Genital mycobacteriosis caused by Mycobacterium marinum detected in two captive sharks by peptide nucleic acid–fluorescence in situ hybridization

Abstract Mycobacterium marinum is a prevalent nontuberculous mycobacterium (NTM)‐infecting teleosts. Conversely, little is known about mycobacteriosis in elasmobranchs, and M. marinum infection has never been reported from the subclass. This study investigated the histopathological characteristics and localization of this mycobacterium through molecular analysis of two captive sharks, a scalloped hammerhead Sphyrna lewini and a Japanese bullhead shark Heterodontus japonicus, exhibited in the same aquarium tank. We detected genital mycobacteriosis caused by M. marinum infection using molecular analyses, including polymerase chain reaction (PCR) and DNA sequencing targeting the 60 kDa heat‐shock protein gene (hsp65), and peptide nucleic acid–fluorescence in situ hybridization (PNA‐FISH) targeting the 16S rRNA gene. Both sharks showed granulomas in connective tissues of the gonads without central necrosis or surrounding fibrous capsules, which is unlike the typical mycobacterial granulomas seen in teleosts. This study reveals that elasmobranchs can be aquatic hosts of M. marinum. Because M. marinum is a representative waterborne NTM and a potential zoonotic agent, cautious and intensive research is needed to overcome a lack of data on the relationship between NTM and the aquatic environment in association with this subclass of Chondrichthyes.

susceptible to NTM, most commonly M. marinum, M. fortuitum and M. chelonae (Gauthier & Rhodes, 2009). In addition to its impact on fish health, M. marinum is a potential zoonotic since it produces granulomatous lesions in human skin (Tomas et al., 2014), usually following exposure to infected fish surfaces (Jernigan & Farr, 2000) and is also a risk factor for pneumonia (Oh et al., 2018). In comparison with the wealth of information on piscine mycobacteriosis in teleosts, there is a remarkable paucity of reports about this disease in elasmobranchs.
Considering that water and biofilms are the natural habitats of Mycobacterium spp. (Pedley et al., 2004), accurate detection of causative pathogens in clinical cases of mycobacterial contamination in a fish-rearing environment is a frequent problem. As such, there is a need to develop a tool that can localize gene expressions of Mycobacterium spp. using tissue sections from infected individuals.

Fluorescence in situ hybridization (FISH) is an important candidate
method that is based on the specific binding of nucleic acid probes to specific regions on targeted DNA in a tissue sample. Peptide nucleic acid (PNA) probes have been applied to the detection of mycobacteria using clinical samples from humans (Kim et al., 2015;Lefmann et al., 2006; because the relative hydrophobic character of PNA, as compared with DNA and RNA, allows better diffusion of the PNA probes through the cell wall of mycobacteria (Stender, Mollerup et al., 1999). However, nothing is known about FISH protocols using PNA probes to detect specific mycobacterial species in piscine mycobacteriosis.
The scalloped hammerhead Sphyrna lewini (Griffith & Smith) occurs circumglobally in warm-temperate and tropical coastal seas (Compagno, 1984); it is one of the most threatened sharks and was listed as Critically Endangered (CR) by the IUCN Red List of Threatened Species in 2019 (Rigby et al., 2019). The Japanese bullhead shark Heterodontus japonicus Miklouho-Maclay & Mcleay occurs in continental shelf waters of the northwestern Pacific, most commonly over rocky, kelp-covered or sandy bottoms (Compagno, 2001). Both shark species have been successfully maintained in Japanese aquaria for exhibition and study.
Here, we present the first reported cases of M. marinum infection from two sharks that died in captivity. The infections were confirmed by peptide nucleic acid-fluorescence in situ hybridization (PNA-FISH).

