Neutrophils in host defense: new insights from zebrafish

Review on the use of zebrafish to study the role of neutrophils in host defense against bacterial and fungal pathogens.

Abstract

Neutrophils are highly motile phagocytic cells that play a critical role in the immune response to infection. Zebrafish (Danio rerio) are increasingly used to study neutrophil function and host-pathogen interactions. The generation of transgenic zebrafish lines with fluorescently labeled leukocytes has made it possible to visualize the neutrophil response to infection in real time by use of optically transparent zebrafish larvae. In addition, the genetic tractability of zebrafish has allowed for the generation of models of inherited neutrophil disorders. In this review, we discuss several zebrafish models of infectious disease, both in the context of immunocompetent, as well as neutrophil-deficient hosts and how these models have shed light on neutrophil behavior during infection.

Introduction

Neutrophils are the most abundant type of circulating leukocyte in humans and zebrafish and are typically the first responders recruited to sites of tissue injury or infection [1, 2]. To traffic to a site of infection, neutrophils mobilize from hematopoietic tissue and travel through the vasculature. Once at the site of infection, these highly motile cells play a critical role in initial defense through phagocytosis of microbes, secretion of granule proteins and other antimicrobials, production of ROS, and release of NETs [3, 4]. Neutrophils also mediate the proinflammatory response to infection by releasing cytokines that recruit and activate other immune cells. The crucial importance of neutrophils in host defense against infection is demonstrated by enhanced susceptibility of neutropenic patients to a wide range of bacterial and fungal infections [2].

The progression of infectious disease is determined by dynamic and complex interactions between host defense systems and pathogen virulence factors. Whereas detailed in vitro analyses have provided critical insight into this delicate interaction, its complexity is more easily studied in vivo, especially when host and pathogen are amenable to genetic analysis. There has been increasing use of the larval zebrafish (D. rerio) to study infectious disease, as its optical accessibility and potential for genetic manipulations allow visualization of the immune response to infection inside a living, intact vertebrate host. Genetic tools in zebrafish allow for the generation of transgenic lines with fluorescently labeled cell populations, including neutrophils [5, 6] and macrophages [7]. The vertebrate innate immune system is highly conserved in zebrafish and includes: complement [8], TLRs [9, 10], macrophages [11], and neutrophils [1, 12]. The adaptive immune system is not functionally mature until 4–6 weeks postfertilization [13–15], and thus, the study of host-pathogen interactions in the embryonic and larval stages of development allows for an understanding of the specific contribution of innate immunity to host defense. Importantly, transient manipulation of gene expression is readily available in the larval zebrafish, including injection of synthetic mRNA for overexpression and injection of antisense morpholino oligonucleotides for gene suppression. Recently, the generation of null mutants has become easier and faster with the application of gene-editing tools, including CRISPR/ Cas [16] and TALENs [17].

Recent advances with the use of zebrafish to study macrophages during host-pathogen interactions were reviewed recently [18]. Here, we will focus on the role of neutrophils in host-pathogen interactions and how studies that use zebrafish have enhanced and complemented work done in mammalian models.

ZEBRAFISH NEUTROPHIL BIOLOGY AND IN VIVO VISUALIZATION OF NEUTROPHIL BEHAVIOR

Zebrafish neutrophil development

Myelopoiesis in zebrafish is genetically controlled by the transcription factor Pu.1/Spi1 [19]. Pu.1-expressing myeloid precursors are first detected ∼12 hpf [20]. Neutrophils originate in the developing larval zebrafish from the primitive and definitive waves of hematopoiesis. During the primitive wave, myeloid precursors arise in the rostral blood island (also known as the anterior lateral plate mesoderm) and migrate over the yolk sac, where they differentiate into primitive macrophages by 22 hpf [11]. A subset of these primitive macrophages then differentiates further into neutrophils by 33 hpf [12]. Definitive or multilineage hematopoiesis begins as early as 24 hpf when a transient population of pluripotent erythromyeloid progenitor cells differentiates within the posterior blood island, an area that later expands to become the CHT [21, 22]. These cells are soon replaced by another group of cells that migrates from the aorta-gonad-mesonephros through the blood to start colonizing the CHT by 48 hpf [23], where they also give rise to neutrophils [22]. Neutrophil distribution in the developing larva includes a population of neutrophils in the CHT, as well as a population of randomly migrating neutrophils in the head mesenchyme (Fig. 1A and Supplemental Video 1). By 4 dpf, the kidney marrow begins to mature and will become the site of definitive hematopoiesis in adult fish [24].

Figure 1. Live, in vivo imaging of neutrophil response to infection.

