Illuminating Macrophage Contributions to Host-Pathogen Interactions In Vivo: the Power of Zebrafish

Macrophages are a key cell type in innate immunity. Years of in vitro cell culture studies have unraveled myriad macrophage pathways that combat pathogens and demonstrated how pathogen effectors subvert these mechanisms. However, in vitro cell culture studies may not accurately reflect how macrophages fit into the context of an innate immune response in whole animals with multiple cell types and tissues. Larval zebrafish have emerged as an intermediate model of innate immunity and host-pathogen interactions to bridge the gap between cell culture studies and mammalian models.

ABSTRACT

Macrophages are a key cell type in innate immunity. Years of in vitro cell culture studies have unraveled myriad macrophage pathways that combat pathogens and demonstrated how pathogen effectors subvert these mechanisms. However, in vitro cell culture studies may not accurately reflect how macrophages fit into the context of an innate immune response in whole animals with multiple cell types and tissues. Larval zebrafish have emerged as an intermediate model of innate immunity and host-pathogen interactions to bridge the gap between cell culture studies and mammalian models. These organisms possess an innate immune system largely conserved with that of humans and allow state-of-the-art genetic and imaging techniques, all in the context of an intact organism. Using larval zebrafish, researchers are elucidating the function of macrophages in response to many different infections, including both bacterial and fungal pathogens. The goal of this review is to highlight studies in zebrafish that utilized live-imaging techniques to analyze macrophage activities in response to pathogens. Recent studies have explored the roles of specific pathways and mechanisms in macrophage killing ability, explored how pathogens subvert these responses, identified subsets of macrophages with differential microbicidal activities, and implicated macrophages as an intracellular niche for pathogen survival and trafficking. Research using this model continues to advance our understanding of how macrophages, and specific pathways inside these cells, fit into complex multicellular innate immune responses in vivo, providing important information on how pathogens evade these pathways and how we can exploit them for development of treatments against microbial infections.

INTRODUCTION

The first line of defense against infecting pathogens is the innate immune response. Macrophages are a key cell type in this response. However, macrophages do not have as much microbicidal activity as some other immune cells, such as neutrophils (1), and the pathways that regulate this activity are incompletely understood (2). Macrophages also have immunomodulatory functions, such as promoting tissue homeostasis and resolution of inflammation (3). It is therefore appreciated that in some infection contexts, macrophages can actually promote infection persistence (4, 5). Altogether, the roles and functions of macrophages in vivo against many pathogens are incompletely understood.

A major barrier to understanding cell function in vivo is the inability to observe and measure cell behavior in real time in a live, intact host and to correlate these behaviors with infection outcome. While imaging modalities in mice are improving, these experimental setups are complicated and expensive. Generally, researchers have to choose between high-resolution, invasive, short-term imaging (6, 7) and lower-resolution imaging that allows longitudinal tracking (8, 9). Because of these limitations, the larval zebrafish model has emerged as an intermediate vertebrate model in which to study host-pathogen interactions (10). Larval zebrafish are optically transparent and relatively small (∼5 mm long), and therefore, the entire animal can be imaged using high-resolution light microscopy, even for multiple days in a row. Transgenic lines fluorescently marking specific cell types—such as macrophages (11, 12)—have been established, which can be combined with fluorescently labeled injected pathogens for high-resolution imaging of host cell responses and pathogen growth or control. This has allowed visualization of processes that were not previously observable in vivo, such as replication of pathogens inside macrophages (13), recruitment of intracellular proteins to pathogen compartments, such as autophagic machinery (14, 15), and the effect of infection on cell activation, including calcium flux through the cell (16, 17).

The immune system of larval zebrafish is relatively conserved to that of humans, and zebrafish are used as a host to investigate infections relevant to humans (18). These include both bacterial infections, such as those caused by mycobacterial species (19), and fungal infections, such as those caused by Aspergillus fumigatus and Cryptococcus neoformans (20). Adaptive immunity does not develop in zebrafish until ∼4 to 6 weeks postfertilization (21), and experiments in the early larval stage therefore interrogate innate immune cell function in isolation. Larvae are typically infected with pathogens via microinjection, and experimental outcomes can be monitored by survival assays, CFU assays, or high-resolution imaging. Such experiments have demonstrated that zebrafish macrophages have many of the same functions as human macrophages. For example, zebrafish macrophages can phagocytose multiple injected pathogens, including both bacteria, such as Salmonella (22), Staphylococcus aureus (23), Streptococcus iniae (24), Pseudomonas aeruginosa (25), and Mycobacterium marinum (26), and fungi, such as Candida albicans (27), A. fumigatus (28), and C. neoformans (13).

