids. Padgett-Vasquez and colleagues (2008) reported that juvenile growth was not affected when dietary calcium was limited and levels in the water were <1 ppm, indicating that the uptake of calcium from the water is very effective and most likely accomplished through the gills. Nutritional requirements of micro-minerals have not received much attention, with most microminerals evaluated exclusively to understand their role in toxicity. Nutritionally, zinc has been implicated as an important dietary micromineral for its role in inducing cell differentiation in the gills of zebrafish (Zheng et al. 2010). Adult zebrafish fed diets depleted in boron produced gametes that, when fertilized, failed to produce viable embryos, as compared with those produced by parents fed boron-supplemented diets (Rowe and Eckhert 1999). Mehrad and colleagues (2011) reported that inclusion of ascorbic acid at levels of 1000 mg/kg of feed increased growth rates, body weight, and fecundity in zebrafish. Vitamin D and its metabolites circulate in the blood of the zebrafish, and receptors for this vitamin are widely distributed in many tissues (Craig et al. 2008), suggesting its potential importance as a dietary nutrient (Miller et al. 2010). Lebold and colleagues (2011) fed zebrafish diets that were deficient in vitamin E (alpha-tocopherol); survival of offspring of deficient adults was lower, malformations resulted, and fatty acid profiles were altered, indicating the importance of vitamin E as a nutrient. In addition, Utomo and colleagues (2008) determined that the provision of dietary vitamin E increased fecundity, fertilization rate, and hatching rate. Vitamin E also protected against the toxic effects of PCB 126 in developing embryos (Na et al. 2009). Retinoic acid, which is derived from vita-min A, induced successful development of the zebrafish intestine (Nadauld et al. 2005). The role of pigments such as carotenoids (other than as retinoids) or xanthophylls has not been adequately investigated; the potential value of these pigments as provitamins and/or antioxidants (Haga et al. 2008) requires further study.

Very good evidence of the qualitative requirement of some vitamins does exist. However, quantitative requirements of vitamins, whether alone or in combination, can change depending on the initial nutritional, physiological, and reproductive state of the individual as well as environmental conditions during experimental trials.

Feed Management

Feed management, the criteria used in feeding a diet, has received widespread consideration as a factor affecting daily nutritional requirements of zebrafish (Lawrence 2007; Reed and Jennings 2010; Westerfield 2007). Despite the acknowledged importance of feed management practices, the amount of labor and time often dictates feed management criteria and not necessarily the nutritional requirements of the organism. Yet proper feed management practice is important and should be designed with consideration of the nutrient and physical properties of the diet. Most studies report both the feed amount and daily/weekly feeding regime (feed ration and rates), but direct effects of feed management criteria on specific outcomes have seldom been investigated. For the determination of specific nutrient requirements, appropriate feed management criteria must be determined. These criteria include feed ration, feeding rate, feeding time, multiple feedings/rations, preexposure/sensitization, sex differences and differentiation, and the role of intestinal flora.

Feed Ration

Feed ration is the amount of diet proffered per individual or per group of individuals. Feed rations are usually proffered as a specific quantity (e.g., grams per individual or percent of body weight). Feed ration can be based on an “as fed” basis (total weight of the diet proffered) or by dry weight (moisture removed from diet). Both expressions of feed quantity are important and should be reported. In most nutrition studies, an ad libitum ration is usually provided. However, an ad libitum ration can introduce new factors that confound the results, including leaching on uneaten food and mixing of feeds with fecal material in the bottom of the tank. An alternative would be to provide a ration that would be less than satiation. A subsatiation ration yields lower growth rates but makes it possible to have a direct measure of feed intake. Whenever possible, the amount of feed ingested, generally over a predetermined time period, should be reported. The ability to quantify feed intake is essential in determining daily quantitative nutrient requirements. Accurate measurements will allow the determination of the amount of a specific nutrient consumed under specific culture conditions. Quantitative determination of feed intake by zebrafish will be a challenging task and should be estimated based on several measurements conducted over a period of days to compensate for daily variation. Accurate estimates of feed intake (addressed below) should be an area of active research for zebrafish nutritionists.