(A) Neutrophil (green) distribution at 3 dpf in a Tg(mpx:dendra2) larva. Neutrophils in the CHT, as well as randomly migrating neutrophils in the head, are shown in Supplemental Video 1. The box indicates the region around the OV (imaged in C and D). Original scale bar, 500 μm. (B) Commonly used sites of infection in embryonic and larval zebrafish. (C) Live imaging of a 3 dpf Tg(mpx:dendra2) larva infected with Pseudomonas aeruginosa (Pa) strain PAK (pMKB1::mCherry) in the OV (white dotted line). Four frames were extracted from a 3 h movie (see Supplemental Video 2) and show neutrophil (green) recruitment and phagocytosis of P. aeruginosa (magenta). Original scale bar, 40 μm. (D) Live imaging of a 3 dpf Tg(mpx:dendra2) larva infected with CellTracker Red-labeled Streptococcus iniae (Si) strain 9117 in the OV (white dotted line). Four frames were extracted from a 4 h movie (see Supplemental Video 3) and show neutrophil (green) recruitment and phagocytosis of S. iniae (magenta). Original scale bar, 40 μm.

Zebrafish macrophages and neutrophils are fully functional in the early developmental stages, capable of phagocytosis by 28–30 hpf [11, 12] and of generating a respiratory burst by 72 hpf [25]. They are also capable of NETosis [26], the release of decondensed chromatin, histones, and granule proteins into the extracellular space to trap and kill microbes physically by exposing them to high, local concentrations of antimicrobial molecules [27, 28]. Direct visualization of neutrophils in response to wounding has led to new insight into neutrophil behavior at sites of tissue damage. Indeed, a study of the neutrophil response to wounds in zebrafish provided the first direct demonstration that resolution of neutrophil infiltration at wounds can occur by neutrophil reverse migration and not apoptosis [5]. Reverse transmigration of neutrophils from a site of inflammation has also been reported with primary human neutrophils in vitro [29] and more recently in a mouse model of ischemia-reperfusion injury [30]. This highlights the advantage of the use of zebrafish as a model for visualizing neutrophil behavior in vivo and supports the idea that zebrafish represent an important model system that complements studies that use mouse and human neutrophils.

Tools for imaging neutrophils in zebrafish

Neutrophils can be imaged in whole, fixed zebrafish by use of Sudan Black, which stains the rod-shaped cytoplasmic granules irreversibly [12], or immunostaining. Differentiated neutrophils and macrophages express the pan-leukocyte marker L-plastin (lcp1) [31], and immunolabeling of L-plastin has been used to visualize leukocytes in fixed larvae [32]. For live in vivo imaging, the generation of transgenic lines that label zebrafish neutrophils has provided an important tool to image neutrophil behavior. Transgenic lines with fluorescently labeled leukocytes have been made possible, in part, by the Tol2 system, which allows for the efficient integration of exogenous DNA into the genome [33]. Photoconvertible proteins, such as Dendra2 [34, 35] or Kaede [36, 37], allow for the tracking of individual host cells over the course of infection. Fluorescent transgenic lines by use of the pu.1/spi1 promoter have been used to drive expression in all myeloid cell precursors [38, 39]. Reporter lines specific for macrophages or neutrophils have also been generated in zebrafish. Myeloperoxidase and lysozyme C are components of mammalian neutrophil granules [40], and the mpx [5, 6] and lyz [31, 41] promoters are generally neutrophil-specific in zebrafish. Although it has been shown that lyz marks neutrophils and macrophages at 32 hpf [42], lyz primarily labels neutrophils by 48 hpf [31, 41]. An additional transgenic line with expression in only a subset of neutrophils was generated by use of an enhancer trap screen [31]. Labeled neutrophils in this line are able to respond to wounds and bacterial infection and make up ∼50% of the neutrophil population. This suggests that different subsets of neutrophils serve distinct functions in zebrafish innate immunity. With the generation of a macrophage-specific reporter line [7, 43], double transgenic fish allow for the observation of neutrophil-macrophage interactions during inflammation and infection in intact hosts [7].

The use of fluorescently labeled microbes and phagocyte transgenic lines allows for the spatial and temporal imaging of host-pathogen interactions in a living animal (Fig. 1C and D and Supplemental Videos 2 and 3). As in mammalian models, localized and systemic infections can be modeled in the larval zebrafish, depending on the initial site of infection. A number of different anatomic sites in the larval zebrafish have been used to investigate the host response to systemic infection (Fig. 1B), but the most commonly used injection sites are the CV and DC. Localized infection in the HBV [44], OV [12], DM [45, 46], and PC [47] allows imaging of phagocyte recruitment. Visual analysis of phagocyte function in vivo is enhanced by the use of fluorescent reporter probes. During phagocytosis, pHrodo dye can be used to label microbes and will fluoresce when exposed to acidic environments, such as the leukocyte phagosome [48]. There are also probes for live imaging of the production of ROS [49, 50] and RNS [51, 52].

