Here, we present a protocol for the generation and imaging of a localized bacterial infection in the zebrafish otic vesicle.
The aquatic pathogen, Streptococcus iniae, is responsible for over 100 million dollars in annual losses for the aquaculture industry and is capable of causing systemic disease in both fish and humans. A better understanding of S. iniae disease pathogenesis requires an appropriate model system. The genetic tractability and the optical transparency of the early developmental stages of zebrafish allow for the generation and non-invasive imaging of transgenic lines with fluorescently tagged immune cells. The adaptive immune system is not fully functional until several weeks post fertilization, but zebrafish larvae have a conserved vertebrate innate immune system with both neutrophils and macrophages. Thus, the generation of a larval infection model allows the study of the specific contribution of innate immunity in controlling S. iniae infection.
The site of microinjection will determine whether an infection is systemic or initially localized. Here, we present our protocols for otic vesicle injection of zebrafish aged 2-3 days post fertilization as well as our techniques for fluorescent confocal imaging of infection. A localized infection site allows observation of initial microbe invasion, recruitment of host cells and dissemination of infection. Our findings using the zebrafish larval model of S. iniae infection indicate that zebrafish can be used to examine the differing contributions of host neutrophils and macrophages in localized bacterial infections. In addition, we describe how photolabeling of immune cells can be used to track individual host cell fate during the course of infection.
Streptococcus iniae is a major aquatic pathogen that is capable of causing systemic disease in both fish and humans1. While S. iniae is responsible for large losses in the aquaculture industry, it is also a potential zoonotic pathogen, capable of causing disease in immunocompromised human hosts with clinical pathologies similar to those caused by other streptococcal human pathogens. Given its similarities with human pathogens, it is important to study S. iniae disease pathogenesis in the context of a natural host. An adult zebrafish model of S. iniae infection revealed robust infiltration of host leukocytes to the localized site of infection as well as a rapid time to host death, a time too short to involve the adaptive immune system7. In order to gain an in-depth look into the innate immune response to S. iniae infection in vivo, it is necessary to use a model that is more amenable to non-invasive live imaging.
The larval zebrafish has a number of advantages that make it an increasingly attractive vertebrate model for studying host-pathogen interactions. Zebrafish are relatively inexpensive and easy to use and maintain compared to mammalian models. Adaptive immunity is not functionally mature until 4-6 weeks post fertilization, but larvae have a highly conserved vertebrate innate immune system with complement, Toll-like receptors, cytokines, and neutrophils and macrophages with antimicrobial capabilities including phagocytosis and respiratory burst2-6,8-11. In addition, the genetic tractability and optical transparency of the embryonic and larval stages of development allow for the generation of stable transgenic lines with fluorescently labeled immune cells making it possible to examine host-pathogen interactions in real time in vivo. The generation of these transgenic lines using a photoconvertible protein such as Dendra2 allows for the tracking of individual host cell origin and fate over the course of infection12.
When developing a zebrafish larval infection model, the chosen site of microinjection will determine whether an infection is initially localized or systemic. Systemic blood infections into the caudal vein or Duct of Cuvier are most commonly used to study microbial pathogens in zebrafish and are useful for studying interactions between host and microbial cells, cytokine responses, and differences in virulence between pathogen strains. For slower growing microorganisms, early injection into the yolk sac of an embryo at the 16-1,000 cell stage can be used to generate a systemic infection13,14, with the optimal developmental stage for microinjection of a slow-growing microorganism found to be between the 16 to 128 cell stage15. However, yolk sac injections of many microbes at later stages of host development tend to be lethal to the host due to the nutrient-rich environment for the microbe and lack of infiltrating leukocytes16-18.