| Spontaneous mycobacteriosis and environmental samples
An approximately 10-year-old wild-caught male scalloped hammerhead shark and a wild-caught adult female Japanese bullhead shark of uncertain age were housed at the Shimane Aquarium in Shimane Prefecture, Japan. They were exhibited in a 1 million-L tank exhibit along with other fish, including multiple teleost and elasmobranch species, and one sea turtle. Water pumped from the local sound (Sea of Japan), treated with pressure filtration, is used as the aquarium's water source. After 9 years and 9 months in captivity, the hammerhead shark presented anorexia for 1 month and was subsequently found dead in the tank. One month later, the bullhead shark was also found dead, but without any remarkable clinical signs. The captivity duration of this bullhead shark was unknown because Japanese bullhead sharks in the tank were not individually identified. Autopsies revealed marked hemocelom, multifocal congestions in the bilateral testes of the hammerhead shark and multifocal haemorrhages in the bilateral epigonal organ To assess the microbial community distribution in the rearing environment, samples of water (1 L) and filtration sand, as well as swabs of biofilms from a wall and a water intake were provided. All samples were sent to the Laboratory of Aquatic Medicine at Nippon Veterinary and Life Science University in Tokyo.

| Histopathological analysis
Samples from the brain, alimentary canal, liver, kidney, spleen, pancreas and reproductive organs were collected from both sharks, and samples from the epigonal organ, eyes and skin were collected from the hammerhead shark. The reproductive organs from the bullhead shark were frozen on site and then fixed in 10% neutralbuffered formalin at our laboratory; the other remaining samples were fixed in 10% formalin on site. All samples were routinely processed to create paraffin-embedded tissue blocks, sectioned at 2μm thickness, mounted on frosted glass slides, stained with haematoxylin and eosin (HE), Giemsa, Gram and Ziehl-Neelsen (ZN) stains and subjected to histopathological examination.

| Tissue samples
Isolation of the etiological agent of mycobacteriosis was attempted from frozen organs, including the reproductive organs, kidney, spleen and liver, in both cases, and the epigonal organ from the hammerhead shark. Pretreatments were performed based on four methods, as follows: (1) non-treatment; (2) decontamination by hydrochloric acid (HCl) solution, at a final concentration of 0.5 N (Palomino & Portaels, 1998); or (3) decontamination by N-acetyll-cysteine-sodium hydroxide (NALC-NaOH) solution, based on diagnostic methods in a handbook by the (European Centre for Disease Prevention and Control, 2018), containing 2.0% NaOH and 0.5% NALC; or (4) dilution by Middlebrook 7H9 Broth. Thereafter, the samples were inoculated on BD™ Middlebrook 7H11 Agar supplemented with 10% OADC enrichment and 0.05% Tween 80 and 2% Ogawa medium (Kyokutou, Japan). These were allowed to incubate at 25 or 30°C for 8 weeks. At 28 and 56 days, a loopful of a visible colony was stained with ZN stain, and ZN-positive colonies were purified.

| Environmental samples from the rearing tank
We performed NTM isolation from the tank water, the filtration sand, and the swabs of biofilm sampled from a wall and a water intake, following the previous methods of Thomson et al. (2008), Parashar et al. (2004) and Thomson et al. (2013), respectively, with small modifications. Following pretreatment and inoculation, the colonies were counted and purified as per the isolations of the tissue samples.

| Polymerase chain reaction (PCR)
The kidney, liver and spleen of both cases, and the testes and epigonal organ of the hammerhead shark were collected and fixed in ethanol. The ovary, uterus and shell gland of the bullhead shark were sampled and frozen. Mixed genomic DNA was extracted from the ethanol-fixed or frozen samples, and the pellets of isolates were cultured from the tissues or environmental samples using a QIAamp DNA Mini Kit (Qiagen). The DNA was tested for the presence of mycobacterial DNA by PCR on the C1000TM Thermal Cycler (Bio-Rad) using GoTaq® DNA Polymerase (Promega) and mycobacterial universal primers targeting the 16S ribosomal RNA (16S rRNA), 65 kDa heat-shock protein (hsp65), the beta subunit of RNA polymerase (rpoB) and superoxide dismutase (sodA), through the cycles described by Fukano et al. (2017). The primer sets are listed in Table S1. Products were separated on a 2% agarose gel and the amplicon was purified using a NucleoSpin Gel and PCR Clean-up Kit (Macherey-Nagel).
The DNA showing a high percentage of nucleotide sequence identities to M. marinum was selected to confirm the presence of insertion sequences IS2404 and IS2606, which relate to mycolactoneproducing mycobacteria (MPM), using the primer sets MU5-MU6 and PU4F-PU7Rbio for IS2404, and MU7-MU8 for IS2606 (Table S1). The cycling conditions with the primer sets MU5-MU6 and MU7-MU8 are described by Stinear et al. (1999). The cycling condition with the primer set PU4F-PU7Rbio is described in Phillips et al. (2005), except that an initial hold step of 95°C for 10 min was