In this review, I delve into the function of macrophages in response to pathogens, relying primarily on studies that have taken advantage of the imaging capabilities of the zebrafish model. These studies have identified pathways in macrophages that control pathogens, determined how certain pathogens subvert these responses, and identified infection contexts where macrophages actually promote pathogen survival and dissemination (Fig. 1).

FIG 1.

Macrophage effector mechanisms elucidated in larval zebrafish. Using larval zebrafish, researchers have identified new pathways in macrophages that control pathogens as well as multicellular macrophage responses and mechanisms by which pathogens subvert these effector mechanisms. Selected key findings from these studies are illustrated here. (A) Phagocytosis of pathogens and their intracellular fate can be visualized inside an intact living host in larval zebrafish. Through such imaging studies, pathways such as selective autophagy and Lc3-associated phagocytosis, as well as new initiators of ROS production, were found to be activated in macrophages in response to specific pathogens. However, other pathogens can subvert these responses, inhibiting phagocytosis through changes in pathogen shape, size, and structure, or escaping macrophage-mediated killing and utilizing these cells as an intracellular niche for persistence and growth. (B) Multicellular innate immune mechanisms are also visualized and easily studied in zebrafish. (Panel i) Granuloma formation occurs in larval zebrafish after infection with M. marinum in the absence of an adaptive immune system. Through multiple studies, the formation, development, and maintenance of this structure are being elucidated. Macrophages in granulomas can express epithelial markers, disseminate out of the granuloma to seed secondary sites of infection, undergo necrosis in the granuloma, releasing extracellular bacteria, and replenish the granuloma with new cells. (Panel ii) Macrophage aggregation is also found in other infection contexts in larval zebrafish; however, whether these clusters of cells are structurally similar to granulomas is unknown. (Panel iii) Live imaging in zebrafish can also visualize the transfer of pathogens between immune cells. This phenomenon is observed in multiple infections, including transfer from tissue-resident macrophages to inflammatory monocytes and from neutrophils to macrophages.

PATHWAYS IN MACROPHAGES THAT CONTROL PATHOGENS

A canonical role for macrophages in response to infection is to promote control of the pathogen by phagocytosis and killing. However, the pathways in macrophages that promote this killing are incompletely understood. Additionally, pathways identified in in vitro tissue culture experiments may not have the same function in vivo. Zebrafish are a powerful experimental system in which to address these questions in vivo. After infection with many pathogens, including M. marinum (29), Mycobacterium leprae (30), Listeria monocytogenes (31), and A. fumigatus (28), macrophages provide protection to the host, demonstrated by decreased survival and/or increased pathogen burden if macrophages are depleted. The ease of innate immune cell depletion in larval zebrafish also has allowed for the identification of redundant roles of macrophages. For example, in a P. aeruginosa infection, while macrophages are not required for host survival, if neutrophils are defective, macrophages can provide some protection (32).

Work in zebrafish has also identified intracellular macrophage responses to infection that promote pathogen control (Fig. 1A). For example, a gene that was originally identified in zebrafish as a macrophage lineage-specific marker, mpeg1 (11), was later found to encode a perforin protein that may directly target the membranes of bacterial pathogens (33). A common pathway implicated in the response to many pathogens in macrophages is autophagy, which can be easily studied in real time in transgenic larval zebrafish with a green fluorescent protein (GFP)-tagged Lc3 protein, a common autophagosome marker (34). One way in which autophagy can be involved in macrophage responses is by targeting pathogens that escape the phagosome. This response involves selective autophagy receptors that recognize such pathogens and deliver them to autophagic machinery (35). Zhang et al. found that after injection of M. marinum, the bacteria colocalize with GFP-Lc3 inside macrophages and that knockdown of the receptor p62 or optineurin decreased this colocalization and led to higher bacterial burdens, implicating this pathway in control of this pathogen (14). Shigella is similarly targeted by autophagy pathways intracellularly in zebrafish, although whether this occurs in macrophages is unclear (36). However, Lc3 and autophagy pathways can also be involved differently, in a process called Lc3-associated phagocytosis (LAP). In response to a different pathogen, Salmonella enterica serovar Typhimurium, LAP is required for bacterial control. Again, this bacterium colocalizes with GFP-Lc3 in macrophages, but this colocalization is dependent on Rubicon (15), a protein that acts as a molecular switch between autophagy and LAP (37).