Feeding Rate

Feeding rate, the frequency of providing a feed ration, is the ration per individual per unit time (e.g., grams per individual, grams per unit of body weight, or grams per percent body weight per feeding). Feeding is commonly reported on a per day basis. Unfortunately, feeding rates are often practically determined by the level of an investigator’s costs (i.e., the labor involved in conducting the experiment). However, feeding rate is suggested to have a significant effect on zebrafish nutrition, particularly if formulated diets are used (Carvalho et al. 2006; Eaton and Farley 1974; Goolish et al. 1999). Feed rations can be provided once per day or several times over the day or night. There is considerable interest as to the appropriate feeding rate, and feeding rate should be investigated for various age and size classes of zebrafish. Nutrient leaching is an important consideration when using formulated diets in water; consequently, feeding rates must be adjusted to reduce the time that a diet is in the water before it is ingested. Multiple daily feedings combined with a diet that possesses good physical integrity may be the best approach to optimize feed rate. Feed ration and feeding rates should not be confused with feed intake because feed intake requires accurate measurements of the amount of fed diet that is actually consumed (ingested).

Feeding Time

Feeding time refers to the time of day (or night) when the diet is provided. Diets can be fed to zebrafish at any time of the day or night; however in most labs, feeding time is determined by convenience (i.e., when labor is available). No general consensus exists about the appropriate time to feed in relation to feeding rate in most zebrafish studies. It is possible that feeding in the laboratory may be optimized at periods associated with changes in the light-dark cycle or based on circadian rhythms, but data to support optimal feeding times are not available. There is evidence that fish behavior can be entrained to specific feeding times (Comperatore and Stephan 1987) and may ultimately affect feed intake. The specific time(s) of feeding should be standardized and reported for each study. The combined effects of feeding time and feed ration (number of times and ration of feeding per unit time) needs further study.

Multiple Feedings/Rations

If more than one ration is fed per day, it is important to consider the amount of feed fed relative to the time of feeding. Although the effect of time of feeding has yet to be well defined in zebrafish, ingestion of a ration may vary according to the time of the day. Thus, a ration would need to be modified to optimize ingestion at specific feeding times. There are no reports on preferential feeding times in zebrafish; however, zebrafish have been shown to develop food-anticipatory activity when fed on a fixed schedule (Blanco-Vives and Sánchez-Vásquez 2009).

Preexposure/Sensitization

The nutritional and physiological state of a zebrafish at the beginning of a nutrition study will most likely affect the outcome of the determination of the requirement for a specific nutrient. Zebrafish can store significant nutrient reserves in their alimentary canal organs, muscle tissue, and gonads. For example, many minerals and fat-soluble vitamins can be stored in tissues and released as these cellular stores are mobilized. As such, investigations may require a conditioning period wherein the nutrient reserves are depleted prior to the initiation of an experiment. Consequently, a period of starvation (or feeding with a diet limited in a specific nutrient) may be required to limit the tissue availability of a specific nutrient in a subsequent experimental trial. Whenever possible, nutrient availability in these organs before and after nutrient requirement determinations or nutrient restriction should be determined, particularly for vitamins and microminerals.

Sex Differences and Differentiation

It is well known that growth rates, asymptotic size, and body composition differ between sexes. Sexual dimorphism is apparent within several weeks of growth, and proportional allocation of energy to somatic or gonadal growth will most likely differ between the sexes. Basic nutrient requirements should be very similar, but caloric and some nutrient requirements could change depending on the reproductive state of the individual. Further studies would be valuable.

Sex differentiation is not fully understood in zebrafish. All zebrafish initially have undifferentiated ovary-like gonads, but by 30 days posthatch, all oocytes disappear from male gonads, and males undergo testicular differentiation (Takahashi 1977). Studies have shown that the sex of a zebrafish can be induced or changed by chemical manipulation (Andersen et al. 2003; Uchida et al. 2004). Consequently, some practical ingredients in semipurified diets may contain non-nutrient chemicals (e.g., isoflavones) that induce undesired sex changes in zebrafish, possibly resulting in altered basic metabolic pathways and changes in nutrient (or caloric) requirements. Therefore, care must be exercised in the choice of dietary ingredients that contain those non-nutrient compounds found to exert such unintended effects. Use of purified ingredients in diet formulations should largely eliminate this concern.