Depletion of neutrophils or macrophages can provide important information about phagocyte function during infection [53]. Neutrophils or macrophages can be depleted by use of zebrafish lines that express the bacterial nitroreductase under control of neutrophil- or macrophage-specific promoters. The Escherichia coli nitroreductase converts the drug metronidazole to a cytotoxic product, inducing cell death in expressing cells to achieve tissue-specific ablation [54, 55]. Importantly, it has been shown that specific ablation of neutrophils or macrophages by use of this technique has no effect on the other population [56]. Antisense morpholino oligonucleotides can also be used to deplete specific phagocyte populations. Morpholinos are nucleic acid analogs that act to block splicing or initiation of translation of the target mRNA. Limitations of morpholino technology include lack of tissue specificity and transient effects, typically wearing off after 3 dpf. Morpholinos targeting pu.1 deplete neutrophils and macrophages and are often used to examine the effect of myeloid cell loss on host defense. A morpholino targeting irf8 specifically depletes macrophages but not neutrophils. However, a caveat of this approach is that depletion of irf8 results in an expansion of the neutrophil population [57]. A morpholino target that allows specific depletion of neutrophils has not been reported. Several groups have used a morpholino targeting gcsfr/csf3r to reduce neutrophil numbers [51, 58], but gcsfr/csf3r is also expressed in a subset of monocytes and macrophages [59], and expression of mpeg1 is dependent on gcsfr/csf3r signaling [7]. Thus, although a gcsfr/csf3 morpholino largely reduces the neutrophil population, it also affects a subset of macrophages [60, 61].

MODELS OF PRIMARY IMMUNODEFICIENCY IN ZEBRAFISH

Defects in neutrophil function in humans lead to severe, recurrent bacterial and fungal infections. Human neutrophil disorders include neutrophil mobilization and adhesion defects, altered actin-based motility, and impaired microbial killing. The following is a summary of how zebrafish have been used to model human immune deficiencies that affect neutrophil function.

WHIM syndrome

WHIM syndrome is a primary immunodeficiency, a type of severe, chronic neutropenia with apoptosis of mature myeloid cells in the bone marrow [62, 63]. SDF1 is expressed in bone marrow stromal cells and binds the receptor CXCR4 [64], which is expressed on mature neutrophils in the bone marrow [65]. In WHIM syndrome, gain-of-function mutations associated with truncation of CXCR4 prevent CXCR4 internalization and result in the active retention of neutrophils in the bone marrow [65]. In a mouse model of WHIM syndrome, expression of the CXCR4 WHIM mutations in hematopoietic stem cells impairs neutrophil release from the bone marrow and increases neutrophil apoptosis [66], causing neutropenia and recapitulating the human disease. A zebrafish model of WHIM syndrome is also associated with reduced circulating neutrophils and recapitulates the neutropenia, although there was no reported effect on neutrophil apoptosis [67]. Endogenous expression of Sdf1a mRNA is concentrated in areas of neutrophil production in zebrafish embryos, both in the head and the CHT. Expression of the human WHIM mutations [68] in zebrafish neutrophils impairs Cxcr4b internalization and leads to retention of neutrophils in the CHT. Thus, WHIM neutrophils are not recruited to sites of wounding or infection, resulting in increased susceptibility to infection. Morpholino-mediated depletion of Sdf1a rescues neutrophil retention in the bone marrow [67], indicating that as in humans and mice, Sdf1a in zebrafish mediates the neutrophil retention signal. Thus, the zebrafish model of WHIM syndrome complements human and mouse WHIM models and provides the advantage of being amenable to direct visualization of the effects of altered chemokine signaling on neutrophil trafficking in vivo.

LAD

LAD represents a group of human disorders characterized by an increase in circulating neutrophils (neutrophilia) and recurrent infections [69]. There are different types of LAD, depending on the underlying genetic defect (reviewed in ref. [70]). LAD IV [71] is a recently reported type of LAD identified in 2 patients with a dominant inhibitory mutation in the hematopoietic tissue-specific Rho GTPase, Rac2 [72–75], which is the GTP-binding protein in neutrophils [76] and plays important roles in neutrophil activation by regulating the actin cytoskeleton, cell migration, and NADPH oxidase activity [77]. Rac2-deficient mice display neutrophilia and impaired emigration out of the vasculature to sites of infection, resulting in increased susceptibility to infections, such as Aspergillus fumigatus [78]. The phenotype of Rac2-deficient mice is similar to the phenotype of human patients with a dominant inhibitory mutation in Rac2 (Rac2D57N). Neutrophils from these patients have impaired chemotaxis [73, 75], azurophilic granule secretion [72], and defects in superoxide generation [73, 74]. A zebrafish model of LAD IV (LAD fish) also recapitulates the neutrophilia and increased susceptibility to infection [79] seen in mammalian disease. Zebrafish neutrophils expressing the dominant inhibitory Rac2D57N mutation show defects in cell polarization and migration [79]. The zebrafish model provides new insight into the mechanisms that contribute to neutrophilia. LAD fish exhibit an increase in neutrophil mobilization from hematopoietic tissue and a defect in egress from the vasculature to sites of tissue damage, resulting in neutrophilia. Interestingly, neutrophil retention in the CHT in WHIM fish is attenuated by the expression of Rac2D57N in neutrophils [79], suggesting a critical role for Rac2-Cxcr4 signaling in retaining neutrophils in hematopoietic tissue. Thus, the zebrafish model of LAD recapitulates the human disease and provides new insight into the signaling mechanisms that mediate neutrophil retention in hematopoietic tissue.