Reactive oxygen species (ROS) production has also long been known to be involved in the response against pathogens, but the role of ROS in vivo in macrophages and the various pathways that can induce ROS are not fully elucidated. In zebrafish, two new controllers of ROS production have been identified and tied to targeting of pathogens by macrophages. The first is CFTR, the cystic fibrosis (CF) transmembrane conductance regulator, mutations in which are the cause of CF (38). For a long time, the only cause of disease and susceptibility to infection in CF patients was thought to be a lack of mucociliary clearance in the lungs, but it has recently been appreciated that this defect may directly affect immune cell function as well (39, 40). Bernut et al. recently found that in response to Mycobacterium abscessus, CFTR deficiency decreases ROS production by macrophages, which subsequently decreases the control of intracellular bacteria and increases macrophage death (41). Experiments in zebrafish also identified a new inducer of mitochondrial ROS (mROS) production in macrophages, Irg1 (22). mROS promotes killing of S. Typhimurium intracellularly in macrophages, and knockdown of Irg1 leads to a higher intracellular bacterial burden in macrophages and decreased host survival (22).

Because macrophages are a key first responder to many pathogens, other studies have implicated macrophages as a cell type that could be targeted with host-directed therapies to increase pathogen control. For example, the antifungal drug voriconazole requires the presence of macrophages for efficacy against A. fumigatus infection in larval zebrafish (42). Matty et al. took this one step further, performing a chemical screen for drugs that specifically inhibit M. marinum growth inside live hosts (16). This screen identified the drug clemastine, which increases the microbicidal activity of macrophages and therefore intracellular control of M. marinum. They found that a target of clemastine is P2RX7, a calcium channel. By targeting this channel, clemastine increases calcium flux into the cell, activating the inflammasome. Alterations in calcium flux are another example of a phenotype that would have been difficult to observe in vivo without the live imaging and genetic tools available in the larval zebrafish model. Depleting macrophages, mutating the inflammasome adaptor ASC, and decreasing inflammasome activation by infecting with a strain of M. marinum that lacks the ESX-1 type VII secretion system all decreased the ability of clemastine to decrease bacterial burden, further implicating macrophages and inflammasome activation in control of M. marinum (16).

A well-known multicellular macrophage response to infection with an important human pathogen, Mycobacterium tuberculosis, is the granuloma, a structure made up of macrophages and other immune cells that contain the bacteria (43). Larval zebrafish were instrumental in demonstrating that mycobacterial granulomas can form in the absence of the adaptive immune system. Infection of larvae with M. marinum, a mycobacterial species that is a natural pathogen of fish, leads to the formation of granuloma-like structures, even though these larvae are less than 5 days old and do not yet possess any adaptive immune cells, like T cells (26). The best evidence that these structures are real granulomas is that the bacteria within them recognize the structures as a granuloma environment. M. marinum organisms within these macrophage aggregates turn on the expression of genes that are expressed only in granulomas, not in single macrophages or under any in vitro conditions (26).

Work in zebrafish identified some of the mechanisms by which this structure develops and is maintained (Fig. 1B, panel i). Macrophages that phagocytose M. marinum later burst, releasing bacteria extracellularly, which are then taken up by new macrophages (44). A constant influx of new macrophages was found to be required to maintain the granuloma, replenish dying necrotic macrophages, and prevent bacterial extracellular growth (45). Additionally, Cronan et al. found that these macrophages reprogram to express epithelial markers in the granuloma and form a more ordered structure, although this process actually inhibited immune cell access and promoted bacterial growth (46). Surprisingly, experiments in larval zebrafish visualizing the response to many pathogens in vivo identified similar “clusters” or “aggregates” of macrophages in response to many different microbes, including both other bacteria, such as Burkholderia cenocepacia (47) and S. iniae (48), and fungi, such as A. fumigatus (49), C. neoformans (13), and Mucor circinelloides (50). It is unknown if these clusters have the same complex structure as established and developed mycobacterial granulomas, but these experiments do identify clustering as a conserved early macrophage response (Fig. 1B, panel ii).

