Detection of Autofluorescent Mycobacterium Chelonae in Living Zebrafish

Abstract

Mycobacterium chelonae is widespread in aquatic environments and can cause mycobacteriosis with low virulence in zebrafish. The risk of infection in zebrafish is exacerbated in closed-recirculating aquatic systems where rapidly growing mycobacteria can live on biofilms, as well as in zebrafish tissues. We have discovered a method of identifying and visualizing M. chelonae infections in living zebrafish using endogenous autofluorescence. Infected larvae are easily identified and can be excluded from experimental results. Because infection may reduce fertility in zebrafish, the visualization of active infection in contaminated eggs of transparent casper females simplifies screening. Transparent fish are also particularly useful as sentinels that can be examined periodically for the presence of autofluorescence, which can then be tested directly for M. chelonae.

Introduction

Transgenic animals with endogenous fluorescent tags are increasingly being used in many model systems for understanding a range of biological phenomena. Embryonic zebrafish, Danio rerio, marked with cell-type-restricted fluorescent proteins are particularly useful for evaluating cell fate during development since these vertebrates are optically transparent and develop ex utero. Therefore, it is critical that fluorescence output reflects the underlying nature of the biological process of interest. We have employed fluorescence-tagged transient (MLC1a:GFP¹), stable (Ins:eGFP²), and conditional (Ins-NTRmCherry³) transgenic zebrafish as the basis for understanding development and disease. We recently transferred the Ins:nfsB-mCherry transgene⁴ into transparent casper zebrafish, generating casper Insulin-Nitroreductase-mCherry (casperINC) transgenics. Adult zebrafish incubated in water that contains the prodrug metronidazole is converted by the beta-cell-specific nitroreductase enzyme into a toxic compound that selectively destroys beta cells.³ Beta cell regeneration occurs during a 2-week time course, a process that does not occur in diabetic humans or aging murine diabetic models. During development of our zebrafish model system, we discovered an unusual in vivo fluorescence in the transparent casperINC adults that could not be attributed to the transgenic construct. Through analysis of potential contributors to the aberrant fluorescence signal, we determined that zebrafish infected with the rapidly growing Mycobacterium chelonae contained foci of infection that generated a broad emission spectrum of green and red autofluorescence, leading to the exclusion of these infected fish from the study.

Mycobacteriosis is a recognized disease of laboratory zebrafish.^5,6 Although infections of Mycobacterium haemophilum⁷ and Mycobacterium marinum⁸ are associated with significant mortalities, infections with less virulent species like M. chelonae may also have significant impact on research that relies on the zebrafish model. M. chelonae is not typically associated with mortalities, even when fish are experimentally exposed by intraperitoneal injection.⁹ The concern is that these underlying infections are more likely to go unrecognized and may influence the outcome of an experiment, especially where fish are evaluated over weeks and months.¹⁰ Complete eradication of all mycobacteriosis from a colony is a challenge yet to be met¹¹ and strict biosecurity is recommended to keep the more virulent species out of the aquatic system. However, the effort required for complete elimination of all mycobacterial species may exceed expectations for a typical facility. With this in mind, it becomes very important to recognize and evaluate the potential influence these infections have on zebrafish research.

Little is known about subclinical infections in zebrafish and their impact on zebrafish studies.¹² In many cases, the experimental endpoints do not include a histological or microbiological examination. If discovered, then an infected group of fish may be excluded from any further analysis as well as from the final publication. For example, when mycobacteria were observed histologically in two experimental groups, infected animals were excluded from the report.¹³ Underlying infections are more likely to be associated with mortalities in fish that are stressed.¹⁴ If a procedure involves frequent handling, crowding, or suboptimal husbandry conditions, then unanticipated mortalities may occur. When the medaka Oryzias latispes was infected with M. marinum and exposed to a carcinogen, a greater number of liver tumors were observed compared with uninfected controls.¹⁵ In addition, the zebrafish transcriptome is altered in response to M. marinum,^16,17 highlighting the potential variation that infection can cause at the level of gene expression. These studies have provided an incomplete picture of the impacts of mycobacterial infections. Here we show that a simple method for detection of in vivo fluorescence due to M. chelonae infection of larvae or transparent adults can be used to identify mycobacterial infection.