WAS and X-linked neutropenia

Leukocyte trafficking to infection requires the ability of leukocytes to sense directional cues and modify their actin cytoskeleton to form an F-actin-rich leading edge. WASp is a key regulator of actin dynamics [80] that binds the actin-nucleating protein Arp2/3, resulting in actin polymerization at branch points along actin filaments [81]. Mutations in WASp result in a primary immunodeficiency, known as WAS, which is characterized by recurrent infections, autoimmunity, and bleeding disorders [82]. Zebrafish have 2 WASp orthologs in hematopoietic cells, zWASp1 and zWASp2, which are 52% and 41% similar to human WASp, respectively [83]. Morpholino-mediated depletion of zWASp1 results in impaired directed migration of neutrophils and macrophages to a tail wound as a result of defects in pseudopod selection and directed migration [83]. To validate the morpholino studies, a zWASp1 mutant was generated [83] by TILLING [84]. As in zWASp1 morphants, the zWASp1 null mutant has a normal number of neutrophils and macrophages, but these leukocytes exhibit impaired recruitment to a tailfin wound [85] with defects in pseudopod selection similar to in vitro studies [86, 87]. The WASp null mutant also has increased susceptibility to bacterial infection with Staphylococcus aureus, recapitulating the infection phenotype of WAS patients [85]. Constitutively activating mutations in the GTPase binding domain of WASp result in X-linked neutropenia [88] and have also been modeled in zebrafish larvae [85] to provide an in vivo model for human X-linked neutropenia.

CGD

CGD is an immunodeficiency characterized by inflammatory granulomatous lesions and enhanced susceptibility to bacterial and fungal infections. CGD is associated with mutations in the PHOX complex that impair the ability of leukocytes to produce an oxidative burst [89, 90]. Elements of the PHOX complex include the membrane-bound components p22^phox and gp91^phox, cytosolic components p47^phox and p67^phox, the regulatory units p40^phox, and the small GTPase Rac1 or Rac2 [90]. Upon phagocyte activation, the cytosolic and membrane-bound components form the intact PHOX complex, leading to the generation of superoxide, which is converted to hydrogen peroxide by superoxide dismutase [90]. Hydrogen peroxide, in the presence of myeloperoxidase, can be converted further to other antimicrobial ROS, such as hypochlorous acid [40]. These ROS activate neutrophil granule proteins and contribute to microbial killing [91]. Although there is currently no transgenic model of CGD in zebrafish, the use of morpholinos targeting different components of the PHOX complex have been used successfully to model CGD [41, 92]. These studies demonstrate that phagocyte-mediated killing of Candida albicans [92] and Mycobacterium marinum [41] are dependent on their ability to generate an oxidative burst.

INSIGHTS INTO NEUTROPHIL-PATHOGEN INTERACTIONS BY USE OF ZEBRAFISH

Zebrafish have been used to model bacterial and fungal disease. For a comprehensive list of zebrafish infection models, see Table 1. Here, we focus on studies that address how neutrophils contribute to host defense to specific infections.

TABLE 1.