MACROPHAGE PROINFLAMMATORY GENE EXPRESSION

A major role of macrophages in response to infection is the production of cytokines. Zebrafish macrophages can polarize into different phenotypes along an M1-M2 axis, similar to mammalian macrophages, producing either proinflammatory or anti-inflammatory proresolution cytokines, with proinflammatory phenotypes dominating upon infection (51, 52). In some cases, these proinflammatory cytokines have positive effects, such as increasing the intracellular control of bacteria by macrophages or recruiting further immune cells, such as neutrophils. However, in other cases these cytokines can have negative effects, such as increasing damage to host tissues and cells. Both of these responses are easily observed in larval zebrafish in real time and have been popular areas of study.

Tumor necrosis factor alpha (TNF) is perhaps the best-studied cytokine produced by macrophages in larval zebrafish and is part of a major signaling axis controlling the balance of host inflammation and pathogen control. TNF can restrict the growth of both M. marinum (53) and M. abscessus (54) in macrophages in larval zebrafish. However, in several sequential studies, Tobin et al. found that either too low or too high TNF levels can be detrimental to the host and that levels of TNF can be controlled by the balance between the proinflammatory LTB4 and the proresolution LXA4 eicosanoids (55, 56). Too little TNF leads to decreased control of bacteria by macrophages, with more intracellular bacterial growth and eventually cell lysis, granuloma breakdown, and excessive extracellular bacterial growth (53, 55). However, while increasing the level of TNF can help macrophages to control intracellular M. marinum replication early in infection, it also eventually leads to macrophage death and lysis (56). This programmed necrosis occurs through an interorganellar circuit involving mROS, lysosomal ceramide, the cytosolic protein BAX, calcium flux into mitochondria, and cyclophilin D-mediated necrosis (57, 58). Black et al. also found that macrophages are the major producers of ROS after M. marinum infection and that this ROS leads to increased host cell death (59). Both groups reported that treatment with a ROS scavenger can decrease host cell death and increase control of bacterial growth (57, 59).

Studies on the role of eicosanoids in the immune response to M. marinum also provided a clear example of how discoveries in larval zebrafish can translate to human infectious disease treatment. In two studies, Tobin et al. investigated the effect of multiple single nucleotide polymorphisms (SNPs) in the human LTA4H locus, some intronic (55) and one in the promoter of the gene (56), on the mortality of patients infected with meningeal tuberculosis. Heterozygote status at several of these SNPs was significantly associated with decreased mortality of patients (55, 56). An SNP in the promoter of LTA4H was found to influence the expression of this gene in human cells and therefore is likely to alter the proinflammatory LTB4-proresolution LXA4 balance and downstream TNF production in these cells as well, through a mechanism similar to that in zebrafish cells (56). This promoter SNP also influences how patients respond to a specific therapy given to meningeal TB patients—dexamethasone (56). Dexamethasone is an anti-inflammatory drug, and Tobin et al. found that it preferentially improved the survival of patients homozygous for the LTA4H promoter SNP associated with higher LTA4H expression and therefore higher proinflammatory TNF levels (56).

However, the finding that too much proinflammatory cytokine expression can be detrimental to the host is not limited to M. marinum infection or always due to uncontrolled bacterial extracellular growth. Macrophages in M. marinum granulomas can also produce proangiogenic cytokines such as VEGF and ANG-2, which promote bacterial growth and dissemination (60, 61). In infections with another mycobacterial species, M. leprae, Madigan et al. were able to visualize the nerve damage that this infection causes, including myelin breakdown and removal from axons (30). Surprisingly, depletion of macrophages abrogated this damage, and the authors elucidated a pathway whereby a phenolic glycolipid on the surface of the bacteria activates infected macrophages to produce nitric oxide (30). This nitric oxide then induces mitochondrial damage in axons near bacterium-macrophage aggregates (30).