A localized infection usually results in directed migration of leukocytes towards the site of infection that can be easily quantified with non-invasive imaging. This type of infection can allow for dissection of the mechanisms that mediate leukocyte migration as well as investigation of different migratory and phagocytic capabilities of various leukocyte populations. Localized infections are also useful when examining differences in virulence between bacterial strains as well as studying microbe invasion mechanisms since physical host barriers must be crossed for a localized infection to become systemic. Zebrafish are typically raised at temperatures of 25-31 °C19, but they can also be maintained at temperatures as high as 34-35 °C for studies of the invasiveness of certain human pathogens with strict temperature requirements for virulence20,21.
Many different sites have been used to generate an initially localized bacterial infection including the hindbrain ventricle22, dorsal tail muscle18, pericardial cavity23, and otic vesicle (ear)5,16,24. However, it has been found that injection of bacteria into tail muscle can cause tissue damage and inflammation independent of the bacteria, which may skew results when investigating leukocyte response13. Although less damage is associated with injection into the hindbrain and although it is initially devoid of leukocytes in young embryos, the hindbrain ventricle steadily gains more immune cells over time as microglia take up residence. The hindbrain ventricle is also a more difficult location to image. The otic vesicle is a closed hollow cavity with no direct access to the vasculature25,26. It is normally devoid of leukocytes, but leukocytes can be recruited to the otic vesicle in response to inflammatory stimuli such as infection. It is also a preferred site of microinjection of bacteria in zebrafish aged 2-3 days post fertilization (dpf) because of the ease of imaging and the visualization of the injection. Therefore, we chose the otic vesicle as our site of localized bacterial infection.
Adult and embryonic zebrafish were maintained in accordance with the University of Wisconsin-Madison Research Animal Resources Center.
1. Preparing Microinjection Needles
2. Preparing Larval Injection Dishes
3. Preparing S. iniae Inoculum
4. Labeling S. iniae with a CellTracker Red Fluorescent Dye
5. Preparation of Zebrafish Larvae for Infections
6. Otic Vesicle Injection of S. iniae into Three Day Old Larvae
7. Sudan Black Staining of Neutrophils
Note: The following steps can be done at room temperature in a small Petri dish (35 x 10 mm2) on an orbital shaker unless otherwise stated. For each step, the amount of reagent used is approximately 2 ml per dish, using just enough liquid to completely cover the larvae.
8. Enumeration of Viable Bacteria from Infected Larvae
9. Fixation of Larvae for Imaging
10. Preparation of Larvae for Live Imaging
11. Confocal Imaging of Infection
12. Photoconversion of Dendra2-labeled Leukocytes at the Otic Vesicle
Note: Dendra2 can be photoconverted from green to red fluorescence by focusing a 405 nm laser (50-70% laser power should be sufficient) on the region of interest (ROI) for 1 min. Below is the step-by-step protocol used for the FV-1000 laser scanning confocal system:
Microinjection of S. iniae into the otic vesicle (Figure 1 and Figure 2) results in an initially localized host response. When injected correctly, the bacteria should only be seen in the otic vesicle and not in the surrounding tissue or blood. This can be visualized during microinjection using phenol red dye (Figure 1A). Alternatively, if labeled bacteria are injected, a quick scan of infected larvae immediately post injection can confirm the bacteria are only in the otic vesicle and not the surrounding tissue (Figure 1B). Although a dose as little as 10 CFU wild type S. iniae is able to establish a lethal infection within 24-48 hr post infection (hpi), injection of 1,000 CFU of an avirulent strain, cpsA, does not result in lethal infection and that strain seems unable to proliferate in the host24. Thus, this localized infection model is able to differentiate between bacterial strains of altered virulence.
