The Zebrafish as a Model for Gastrointestinal Tract - Microbe Interactions

Summary

The zebrafish (Danio rerio) has become a widely used vertebrate model for bacterial, fungal, viral, and protozoan infections. Due to its genetic tractability, large clutch sizes, ease of manipulation, and optical transparency during early life stages, it is a particularly useful model to address questions about the cellular microbiology of host-microbe interactions. Although its use as a model for systemic infections, as well as infections localized to the hindbrain and swimbladder have been thoroughly reviewed, studies focusing on host-microbe interactions in the zebrafish gastrointestinal tract have been neglected. Here, we summarize recent findings regarding the developmental and immune biology of the gastrointestinal tract, drawing parallels to mammalian systems. We discuss the use of adult and larval zebrafish as models for gastrointestinal infections, and more generally, for studies of host-microbe interactions in the gut.

1. Introduction

Due to its fecundity, genetic tractability, small size, rapid development, and optical transparency during early development, the zebrafish (Danio rerio) has emerged as one of the most well-used vertebrate model organisms in cellular microbiology. Several reviews have focused on zebrafish as a model for infectious disease (Tobin et al., 2012; Sullivan et al., 2017; Duggan and Mostowy, 2018), but all were focused on systemic disease models or localized infections of the swim bladder or hindbrain; none discussed infections of the gastrointestinal (GI) tract. Over the last few years however, literature in this area has grown considerably, as has literature focusing on the microbiota, and characterization of the transcriptional profiles and cell types of the zebrafish GI tract, allowing us to draw comparisons to the mammalian GI tract. These studies have made it clear that the development and physiology of the zebrafish GI tract bear many similarities to humans. For example, ~70% of human genes have a corresponding orthologue in zebrafish (Howe et al., 2013), and the transcriptional regulatory networks controlling intestinal development and physiology are highly conserved between fish and mammals (Lickwar et al., 2017). These findings highlight the utility of the zebrafish model to study in vivo host-microbe and host-pathogen interactions in the GI tract. With this review, we summarize recent literature highlighting structural and molecular similarities between zebrafish and mammalian GI tract architecture and immune surveillance. We further discuss examples of studies looking at host-microbe and host-pathogen interactions in the zebrafish GI tract, highlighting experimental features and summarizing the most important findings. We hope this review will highlight the utility of the zebrafish model for GI tract- microbe studies, bridging the gap from whole animal studies to single cell and molecular, mechanistic experiments.

2. Organization and development of the zebrafish intestine

2.1. Intestinal segmentation and architecture in zebrafish

The primary functions of the intestine include digestion and absorption of nutrients, and the elimination of waste products. The zebrafish intestine is highly homologous with the mammalian intestine in its development, organization, and function (Pack et al., 1996; Wallace et al., 2005; Ng et al., 2005; Carten and Farber, 2009). Gene expression and transcriptional regulation in intestinal epithelial cells are highly conserved along the segmental regions of the mammalian and zebrafish intestines (Lickwar et al., 2017), and many metabolic functions are also conserved (Schlegel and Gut, 2015; Quinlivan and Farber, 2017). Both zebrafish and mammalian GI tracts are covered by a protective mucus layer that predominantly consists of the gel-forming mucin Muc2 as its structural component (Johansson et al., 2011; Jevtov et al., 2014). Mucin is primarily secreted by goblet cells, which in adult zebrafish are distributed throughout the gut (albeit more abundant in the middle and posterior gut) (Wallace et al., 2005; Ng et al., 2005).

There are, however, some key functional and architectural differences between zebrafish and mammalian GI systems to consider (Figure 1). The mammalian gastrointestinal tract is generally composed of five distinct parts: (1) the stomach, which partially digests food by mixing it with acid and digestive enzymes, (2) the duodenum, which aids chemical digestion, (3) the jejunum, which absorbs nutrients, (4) the ileum, which absorbs bile salts, and (5) the colon, which absorbs water and salts. The zebrafish intestine is divided into three histologically defined segments, including (1) the anterior intestinal bulb, (2) the middle intestine, and (3) the posterior intestine. This nomenclature for the zebrafish gut is typically maintained from larval to adult stages. In contrast to mammals, zebrafish lack the typical signatures of a stomach organ and the intestinal bulb undergoes no acidification. The passage of microbes through the acidic environment of the human stomach, which can reach pH values as low as 1.4 (Dressman et al., 1990), serves as a regulatory cue for some GI pathogens (Ramos-Morales, 2012), and its absence may impact the outcome of infection. Zebrafish do, however, express several digestive enzymes that are functionally equivalent to mammalian gastric markers, including rennin, nothepsin, several cathepsins, and the lipase Lipf (Wallace et al., 2005; Wang et al., 2010). The different regions of the adult zebrafish gut have distinct functions analogous to the mammalian small intestine and colon. Functionally, the anterior intestinal bulb likely plays a role in bile salt recovery (Lickwar et al., 2017), while the anterior and middle regions may aid in lipid and protein absorption (Ng et al., 2005; Brugman, 2016). Like the mammalian colon, the posterior region of the zebrafish gut is responsible for ion and water absorption (Wallace et al., 2005). Established markers of the mammalian small and large intestine (e.g., villin and fabp2) are differentially expressed along the length of the zebrafish intestine (Wang et al., 2010). Transcriptional domains identified in larval zebrafish are maintained in adults, and a model delineating five transcriptional and functional domains has been proposed, and includes ada for the anterior, duodenum-like region, fabp2 and rbp2a for the anterior, jejunum-like region, fabp6 and slc10a2 for the middle, ileum-like region, and lamp2 for the middle to posterior, colon-like region (Lickwar et al., 2017).