The recruitment of new immune cells is a major role of cytokine production, and this can also be easily visualized and quantified in the larval zebrafish model. In addition to promoting intracellular control of M. abscessus, TNF produced by macrophages also promotes production of the neutrophil chemokine Cxcl8 and the recruitment of neutrophils to help contain the infection (54). Neutrophil recruitment to M. marinum granulomas is also important for long-term pathogen containment, and signals from dying macrophages are important for this recruitment (62). Interleukin-1β (IL-1β) is another major cytokine which can recruit more immune cells, such as neutrophils, and macrophages are also the primary producers of IL-1β after Escherichia coli infection (63). In humans and mice, IL-1β must be cleaved from a “pro-” form to be activated, a process that is executed by inflammasome and caspase-1 activation. While inflammasome activation is known to occur in zebrafish (31), this process is poorly understood, and most studies investigating IL-1β rely solely on expression or reporter lines as a readout (64).

PATHOGEN SUBVERSION MECHANISMS

While macrophages have many pathways that can target and control pathogens, pathogens have also evolved mechanisms by which to evade these responses. It has been fairly simple in murine and other animal models to test whether a specific pathogen factor is important for virulence by tracking host survival after infection with mutant strains. However, which of these factors and phenotypes directly affect pathogen interactions with macrophages in vivo has not been so clear, and this is another area of research for which the larval zebrafish host model is ideal.

One mechanism that pathogens use to evade macrophages is that of inhibiting their phagocytosis by macrophages (Fig. 1A). By this reasoning, “cording,” which is a type of extracellular growth that can be seen in multiple mycobacterial species and easily observed in larval zebrafish, can be considered a virulence mechanism, as it creates structures that are too large to be phagocytosed by macrophages (53, 65). Similarly, macrophages prefer to phagocytose C. neoformans yeast cells with smaller polysaccharide capsules, but C. neoformans increases cell size as well as the size of the polysaccharide capsule over the course of infection, inhibiting this uptake (13, 66).

Pathogens can also evade macrophage microbicidal mechanisms after they are phagocytosed. A major intracellular microbicidal mechanism of macrophages is trafficking pathogens to acidic compartments, including by fusion of phagosomes with lysosomes. Levitte et al. visualized this trafficking in larval zebrafish, reporting that ∼50% of injected M. marinum organisms are trafficked to acidic compartments in macrophages in vivo (67). However, by visualizing larvae infected with single bacteria, they found that even if this trafficking occurs, in ∼60% of infections the bacteria survive and may even replicate (67). This acidic tolerance of the bacteria depends in part on the bacterial factor MarP, a membrane serine protease (67).

Simply existing inside host cells can alter microbe cell biology in ways that increase pathogen survival during infection. For example, Adams et al. used the transparency of larvae and ease of drug delivery via the larval water to find that treatment of M. marinum-infected larvae with antitubercular drugs cannot clear all bacteria (68). Instead, a subset of bacteria tolerate the drug and persist by activation of drug efflux pumps (68). This tolerance depends on the bacteria having an intracellular cycle inside macrophages—macrophage depletion decreases the percentage of bacteria that tolerate the drug (68).

Another way for pathogens to evade macrophage responses is to prevent recognition by macrophages and to recruit new immune cells that are less microbicidal, a mechanism utilized by M. marinum, as elucidated in two publications by Cambier et al. (69, 70). M. marinum cells have phthiocerol dimycoceroserate (PDIM) lipids on their surfaces, and these lipids mask pathogen-associated molecular patterns on the bacteria (69). If the bacteria lack these PDIM lipids, macrophages at the site of infection are induced to express inducible nitric oxide synthase, and microbicidal macrophages are recruited through Toll-like receptor–Myd88 signaling (69). However, M. marinum organisms also have another type of lipid on their surfaces, phenolic glycolipids (PGLs), and these PGLs promote recruitment of a different type of cell—Ccr2⁺ inflammatory monocytes—that are less microbicidal and promote bacterial persistence (69). This recruitment depends on the initially responding tissue-resident macrophages, in which PGLs activate STING to induce production of the chemokine Ccl2 and recruitment of Ccr2⁺ cells (70). M. marinum, through ESX-1-dependent virulence mechanisms, can also activate a different cell type—thrombocytes—to interact with granuloma macrophages and increase bacterial burden (71).