Materials and Methods

Zebrafish care, maintenance, and strains

Zebrafish were raised in an eight-rack Aquatic Habitats integrated system using a single pump, reverse osmosis (RO) water reservoir and an Emperor Aquatics quad UV sterilization system. The particulate and charcoal filters were changed once every 6 weeks and sump filters were changed weekly. RO water filters and UV lights were changed every 6 months. Tanks were cleaned daily in a commercial cage wash after scrubbing in bleach. Salinity was manually maintained at 800 μS with Instant Ocean and pH 7.6 was balanced using sodium bicarbonate. Temperature ranged between 27.5°C and 28°C. Light–dark cycle settings were 14 h on and 10 h off. Adult zebrafish were fed decapsulated artemia once daily after culturing for 48 h, supplemented with Zoe Marine vitamins. Aquatic Habitats Adult Zebrafish Diet was fed on weekday afternoons. New fish were introduced into the system as bleached embryos. After 5 days in dishes, larvae were transferred to the system in baskets at a density of 18–20 per basket, where they were fed a series of Zeigler larval food that ranges LD100, 150, and 250 during a 3-week period after which they were transferred to 3-L tanks at a density of 10–15 per tank or 10-L tanks at a density of 25–40 per tank. After breeding, adult zebrafish (4 months old) were maintained at a density of 6 per 3-L tanks, typically containing four males and two females. Ins:eGFP² transgenics were originally generated in our laboratory and are available from the ZIRC stock center [Tg(1.0ins:eGFP)sc1]. The casper strain was a generous gift of Dr. R.M. White¹⁸ and has two mutations, one in the mitfa gene (nacre) and another in the mpv17 gene (roy). The *AB strain was maintained in an independent zebrafish facility at Duke University. The Insulin-nitroreductase-mCherry plasmid (Ins:nfsB-mCherry),⁴ a generous gift from Mike Parsons, was injected into casper embryos using tol2-mediated transposition¹⁹ to create casperINC zebrafish. Transgenic adults were maintained as heterozygotes.

Imaging and histology

A Leica MZFl III microscope attached to a Hamamatsu Orca camera was used for imaging Tricaine-anesthetized²⁰ zebrafish. The mCherry Chroma filter set No. 49008 ET for Leica and Leica GFP-2 set from Chroma were used to detect red or green fluorescence, respectively. To evaluate mycobacterial infection, adult zebrafish were euthanized, fixed in Dietrich fixative, and placed on a laboratory rocker for a minimum of 24 h. Fixed fish were decalcified in 10% trifluoroacetic acid, processed for paraffin sectioning according to standard techniques, and stained with Ziehl–Neelsen acid-fast stain.²¹ Cultured bacteria were examined with Zeiss Imager M1 AX10 microscope using a 63× water-dipping objective and GFP filter set. Images were captured with a Zeiss Axiocam HSM digital camera and processed using AxioVision software. Figures were prepared using Improvision Openlab and Adobe Photoshop software on Apple computers.

Mycobacterial culture and identification

Mycobacteria were isolated from zebrafish on two occasions. In the case submitted February 2, 2011, spleen and liver were removed from six INC fish and placed in a sterile 1.5 mL tube with 100 μL of 0.9% saline. In the case submitted June 2, 2012, zebrafish were examined under fluorescence to determine infection status. A single fish was selected with an infection of the testes and the infected tissue was removed and placed in sterile saline as explained just now. Tissues were ground with a microtube pestle and 50 μL of this suspension was plated onto Middlebrook 7H10 agar supplemented with OADC (BD) and incubated at 28°C for 5–7 days. Individual colonies were subcultured in Middlebrook 7H9 supplemented with OADC for 7 days. Aliquots were frozen and this strain was designated H11-05. DNA was extracted from a bacterial pellet using the UltraClean microbial DNA isolation kit (Mo Bio Laboratories). PCR was used to amplify part of the heat shock protein 65 (hsp65) gene as described.²² The same forward primer was used to sequence hsp65 with the ABI BigDye Terminator Cycle Sequencing Ready Reaction Kit v3.1 on an ABI PRISM^® 3730 DNA Analyzer (Applied Biosystems).