The role of neutrophils in zebrafish models of infection

Pathogen	Age of zebrafish	Route of infection	Role of neutrophils
Bacterium
Aeromonas hydrophila	Adult	Immersion [93]; i.p. [94, 95]	N/A
	Larva	Immersion [96, 97]	N/A
Aeromonas salmonicida	Adult	Immersion after woundinga [98]	N/A
Aeromonas veronii	Larva	Immersion [99]	N/A
Bacillus sp.	Adult	Mixed with feed [100]	N/A
Bacillus sphaericus	Adult	Immersion [101]	N/A
Bacillus subtilis	Embryo	CV [11]; DC [102]	Recruitment and phagocytosis [46]
	Larva	DM [46]
Bacillus thuringiensis	Adult	Immersion [101]	N/A
Bartonella henselae	Embryo	CV, DC, Y [103]	Recruitment and phagocytosis [103]
Burkholderia cenocepacia	Adult	i.p. [104]	N/A
	Embryo	CV [105, 106]	Some recruitment and phagocytosis, intracellular niche for bacteria [105]
Citrobacter freundii	Adult	Immersion [107]	N/A
Edwardsiella ictaluri	Adult	i.m. [108]; immersion [108]	Leukocyte recruitment [108]
Edwardsiella tarda	Adult	i.m. [109, 110]; immersion [111, 112]; i.p. [113, 114]	Infection causes up-regulation in genes involved in neutrophil recruitment [112]
	Embryo	CV [115, 116]; DC [117]; immersion [111, 115]	N/A
Enterococcus faecalis	Embryo	DC [118]	Phagocytosis, pu.1 morphants have increased mortality [118]
E. coli	Embryo	CV [11, 12, 119]; DC/PC [47, 120–122]; mesenchyme between eye and OV [123]; notochord [124]	Recruitment [12, 47, 124] mediated by Cxcl8 [123]; phagocytosis [12]; degranulation, production of IL-1β [124]; pu.1 morphants have increased mortality [47]
	Larva	Immersion [97, 125]; OV [12]
Flavobacterium columnare	Adult	i.m., immersion after wounding,a i.p. [126]	N/A
	Embryo	Immersion [127]	N/A
Flavobacterium johnsoniae	Adult	i.m., immersion after wounding,a i.p. [126]	N/A
Francisella spp.	Adult	i.p. [128]	N/A
	Embryo	DC, DM, OV [129]	Recruitment, phagocytosis is more efficient when microbe is attached to surface [129]
Haemophilus influenzae	Embryo	DC [117]	N/A
Lactococcus garvieae	Adult	i.p. [130]	Infiltration into optic lobes, lamina propria, and gills [130]
Leptospira interrogans	Embryo	CV, HBV [131]	Recruitment and phagocytosis by phagocytes [131]
Listeria monocytogenes	Adult	i.p. [132]	N/A
	Embryo	CV [133]	Limited phagocytosis, engulfed bacteria are always in vacuoles, usually degraded, neutrophils contact infected macrophages [133]
Listonella anguillarum	Adult	i.p. [134]	Up-regulation of neutrophil genes following infection (e.g., mpx and lyz) [134]
	Larva	Immersion [135]	N/A
Mycobacterium abscessus	Embryo	CV [136]	Recruitment to macrophage granulomas, phagocytosis [136]
M. marinum	Adult	Immersion [137]; oral intubation [137]; i.p. [137–150]	Infection induces up-regulation of neutrophil genes following infection [139–141]
	Embryo	CV [41, 44, 116, 147, 151–162]; HBV [41, 44, 141, 147, 150, 152–154, 156–158]; Y [162, 163]	Rare recruitment and phagocytosis at infection site [41]; recruitment and scavenger role at granuloma [41]; pu.1 morphants have increased mortality [154]; RNS production [25]; oxidative burst [41]
	Larva	Tail fin [164]	Rare recruitment [164]
Mycobacterium peregrinum	Adult	Immersion, i.p., oral intubation [137]	N/A
P. aeruginosa	Embryo	CV [165–167]; DC [117, 168, 169]; HBV [117]	Recruitment [117]; phagocytosis [117, 165, 168]; pu.1 morphants have increased mortality [165, 168]
	Larva	Immersion [96, 97, 125]; OV [79, 170, 171]; Y [170]	Recruitment [79, 170, 171]; phagocytosis [170]; LAD model supports critical role of neutrophils for survival [79]
Pseudomonas fluorescens	Larva	Immersion [99]	N/A
Salmonella arizonae	Embryo	CV [44]	N/A
Salmonella enterica serovar Typhimurium	Embryo	CV [116, 119, 172–175]; HBV [49, 51]; OV [176]; DM [176]; Y [177]	Recruitment [49, 51, 176]; infection induces neutrophil production [51]
	Larva	Immersion [178]	N/A
Shigella flexneri	Larva	CV [179]	Some phagocytosis, efficient killing of engulfed bacteria, scavenger role [179]
S. aureus	Embryo	CV [180]; eye [180]; HBV [180]; DC [56, 180, 181]; PC [180–182]; Y [181, 182]	Recruitment [180]; phagocytosis [56, 180]; possible intracellular niche for bacteria [56]; neutrophil ablation increases susceptibility to infection [5]; pu.1 and irf8 morphants have increased mortality [56]
Staphylococcus epidermidis	Embryo	Y [163]	N/A
Staphylococcus chromogenes	Adult	Immersion after woundinga [107]	N/A
Stenotrophomonas maltophilia	Adult	i.p. [183, 184]	N/A
Streptococcus agalactiae	Adult	i.m. [185]; i.p. [186–188]	N/A
Streptococcus equi	Adult	i.m., i.p. [189]	Leukocyte recruitment [189]
S. iniae	Adult	i.m., immersion after wounding,a i.p. [190–193]	Leukocyte recruitment [190, 191]
	Larva	OV [43]	Recruitment, phagocytosis, LAD model supports critical role of neutrophils for survival [43]
Streptococcus pnuemoniae	Adult	i.m., i.p. [194]	N/A
	Embryo	DC [195]	Phagocytosis by Spi1+ cells, pu.1 and WASp morphants have increased mortality [195]
Streptococcus pyogenes	Adult	i.p. [191, 196–198]; i.m. [45, 191, 196–200]	Lack of infiltrating leukocytes [191, 196–198]
	Embryo	Y [45]	N/A
	Larva	DM [45]	Lack of neutrophil recruitment [45]; recruitment to SLS mutant [45]
Streptococcus suis	Adult	i.p. [201–206]	N/A
Vibrio alginlyticus	Adult	i.m. [207]	N/A
Vibrio anguillarum	Larva	Immersion [208]	N/A
Vibrio vulnificus	Adult	i.p. [209]	N/A
Vibrio cholerae	Adult	Oral gavage, immersion [210]	N/A
	Larva	Immersion [210]	N/A
Yersinia ruckeri	Embryo	CV [211]	N/A
Intestinal microbiota	Larva	Immersion [96, 99, 125, 212–214]	Presence of microbiota results in: establishment of neutrophil homeostasis in the gut [99, 212, 213]; enhanced neutrophil recruitment to wound [212, 214]; neutrophil priming to produce H2O2 upon PMA stimulation [214]
Human intestinal microbiota	Larva	Injection into gut, immersion [215]	N/A
Mouse intestinal microbiota	Larva	Immersion [96]	N/A
Fungus
A. fumigatus	Embryo	HBV [216]	Recruitment to hyphae but not conidia, LAD model supports critical role of neutrophils for survival [216]
C. albicans	Adult	i.p. [217, 218]	N/A
	Embryo	HBV [92, 217, 219]; Y [217]	Recruitment [219, 220] mediated by NADPH oxidase [219]; phagocytosis and killing of yeast cells [92]
	Larva	Swimbladder (injection) [219]; swimbladder (immersion) [220]	Recruitment [220]
Penicillium marneffei (spores)	Larva	Somatic muscle [7]	Phagocytosis [7]

^aAs described in ref. [190].