Role of Intestinal Flora

Intestinal flora are important in development of the alimentary canal and associated anlagen. At 5 days postfertilization, the intestine is fully functional and zebrafish begin to feed exogenously. At this time, lipid and protein macromolecule uptake occurs in an intestine that contains normal microflora (Farber et al. 2001; Wallace et al. 2005). The initial colonization of the zebrafish intestine occurs concurrently with formation, and proper differentiation of intestinal cells is dependent on the presence of microflora obtained from their environment (Bates et al. 2006). If appropriate microflora are not present in the intestine, specific aspects of cell differentiation are arrested, and the uptake of protein macromolecules and possibly other nutrients is affected (Bates et al. 2006). Rawls and colleagues (2004) reported that 212 genes in the zebrafish are regulated by the intestinal microflora, suggesting considerable interactions between specific microflora and cell signaling pathways and the onset of cell differentiation. Functions of specific microflora taxa in the zebrafish intestine are not known, but variations in the microflora taxa occur among facilities (Roeselers et al. 2011). Interestingly, micro-flora of fish from rearing facilities that had historical connections are similar. Despite variation in domesticated fish, the microflora of domesticated fish remains strikingly similar to that of recently caught wild fish (Roeselers et al. 2011). Therefore, selection for specific bacterial taxa appears to be independent of the host environment or domestication status. Furthermore, it is suggested that the level of feed ration—and potentially diet content—is likely to affect host physiologic processes as a function of the microbial community (Rawls et al. 2006). The role of bacteria in the intestine and their response or the consequence of their response to specific dietary nutrients need further investigation.

Outcomes and Metrics

Outcome assessment requires that specific outcomes be evaluated in reference to the determination of specific nutrient requirements. The specific outcome being assessed will be dependent on the question or goal of the study, but several outcomes are used consistently in many studies, thereby allowing nutritional comparisons. As such, there are specific metrics that ideally should be reported, some of which are currently not within the realm of determinations with zebrafish due to their small size.

Feed Intake

It would be desirable to report feed intake when possible. The absolute requirements for specific nutrients cannot be accurately determined unless this intake is measured. It can be reported as wet or dry matter consumed or in terms of organic material, inorganic material, energy, or a specific nutrient. If zebrafish are fed ad libitum, then care should be used in estimating feed intake because nutrient leaching will adversely affect the recovery of any uneaten food collected for analysis. Despite the need for feed intake measurements, this task will be very difficult with zebrafish. Their small size makes it extremely challenging to determine intake for individuals. Most labs provide a ration ad libitum, so intake cannot easily be determined. It may be possible to evaluate feed intake for a defined population (representing an average per individual), but social interaction (e.g., competition) would have to be considered as a potentially confounding factor. It may also be possible to feed a subsatiation ration, thereby ensuring a direct and accurate measurement of feed intake. Oka and colleagues (2010) were able to estimate intake in overfed, maintenance, and calorie-restricted diet treatments by feeding known numbers of Artemia nauplii and counting the number of nauplii remaining after feeding; however, one must appreciate the amount of time and care that was required to determine feed intake in this or similar studies. In short, the ideal desire to have this important metric will be mitigated by the realistic difficulty in measuring feed intake. Most studies may only be able to provide estimates (if at all); however, even estimates would provide important information for nutrition studies.

Digestibility

Feed intake of wet or dry diet does not necessarily indicate that a specific nutrient was available biologically to the individual. Thus, digestibility (often referred to as absorption efficiency) determinations are required for the evaluation of a specific nutrient or nutrient class. Digestibility refers to the amount of nutrient absorbed by the individual and is generally calculated as the amount of nutrient consumed minus the amount of nutrient retained in the feces. Direct determination (gravimetric collection of all feces for analysis) is an effective determinant of apparent dry matter digestibility in many species, but would be a time-consuming and difficult procedure in zebrafish due to the small size of the fish and the solubility of the feces (as well as the difficulty in collection). Indirect measures of digestibility might include the use of specific markers, such as acid-washed diatomaceous earth, chromic oxide, or other indigestible markers. Studies must also consider that various forms (organic vs. inorganic conjugates, practical vs. purified) and sources (different vendors, batches, or lots) of specific nutrients or nutrient classes may vary in digestibility. Consequently, empirical determinations of specific organic and inorganic nutrients are required to fully estimate nutrient digestibility.