Injection of independent M. chelonae strains into *AB wild-type embryos

The H11-05 strain or the independently isolated M. chelonae strain ZF178²³ was propagated in Middlebrook 7H9 liquid culture supplemented with 10% OADC and 0.25% Tween 80 (20% v/v) at 28°C. Embryos were collected from *AB wild-type adults located in an independent facility from the one where infected zebrafish were first identified, dechorionated at 1 dpf, and maintained until 3 dpf in 1.5×phenylthiourea in filtered fish water. On the day of infection, 1 mL of each liquid culture was centrifuged and resuspended in one-half-volume 7H9 media prior to homogenization with a 27½-gauge needle. To aid in visualization, the inoculum was prepared for injection by the addition of 0.125% phenol red. Prior to manipulation, embryos were anesthetized in pH 7-buffered MS-222 (Tricaine) at 150 ppm. Microinjections were performed with an Eppendorf Transjector 5246 into the caudal vein of 3 dpf embryos as previously described.²⁴ Images were taken at 3 dpi. Bacterial burden was 150–250 CFU per fish as determined by plating onto 7H10 plates and maintained at 28°C until colonies were visible. Single bacterial colonies from H11-05 or ZF178 were mixed into 10 μL sterile-filtered PBS. Smears were made on glass slides and then imaged for red/green fluorescence. Microscopy was performed using a Zeiss Axio Observer Z.1 equipped with an X-Cite 120 Fluorescence Illumination System (EXFO) utilizing a 120 W mercury lamp, Cy3 or GFP filters, and bright-field optics. The objectives used included a Zeiss 2.5× Fluar and a 20× LD Plan Neofluar.

Results

We previously developed a conditional ablation system in adult zebrafish where the destruction followed by expansion of pancreatic beta cell mass occurs within a 2-week period.³ Metabolic demands on animal physiology largely determine adult beta cell growth and function.²⁵ To evaluate changes in beta cell mass in adult zebrafish, we generated a transparent, transgenic zebrafish where mCherry-derived red fluorescence expressed only in beta cells was recorded in living adults as regeneration progressed. During the course of this study, we observed spurious, red fluorescent nodules that were often associated with the gonads of male (n=6, Fig. 1A) or female (n=8, Fig. 1B) casperINC adults as well as casper InsGFP larvae (Fig. 1C). The nonpancreatic tissue that contains this anomalous red fluorescence was also positive for green fluorescence (Fig. 1, merge). The fish behaved and appeared normal and were housed in a single 10-L tank. We could not detect fluorescence in anesthetized, pigmented wild-type adults located in adjacent tanks, either because melanophores and iridophores blocked the view of internal organs or because they were not infected. Adult casperINC females presented with visible red and green fluorescence located within and around eggs (Fig. 1B). In addition, we observed red as well as green autofluorescence in InsGFP² larvae, suggesting that the aberrant fluorescence was not due to the transgene (Fig. 1). InsGFP transgenics express green (eGFP) instead of red (mCherry) fluorescence only in beta cells. Cords of red and green autofluorescence lining the swim bladder and projecting caudally along the gut were present in infected InsGFP larvae (Fig. 1C). Red/green fluorescent nodules were present whether the zebrafish expressed a red or green fluorescent protein. Larvae were euthanized after imaging and all dishes and pipets were soaked in bleach and cleaned in a commercial cage wash to prevent contamination. On dissecting the red/green nodules present in casperINC adult male fish, we discovered a bright red-and-green fluorescent mass that could be excised from between the gonads and swim bladder. To determine whether the red/green fluorescence was due to an infectious agent, preserved male and female zebrafish were prepared and sent to ZIRC Health Services (http://zebrafish.org/zirc/health/) for histological analysis. Granulomatous nodules that contain acid-fast bacilli indicative of mycobacteriosis were found in both male paraffin sections (Fig. 2A).

FIG. 1.

Mycobacterium chelonae red and green autofluorescence in adult casperINC zebrafish. Living, transgenic adult and larval zebrafish were anesthetized and imaged with a Leica MZFl III microscope. (A) Eight-month-old male casper transgenic. mCherry-positive beta cells mark the pancreas (arrow), while additional red (left panel) and green (middle panel) fluorescence is due to mycobacterial autofluorescence (yellow granulomas; merge). (B) Mycobacterial infection in female casperINC zebrafish is apparent as red/green autofluorescence surrounding eggs (Merge-yellow). (C) Autofluorescence in infected, 6-day-old casper InsGFP larvae is both red and green (Merge-yellow). Arrowhead: swim bladder; arrow: GFP+transgene expressed only in beta cells; *: gall bladder; bracket: yellow mycobacterial nodules.

FIG. 2.