| DNA sequencing and phylogenetic analysis
Purified products were sequenced by the FASMAC Corporation using the specific primers described in Table S1. The resulting sequences were assembled and aligned using ClustalW in MEGA X software (Kumar et al., 2018). Sequence identities were established using the NCBI Basic Local Alignment Search Tool (BLAST) software (http://www.ncbi.nlm.nih.gov/blast/ Blst.cgi). The phylogenetic tree of the hsp65 sequences was reconstructed using the neighbourjoining method with Kimura's 2-parameter model, implemented in MEGA X software and evaluated further in a bootstrap analysis of 1000 replicates.

| Bacterial and tissue samples
To estimate the hybridizability and specificity of the designed PNA probes and to validate the PNA-FISH procedure, mycobacterial isolates from the tissues of the current clinical cases and the reference strains M. marinum ATCC BAA-535 and M. marinum JCM 17638 T were used. To detect pathogens in the shark tissues, paraffin-embedded tissue samples of the current cases, including the testes and epigonal organ of the hammerhead shark, and the ovary, uterus and shell gland of the bullhead shark, were used. As a negative tissue control, the ovary of another hammerhead shark, in which mycobacterial infection was not detected by ZN staining and by PCR assay, was also tested.

| PNA probes
With the genome and partial sequence database using BLAST, and the results of sequence analyses of DNA extracted from the present samples, we designed four PNA probes ( Table 2), each of which was designed to detect the targeted 16S rRNA of the specific NTM species among the four candidate etiological agents detected as the genomic evidence from the tissue samples and/or the isolated strains in this study, including M. marinum, Mycolicibacterium llatzerense (NJB1901), M. hodleri (NJB19061) and M. obuense (NJB19062). Each probe was designed not to cross-hybridize with the non-targeted genomic sequences of the other NTM candidates, and therefore, it acted as a negative control for the other three NTM species like a sense probe. The probes were chosen with regard to purine content, avoiding hairpin formations and inverted repeats. The N-terminus of all probes were labelled with cyanine 3 (Cy3) fluorescent dye. All designed probes were custom-synthesized by PANAGENE Inc.

| FISH procedure
The preparation of tissue sections and bacterial samples for FISH was performed as described in Lehtola et al. (2006) and Lefmann F I G U R E 2 Ovary, Japanese bullhead shark. Oedema of the ovarian surface (arrowhead). et al. (2006), respectively. Both samples were preheated to the hybridization temperature of 60°C. The hybridization protocol was performed as previously reported in Lefmann et al. (2006) by applying a 50μl aliquot of hybridization solution containing a fluorescent probe with a final concentration of 1.5 mol/L to each sample and then incubating the samples at 60°C for hybridization and washing.
Following the hybridization procedure, slides were stained for 10 min in 600 μM 4′,6-diamidino-2-phenylindole (DAPI; Life Technologies), washed in DW for 5 min, air dried and then mounted with 1 drop of Vectashield (Vector Laboratories). Slides were covered with a glass coverslip, and the edges were sealed with clear nail polish. As a negative control, the hybridization procedure was performed without a probe added.

| Microscopy
Microscopy was performed with a fluorescence microscope (Keyence BZ-X700). TRITC OP-87764 and DAPI OP-87762 filter sets (Keyence) were used to analyse the Cy3 and DAPI signals, respectively. Digital images were captured under the fluorescence microscope, and image analysis was performed with BZ-H3 software (Keyence).