Figure 1.

Comparison of cell types and structures between the mammalian (A) and zebrafish (B) gastrointestinal tract. The zebrafish intestine is organized into irregular folds as opposed to villi, and lacks crypts. Mammalian and zebrafish intestines share stem cells, enterocytes, enteroendocrine cells, and goblet cells. Paneth cells and M-cells are absent in zebrafish.

The intestinal architecture of zebrafish is also less complex than that of mammals as it consists only of the mucosa, muscularis externa, and serosa layers. Underneath the mucosa and epithelium is the lamina propria containing blood capillaries, muscle fibers, and lymphatic vessels. Surrounding this layer is the muscularis externa composed of circular and longitudinal smooth muscle fibers and enteric neurons (Wallace and Pack, 2003). The lining of the mucosa is folded into large, randomly-shaped intestinal folds instead of the mammalian villi and intestinal crypts (crypts of Lieberkühn) (Pack et al., 1996; Wallace et al., 2005; Wang et al., 2010). Since zebrafish lack crypts of Lieberkühn, which in mammals are the source of intestinal stem cells, cell division instead occurs at the base of the folds and cells migrate to the tip of the folds where they become apoptotic (Wallace et al., 2005; Wang et al., 2010). The time course for cell migration in the anterior intestine is 5–7 days, while migration takes 7–10 days in the middle intestine (Wallace et al., 2005). Goblet cells are still present in zebrafish, but distributed throughout the mucosa rather than localized to crypts (Pack et al., 1996; Wang et al., 2010).

The enteric nervous system (ENS) is a functionally important component of the GI system both in zebrafish and mammals, and consists of enteric neurons and glial cells. These cells modulate several key functions including gut peristalsis, hormone secretion, water balance, and absorption. In contrast to the more architecturally complex mammalian ENS which is composed of two plexuses (i.e., myenteric and submucosal), each with their own interconnected ganglia, the zebrafish ENS develops into a single myenteric layer of neurons, glia, and other cell types (e.g., interstitial cells of Cajal) without any clearly defined ganglia (Wallace et al., 2005).

2.2. Development of the zebrafish gastrointestinal tract

Intestinal development in zebrafish larvae can be categorized into 3 major stages (Ng et al., 2005). Stage 1 is defined by formation of the lumen; a thin rod of endodermal cells that undergo anterior-to-posterior differentiation and proliferation (Wallace and Pack, 2003); by 60 hours post-fertilization (hpf) the lumen has expanded rostrocaudal (Kimmel et al., 1995; Alvers et al., 2014). During stage 2, the intestinal epithelium becomes polarized via cell type differentiation (Pack et al., 1996; Wallace et al., 2005), and by 76 hpf, enteroendocrine cells are present throughout the anterior to posterior regions of the intestine. Enteroendrocrine cells produce and release active compounds (e.g., hormones) into the surrounding tissues as well as modulate enzyme secretions and muscle contractions (Wallace et al., 2005; Takashima et al., 2013). At 76 hpf, the beginning of stage 3, intestinal folds have developed in the anterior and middle intestinal regions, and peristaltic contractions have begun (Pack et al., 1996; Wallace et al., 2005). Growth and differentiation of the digestive tract continues, and by 5 days post-fertilization (dpf), the majority of the system is functional (Wallace and Pack, 2003; Ng et al., 2005); at this point the larval zebrafish digestive tract is comprised of the mouth, pharynx, esophagus, intestinal bulb, intestine, and anal opening. Extensive folding can be found within the anterior intestine, but at 5 dpf the posterior intestinal regions have no folds (Ng et al., 2005), and by 6 to 8 dpf, cell proliferation begins to decrease (Cheesman et al., 2011). As the larval zebrafish ages, folding continues and the folds themselves become shorter in the caudal direction (Menke et al., 2011; Li et al., 2019), and the lumen begins to widen at the anterior end and becomes progressively smaller towards the posterior region (Wallace et al., 2005).

Enteric precursors originating from the neural crest migrate into the intestinal tract at 32 hpf and reach the posterior end by 66 hpf (Olden et al., 2008; Heanue et al., 2016). By 4 dpf the zebrafish gut exhibits spontaneous, coordinated contractions and the digestive system is fully functional by 7 dpf (Holmberg et al., 2007). We refer readers to excellent recent reviews of ENS function and its interactions with the various other cells types of the zebrafish intestine (Ganz et al., 2016; Ganz, 2018).

3. Immune surveillance of the gastrointestinal tract

Zebrafish have made valuable contributions to our understanding of vertebrate immunity (Martins et al., 2019). Multiple comprehensive reviews discuss the teleost immune system (Traver et al., 2003; Trede et al., 2004; Sullivan and Kim, 2008a; Meeker and Trede, 2008; Uribe et al., 2011; Renshaw and Trede, 2012; Rauta et al., 2012), and zebrafish immunity in the context of infectious disease (Sullivan and Kim, 2008b; Meijer and Spaink, 2011; Masud et al., 2017). The zebrafish and human immune systems share many similarities, including both innate and adaptive components (Traver et al., 2003; Renshaw and Trede, 2012).