Rapidly growing M. chelonae infection in casper or *AB zebrafish retains autofluorescence in culture. (A) Adult casper Female: granuloma (arrow) within the ovary (400× magnification). (B) Adult casper Male: granuloma (arrow) containing acid-fast bacilli (purple) within abdominal fat (400× magnification). (C) Wet mount of M. chelonae that expresses green autofluorescence in culture (630× magnification). Bright-field image overlaid with fluorescence image and green autofluorescence alone. Scale bar=10 μm (C). Red or green autofluorescence from regrowth of M. chelonae H11-05 and an independent M. chelonae strain ZF178 isolated and then cultured for injection into zebrafish embryos.

Adult zebrafish that contain red/green fluorescent nodules were subsequently tested for their mycobacterial strain and culture characteristics. Touch imprints made from tissues dissected from zebrafish contained characteristic acid-fast rods (data not shown). A suspension of spleen and liver was plated on 7H10 media and incubated for 5–7 days at 28°C. Visible colonies were subcultured and designated H11-05. Extracted DNA was used as a template to amplify the mycobacterial hsp65 gene, generating a fragment 650 base pairs in length. The fragment was sequenced and identified as M. chelonae using a BLAST search in GenBank (www.ncbi.nlm.nih.gov/genbank). DNA was also extracted directly from the spleen and liver suspension and PCR was used to amplify the hsp65 gene. Sequencing yielded identical results to that of the isolated H11-05 colony, suggesting no evidence of a mixed infection. In another batch of infected fish, animals were examined for autofluorescence. Testes that exhibit green fluorescence and numerous granulomas were dissected. A touch imprint was made of this tissue, and the remainder was plated directly onto 7H10 media. DNA sequencing confirmed the identity of this species as M. chelonae. After 5 days of culture at 28°C, wet mounts were prepared from colonies and observed under a microscope. Characteristic 1.5–2-μm rods were present (Fig. 2B) and many emitted green fluorescence. To further test the infectious nature of independent isolates in a wild-type zebrafish strain, we compared two different M. chelonae strains injected into *AB zebrafish embryos for red/green autofluorescence.

M. chelonae strain ZF178²³ or the M. chelonae H11-05 isolates exhibited growth in culture at 28°C after 7 days. Both strains exhibited red as well as green fluorescence in histological smears (Fig. 2C). Embryos from a mating of *AB wild-type adults maintained in a separate zebrafish facility were inoculated with either strain of M. chelonae (H11-05 and ZF178). On day 3 post-fertilization, homogenized inoculum was injected at ∼200 CFU/anesthetized embryo. Green and red fluorescence was observed after 3 more days (Fig. 3). Granulomas could not be detected in the 6-day, uninjected control *AB larvae (Fig. 3A). In contrast, both the H11-05 strain (Fig. 3B) as well as the independently isolated ZF178 strain (Fig. 3C) produced red and green fluorescence in fish. Because M. marinum strains used by investigators in the facility where the injections were performed are negative for red/green fluorescence (data not shown), we conclude that M. chelonae generates a characteristic red-and-green fluorescence whether growing in culture or after infection of different zebrafish strains, irrespective of the presence of fluorescent transgenes.

FIG. 3.

Living 6-day-old *AB wild-type zebrafish larvae infected with M. chelonae isolates. Hatched embryos were inoculated at 3 dpf with strain H11-05 or a previously isolated and characterized M. chelonae strain ZF178 from an independent source. Control or infected larvae were anesthetized and then imaged in the trunk region at 6 dpf. (A) Uninjected control: Red and green autofluorescence. Merge: overlay of red and green images indicates M. chelonae infectious nodules (yellow). (B) M. chelonae strain H11-05: red and green autofluorescence from M. chelonae isolated and characterized from zebrafish described in this report. (C) M. chelonae strain H11-05: red and green autofluorescence from an independent isolate. Merged autofluorescent signals appear yellow. Red and/or green fluorescent images were overlaid on bright-field images and indicate granulomas. Scale bars=100 μm.

Discussion

We discovered that M. chelonae-infected zebrafish can be screened and eliminated from aquatic systems or experimental results after observation under a fluorescence-equipped microscope. We determined that a recognition of mycobacteriosis caused by the rapidly growing mycobacteria M. chelonae can be made directly in living zebrafish larvae or transparent adults. The observation of red/green autofluorescence from M. chelonae is a potential source of variation in experimental protocols as unrecognized infection could be misinterpreted as fluorescently labeled target tissue. In the event of an anomalous finding, a simple test for infected embryos or larvae could be used to observe both red and green fluorescence under a dissecting microscope with red (mCherry or Cy3) and green (GFP) fluorescence filters. These are commonly used filters in the zebrafish community. A potential application of this discovery might be to monitor infections in aquatic facilities using casper adults as sentinels for the presence of autofluorescent M. chelonae. If the goal is to screen zebrafish to minimize nonprotocol variation using autofluorescence, then further identification of species may not be required. However, it is still important to discriminate the strain of mycobacteria due to differences in severity of disease by species and the potential for zoonotic transmission.