| Histopathological analysis
In both testes of the hammerhead shark, foamy macrophages infiltrated into the serosal connective tissues in the resorption zone mixing with large numbers of infiltrating lymphocytes ( Figure 3). Numerous acid-fast bacilli (AFB) were observed within the macrophages comprising the granulomatous inflammation by using ZN stain (Figure 3, inset), but the AFB did not infiltrate into the epigonal organ. Thus, the testes of the hammerhead shark were morphologically diagnosed as: 'granulomatous serositis with intralesional AFB'.
Although it was somewhat difficult to distinguish inflammation In the hammerhead shark, in addition to mycobacteriosis, we diagnosed panophthalmitis, corneal trephination, enteritis, dermatitis, F I G U R E 3 Testis, scalloped hammerhead shark. In the resorption zone, foamy macrophages infiltrated into the connective tissue mixing with large numbers of infiltrating lymphocytes, shown using HE staining. E, epigonal organ; PoM, postmeiotic zone; R, resorption zone. Bar = 200 μm. Inset: Numerous acid-fast bacilli in cytoplasm of the macrophages (arrowhead), visible with Ziehl-Neelsen stain.
focal lymphocytic meningitis and suspected collagenofibrotic glomerulopathy. In the bullhead shark, we additionally diagnosed focal lymphocytic gastritis, focal lymphocytic interstitial nephritis, pancreatitis and necrotic splenitis. However, in both cases, no apparent pathological features of mycobacterial suspicious lesions were detected other than in the testes or ovaries.

| Tissue sampling
In total, three AFB strains were isolated from the testis of the hammerhead shark (NJB1901) and the ovary of the bullhead shark (NJB19061 and NJB19062). NJB1901 was cultured using Middlebrook 7H11 agar without pretreatment; on day 21 after inoculation at 30°C, it formed a smooth, white colony. NJB19061 and NJB19062 were cultured using the same agar without pretreatments; on day 14 after inoculation at 25 and 30°C, respectively, each formed a smooth, yellow colony. All purified colonies of the three isolates showed rapid growth on day 3, with Middlebrook 7H11 agar, at 25°C (NJB19061) or 30°C (NJB1901, NJB19062).

| Environmental samples from the rearing tank
Seven isolates of mycobacteria were cultured from the environmental samples. The colony characteristics of the isolates are presented in Table S2.

| Housekeeping genes of mycobacterium
Amplicons of the appropriate molecular weight were produced by PCR of the DNA extracted from four parts of the bilateral testes testis II of the hammerhead shark, based on the high degree of molecular homology (Table 1).
Polymerase chain reaction of DNA extracted from the isolates produced amplicons of the appropriate molecular weights using primers to amplify the 16S rRNA, hsp65, rpoB and sodA genes. The results of sequence analyses of the isolates showed that NJB1901, NJB19061 and NJB19062 were identical with M. llatzerense, M. hodleri and M. obuense, respectively (Table 1).
The nucleotide sequences of the isolates from the environmental samples were identical to some NTM species other than those detected in the tissue samples (Table S2). The nucleotide sequence data are available in the DDBJ/EMBL/GenBank databases under accession nos. LC713220-LC713257. Figure 5 shows the results of the phylogenetic analysis with the isolates and the genomic evidence from tissue samples using a 401-bp sequence of the hsp65 gene.

| IS2404 and IS2606
No apparent bands were detected for the insertion sequences

| Probe design and evaluation
Probes of 15 bp each were designed to match 100% with the 16S rRNA of each target Mycobacterium species (Table 2)

| Mycobacteria in tissue sections
After probe evaluation using FISH for the type strains and the isolates, PNA-FISH was used to visualize mycobacteria in fixed  in the histopathological examination. At high magnification, rodor dot-shaped red signals were scattered and accumulated immediately adjacent to nuclei, which were stained by DAPI as blue signals ( Figure 10). Ovary sections of the bullhead shark were stained with probe MM, demonstrating the distribution of red signals within the area of macrophage-like cell aggregates with intracellular AFB, likewise revealed by ZN stain (Figure 11). High magnification of the ovary section showed the accumulations of rod-or dot-shaped red signals, as seen in the testis section ( Figure 12).
Conversely, the PNA probes designed for other candidate NTM (i.e., ML, MH and MO) did not show positive signals indicating the presence of the targeted NTM in the tissue sections of both samples. In the negative controls, no signal was detected in the tissue sections examined ( Figure S2). Also, no signal was determined in other tissue sections, including the epigonal organ of the F I G U R E 5 Neighbour-joining tree generated from a concatenated 401-bp sequence of the hsp65 gene from two parts of the testes of the scalloped hammerhead shark (Case 1: Testis I, testis II), the ovary of the Japanese bullhead shark (Case 2: Ovary), and the three isolates cultured from the sampled tissues (NJB1901, NJB19061, NJB19062) with Kimura's two-parameter distance correction model. Bootstrap values are indicated at nodes as a percentage of 1000 replicates.
hammerhead, and the shell gland and uterus of the bullhead shark (data not shown).