3.1. Innate immunity and the GI tract

Larval zebrafish solely depend on their innate immune system to fend off invading microbes since the adaptive immune system is not functional until 4–6 weeks post-fertilization (Willett et al., 1999; Lam et al., 2004). During this time, colonization of the larval gut by commensals can prime innate immune cells against infections through a Toll-like receptor (TLR) and myeloid differentiation primary response 88 (MyD88) dependent pathway (Galindo-Villegas et al., 2012). One of the benefits of zebrafish is their optical transparency, allowing for the use of transgenic lines with fluorescently-marked innate immune cells like macrophages (Ward, 2003; Redd et al., 2006; Hall et al., 2007; Ellett et al., 2011; Walton et al., 2015; Sanderson et al., 2015; Nguyen-Chi et al., 2017), neutrophils (Mathias et al., 2006; Renshaw et al., 2006; Buchan et al., 2019), and eosinophils (Balla et al., 2010). These resources have given valuable insights into the role of innate immune cells in intestinal homeostasis and pathogenesis.

Intestinal macrophages play a role in microbiome homeostasis and complement regulation. Loss of intestinal macrophages in adult zebrafish irf8 mutants causes dysbiosis of the resident gut microbiota, and a reduction in the complement component C1q (Earley et al., 2018). Neutrophils are recruited to the intestine in response to inflammatory stimuli such as the pathogen-associated molecular pattern (PAMP) lipopolysaccharide (LPS). Immersion of larvae in LPS activates MyD88 and tumor necrosis factor-alpha (TNF-α), and leads to intestinal neutrophil influx (Bates et al., 2007). Colonization of larvae by a conventional microbiota promotes production of alkaline phosphatase by the gut epithelium, which detoxifies microbiota-derived LPS and prevents excessive inflammation to maintain intestinal homeostasis (Bates et al., 2007). Zebrafish TLRs that recognize PAMPs share key structural similarities with mammalian TLRs, but some TLRs may differ in functional aspects (Palti, 2011). For example, in mammals TLR4 responds to LPS, while zebrafish TLR4 paralogs (Tlr4a and Tlr4b) fail to recognize and respond to LPS (Sullivan et al., 2009). Since zebrafish are, however, able to mount a MyD88-dependent inflammatory response upon LPS stimulation (Bates et al., 2007; Yang et al., 2017), it is likely that a hitherto unidentified TLR could be functionally equivalent to mammalian TLR4. At least 17 TLRs have been identified in zebrafish (Jault et al., 2004; Meijer et al., 2004), suggesting redundant and potentially overlapping functions; we refer readers to another review of TLRs and their known functions in teleost species (Palti, 2011).

Zebrafish neutrophils typically migrate to sites of infection or injury at a faster rate than macrophages (Ellett et al., 2011), thus forming the first line of defense against pathogen insult. The presence of a commensal intestinal microbiota increases the number of circulating neutrophils, enhances neutrophil migratory velocity, and recruitment to extra-intestinal sites of injury (Kanther et al., 2014). Zebrafish neutrophils are functionally largely equivalent to mammalian neutrophils: they are capable of phagocytosis and degranulation, produce cytokines and reactive oxygen species (ROS), and can form neutrophil extracellular traps (NETs) (Lieschke et al., 2001; Harvie and Huttenlocher, 2015; Palie et al., 2007).

The precise role of eosinophils has been difficult to delineate due to a lack of specific markers, but they may play a role in responding to intestinal helminth infections (Balla et al., 2010). Adult zebrafish eosinophils share morphology and differential gene expression with mammalian eosinophils (Lieschke et al., 2001). When challenged with helminth antigens or - infection, eosinophil numbers within the zebrafish intestine increase, indicating conservation of eosinophil-mediated immune responses between zebrafish and mammals (Balla et al., 2010).

Secretion of antimicrobial molecules into the intestinal mucosal layer, both constitutive and in response to TLR activation by PAMPs, is also conserved from mammals to zebrafish. Zebrafish produce antimicrobial peptides (AMPs), of which β-defensins, cathelicidins, hepcidins, and histone-derived peptides are also found in mammals, and piscidins are fish specific (Zou et al., 2007; Katzenback, 2015). Secreted peptidoglycan recognition proteins (PGRPs) (Chang et al., 2007) and antibacterial lectins functionally similar to mammalian pore forming C-type lectins (Brinchmann et al., 2018) also contribute to zebrafish intestinal mucosal immunity. Secreted antimicrobial factors are expressed throughout the zebrafish intestine, both in larvae and adults (Oehlers et al., 2011a).

3.2. Adaptive immunity and the GI tract

The zebrafish adaptive immune system reaches morphological and functional maturity by 4–6 weeks post-fertilization (Willett et al., 1999; Lam et al., 2004) and consists of B- and T-cells capable of antigen receptor rearrangement in response to pathogens via recombination activating genes 1 and 2 (ragl and rag2) (Trede and Zon, 1998; Langenau and Zon, 2005). Other initiators of the adaptive immune response, like major histocompatibility complex class I and II molecules, are also present in zebrafish (Sültmann et al., 1993; Sültmann et al., 1994; Takeuchi et al., 1995; Rauta et al., 2012).

The zebrafish lymphatic system shares many morphological, molecular, and functional features with those of mammals, but lacks some lymphoid tissues, like lymph nodes (Jung et al., 2017) and Peyer’s patches (Boehm et al., 2003; Brugman, 2016). However, there is evidence that the enterocytes of the larval and adult zebrafish mid-intestine are highly endocytic (Oehlers et al., 2011a) and perform a specialized function in luminal antigen sampling. Consistent with that, a labeled antigen derived from a Yersinia ruckeri immersion-vaccine is initially detected in the mid-intestine enterocytes of zebrafish larvae, followed by the spleen (Korbut et al., 2016). In adult zebrafish orally infected with Mycobacterium marinum, bacteria are taken up into vacuoles by antigen-sampling cells, which make up the majority of the epithelium of the posterior mid-intestine. Bacteria are trafficked to the basal side of the intestine, where they subsequently co-localize with leukocytes, before eventually travelling to the liver and spleen (Løvmo et al., 2017).