P. aeruginosa

P. aeruginosa, a gram-negative bacterium, is an important, opportunistic pathogen of humans that causes localized and systemic disease. Most clinical cases of P. aeruginosa infection are associated with some compromise in immune function as a result of illness, injury, or underlying immune disorders, such as neutropenia [221]. Chronic infection caused by P. aeruginosa is the leading cause of pulmonary infection in cystic fibrosis patients [221]. In mammalian P. aeruginosa infection, neutrophils are known to be critical for host protection. Mice depleted of neutrophils succumb to localized infection within 2 days [222]. Human neutropenic hosts are also hypersensitive to Pseudomonas infection [223]. Although a number of plant and animal models have been used to study P. aeruginosa virulence, including Arabidopsis thaliana [224], Caenorhabditis elegans [225], Drosophila melanogaster [226], and mice [222], zebrafish have been developed more recently as a model to study P. aeruginosa infection by use of localized [79, 170] and systemic infection [165]. Zebrafish, like humans, are also relatively resistant to systemic infection with P. aeruginosa, and large inocula are needed to induce infection [165]. Localized injection of Pseudomonas into the OV of zebrafish larvae induces neutrophil recruitment (Fig. 1C; and Supplemental Video 2), and ∼65% of the neutrophils contain phagocytosed bacteria [170]. Neutrophils and macrophages phagocytose and kill P. aeruginosa, and depletion of phagocytes results in increased susceptibility to infection [165]. Neutrophils are critical in the host response to localized Pseudomonas infection, as shown by 2 zebrafish models of neutrophil deficiency: LAD [79] and WHIM syndrome [41]. Although LAD and WHIM fish have normal macrophage responses to infection [41, 43, 79], localized infection with P. aeruginosa results in increased mortality, likely as a result of impaired neutrophil function [41, 79]. This supports a more critical role for neutrophils than macrophages in the host defense against Pseudomonas infection in zebrafish, similar to humans. In a zebrafish model of CGD, larvae deficient in 2 key subunits of the NADPH oxidase complex (gp91^phox and p22^phox) have increased susceptibility to P. aeruginosa infection, suggesting that neutrophil ROS signaling is likely important for host defense to P. aeruginosa [41]. Oxidative killing is similarly important in mice, as p47^phox -deficient animals have an increased bacterial burden [227]. The Pseudomonas-zebrafish infection model has also provided new insight into the signaling mechanisms that mediate neutrophil recruitment to localized infection. Unlike wound recruitment, neutrophil recruitment to OV P. aeruginosa infection does not require tissue-generated hydrogen peroxide [170, 228]. Similar to mice [222], neutrophil recruitment to P. aeruginosa infection requires Cxcr2-Cxcl8 signaling [171].

S. iniae

S. iniae is a natural fish pathogen and causes systemic disease, characterized by meningitis and sepsis in fish and humans. S. iniae causes 30–50% mortality in affected fish ponds, resulting in $100 million in annual losses for the aquaculture industry [229–231]. Localized S. iniae infection induces recruitment of leukocytes to infection sites in mice and zebrafish [43, 190, 191]. In mice and adult zebrafish, infection spreads systemically and results in rapid host mortality [190, 191]. Localized OV infection with S. iniae is also extremely lethal in larvae, with only 10 CFUs capable of causing significant host mortality [43]. In contrast, the capsule-deficient cpsA mutant is nonpathogenic in zebrafish [43, 192, 193], supporting the importance of the polysaccharide capsule in protecting this bacterium from the host immune response. Following OV injection with wild-type S. iniae or the cpsA mutant, neutrophils exhibit a type of behavior referred to as “neutrophil swarming” that has also been described in mice [232], in which there is initial chemotaxis of nearby neutrophils, followed by a second phase of amplified recruitment and clustering of neutrophils. Besides recruitment, macrophages and neutrophils (Fig. 1D and Supplemental Video 3) efficiently phagocytose wild-type S. iniae, as well as the cpsA mutant. This finding is in contrast to results with E. coli injected into fluid-filled cavities, such as the OV, where neutrophils, unlike macrophages, are inefficient at phagocytosing bacteria [46]. However, it has been shown that neutrophils are, in fact, capable of phagocytosing certain microbes, such as S. iniae [43] and P. aeruginosa [41, 170], even in these closed cavities. Thus, it appears that the efficiency of neutrophil phagocytosis may be dependent on the type of pathogen. Interestingly, infection of LAD fish shows that neutrophils are necessary for host survival in response to wild-type S. iniae but not the cpsA mutant. These findings suggest a protective role for neutrophils in wild-type S. iniae but not cpsA mutant infection. However, when neutrophils and macrophages are depleted, there is an increased susceptibility to the cpsA mutant, suggesting an important role for macrophages in this context. Interestingly, the closely related S. pyogenes has developed a mechanism for inhibiting neutrophil recruitment. Studies in mice were complemented by live imaging studies in zebrafish larvae to show that S. pyogenes expresses the cytolytic toxin SLS that limits neutrophil recruitment and optimizes pathogen dissemination [45]. SLS-deficient bacteria recruit more neutrophils and have reduced lethality of infection in zebrafish [45]. This is a nice example of how pathogens develop strategies to limit neutrophil recruitment to promote pathogen growth.