Microinjection sites of initially localized infection are useful for studying leukocyte chemotaxis. Leukocyte recruitment can by quantified by either Sudan Black staining of neutrophil granules (Figure 1A) or formaldehyde fixation of fluorescent transgenic lines (Figure 1B). To visualize the recruitment and phagocytic capabilities of immune cells, we used transgenic lines expressing the green fluorescent protein Dendra2 specifically in macrophages Tg(mpeg1:dendra2)24 or neutrophils Tg(mpx:dendra2)12. When S. iniae is injected into the otic vesicle of 3 dpf larvae, both neutrophils and macrophages are rapidly recruited within the first 2 hpi (Figure 1). However, when performed correctly, microinjection of PBS into the otic vesicle does not result in the same robust recruitment of host leukocytes (Figure 1). In addition to recruitment, live confocal time lapse imaging of fluorescent transgenic lines injected with fluorescently-labeled bacteria reveals the phagocytic capabilities of both neutrophils and macrophages. Red dye-labeled S. iniae can be found inside both neutrophils and macrophages (Figure 2). Using the photoconvertible protein Dendra2 to label neutrophils or macrophages allows for non-invasive photolabeling for tracking individual cell fate over the course of the infection. Macrophages that were recruited to the otic vesicle at 5 hpi were photoconverted and then tracked over the following 24 hr. Although some photoconverted cells remain in the otic vesicle or head region, some can also be found disseminated throughout the body of the larvae (Figure 3).
Figure 1: Leukocyte recruitment to otic vesicle infection with S. iniae. (A) Neutrophil recruitment to S. iniae infection. (i) Successful injection of a phenol red-labeled inoculum into the otic vesicle. (ii–iv) Sudan Black staining of larvae for investigation of neutrophil recruitment at 2 hpi. PBS mock-infected larvae show little recruitment of neutrophils to the otic vesicle (ii) whereas infection with either wild type S. iniae or the cpsA mutant results in robust neutrophil recruitment (iii, iv). Scale bar, 300 µm. (B) Macrophage recruitment to S. iniae infection. (i) Successful microinjection of red-labeled S. iniae (depicted in magenta) into the otic vesicle. (ii–iv) Fluorescent confocal images of microinjected transgenic mpeg1:dendra2 larvae fixed at 2 hpi. PBS mock-infected larvae show little macrophage recruitment (ii), but larvae infected with CellTracker Red-labeled (depicted in magenta) wild type S. iniae or the cpsA mutant show robust macrophage recruitment to the otic vesicle at 2 hpi (iii, iv). Scale bar, 30 µm. Please click here to view a larger version of this figure.
Figure 2: Phagocytosis of S. iniae by phagocytes in the otic vesicle. Transgenic mpx:dendra2 (A) or mpeg1:dendra2 (B) larva infected with red-labeled S. iniae (depicted in magenta) and imaged at 60 min post infection using a laser scanning confocal microscope. Scale bar, 30 µm. Please click here to view a larger version of this figure.
Figure 3: Photoconversion of macrophages at the otic vesicle 5 hpi with S. iniae. Macrophages (depicted in green) at the otic vesicle, designated by the circle (A), were photoconverted (B) using a 405 nm laser on a confocal microscope and tracked over time. By 24 hpi, photoconverted macrophages (depicted in magenta) have migrated as far as the trunk/caudal hematopoietic tissue (C); scale bar, 50 µm. Higher magnifications of the boxed regions in C are shown in (i) and (ii), scale bar 30 µm; arrows point to photoconverted macrophages. Photoconverted cells appear white because of the merged 543 nm red fluorescence and any remaining 488 nm green fluorescence. Please click here to view a larger version of this figure.
The infection method used here is useful for the study of the host immune response to an initially localized infection in 2-3 dpf embryos and larvae. The focus of an inflammatory stimulus, such as infection, in a closed cavity such as the otic vesicle allows for the study of neutrophil and macrophage chemotaxis and phagocytosis. One caveat of injecting bacteria into the otic vesicle is that the ability of neutrophils to efficiently phagocytose bacteria in fluid-filled cavities may be dependent on the particular microbe. Although Escherichia coli and Bacillus subtilis are not easily phagocytosed by neutrophils in the otic vesicle28, we have found that neutrophils are able to phagocytose both Pseudomonas aeruginosa and S. iniae in this location16,24. Localized infection is also useful when studying the invasiveness of various pathogens. In order to cause a systemic infection following injection into a closed cavity such as the otic vesicle, the microbe must be able to traverse physical host barriers. Alternatively, different pathogens may rely on host cells for transportation and dissemination from the initial site of infection. This makes localized injections useful for comparing strains of altered virulence.