Problematically, digestibility analyses will be a challenge in zebrafish and may require population determinations (grouped fish evaluations). Awareness of digestibility and its consequences is important; however, we must be realistic in our efforts to evaluate metrics such as digestibility because accurate determinations are not currently possible and will require the development of new technologies and procedures not currently reported.

Survival and Growth

Survival and growth are common outcomes measured when evaluating daily nutrient requirements. Survival has limited value as a metric for determination of daily nutrient requirements and generally only reflects a response to an extreme limitation or toxicity of a specific nutrient. It is useful in assessing overall conditions of the culture system, and reference feed and control populations should have greater than 90% survival. Organismal growth is the most common metric used to estimate nutrient requirements, but limitations do exist. Weight gain (dry matter or organic matter weight gain) is a good estimation of growth or production and can be expressed as absolute weight gain (grams) or as a percentage of the initial weight (percent increase). For the use of weight gain as a comparative outcome, the assumption is that the gain is not attributed to proportionally different amounts of water content that might be treatment dependent. Accordingly, a proximate analysis of zebrafish tissue should accompany every experiment in which growth is presumably due to the accretion of organic matter. Specific growth rates (percent body weight gain per day, calculated as [(ln final weight – ln initial weight) / time (days)] × 100) can be used to estimate organismal growth rates but are restricted to the known period of exponential growth. These growth rates also have comparative restrictions among experiments, unless the fish are the same size and the water temperature of the experiments is the same. Linear measurement (such as total length) is one of the most common ways to estimate growth, partly because it is usually less invasive and can be conducted by image analysis. Linear measurements are often recommended because larval and early juvenile zebrafish are particularly sensitive to handling stress associated with weight measurements. Length measurements can be coupled with other single-dimension measurements of anatomical variations such as abdominal height. Linear measurements can sometimes be used to develop body condition indices (Siccardi et al. 2009). Despite their popularity, linear measurements do not reflect multidimensional growth and, in some cases, must be limited to single sex evaluations.

Parichy and colleagues (2009) provided a detailed morphological analysis of growth and development of important anatomical features in zebrafish. This excellent treatise can provide insight into various metrics that could be evaluated in response to altered nutritional states.

Production Efficiencies and Energetics

As indicated previously, production represents increases in dry or organic matter and is a common metric used in the evaluation of specific nutrients. In addition to production, the rate or efficiency of production can be a valuable metric for evaluating daily nutrient requirements. A common efficiency measurement in many species is the feed conversion ratio. This is the ratio of weight of the feed proffered (as fed) to the amount of wet fish biomass produced. A low feed conversion ratio is desirable, and an optimal diet would potentially produce the most biomass per unit feed. Production efficiency is usually calculated as dry tissue produced relative to dry food consumed and expressed as a percentage. Production efficiency has excellent utility for estimating the efficiency of converting food to biomass (an alternative would be organic tissue matter produced relative to organic food matter consumed). However, this metric requires accurate measurements of feed intake, which we have indicated is very difficult to assess in zebrafish. Furthermore, a high production efficiency may not always be associated with a healthy fish because an obese fish could score comparatively high in production efficiency but low in health assessment. Consequently, production efficiency would be most useful if coupled with another suitable metric (e.g., body composition, organ pathology). Size constraints in zebrafish will provide a challenge for the use of efficiency metrics.

In many animals, the energetic value of specific foodstuffs can affect consumption and, consequently, production and production efficiency. It has been observed anecdotally that zebrafish can be satiated, but whether satiation is induced by energy intake, protein intake, or another metric is not known. Reduced feed intake caused by the provision of high-energy or high-protein food ingestion can significantly reduce the ingestion of other specific nutrients in the diet, thereby limiting production. Energy levels in relation to protein content or other nutrients may be an important index that affects production and production efficiencies and should be a subject of future investigation. The protein-to-energy ratio (sometimes reported as the energy-to-protein ratio) is an important metric when the proximate nutritional and energetic quality of feeds are evaluated. The protein-to-energy ratio is usually reported as milligrams of protein per kilojoule of diet or milligrams of protein per kilocalorie of diet. As a nutrient, protein nitrogen does not appear to be limiting in most commercial zebrafish diets, but empirical evidence is not available. Not all sources of protein are easily digestible or metabolizable (Ostaszewska et al. 2008), so caution must be used in interpreting the contribution of protein nitrogen in various feeds.