M. marinum

M. marinum is a natural fish pathogen and a close genetic relative of Mtb [233], the etiology of human TB. In adult [148, 234] and larval zebrafish [44] models of infection, M. marinum causes chronic disease with necrotic, granulomatous lesions, similar in structure to those induced by Mtb in humans. In the larval zebrafish model, macrophages not only form granuloma-like aggregates with similar histology to TB granulomas but also induce mycobacterial expression of granuloma-specific genes [44]. The optical accessibility of the larval zebrafish has provided important, new insights into host [41, 149–151, 235] and pathogen [152, 153, 236]. With the allowance of repeated live imaging of infected larvae, the zebrafish model has changed the widespread view of the granuloma as a strictly host-protective structure. In a series of studies, it was found that a mycobacterial virulence determinant, the ESX-1/RD1 secretion system, drives the formation of early granulomas [152] through the secretion of ESAT-6, which induces production of MMP9 by nearby epithelial cells [153]. This localized MMP9 enhances recruitment of macrophages to foci of infection, enhancing granuloma formation [153]. The appearance of granulomas coincides with and contributes to a marked increase in bacterial proliferation [152], in contrast to the previous idea that granulomas are primarily host-protective structures. The larval zebrafish model of M. marinum infection has also helped to uncover a host-protective role for neutrophils in the early granuloma. Although neutrophils are present at sites of TB infection, such as in the cerebrospinal fluid of TB meningitis patients [237] or in the granulomas of pulmonary TB patients [238], previously, there has been a limited understanding of their role in mycobacterial pathogenesis. Some in vitro studies show that isolated human neutrophils kill Mtb [239, 240], whereas others seem to demonstrate their inability to do so [241]. In mouse models, the role of neutrophils in mycobacterial infection is also unclear. A study by Pedrosa et al. [242] indicates that neutrophils play a protective role in systemic Mtb infection that is independent of their phagocytic ability. Although neutrophils are recruited to infection foci, bacilli are only found inside macrophages but not neutrophils [242]. However, depletion of neutrophils results in increased severity of infection [242]. Thus, instead of a direct role for neutrophils in bacterial killing, these findings suggest an immunomodulatory role for neutrophils in the immune response to Mtb [242]. Sugawara et al. [243] also demonstrated a protective role for neutrophils in a rat model where neutrophil depletion results in worsened disease, whereas LPS-induced neutrophilia actually prevents early mycobacterial infection. However, other in vivo studies have suggested that neutrophils have a detrimental effect on the host response to Mtb [154]. Recent work that uses the zebrafish model has helped to clarify the role of neutrophils in mycobacterial infection. Systemic infection with M. marinum in zebrafish embryos results in the recruitment of many macrophages and few neutrophils [41, 244]. Of the neutrophils recruited, very few become infected with M. marinum [244]. Live imaging reveals that neutrophils traffic into granulomas and interact closely with infected macrophages and can become infected by engulfing macrophages that contain M. marinum. Treatment with a pan-caspase inhibitor that blocks apoptosis in zebrafish [245] impairs neutrophil recruitment to granulomas, suggesting that neutrophils respond to signals released from apoptotic macrophages. This scavenging role for neutrophils is surprising, as macrophages have long been thought to be the sole phagocyte responsible for removing apoptotic cells and cell debris during resolution of inflammation (described in zebrafish in [7] and reviewed in ref. [246]). Neutrophils also act as scavengers in a zebrafish model of systemic S. flexneri infection [179]. This illustrates the power of the use of zebrafish to dissect the role of immune cells in a living, intact animal. The importance of neutrophils to host defense against mycobacterial infection has been demonstrated directly by the use of the neutropenic WHIM zebrafish [67], which show increased susceptibility to M. marinum infection [41]. Infection of fish depleted of 2 components of the NADPH oxidase complex (gp91^phox and p22^phox) also results in increased susceptibility to infection compared with controls [41]. This NADPH oxidase-dependent killing of mycobacteria may provide a possible explanation for the long-standing observation of the susceptibility of CGD patients to TB.