Teleost B-cells exhibit additional phagocytic and microbicidal functions not typically seen in higher vertebrates (Li et al., 2006; Øverland et al., 2010), and can also function as initiating antigen-presenting cells (APCs), linking the innate and adaptive immune systems in zebrafish (Zhu et al., 2014; Lewis et al., 2014). In teleosts, intestinal mucosal immunity is largely dependent on B-cells acting as the primary responders to perturbation, but our understanding of their origins and full spectrum of functions in maintaining gut homeostasis is still limited (see Parra et al., 2016 for review).

4. Zebrafish infection models of gastrointestinal pathogens

4.1. Common routes of infection

Several approaches are currently in use to establish infections with GI pathogens in the zebrafish host, each with distinct advantages and disadvantages to consider. One of the most commonly used routes of infection is by immersion of larvae or adults in a suspension containing a defined pathogen concentration (for detailed protocols see e.g. Varas et al., 2019 for larvae and Mitchell et al., 2017 for adults). The former also offers a side-by-side comparison of infection resulting from immersion vs. caudal vein injection. Immersion infection can be used for larvae beginning when the mouth first opens at around 3 dpf (Ng et al., 2005). It is advantageous due to its extremely high throughput and does not require much specialized equipment. Its disadvantage is that the exact dose of pathogen taken up by the fish, although it can be measured, can only be indirectly controlled by varying the concentration of pathogen in the suspension. Immersion cannot be applied to strictly anaerobic microbes. For some GI pathogens, in particular those for which zebrafish are not a natural host, immersion does not lead to robust intestinal colonization (e.g., Tysnes et al., 2012; Stones et al., 2017 etc.). More recently, the protozoan Paramecium caudatum, a natural prey for larval zebrafish, has been adapted as a vehicle for food-borne infection models (Stones et al., 2017; Flores et al., 2019; Fan et al., 2019). The protozoan internalizes bacteria into storage vacuoles and following uptake of Paramecia by the zebrafish and degradation of the protozoan in the foregut, the bacterial load is released into the middle intestine. This approach, although more laborious than simple immersion, still allows for high throughput and mimics human exposure to GI pathogens by consumption of contaminated food. Another advantage is that P. caudatum passages the pathogen through acidifying storage vacuoles prior to its release into the zebrafish gut, and the low pH may prime the pathogen similarly to the low pH environment of the human stomach, as discussed in section 2. A third, more labor intensive route of infection is oral gavage of adults or microgavage of larval zebrafish (see Cocchiaro and Rawls, 2013 for a detailed protocol). A defined amount of pathogen suspension is directly delivered into the foregut via a fine capillary that is inserted into the mouth and esophagus of the animal. This technique comes with several costs: it is time consuming, requires specialized equipment, and a skilled experimenter is necessary to prevent injury or death of the infected animal. Sufficient sample sizes are necessary to compensate for attrition or failure of the inoculum to reach the intestine due to regurgitation (Runft et al., 2014). Lastly, some studies inject GI pathogens into the caudal vein or peritoneum of zebrafish, which introduces the pathogen into the blood stream. Despite the difference in administration, these studies have contributed valuable insights into microbial factors modulating inflammation and tissue damage, which often drive mortality as a result of infection (Dong et al., 2013; Okuda et al., 2014).

Below, we discuss GI pathogens studied in zebrafish to date, along with relevant outcomes from those studies. The developmental stage and route of infection used in each case are listed in Table 1.

Table 1. Summary - Examples of gastrointestinal pathogens and their disease profiles in the zebrafish host.

The table is organized first by pathogen (in alphabetical order), followed by developmental stage, followed by route of infection. Some of the studies summarized below discuss additional infection models which were not included in the table due to space constraints.