Organ Analyses

Because specific nutrients can affect nutrient allocation to specific tissues or organs, evaluation of organs can be useful in determining nutrient requirements (and toxicities). The most prominent organs in terms of both size and relation to nutrition are the liver, intestine, muscle, and gonads. Organ weights can be compared directly among treatments, or indirectly through the use of organ indices or ratios (organ indices and ratios must be compared with caution because incorrect interpretation can occur when allometric relationships are found; Packard and Boardman 1999). Of course, the greatest challenge comes from the dissection of zebrafish organs because weights of these small tissues are difficult to measure accurately. For this reason, histological examination of these tissues may provide insight into the value or requirement of a specific nutrient. Recovery of the entire intact tissue is not required for histological analyses. For this approach to be successful a “normal” (healthy) histological record must be determined for each tissue. Several studies have evaluated these tissues (Oka et al. 2010; Tiso et al. 2009; Wallace et al. 2005; Wallace and Pack 2010), but a definitive “normal” control is difficult to establish. For example, most diets used in zebrafish studies are nutrient dense and have high caloric values. As such, many fish in these cultures may, in fact, be obese and would therefore exhibit histopathologies associated with an obese individual rather than a healthy individual (Oka et al. 2010). Fortunately, it is suggested that zebrafish possess adipocyte demographics that are similar to those found in humans (Flynn et al. 2009). An evaluation of a healthy natural population should be beneficial and the condition associated with the feeding of a reference diet must be determined for comparisons to be valid.

Physiological Metrics

The most common metrics used to evaluate daily nutrient requirements involve growth or growth-related indices (i.e., metrics related to morphological determinants). However, many formulated diets will contain adequate (but not optimal) quantities of nutrients, particularly micronutrients, and significant differences in growth metrics will not be observed. Although growth metrics may not be affected by some nutrients, significant physiological consequences may be manifested in response to dietary levels of these nutrients. Accordingly, differences in metabolic pathways, reproductive success, or in the response to stress and disease (e.g., production of reactive oxygen species) may exist. As previously stated, additional outcomes that measure the functional consequence of limited or toxic levels of specific nutrients will need to be developed in zebrafish, and may include stress or disease tolerance as well as those mechanisms important in normal physiological function.

Reproduction

Reproductive output (fecundity) is very important and is one of the most desired outcomes of zebrafish culture. However, few studies have causally evaluated specific diets (natural or formulated) and their effects on production and fecundity (Markovich et al 2007; Meinelt et al 1907). For females, it is possible to evaluate fecundity based on the number of fertilizable eggs (exhibiting normal development) produced per unit time. For males, the ability to mate and produce sperm to fertilize eggs successfully would be a desired marker. Different labs have different expectations and approaches for determining reproductive output (there are no universally accepted procedures), and a diet that allows a zebrafish to mature quickly and produce viable gametes is desired. Although a short time to maturation (egg to egg culture) is desired for most laboratory approaches, precocious gamete development may provide embryos and offspring that, over generations, may exhibit problems associated with increased rates of growth and maturity (e.g., epigenetic influences). There is inadequate genomic information on offspring assessment at this time.

Molecular Assessment

The current availability of many genomic markers will serve as an important tool for evaluating the effects of nutrition or specific nutrients on many physiological or cellular responses (reviewed by Ulloa et al. 2011), including polygenic responses such as growth (Wright et al. 2006). Chip technologies will allow researchers to evaluate specific nutrients on a variety of metabolic pathways. Proteomic and metabolomic approaches may prove valuable as well (Gómez-Requeni et al. 2010). Epigenetic effects of diet have been observed in cardiovascular performance of zebrafish (Schwerte et al. 2005) and will most likely be important in other physiological systems. Systems biology approaches may allow us to model the interactions among the metabolic pathways. In addition, differences between sexes in terms of their molecular response must be considered (Robison et al. 2008). Application of these various molecular approaches using this important animal model should provide important insight into the effects of diet and the specific nutrients therein on normal and pathological function.

