The scales of the zebrafish: host-microbiota interactions from proteins to populations

Abstract

The interactions between animal hosts and their associated microbiota can be studied at multiple spatial and conceptual scales, with each providing unique perspectives on the processes structuring host-microbe systems. Recently, zebrafish, Danio rerio, has emerged as a powerful model in which to study these interactions at many different scales. Controlled but simplified gnotobiotic experiments enable discovery of the molecules and cellular dynamics that shape host-microbe system development, whereas population level investigations of bacterial dispersal and transmission are beginning to reveal the processes shaping microbiota assembly across hosts. Here we review recent examples of these studies and discuss how the results can be integrated to better understand host-microbiota systems.

Graphical abstract

Introduction

Animals and their resident microorganisms form complex biological systems of multiple interacting elements. The potential importance of these interactions in determining the health of humans and other animals has motivated efforts to better manipulate and predict their behavior. Understanding these interactions is challenging, because they operate at multiple scales, from the level of molecular interactions between microbial and host cells to the transmission of microorganisms across both populations and generations of hosts. Studies of other complex biological systems and naturally occurring ecological communities have concluded that the scale at which they are observed can fundamentally alter interpretations about the nature and importance of various mechanisms and processes driving the system [1,2]. Thus, any approach that focuses solely on interactions occurring at a single scale, such as either direct cell to cell interactions within a single host or changes in microbiota composition across populations of hosts, will likely lead to an incomplete or even misleading understanding of these systems as a whole.

In recent years, the zebrafish, Danio rerio, has emerged as an excellent model system in which multiple levels of interactions can be analyzed together to build a comprehensive understanding of how different mechanisms act across scales. For cellular level studies of host-microbe interactions, the zebrafish model affords sophisticated genetic tools and resources for marking and manipulating different cell types. Equally important is the ease with which hundreds of individual embryos can be derived germ-free and reared gnotobiotically with a defined or absent microbiota [3]. The optical transparency of zebrafish larvae provide the unique opportunity to observe the spatial and temporal dynamics of microbial populations alongside the host’s response [4]. At the population level of host-microbe studies, the characteristic variability of animal microbiomes can be overcome through the large degree of biological replication afforded by zebrafish’s high fecundity, allowing for robust analysis of the effects of host factors on the composition of the microbiota. Finally, combining these features with the control over the hosts’ physical environment, zebrafish husbandry opens up opportunities to study broad scale interactions occurring across multiple individual hosts. Like any model system, the zebrafish has its limitations when it comes to modeling human biology. Zebrafish are aquatic animals that interact with their microbial world through the medium of water rather than air and terrestrial surfaces. They lack lungs and mammary glands, limiting their utility for modeling host-microbe interactions that are shaped by these organs, and they only develop their adaptive immune systems during juvenile stages. They are cold blooded and typically reared at 28 degrees Celsius, thus restricting their capacity to be colonized by microbes specialized to warm blooded animal temperatures. Despite these limitations, they have proven to be a powerful model to explore the interactions between animal hosts and their microbial partners at multiple scales. In this review, we discuss how using the zebrafish to study host-microbe interactions across multiple scales has led to novel findings and promising avenues of future research.

Molecular interactions between microbial and host cells

The zebrafish is a powerful model for studying host-microbe interactions at the tissue, cell, and molecular level to identify mechanisms that maintain homeostasis between hosts and their resident microorganisms and the consequences of this interplay on host organ development and differentiation. Comparisons of germ-free larval zebrafish with those colonized with a complex microbiota have repeatedly demonstrated effects of resident microorganisms on host intestinal cell differentiation [5], epithelial cell proliferation [6,7], intestinal absorption of fatty acids [8], and immune responses [6,9,10]. In addition to inducing changes in the intestinal track and the immune system, both of which come into frequent contact with microorganisms and microbial products, there is mounting evidence of microorganism-dependent impacts on the development of other organs. For example, our group found that microbiota are required for normal expansion of pancreatic β-cells in larval zebrafish and we discovered a single protein produced by specific zebrafish gut bacteria that is sufficient for this expansion in germ-free larvae [11]. Furthermore, homologs of this protein, which we named Beta cell expansion factor A or BefA, are found in microorganisms isolated from human microbiota and show a similar ability to induce β-cell expansion in zebrafish, suggesting this function has a conserved origin. The discovery of BefA provides a clear example of how the gnotobiotic zebrafish system can be used to simultaneously uncover the consequences of host-microbe interactions on animal physiology as well as identify the specific molecular mechanisms underlying those interactions.