Pathogen	Route of infection	Zebrafish developmental stage	Disease phenotype/ finding	Reference
Aeromonas hydrophila	Immersion	Larvae, 4 dpf	Colonization and increased mortality	Saraceni et al., 2016
Aeromonas hydrophila	Immersion	Adults, 4 months	Change in microbiota composition; induction of innate immune response	Yang et al., 2017
Aeromonas hydrophila, Aeromonas veronii	Immersion	Germ-free larvae, 4–5 dpf; Adults	Intestinal colonization; increased mortality; intestinal lesions in adults; Aerolysin-dependent virulence	Ran et al., 2018
Edwardsiella ictaluri	Co-housing	Adults	Skin hemorrhage; abdominal hemorrhage; nephritis; septicaemia; mortality	Hawke et al., 2013
Edwardsilla piscicida	Immersion	Larvae, 5 dpf	Type III secretion system- and type VI secretion system-dependent increase in mortality	Hu et al., 2019
Edwardsilla piscicida	Immersion	Larvae, 4 dpf	Histone H2A and RIP2 are required for induction of antibacterial genes and decrease bacterial burden and mortality in response to infection	Wu et al., 2019
Edwardsiella tarda	Immersion	Larvae, 1 dpf	Increased mortality; systemic infection	Pressley et al., 2005
Edwardsiella tarda	Immersion	Larvae, 1 dpf	Flagella-dependent increase in mortality	Xu et al., 2014
Edwardsiella tarda	Immersion	Healthy adults; Wounded adults (skin abrasion)	(Healthy) Colonization; no mortality (Wounded) Increased mortality; septicaemia; colonization of intestinal smooth muscle tissue	Pressley et al., 2005
Edwardsiella tarda	Injection, i.p.	Adults	Type III secretion system-dependent increase in mortality	Okuda et al., 2014
Edwardsiella tarda	Injection, i.p.	Adults	Invasin-dependent increase in mortality (LD50)	Dong et al., 2013
Escherichia coli (EHEC)	Food-borne	Larvae, 4 dpf	Intestinal colonization; neutrophil recruitment; type III secretion system induction; increased mortality; transmission to naïve fish	Stones et al., 2017
Escherichia coli (Nissle, strain 40)	Immersion	Adults	Intestinal colonization; metabolize glucose and decrease intestinal pH; decrease burden of V. cholerae during co-infection	Nag et al., 2018
Giardia duodenalis	Oral gavage	Adults	Intestinal deposition and retention of cysts, excretion of cysts; no detection of trophozoites; no intestinal damage	Tysnes et al., 2012
Human norovirus (HuNoV GI and GII)	Injection, yolk sac	Larvae, 3 dpf	Viral replication; viral persistence in the intestine and hematopoietic cells; transmission to naïve fish; antiviral treatment	Van Dycke et al., 2019
Picornavirus-1	Naturally occuring	Adults	Intestinal colonization; viral shedding	Altan et al., 2019
Salmonella enterica serovar Typhimurium	Injection, otic vesicle; Injection, yolk	Larvae, 2 dpf	Increased SPI-1- and SPI-2-dependent mortality; caspase-1 activation (WT); Gbp4 inflammasome mediated response to infection	Tyrkalska et al., 2016
Salmonella enterica serovar Typhimurium	Food-borne	Larvae, 5 dpf	Colonization of intestine; neutrophil recruitment; translational fidelity-dependent bacterial fitness in vivo	Fan et al., 2019
Salmonella enterica serovar Typhimurium	Immersion	Larvae, 6 dpf	Colonization of intestine and cloaca; inflammation and swelling of the cloaca; neutrophil recruitment to site of infection	Varas et al., 2017
Salmonella enterica serovars Typhimurium, Enteritidis, Weltevreden	Immersion	Adults	Colonization and replication; shedding and transmission to naïve fish; intestinal inflammation and destruction of intestinal epithelium; diarrhea; increased mortality; protective immunity via bath vaccination with heat-killed bacteria	Howlader et al., 2016
Salmonella enterica serovar Typhimurium	Oral gavage	Adults, 8 months	Colonization of intestine; intestinal inflammation and ulceration; macrophage and lymphocyte infiltration; tissue damage; Increased mortality; spv-dependent pathology	Wu et al., 2017
Vibrio cholerae	Immersion	Larvae, 4 dpf	CRP- and TcpA-dependent intestinal colonization and mortality; rtx, hlyA induction during colonization	Manneh-Roussel et al., 2018
Vibrio cholerae	Immersion	Larvae, 5 dpf	Colonization of the intestine	Runft et al., 2014
Vibrio cholerae	Immersion	Germ-free larvae, 5 dpf	Intestinal colonization; decrease in commensal Aeromonas; type VI secretion system dependent increase in gut peristalsis	Logan et al., 2018
Vibrio cholerae	Oral gavage; Immersion	Adults, 6–9 months	Colonization of intestine; microcolony formation; transmission to naïve fish	Runft et al., 2014
Vibrio cholerae	Immersion	Adults	Intestinal colonization; diarrhea; increased mucin production and excretion	Mitchell et al., 2017
Vibrio parahaemolyticus	Injection, i.p.	Adults, 7–8 months	Increased mortality; necrosis of kidneys, liver and intestinal villi; intestinal bleeding	Zhang et al., 2016
Vibrio vulnificus	Injection, i.p.	Adults	Increased mortality; inflammatory response; protection by AMP epinecidin-1	Pan et al., 2011

4.2. Edwardsiella and Aeromonas infections in zebrafish

Major aquaculture and opportunistic human pathogens include members of the Edwardsiella and Aeromonas genera (Lee and Wendy, 2017). The consumption of food contaminated with these bacteria can cause gastroenteritis in healthy persons, and more severe diarrheal disease in elderly and immunocompromised individuals (Clarridge et al., 1980; Gracey et al., 1982). Edwardsiella species known to infect fish include E. ictaluri, E. hoshinae, E. piscicida, and E. tarda, but only the latter is known to affect humans (Jordan and Hadley, 1969; Hawke et al., 1981; Nomura and Aoki, 1985). Like other fish pathogens, the Edwardsiella and Aeromonas species enter their host through the skin and/or gills (Ventura and Grizzle, 1987; Menanteau-Ledouble et al., 2011). Infection of zebrafish with E. tarda (Pressley et al., 2005) or E. ictaluri (Hawke et al., 2013) induces the pro-inflammatory cytokines IL-β and TNF-α, leads to the development of hemorrhagic septicemia and increases mortality. Furthermore, studies using zebrafish infection models have elucidated E. tarda virulence factors such as (1) the type III secretion system essential for invasion and intracellular replication in phagocytic cells (Okuda et al., 2014), (2) an invasin protein important for haemolytic activity and biofilm formation (Dong et al., 2013), and (3) flagellar components involved in motility, biofilm formation, and adhesion (Xu et al., 2014). As with E. tarda, a type III secretion system dependent increase in mortality is seen with E. piscicida, although a type VI secretion system was identified as an additional virulence factor in the latter (Hu et al., 2019). The same model was used to show that the NOD-like receptor mediated immune response is required for upregulation of antibacterial genes, including beta-defensins and major histocompatibility complex (MHC) related genes, in response to E. piscicida infection (Wu et al., 2019). Failure to induce the pathway leads to increased bacterial burden and mortality in response to infection. Several studies have used zebrafish to investigate the efficacy and mechanism of action of potential novel therapeutics to combat Edwardsiella infections, including vaccines (Guo et al., 2015) and immunomodulatory nanoparticles that fortify resistance against pathogens (Udayangani et al., 2017).