During microinjection, to avoid bacteria settling and clogging the needle, change needles either after each condition or after about 50 larvae. This will help ensure the suspension in the needle is more uniform. If settling of bacteria in the needle seems to be a problem, a mix of PBS, 2% PVP40, and 10% glycerol may help keep a homogenous suspension. To ensure that each larva is being injected with approximately the same number of bacteria, check CFU counts by homogenizing a larva immediately following injection and plating the homogenate on CNA agar. CNA agar will select for the growth of culturable gram-positive bacteria, not only S. iniae, but it will provide an idea of how consistent the injection doses are between each individual larva. With a target inoculum of 100 CFU per larva, there are typically between 75-150 colonies on a CNA plate. The number of non-S. iniae gram-positive colonies growing on a CNA plate is probably very low, since most PBS injected fish usually result in between 0-20 colonies. At later time points during the infection, it is not uncommon for there to be variation of up to half a log in colony counts between larvae infected with the same infectious dose. This could represent slight differences between individual larva in their ability to control the infection or could reflect differences in the amount of culturable gram-positive bacteria in the larvae or E3 medium.
While injecting into the otic vesicle, it is important not to inject too large a volume or with too high a pressure as this may cause the cavity to rupture. Injection volumes should be kept to approximately 1 nl. If the needle is too large, microinjection can cause damage to the surrounding tissue, which may affect the recruitment of host leukocytes to the site of infection or may allow bacteria to leak out of the otic vesicle. It is also important to confirm injection into the correct space. If the needle is inserted too deep, it may poke through the otic vesicle or it may hit blood vessels flowing around the otic vesicle. This may lead to the deposition of bacteria into the blood stream or outside the vesicle, which may alter host responses.
It is also possible that when the needle is extracted, some bacteria may be accidentally deposited outside the otic vesicle as shown by Colucci-Guyon et al.28, but we find this to only be the case in less than 5% of the injections. Trying to inject into the center of the otic vesicle may help avoid this situation. In a successful infection, the otic vesicle should fill up with the phenol red dye, but the dye should not leak out into the surrounding tissues. Any mis-injected larvae should be immediately discarded. Injecting labeled bacteria is a useful way to confirm a successful injection (Figure 1B). One of the disadvantages of using a CellTracker dye is that this dye will become diluted as the bacteria divide, eventually preventing the bacteria from being visualized under fluorescence. We chose a red dye because we used green-labeled neutrophil and macrophage transgenic lines for the majority of our studies.
Dendra2 is a photoconvertible protein derived from octocoral Dendronephthya sp. which can be photoconverted from a green to a red fluorescent state with a 405 nm laser29. This photoconversion is a noninvasive way to mark cells and track their fate over the course of infection. Leukocytes recruited to the site of infection can be photoconverted and followed over time to monitor their dissemination throughout host tissues (Figure 3). Alternatively, leukocytes could be photoconverted prior to infection and then monitored post-microinjection to determine the origin of cells recruited to the site of infection. Labeling bacteria and immune cells allows for the study of leukocyte-pathogen interactions in vivo including recruitment and phagocytosis (Figure 1B and Figure 2). Distinguishing the red-labeled S. iniae from the red fluorescence of a photoconverted leukocyte is difficult, so a different fluorescent CellTracker dye could be used. This would be particularly useful to determine which of the photoconverted immune cells that leave the otic vesicle contain S. iniae.