Interactions between defined microbiota and their hosts

Zebrafish have also proven useful in studying the colonization and dynamics of populations of microorganisms in hosts. By combining gnotobiotic husbandry with a high-throughput transposon mutagenesis screen of zebrafish-associated bacterial strains, Stephens et al. [12] characterized the bacterial gene functions predictive of colonization and population sizes in vivo. Doing so highlighted the importance of bacterial motility in particular, both in the form of chemotaxis and flagellar-mediated motility and adhesion, to intestinal colonization and invasion of already established populations. An additional study utilizing gnotobiotic zebrafish further supported the importance of bacterial adhesion in the establishment of specific bacterial strains, particularly those with potential probiotic properties [13].

The optical transparency of larval zebrafish provides unique opportunities to further study the motility and spatial distributions of bacterial populations in vivo. Early work imaging Mycobacteria marinum infection of macrophages fundamentally changed the idea of tuberculosous granulomas as static structures of host immune containment and demonstrated that they play a role in pathogen dissemination [14,15]. The adoption of light sheet microscopy for live imaging of bacterial dynamics in zebrafish has allowed for measurements of bacterial populations along the entire intestine with high temporal and spatial resolution [4]. Using these techniques to characterize colonization dynamics of fluorescently labeled strains has highlighted a diversity of bacterial colonization strategies. For example, Aeromonas veronii, a common resident of the zebrafish intestine, was found to form clustered aggregates of cells that grew at faster rates than individual planktonic cells, suggesting a strategic advantage to growth in aggregates [16]. Interestingly, a second Vibrio species that exhibited a highly planktonic growth, was found to competitively exclude Aeromonas when the two were co-colonized [17]. This exclusion was largely mediated by the different responses of the species to host intestinal motility: the Aeromonas aggregates were regularly expelled from the intestine by peristalsis whereas the more diffuse populations of Vibrio where able to resist these perturbations and prevent the recovery of Aeromonas. The important role of the host in this competitive exclusion was demonstrated by the finding that both species were able to persist together in ret mutant zebrafish lacking normal intestinal motility. These studies reveal potential trade-offs between planktonic versus aggregated bacterial cell strategies for growth and persistence in the intestine and possibly inter-host transmission.

Interactions among taxa of bacteria not only alter their independent growth dynamics, but also their potential effects on their hosts. A screen of multiple species of zebrafish-derived bacteria in monoassociation with larval zebrafish revealed variation in the overall neutrophil response [18]. This relationship between neutrophil response and bacterial abundance also differed among species, with some, such as Vibrio sp., exhibiting a positive monotic relationship, others, such as Shewanella sp., exhibiting a negative relationship, and finally some, such as Aeromonas sp., showing no clear relationship. Interestingly, each species had disproportional effects on neutrophil response, such that predicting the host response to a complex microbiota requires knowing both the abundance and per capita effect of each species. To a first approximation, these same rules were found to apply to much more complex microbiota that assembled in a mutant zebrafish line, sox10, that has defective peristalsis and develops spontaneous intestinal inflammation [19]. The extent of inflammation varied across mutant siblings reared in a shared aquatic environment and could be explained by the relative abundances of pro-inflammatory Vibrio sp. and anti-inflammatory Escherichia sp. in individual intestines. These studies suggest that inferences from simple, defined communities can be scaled up to the complex ones that naturally inhabit vertebrates.

Interactions between complex microbiota and their hosts

The resident microbiota of zebrafish consist of hundreds of microbial taxa and is typically dominated by members of the Proteobacteria phylum, followed by Fusobacteria and Actinobacteria [20–22]. In addition, zebrafish can also be colonized by a wider range of microorganisms than those normally found associated with them, including taxa isolated from mammalian and human hosts [23]. This property has facilitated analysis of potential human probiotics, including several strains of Lactobacillus, in larval gnotobiotic zebrafish [24]. Foundational work utilizing zebrafish as a model to study host-microbe interactions has shown that the composition and distribution of microbial lineages in the zebrafish microbiota is in large part selected by the zebrafish host. Following the reciprocal transplant of mouse microbiota into germ-free zebrafish recipients, bacterial taxa from the donor mouse microbiota were able to successfully colonize larval zebrafish, but the relative abundances of these strains was altered to match the relative abundance of shared lineages in conventional zebrafish microbiota [20]. Thus, despite differences in the microbial inoculum, the zebrafish gut environment selects for a specific distribution of bacterial lineages. The importance of host habitat selection is further supported by the observation that zebrafish raised in laboratory facilities still retain strong, broad level similarities in microbiota composition to wild caught zebrafish, and distinct from other closely related fishes, despite large differences in geography and environmental exposure [21].

Some of the most dramatic changes that happen in animals occur during their development from embryos to fully mature adults. Initial small scale studies found that zebrafish development was a strong predictor of bacterial diversity [25]. A larger, more densely sampled study found similar changes in bacterial diversity and an increase in inter-individual variation from larval to adult stages, suggesting developmental changes in the processes by which hosts select their microbiota [22]. In addition, the zebrafish microbiota became less similar to the microbial communities in the surrounding environment and food over time. To further investigate this relationship between host development, microbiota, and diet, Wong et al. analyzed the intestinal microbiota of zebrafish maintained on one of three constant diets that differed in their fat content [26]. Dietary fat levels had distinct effects on the intestinal microbiota that changed throughout host development and manifested through changes in the assemblages of multiple bacterial species. Furthermore, dietary fat levels also had a substantial effect on environmental communities of bacteria, suggesting that diet may also change hosts’ environmental exposures.