Members of the genus Aeromonas that affect humans include A. hydrophila, A. caviae, A. veronii, and A. dhakensis, while A. salmonicida is fish-specific (Clarridge et al., 1980; Sacho et al., 1990; Joseph and Carnahan, 1994). Immersing larval or injured adult zebrafish in A. hydrophila upregulates the production of IL-ip and TNF-a in response to colonization, and increases neutrophil recruitment to the wound site (Saraceni et al., 2016). Aeromonas sp. isolated from the zebrafish gut exhibit intrinsic antibiotic resistance and harbor extracellular enzymes such as lipase, hemolysin, proteases, and DNase with the potential to degrade host cells, indicative of their potential to cause disease in the zebrafish host (Hossain et al., 2018). Gut colonization with A. hydrophila alters the adult zebrafish intestinal microbiota composition, concomitantly increasing pathogen abundance and decreasing beneficial intestinal bacteria (Yang et al., 2017). Although both A. hydrophila and A. veronii are able to colonize the intestine and cause mortality (Saraceni et al., 2016; Ran et al., 2018), A. veronii expresses more aerolysin toxin, which causes intestinal lesions and invasion of the intestinal barrier, and is associated with increased virulence compared to A. hydrophila (Ran et al., 2018).

4.3. Vibrio infections in zebrafish

Members of the Vibrio genus are natural inhabitants of warm coastal waters. The consumption of raw or contaminated seafood often leads to vibriosis in humans, a disease characterized by diarrhea, nausea, and/or abdominal cramps (Johnston et al., 1986). Though gastroenteritis is most closely associated with V. cholerae, V. parahaemolyticus and V. vulnificus can also cause intestinal disease, serious wound and soft tissue infections, as well as bacteremia (Johnston et al., 1986; Lee et al., 2003; Tsai et al., 2009). Various groups have attempted to model Vibrio-induced diseases in both mammalian and non-mammalian systems, however the animals used were not natural hosts and required tedious manipulation (Rowe et al., 2014). Vibrios naturally colonize zebrafish, and some strains are natural fish pathogens, so using zebrafish as a host to model disease may offer valuable insights into Vibrio pathogenesis and aid the identification of therapies to combat infection (Runft et al., 2014). The V. cholerae serogroup O1 biotype El Tor responsible for cholera efficiently colonizes the zebrafish intestine and forms microcolonies in close contact with the intestinal epithelial surface similar to those seen in human specimen (Runft et al., 2014). The pathogen is shed into the water and can be transmitted to naïve fish (Runft et al., 2014). Adults immersed in V. cholerae develop cholera toxin-independent diarrhea characterized by increased mucin production and excretion (Mitchell et al., 2017). Although cholera toxin is dispensable for colonization and pathogenesis in zebrafish, accessory toxins including RTX and HlyA are induced during intestinal colonization, and the metabolic regulator CRP is required for their activation inside the zebrafish host (Manneh-Roussel et al., 2018). V. cholerae uses a host-directed type VI secretion system to enhance intestinal contractions in the zebrafish gut and to expel resident intestinal microbiota to allow for pathogen colonization (Logan et al., 2018). Treating zebrafish infected with V. vulnificus with the AMP epinecidin-1 increases host survival, possibly through modulation of immunomodulatory genes (Pan et al., 2011). Inoculating zebrafish with V. parahaemolyticus via intraperitoneal injection causes abdominal hemorrhaging, swelling, and secretion of two major markers of host sepsis, TNF-α and IL-1β (Zhang et al., 2016).

4.4. Salmonella infections in zebrafish

More recently, the use of zebrafish has been expanded to model infection of non-fish pathogens in the GI tract. Salmonella enterica serovar Typhimurium is a major causative agent of human foodborne gastroenteritis (Hoelzer et al., 2011). Immersing zebrafish larvae in S. Typhimurium leads to gut colonization and inflammation of the intestine and cloaca (Howlader et al., 2016; Varas et al., 2017). While the precise molecular mechanisms responsible for inflammation are not fully understood, expression of Salmonella virulence plasmid (spv) genes in adult zebrafish suppresses protective host responses such as expression of type II IFN-γ, IL-12, and TNF-α, and promotes expression of cytokines known to facilitate intracellular pathogen survival such as IL-4, IL-10, and IL-13 (Wu et al., 2017). Furthermore, in vivo clearance of S. Typhimurium is mediated by inflammasome activation in neutrophils (Tyrkalska et al., 2016). Mechanistically, infected zebrafish release CXC chemokine 18 and leukotriene B4 that trigger both the recruitment of neutrophils to the infected site and phagocytosis of S. Typhimurium. Once engulfed, S. Typhimurium activates the Gbp4 inflammasome that modulates the activity of cytosolic phospholipase A2 and production of prostaglandins, ultimately leading to bacterial clearance (Tyrkalska et al., 2016). Strains of S. Typhimurium with non-optimal translational fidelity recruit fewer neutrophils following colonization via the food-borne route, and are outcompeted by a wild-type strain in vivo (Fan et al., 2019).