MACROPHAGES AS AN INTRACELLULAR NICHE FOR PATHOGENS

While macrophages can act to control many microbes, they are not always microbicidal, and in fact, multiple pathogens take advantage of macrophages as an intracellular niche in which to develop, replicate, or evade other immune cells (Fig. 1A).

One bacterial pathogen that can replicate inside macrophages in vivo is B. cenocepacia (47). Live imaging of this infection in larval zebrafish detected replication of bacteria inside macrophages, followed by nonlytic release from the cells and bacteremia (47). Correspondingly, depletion of macrophages led to increased host survival and decreased bacterial burden, underlining the importance of macrophages in progression of this infection (72).

Using macrophages as an intracellular niche is a common mechanism utilized by fungal pathogens, possibly because these pathogens mask their cell surfaces well and are harder for macrophages to recognize. For example, survival and replication of Candida albicans inside macrophages can be observed by live imaging in larval zebrafish (27). While macrophages promote control of a highly virulent H99 strain of C. neoformans, even in these infections, yeast replication can be observed inside macrophages (13, 66). This intracellular niche is even more pronounced in infections with a lower-virulence strain and with C. neoformans spore, as opposed to yeast, inoculations (73). Inside macrophages, C. neoformans spores can also germinate into yeasts and are then released back into the vasculature, resulting in a secondary fungemia in the host days after the initial infection (73). Genetically inhibiting the development of macrophages and proportionally increasing neutrophil development leads to better clearance of C. neoformans (73).

Fungal pathogens can also utilize macrophages as a protective niche against other, more microbicidal, immune cells, such as neutrophils. In A. fumigatus infections, macrophages are the major cell type responsible for inhibiting spore germination into hyphae in vivo (42, 49). By inhibiting this germination, macrophages actually inhibit neutrophil recruitment and neutrophil-mediated killing, both of which are primarily driven by hyphal forms of A. fumigatus (49). As a result, fungal burden can be decreased when macrophages are depleted (49). While macrophages inhibit A. fumigatus spore germination, they do not completely prevent it, and in fact, the majority of germination happens from within macrophages, also making these cells a niche for fungal development (74). Germination is another example of an infection process that can be uniquely visualized in vivo in larval zebrafish. Macrophages provide a similar protective niche for the fungus Talaromyces marneffei (75). Macrophages can protect T. marneffei from myeloperoxidase-dependent neutrophil-mediated killing, and consequently, depletion of macrophages leads to a decrease in fungal burden (75).

MACROPHAGES CAN PROMOTE INFECTION DISSEMINATION

In addition to promoting the survival of some pathogens, macrophages can also promote pathogen dissemination. The dissemination of some pathogens is crucial for causing disease and is a process that is very difficult to visualize in most animal models. Larval zebrafish provide a system in which the entire animal can be imaged at once. For example, C. neoformans infections have a tendency to cross the blood-brain barrier into the central nervous system to cause disease (76), and this dissemination has been visualized in larvae (73). Whether macrophages play a role in this process is still unclear.

Even in cases where the presence of macrophages promotes control of the infection, macrophages can still act as vehicles for pathogens to disseminate away from the infection site. Clay et al. referred to this as a “dichotomous role” of macrophages in response to M. marinum, showing that while depleting macrophages led to increased host death, it also decreased the dissemination of bacteria out of a localized infection site (29) (Fig. 1B, panel i). Davis and Ramakrishnan visualized and quantified this process of macrophages leaving primary granulomas to initiate secondary granulomas, using an M. marinum strain expressing a photoconvertible fluorescent protein (77). With this strain, the authors labeled only bacteria that were present in a primary granuloma and conclusively showed that macrophages containing these bacteria can leave this site and establish secondary granulomas elsewhere in the larvae (77). Altering macrophage migration in the context of infection by mutating the receptor Cxcr3 decreased this dissemination of the bacteria (78). Similar results are seen with other mycobacterial species—macrophages both control and promote the dissemination of M. abscessus (65) and M. leprae (30). The clustering of macrophages in response to the fungus M. circinelloides also was found to correlate with infection dissemination (50).