Zebrafish have provided a powerful high-throughput system in which to screen and study the effects of environmental exposures on vertebrates for many of the same reasons that they make attractive models to study host-microbe interactions [27], and this can further be extended to study the additional effects on the microbiota. For example, Gaulke et al. characterized the microbiota of zebrafish exposed to the antibacterial agent triclosan, and detected significant changes in the composition and diversity of the intestinal microbiota [28]. In addition, triclosan exposure altered the correlation networks of microbial taxa, increasing their connectedness. This suggests that alterations to a host and its environment can not only alter the composition of an individual’s microbiota, but also the nature of the interactions among its resident microbial taxa.

Interactions among microbiota and populations of hosts

Animal microbiota also have the potential to be influenced by processes occurring beyond an individual host, through the migration of microorganisms from biotic and abiotic sources. For example, transmission among hosts is already known to be a key process determining the distributions and spread of pathogens in animals, and zebrafish have increasingly been used as a model to study the colonization and transmission of Vibrio cholera [29], Salmonella [30], and Mycobacterium [31]. Key to the use of zebrafish as a model for pathogen transmission is the ability to raise and house large populations under controlled environmental conditions and the ability to manipulate their exposure to microorganisms in their environment and to other individuals.

Recently, research utilizing zebrafish has attempted to expand the study of microbiota assembly from an individual host-centric to a population-based investigation. Toward this goal we employed a neutral modeling framework to test the null hypothesis that differences among bacterial species and hosts were unimportant to microbiota assembly [32]. Our analysis demonstrated that passive migration and stochastic demographic processes were sufficient to explain a large among of variation in the microbiota of a large population of zebrafish, however the relative importance of these processes decreased with host development. A major component of this development is the maturation of the adaptive immune system, which does not become fully active in zebrafish until they are juveniles [33]. Due to its potential ability to selectively identify and target specific microorganisms, we hypothesized that the adaptive immune system would limit the transmission of microbes from other hosts. Indeed, we found that ragI⁻ mutant zebrafish lacking a functioning adaptive immune system were associated with more neutrally assembled microbiota [34]. However, differences in the composition of the microbiota between wild type and mutant ragI⁻ zebrafish were subtle and weaker overall than the effect of cohousing, a phenomena that has been observed in other model systems. This suggests that although adaptive immunity may constrain colonization of zebrafish, the effects of transmission overall may be great enough to overwhelm individual host factors.

Conclusion

Increasingly expanding the scale at which host-microbiota systems are observed has revealed what appears to be the inherently noisy and variable nature of these systems. At very fine scales, studying molecular and cell-cell interactions, the impacts of an individual bacterial taxa or molecule on the host can be measured as a binary response. In these situations, the variable nature of the microbiota is merely manifested as statistical “noise” that is more an obstacle than a feature of the system. However, beyond these relatively reductionist experimental approaches, we have seen an increased role of stochastic effects, from the potentially random influence of population bottlenecks during the initial colonization of the intestine [12], to the importance of passive bacterial transmission in determining the overall composition of intestinal microbiota [32,34]. Furthermore, where fine scale gnotobiotic studies have identified specific bacterial factors inducing changes in zebrafish hosts [5,7–9,11], broader scale studies can inform the circumstances under which those interactions even have the potential to occur (for example, by determining whether or not the bacterium is even able to colonize a zebrafish [17]), and to what degree the overall effect on the host can be predicted by the population size of individual bacterial species [18,19].

A major challenge for the microbiome field today is translating broad patterns gleaned from studies of highly variable human populations into therapeutic actions for individuals. By serving as a model to study host-microbe interactions from the molecular to the population scales, zebrafish provide a potential path towards tackling this challenge. Live imagining techniques, like those described above, enable studies of temporal and spatial dynamics within an individual zebrafish, and the flexible and controllable husbandry of zebrafish allows for the design of experiments to study fine scale interactions occurring among populations of zebrafish and between zebrafish and their environment. By exploiting the zebrafish model in this way, future research can connect insights on multiple scales at the same time, thereby providing a more comprehensive understanding of host-microbe systems as a whole.

Highlights.

• Interactions between animal hosts and their microbiota occur at multiple scales

• The zebrafish is a powerful model to study many scales of host-microbe interactions

• Gnotobiotic experiments provide fine scale resolution to identify molecular interactions

• Live imaging reveals dynamics of host-microbe interactions with cellular resolution

• Population level studies explore the host and microbe traits that shape microbiota