4.5. E. coli infections in zebrafish

Another human enteric pathogen successfully modeled in zebrafish is enterohemorrhagic Escherichia coli (EHEC), which colonizes the intestine and causes bloody diarrhea in the human host. Following foodborne delivery via P. caudatum, EHEC colonizes the middle intestine of zebrafish larvae despite the presence of endogenous microbiota, and by 4 days post-infection survival rates decrease by ~40% (Stones et al., 2017). In humans and cattle, intestinal colonization by EHEC and the formation of attaching and effacing lesions on enterocytes is mediated by a set of virulence genes encoded by the locus of enterocyte effacement (LEE), (Phillips et al., 2000; Elliott et al., 2000). Following foodborne infection of larval zebrafish, EHEC induce the LEE during early colonization of the GI tract, and LEE induction is required for optimal colonization and pathogenesis (Stones et al., 2017). The ability to colonize the zebrafish intestine, however, is not limited to pathogenic E. coli strains; Commensal E. coli isolates from the gut of healthy human volunteers also colonize zebrafish, and can reduce colonization by V. cholerae by decreasing the intestinal pH as a result of glucose metabolism (Nag et al., 2018).

4.6. Non-bacterial GI infections in zebrafish

Gastrointestinal microbes studied in zebrafish are not limited to bacteria; Eukaryotic microorganisms, such as the protozoan Giardia duodenalis can cause intestinal damage in humans and a zebrafish model of inoculation with cysts has been established (Tysnes et al., 2012). Excystation and infection of the zebrafish intestine by trophozoites, however, remains to be shown. Altan et al. recently described a naturally occurring picornavirus in zebrafish (Altan et al., 2019). Picornavirus-1 ZfPV-1 selectively infects a subset of enterocytes and cells in the lamina propria, providing a natural model to study virus-GI tract interactions in a vertebrate host. Human norovirus persists in the larval zebrafish intestine and hematopoietic cells following yolk injection, replicates, and can be transmitted to naïve hosts (Van Dycke et al., 2019).

Zebrafish were initially used to study natural fish pathogens with the intention to limit the impact of infection on aquaculture, but the studies discussed above highlight a rapidly expanding movement to use zebrafish to model zoonotic and human bacterial infections, as well as protozoan and viral infections.

5. Zebrafish as a model to study the intestinal microbiome

The gut microbiome is the collective population of microorganisms that reside in the host GI tract. Many members of the microbiome have beneficial roles, such as regulating digestion, nutrient absorption, and immune system maturation (Umesaki et al., 1999; Semova et al., 2012; Thaiss et al., 2016). Perturbation of the gut microbiome has been linked to a wide range of disease states including neurodevelopmental disorders, obesity, and inflammatory bowel diseases (IBD) (Turnbaugh et al., 2008; Frank et al., 2011; Luna et al., 2017; Felice and O’Mahony, 2017). The application of metagenomics has enabled the differentiation of microbes preferentially residing in healthy versus diseased hosts, but it is presently unclear if there is a conserved ‘core’ microbiome, or how host genetics, the microbiome, and environmental factors interact to determine physiological and pathophysiological states of the host.

Zebrafish are a powerful model to address these unknowns and begin to unravel complex host-microbiota interactions in a more controlled, simple model system compared to other vertebrates used in microbiome studies. For example, derivation of germ-free (GF) larvae is simpler, more expedient, and more cost-effective than using murine GF models (Pham et al., 2008). Studies using conventionally raised (i.e., born and raised in the presence of their normal microbiota), GF, and conventionalized (i.e., derived GF and then colonized with microbiota) larvae are beginning to elucidate the mechanisms governing host-microbiota homeostasis as well as the effects of the microbiota on host organ development and differentiation. Comparative studies of conventional and GF zebrafish larvae demonstrate that the microbiota influences host gene expression, particularly genes related to intestinal epithelial proliferation, nutrient metabolism, and the innate immune response (Rawls et al., 2004; Reikvam et al., 2011). Other studies support the hypothesis that host factors have a dominant effect on determining the intestinal microbiome composition (Rawls et al., 2006; Roeselers et al., 2011). For example, transplantation of intestinal microbiota from donor mice into GF zebrafish larvae reveals that, while the resulting GI microbiota in zebrafish resembles that of the donor mice, the number of microbial lineages present in the recipient community changes drastically. This suggests host-mediated selective pressures restrict the microbiome composition (Rawls et al., 2006). This observation is further supported by microbiome profiling of recently caught and domesticated zebrafish reared in multiple research facilities; there exists a ‘core’ microbiome that is shared among zebrafish lineages regardless of environmental exposure (Roeselers et al., 2011). In vivo analyses indicate the intestinal microbiome is required for intestinal epithelium development and differentiation in zebrafish (Bates et al., 2006). Several developmental deficiencies have been observed in GF larvae, including the absence of brush border intestinal alkaline phosphatase activity, immature glycan expression, a decrease in secretory cell numbers, and altered gut motility. These effects can be reversed with the re-introduction of microbiota from conventional donors (Bates et al., 2006). For an in-depth review of the gut microbiome and its analysis in zebrafish and other teleosts, see Tarnecki et al., 2017.