Fungal disease

C. albicans and A. fumigatus undergo morphologic changes that induce specific immune responses. In zebrafish larvae, C. albicans yeast can germinate to form hyphae, but this switch in morphology is inhibited after phagocytosis of the yeast by neutrophils and macrophages [92]. The NADPH oxidase complex is necessary to control this change in morphology [92, 219]. However, the specific role of neutrophils in controlling these morphologic changes remains unclear. A morphology-specific innate immune response occurs to the opportunistic human pathogen A. fumigatus in zebrafish [216]; macrophages are recruited to spores and hyphae, whereas neutrophils are recruited only to hyphae. The infectious particles of Aspergillus are airborne asexual spores called conidia, and because of the prevalence of Aspergillus in the environment, the average individual is thought to inhale as many as several hundred conidia per day [247]. Immunocompromised populations, particularly neutropenic patients, are at risk for developing the often-fatal disease, Invasive Aspergillosis, characterized by conidial germination into tissue-penetrating hyphae [248, 249]. Like humans and mice, immunocompetent fish are not susceptible to A. fumigatus infection [216], but depletion of neutrophils and macrophages in zebrafish increases susceptibility to infection and results in conidial germination into tissue-penetrating hyphae, similar to aspergillosis in humans. LAD zebrafish with impaired neutrophil function develop invasive disease and have impaired survival with A. fumigatus [216]. Interestingly, the inability of neutrophils to phagocytose A. fumigatus conidia is pathogen-specific, as neutrophils respond to and phagocytose C. albicans yeast forms [92]. Schaffner et al. [250] also demonstrated distinct roles for neutrophils and monocytes in targeting hyphae and conidia, respectively, in a mouse model of Aspergillus infection, which is consistent with observations made in zebrafish. Accordingly, neutropenic mice are unable to control hyphal growth [250]. Isolated human neutrophils also show a higher induction of NETosis in response to hyphae compared with conidia [251], and hyphal growth is inhibited by NET production [252]. Taken together, the zebrafish model of fungal disease highlights the conservation of the innate immune response to different morphologic forms of Aspergillus species, and the optical accessibility provided by zebrafish has helped to uncover the distinct responses of neutrophils to different morphologic forms of fungi in vivo.

CONCLUDING REMARKS

Zebrafish represent a powerful model system to study the role of neutrophils in host-pathogen interactions that complements studies in mice and humans. Challenges remain in dissecting how neutrophils modulate disease, as neutrophils can have beneficial and detrimental effects. For example, in a mouse model of S. aureus infection, neutrophils crawling in capillaries around the site of infection have a potentially beneficial role early on in the infection in limiting bacterial dissemination, but continued presence at the infection site results in increased tissue injury [253]. It has been shown in mice and zebrafish that neutrophils can traffic away from sites of tissue damage in a process known as neutrophil reverse migration [5, 30]. Although the contribution of neutrophil reverse migration in the context of infection remains unclear, the genetic tractability and optical transparency of zebrafish larvae can help elucidate novel mechanisms of neutrophil behavior in response to tissue damage and infection. It is possible that reverse migration of neutrophils may serve as a mechanism of systemic activation of the immune response [34]. Alternatively, infected neutrophils may provide an intracellular niche for persisting microbes, serving as "Trojan Horses" to disseminate bacteria from the initial infection site, as has been suggested with S. aureus infection in zebrafish [56]. The use of labeled bacteria and transgenic fish lines with labeled phagocytes will provide further insight into the temporal and spatial interactions between neutrophils and macrophages during infection. Finally, zebrafish models of neutropenia and other human neutrophil disorders have made it possible to elucidate the specific role of neutrophils in host defense to different types of bacterial and fungal infections. Future studies will be needed to determine if these models can be used to advance treatments that target innate immunity and infection. Recent studies that use small molecule screening in zebrafish mycobacterial models have raised the possibility that zebrafish infection models can be used to identify new therapeutic strategies to treat infectious disease [254]. It will be particularly intriguing to understand how the inflammatory status affects response to a particular therapy, with the aim toward host-directed therapeutic strategies.

E.A.H. performed the experiments and wrote the paper; A.H. cowrote the paper.

Glossary

Cas

CRISPR-associated

CGD

chronic granulomatous disease

CHT

caudal hematopoietic tissue

CRISPR

clustered regularly interspaced short palindromic repeats

CV

caudal vein

DC

duct of Cuvier

DM

dorsal muscle

dpf

days postfertilization

ESAT-6

early secretory antigenic target of Mycobacterium tuberculosis

ESX-1

early secretory antigenic target 6 system 1

HBV

hind brain ventricle

hpf

hours postfertilization

irf8

IFN regulatory factor 8

LAD

leukocyte adhesion deficiency

lyz

lysozyme C

MMP9

matrix metalloproteinase 9

mpeg1

macrophage expressed gene 1

mpx

myeloid peroxidase

Mtb

Mycobacterium tuberculosis

NET

neutrophil extracellular trap

OV

otic vesicle

PC

pericardial cavity

PHOX

phagocyte NADPH oxidase complex

Rac1/2

Ras-related C3 botulinum toxin substrate 1/2

RD1

region of difference 1

RNS

reactive nitrogen species

ROS

reactive oxygen species

SDF1

stromal cell-derived factor 1

SLS

streptolysin S

TB

tuberculosis

TALEN

transcription activator-like endonucleases

WAS

Wiskott Aldrich syndrome

WASp

Wiskott-Aldrich Syndrome protein

WHIM

warts, hypogammaglobulinemia, infections, and myelokathexis

Y

yolk sac

DISCLOSURES

The authors declare no competing financial interests.