RELEASE AND TRANSFER OF PATHOGENS FROM AND INTO MACROPHAGES

A relatively newly discovered phenomenon is the nonlytic release of pathogens from macrophages, in which both the host cell and pathogen remain alive afterwards. This process was first described for the fungal pathogen C. neoformans and termed “vomocytosis.” It was observed in both murine and human macrophages in vitro (79) and from soil amoebae (80). However, whether this actually occurred inside whole animal hosts was unclear until it was visualized inside a larval zebrafish host (66). Larval zebrafish experiments also helped to identify the mitogen-activated protein (MAP) kinase ERK5 as a suppressor of this mechanism (81). Infection experiments in larval zebrafish have also expanded the number of pathogens that are thought to be able to be released this way. We now know that bacteria can also be expelled from macrophages, as B. cenocepacia was observed to nonlytically escape from these cells (47).

Experiments in larval zebrafish have also expanded the types of nonlytic pathogen transfer known to occur in vivo. In addition to extracellular release of pathogens, direct transfer of pathogens from one macrophage to another has been observed in vivo for both A. fumigatus (82) and M. marinum (70). In the case of M. marinum, this transfer occurred from tissue-resident microbicidal macrophages to recruited, permissive inflammatory monocytes, thereby promoting bacterial persistence (70) (Fig. 1B, panel iii).

One such type of transfer that was recently described in detail involves macrophages as the recipient cell (83). Although macrophages are the primary phagocytosers of both A. fumigatus and T. marneffei fungi (28, 75), when neutrophils do phagocytose these spores, the spores are later transferred to macrophages through direct cell-cell contacts and phagosome transfer (83) (Fig. 1B, iii). This study suggested that this transfer is due to manipulation of host cell mechanisms by the fungi, as it is dependent on the fungal cell wall component β-glucan and benefits the fungi by transferring the spores out of highly fungicidal neutrophils into a macrophage niche, in which they can survive (83). However, in most cases, it remains to be determined how these transfers are controlled and whether transfer benefits the host, the pathogen, or both.

CONCLUSION

While much has been learned about macrophage behavior and function from in vitro studies, it has never been clear which processes actually occur in the context of whole animals. The environment that macrophages experience in whole organisms is quite different from the environment in a petri dish, including differences in the extracellular matrix, three-dimensional (3D) architecture, and interactions with other cell types. While cell culture methods can incorporate many of these components through 3D cultures, microfluidic devices, and coculture experiments to better recapitulate the in vivo environment, they are still lacking much of the complexity of a whole organism. The larval zebrafish model is an excellent intermediate vertebrate animal model that combines the complexity of a conserved immune system within whole tissues and myriad cell types with the ability to image and quantify cell behaviors at high resolution in intermediate-term experiments. Many of the functions of macrophages discussed in this review would be simple to determine from in vitro tissue culture studies but almost impossible to visualize in a mouse, highlighting the utility of the larval zebrafish model in investigating host responses in the context of an entire living vertebrate organism at a cellular and subcellular resolution.

There is still much that we do not understand regarding macrophage responses to pathogens and why macrophages sometimes are able to control infections and sometimes actually promote pathogen persistence. Further elucidation of these roles in response to different pathogens should inform our understanding of each of these infections, including how we think about developing new therapies to treat them. Antimicrobial drug resistance is increasing, and our ability to design novel drugs that target these pathogens cannot currently keep pace. Immune cells are a promising target for host-directed therapy development, as demonstrated by the success of T-cell-directed therapies in cancer treatment. Macrophages may be a promising target for such therapies, as these cells often live on the border between helping the host and helping the pathogen. Future studies further investigating the pathways in macrophages that control pathogens, as well as how macrophages may synergize with existing drugs, could provide possible avenues for such treatments.

Biography

Emily E. Rosowski is an assistant professor at Clemson University who is broadly interested in host-pathogen interactions. She received her Ph.D. in 2013, from Massachusetts Institute of Technology, studying how the intracellular parasite Toxoplasma gondii modulates host innate immune pathways in the lab of Dr. Jeroen Saeij. She then performed postdoctoral research with Dr. Anna Huttenlocher at the University of Wisconsin–Madison, learning the larval zebrafish infection model and switching her research focus to investigating innate immune mechanisms, in particular in response to the fungal pathogen Aspergillus fumigatus. Dr. Rosowski is a recipient of a K22 Career Transition Award from NIAID and the NIH and started her own lab at Clemson in the Department of Biological Sciences in January 2019.