The mammalian and zebrafish gut microbiome share six bacterial divisions, but the human GI tract is primarily dominated by Firmicutes, Bacteroidetes, and Actinobacteria while Proteobacteria are the predominant phylum in zebrafish (Rawls et al., 2004; Eckburg et al., 2005; Bates et al., 2006). Several studies demonstrate the GI tract of zebrafish can be colonized with aerobic and anaerobic species derived from human fecal communities (Toh et al., 2013; Arias- Jayo et al., 2018; Valenzuela et al., 2018). The successful colonization of the larval gut with anaerobic species, including Lactobacillus paracasei and strict anaerobes such as Eubacterium limosum and Akkermansia muciniphila, in particular, was an important stepping stone in establishing zebrafish as a surrogate host for studies on the human gut microbiome (Toh et al., 2013; Arias-Jayo et al., 2018).

The emerging consumer interest in probiotics as a supplement to promote health and well-being has further propelled microbiome studies in zebrafish. The lactic acid bacteria Lactobacillus plantarum is of particular interest and is currently commercialized as a beneficial probiotic that alleviates stress and anxiety in humans (Chong et al., 2019). In zebrafish, L. plantarum protects against stress-induced dysbiosis of the gut microbiota and ameliorates anxiety-related behavior. These changes in microbiome composition and behavior correlate with an increase in expression of gamma-aminobutyric acid receptors, and of serotonin transporters in zebrafish brains following administration of L. plantarum (Davis et al., 2016). Colonization of adult zebrafish with the probiotic Lactobacillus rhamnosus decreases the expression of host genes associated with dietary lipid metabolism (Falcinelli et al., 2017). Consequently, shifts in microbiome community structure resulting from high dietary fat intake and associated weight gain are ameliorated by supplementation with L. rhamnosus, highlighting its potential to attenuate diet-related metabolic disorder (Falcinelli et al., 2017).

6. Zebrafish models of inflammatory bowel diseases

While the etiologies of multifactorial IBDs like ulcerative colitis and Crohn’s disease are unknown, two key features of the pathology are intestinal epithelial damage and intestinal dysbiosis. The intestinal epithelium provides a physical barrier between the intestinal lumen and surrounding tissues, and works in concert with immune cells to sense and respond to both normal microbiota and opportunistic infection. Therefore, factors such as host genetics and environmental conditions that impair intestinal epithelial homeostasis and disrupt the interplay with resident microbiota can significantly impact IBDs (Bellaguarda and Chang, 2015). Novel murine models of intestinal epithelial damage and inflammation have advanced our understanding of IBDs, but several limitations associated with murine models such as cost, imaging capacity, and genetic manipulation still constrict elucidation of how the aforementioned disease modifiers converge and affect pathogenesis in vivo (Kiesler et al., 2015).

The zebrafish model circumvents murine-associated limitations and provides the opportunity to rapidly explore factors associated with increased inflammation and dysbiosis in response to intestinal epithelial damage and is thus relevant to IBDs in humans. Zebrafish models of chemically-induced IBD mimic some key aspects of the pathological condition in humans, including enterocolitis, intestinal epithelial damage, disruption of intestinal architecture, and shifts in microbiome composition (He et al., 2013). For example, intra-rectal administration of oxazolone in adult zebrafish causes disruption of the intestinal fold architecture, depletes goblet cells, increases immune cell infiltration of the gut, and upregulates pro-inflammatory cytokines (Brugman et al., 2009). Likewise, immersion of larval and adult zebrafish in 2,4,6-trinitrobenzenesulfonic acid (TNBS) induces intestinal inflammation, inhibits peristalsis, and disrupts epithelial integrity by impairing the tight junctions between cells. This suggests TNBS disrupts intestinal barrier function and promotes features observed in IBD patients (Fleming et al., 2010; Oehlers et al., 2011b). It has been proposed that increased intestinal epithelial cell shedding after damage compromises barrier integrity, which in turn fuels more inflammation (Blander, 2016). In support of this, leukocytes mobilize from the caudal hematopoietic tissue to the periphery and accumulate around the damaged intestine following TNBS insult in zebrafish (Oehlers et al., 2011b). Disruption of the intestinal barrier by TNBS promotes microbial infiltration of the lamina propria, initiating a cascading pro-inflammatory response characterized by secretion of TNF-α, via activation of TLR3, MyD88, TIR-domain-containing adapter-inducing interferon-β (TRIF), and nuclear factor kappa-light-chain-enhancer of activated B-cells (NF-kB) (He et al., 2013). Together, these data support the idea that zebrafish provide a high-throughput model to interrogate genetic pathways and drug candidates that can restore epithelial barrier function and modulate immune cell function to protect against microbial infiltration during IBD.

7. Perspective

Zebrafish have been a well-used model for infection biology for many years. More recent research highlighting similarities between the zebrafish gut and the GI tract of mammals has expanded to include studies on GI tract-pathogen interactions. The zebrafish as a model organism particularly lends itself to questions that cannot be addressed easily in other model systems thanks in part to the ease of intravital imaging and raw statistical power. Its use in this context may help unravel new details of the cellular microbiology underpinning infection, lead to the identification of novel virulence factors and aspects of their regulation within the GI niche, and aid the discovery of effective therapies by enabling high-throughput drug screening in the live host. Future studies that use zebrafish to study host-microbiome interactions will continue to identify host-intrinsic and - extrinsic factors and selective pressures critical for establishing and shaping the host microbiome. Such discoveries may aid in the development of novel therapies for combating microbiome- associated diseases.