Future applications of the otic vesicle infection method could include measuring the speed and directionality by which neutrophils and macrophages move in response to various infections. The infection kinetics can also be studied to see which cell types are the first to arrive to a site of infection and which cell types are most robustly recruited in addition to where the cells originate or where they disseminate30. In addition to studying the recruitment to an initially localized infection, this infection model can be used to study the function of host immune system components during initial infection. Antisense morpholino oligonucleotides that target specific RNAs can be used to knock down expression of host immune components including Toll-like receptors, cytokine receptors and leukocytes, and CRISPR-Cas or TALEN technology can be used to create genetic mutants. Gene expression using RNAseq or qPCR can also be used to characterize the expression of certain host genes in response to infection.
The authors have nothing to disclose.
The authors would like to thank lab members for zebrafish care and maintenance. This work was supported by National Institutes of Health, National Research Service Award A155397 to E. A. Harvie and NIH R01GM074827 to Anna Huttenlocher.
1.7 ml eppendorfs | MidSci | AVSS1700 | |
14 ml falcon tube | BD Falcon | 352059 | |
27 G x 1/2 in. needle | BD Biosciences | 305109 | |
96 well plate | Corning Incorporated | 3596 | |
Agar | BD Biosciences | 214030 | |
CellTracker Red | Molecular Probes, Invitrogen | C34552 | |
CNA agar | Dot Scientific, Inc | 7126A | |
Disposable transfer pipets | Fisher Scientific | 13-711-7m | |
Dissecting Scope | Nikon | SMZ745 | |
DMSO | Sigma Aldrich | D2650 | |
Ethanol 200 proof | MDS | 2292 | |
Fine tweezers | Fine Science Tools | 11251-20 | |
Gel comb | VWR | 27372-482 | 4.2 mm width, 1.5 mm thick |
Glass bottom dishes | Custom made by drilling a 16–18 mm hole in the center of a 35-mm tissue culture dish bottom and placing a 22-mm round #1 coverslip in the hole and sealing with a thin layer of Norland Optical Adhesive 68 cured by UV light. | ||
Glycerol | Fisher Scientific | G33-4 | |
High melt agarose | Denville Scientific, Inc. | CA3510-6 | |
Hydrogen peroxide | Fisher Scientific | H325 | |
Laser Scanning Confocal Microscope | Olympus | with FV-1000 system | |
Low melt agarose | Fisher | BP165-25 | |
Magnetic stand | Tritech (Narishige) | GJ-1 | |
Microinjection system | Parker | Picospritzer III | |
Microloader pipet tips | Eppendorf | 930001007 | |
Micromanipulator | Tritech (Narishige) | M-152 | |
Micropipette puller | Sutter Instrument Company | Flaming/Brown P-97 | |
Nanodrop spectrophotmeter | Thermo Scientific | ND-1000 | |
N-Phenylthiourea (PTU) | Sigma aldrich | P7629 | |
Paraformaldheyde | Electron Microscopy Sciences | 15710 | |
Petri Dishes | Fisher Scientific | FB0875712 | 100 mm x 15 mm |
Phenol | Sigma Aldrich | P-4557 | |
Phenol Red | Ricca Chemoical Company | 572516 | |
Phosphate Buffered Saline | Fisher Scientific | BP665-1 | |
Potassium hydroxide | Sigma Aldrich | P-6310 | |
Pronase | Roche | 165921 | |
Protease peptone | Fluka Biochemika | 29185 | |
Small cell culture dish | Corning Incorporated | 430165 | 35 mm x 10 mm |
Sudan Black | Sigma Aldrich | S2380 | |
Thin wall glass capillary injection needles | World Precision Instruments, Inc. | TW100-3 | |
Todd Hewitt | Sigma Aldrich/Fluka Analytical | T1438 | |
Tricaine (ethyl 3-aminobenzoate) | Argent Chemical Laboratory/Finquel | C-FINQ-UE-100G | |
Triton X-100 | Fisher Scientific | BP151-500 | |
Tween 20 | Fisher Scientific | BP337-500 | |
Yeast extract | Fluka Biochemika | 92144 |