Blue autofluorescence has previously been reported as a diagnostic tool for M. tuberculosis in humans.^26,27 Patiño et al.²⁸ examined fluorescence emitted from additional mycobacterial species, including M. smegmatis, M. fortuitum, and M. marinum, using laser excitation at a wavelength of 405 nm with emission peaks centered at 470 nm. The source of this fluorescence may be due to the F420 and FO coenzymes important for mycobacterial electron transfer reactions, capable of generating a bright blue-green fluorescence that is enhanced by heat treatment.²⁸ Using a red exciter ET560/40X and emitter ET630/75m and/or a green exciter ET480/40X and emitter 510lp to analyze anesthetized casper adults or larvae, we visualized red/green nodules of M. chelonae in vivo and in isolated cultures of M. chelonae. These filter sets are commonly used for tracking tissue-specific expression in living embryos. Because many reporter transgenes are engineered to generate green or red fluorescence in zebrafish models, this rapid assay for the presence of M. chelonae can be used to evaluate unexpected results. We found that M. chelonae generates a characteristic red-and-green fluorescence whether growing in culture or after infection of either casper or *AB fish, irrespective of the presence of fluorescent zebrafish transgenes. Actions to eliminate infections typically rely on euthanization. Due to mycobacterial autofluorescence, any infected larvae or transparent casper adult sentinel can easily be eliminated. Because the virulence of M. chelonae in zebrafish colonies is low and often goes undetected,⁶ the ability to easily correlate infection with mycobacterial autofluorescence in living tissue provides a useful tool for eliminating spurious results or infected zebrafish. We have not examined animals other than zebrafish infected with this mycobacterial strain. However, it may be possible to identify granulomas in different species using the characteristic red/green autofluorescence we have attributed to M. chelonae.

We removed any zebrafish positive for red/green fluorescence because we were concerned about the potential for infection of other fish in the facility as well as the possibility of zoonotic transfer of mycobacteria to fish room workers. M. marinum, a nontuberculous mycobacterium that causes skin infections in humans, is of particular concern because 84% of humans who were infected with this species had contact with home aquaria.²⁹ We have also linked identical strains of M. marinum in fish with a human infection.⁸ Humans with immune deficiencies are at risk and, although long-term antibiotic therapy eliminates infection, the treatment success relies on identification of the mycobacterial strain. M. chelonae has been documented as a human pathogen in immunocompromised patients and treated with antibiotics, primarily clarithromycin.³⁰ However, no link has been made between fish and humans for M. chelonae transmission, and the occurrence in both species may represent epidemiologically unlinked strains. Interestingly, a recent systematic survey for pedicure-associated nontuberculous mycobacterial outbreaks in humans determined that 91% of infections were caused by M. chelonae.³¹ Biofilms from footbaths were processed and precise isolation and speciation was successful in 50% of the samples. Cultured isolates were resistant to multiple antibiotics. However, 12 of 13 isolates from human samples were susceptible to clarithromycin. Treatment of zebrafish that have mycobacterial infections with antibiotics is not recommended. However, if it is considered as a means to maintain a particularly valuable line in order to spawn a next generation, then potential ramifications including antibiotic resistance and prolonged human exposure should be evaluated.

Acknowledgments

This work was supported by a grant from the Juvenile Diabetes Research Foundation to L.G.M. and J.B.M. The authors appreciate all the members of the Sarah W. Stedman Nutrition and Metabolism Center under the direction of Chris Newgard for continuing support. C.M.W. thanks Hadi L. Jabbar Al-Hasnawi for assistance in the initial characterization of the H11-05 isolate and Jeffrey D. Amack at SUNY Upstate Medical University for the use of the Zeiss microscope. C.M.W. was funded in part from a subaward from NIH-NCRR grant 5R24RR01386-07. The Zebrafish International Resource Center (ZIRC) is supported by the NIH Office of Research Infrastructure Programs.

Disclosure Statement

No competing financial interests exist.