Statistical Design and Evaluation

Nutritional research must aptly include good and appropriate statistical design and analysis. A good design must have identified a priori outcomes to be evaluated by standard statistical assessments. As is often the case, the desire to evaluate various parameters a posteriori may also exist, and many important observations can be made during the course and at the final stages of nutrition research. A full treatise on experimental design is well beyond the scope of this paper, but several important considerations should be mentioned.

Although group or pooled observations provide insight into specific outcomes, providing sufficient replicatees is necessary. Whenever possible, individual observations should be part of the experimental design in nutrition research because individual observations increase the power of the statistical test. Zebrafish can be variable in their responses, so knowledge of individual variation can help define the precision and accuracy of outcome assessments. Individual observations can allow the use of analysis of covariance to determine the effect of body size and treatment on specific organ outcomes. However, individual observations are time consuming and are not always easily accomplished in all laboratories. For those of us who are not gifted with statistical understanding, whether using individuals or pooled observations, consultation with appropriate experts in the field is desirable in both the design and evaluation of nutritional outcomes.

The future of nutritional outcomes resides in the use of appropriate molecular tools, including genomics, transcriptomics, proteomics, metabolomics, and systems biology. The cost of these tools often dictates the number of fish (or tissue samples) that can be analyzed. For use of these molecular tools, consultation with a statistician would be of benefit for any investigator evaluating nutrition.

Conclusions: What Is the Importance of Diet Development in Zebrafish?

“The ability to replicate research is fund amental for good science” (Barnard et al. 2009).

Barnard and colleagues (2009) explain clearly the rationale for standardized diets in laboratory animals. For the last three decades, mammalian animal model researchers have recognized that “good” science requires limiting experimental variability ideally to that which exists endogenously in the organism. As these researchers and other scientists have recognized, through the use of a standardized diet, control over one of the most important sources of environmental variability is achieved. The current and inevitable variations arising from the use of commercial or other live and formulated diets among zebrafish researchers will soon encounter resistance from funding agencies that promote this model for biomedical research. Future studies will most likely require that strict dietary regimes be established for zebrafish stocks so that the foundation for comparable data sets can be determined. Similar dietary regimes are available for mammalian models and, in the United States, have been approved by institutions including the American Institute for Nutrition and the National Academy of Sciences. Additionally, a recent contribution by the National Research Council (2011) will aid in the development of appropriate diets.

The future benefits of scientific studies conducted with zebrafish depend, in part, on the development of a standardized reference diet. Such a diet will not meet the requirements of every research program but can be modified for the specific purposes of the researcher. From these reference diets, open-formulation diets can be produced and serve as a dietary standard across all studies, particularly across various strains within and among zebrafish laboratories. There is adequate information among cyprinid diets, and, with additional research, a standard reference diet can be a reality. Commercial diets (closed formulation, ingredients are proprietary and can potentially change without notification) have some utiliy but are not adequately controlled to allow for comparisons of datasets. As indicated in the Guidelines for the Uses of Fishes in Research (American Fisheries Society et al. 2004), commercial feeds are formulated not on a specific and consistent list of ingredients but rather to meet broad nutrient requirements for protein, carbohydrate, and fat. Specific ingredients will vary from batch to batch. Consequently, the capability of these feeds to meet specific nutrient requirements consistently is not possible, and the impact of this variability on outcomes measured for experimental organisms must be considered in the design of studies (Barrows and Hardy 2001).

This paper is an introductory primer to the basic methodology needed to achieve an understanding of zebrafish nutrition. Currently, we are in the very early stages of understanding zebrafish nutrition and diet development; however, a standard diet is essential for conducting further research if zebrafish continue to serve as a biomedical research model. A standardized diet can be used while nutrient requirements are being determined, which will ultimately refine the composition of the standardized diet. This will enhance its value as a means to introduce environmental control so that the reported results of experiments will no longer be in question. Refinement will assist in defining the roles of specific nutrients in organismal, physiological, and molecular responses.

Investment in the formulation and testing of potential standardized diets, followed by determination of the nutrient requirements of zebrafish, is no longer just an idea; it is timely, relevant, and essential. The ultimate definition of a nutritionally complete diet will be important for maximizing the benefit of this important scientific model for all levels of research.