| DISCUSS ION
In the present two cases involving a scalloped hammerhead shark and a Japanese bullhead shark that had been maintained in a public display aquarium, the sequence analyses of the extracted DNA from testis I and ovary tissue sections using the hsp65 gene detected genomic evi- shottsii. However, in most M. marinum strains reported previously, the insertion sequences IS2404 and IS2606 relating to mycolactone were not detected by PCR, unlike detection in MPM (Rhodes et al., 2005;Stinear et al., 1999;Trott et al., 2004). Therefore, PCR analysis targeting IS2404 and IS2606 was used to distinguish most types of strains of M. marinum from MPM based on the absence of both these sequences.
The present study found that the hsp65 gene was detected in DNA extracted from testis I and the ovary, but insertion sequences IS2404 and IS2606 were not detected in those tissues using primer sets that had been previously applied to clinical tissue samples and/ or formalin-fixed paraffin-embedded samples (Jacobs, Howard et al., 2009;Phillips et al., 2005;Sakyi et al., 2016), as observed with the reference M. marinum strains examined. These features indicated that the DNA extracted from testis I and the ovary could be identified as M. marinum.
The phylogenetic analysis using the 401-bp sequence of the hsp65 gene also supported this identification, because these nucleotide samples ( Figure 5: Case I Testis I, and Case 2 Ovary) were classified into the same cluster in which the other M. marinum reference strains, such as M. marinum CCUG20998 and E11, were included.
In this study, PNA-FISH with MM probe revealed the existence of M. marinum by detecting its partial sequence in tissues of the reproductive organs in clinical cases in two shark species. Thus, the current study demonstrates that PNA-FISH is useful for the diagnosis of mycobacterial infection in fish and in humans (Kim et al., 2015;Lefmann et al., 2006;. From these overall results, we conclude that the current cases could be diagnosed as mycobacteriosis, with M. marinum as the likely pathogen, and NJB1901, NJB19061 and NJB19062 might be contaminating environmental microorganisms. This study is the first to describe cases of M. marinum infection in an elasmobranch.  (Karsten & Rice, 2006).
Mycobacterium marinum is a representative NTM agent in aquatic animals; it is considered an important candidate infectious agent of aquatic zoonosis because M. marinum has been isolated from human lesions, including dermatitis (Tomas et al., 2014), following exposure to contaminated aquatic materials (Jernigan & Farr, 2000), and is implicated in pneumonia as well (Oh et al., 2018). Based on these features it is vital to know the prevalence and distribution of NTM in aquatic environments and organisms, in accordance with the 'One Health' approach to control NTM infections in humans (Thirunavukkarasu et al., 2017).
However, based on a paucity of research, the susceptibility of elas-  (Streit et al., 2006).
The progress of granuloma pathology is usually dependent on the duration and amount of infecting bacteria (Swaim et al., 2006) F I G U R E S 7-1 2 Visualization by fluorescence in situ hybridization (FISH), superimposed with MM probe (red) and DAPI stain (blue) of Mycobacterium marinum JCM17638T (Figure 7) and of M. marinum ATCC BAA-535 ( Figure 8).
Rod-or dot-shaped red signals were apparent at high magnification. Bar = 20 μm. (Figures 9 and 10) Testis, scalloped hammerhead shark. Visualization by FISH superimposed with MM probe (red) and DAPI stain (blue) (Figure 9). The red signals were located in the resorption zone, where acid-fast bacilli were evident from Ziehl-Neelsen staining. Bar = 500 μm ( Figure 10). Rodor dot-shaped red signals beside the nuclei, blue signal, at high magnification. Bar = 20 μm. (Figures 11 and 12) Ovary, Japanese bullhead shark. Visualization by FISH superimposed with the MM probe (red) and DAPI stain (blue) (Figure 11). The rod signals were located at the connective tissue near follicles. Bar = 500 μm ( Figure 12). Rod-or dot-shaped red signals beside the nuclei, blue signal. Bar = 20 μm. and is influenced by other factors, such as differences among host species (Antuofermo et al., 2017;Swaim et al., 2006) and the bacterial strain (van der Sar et al., 2004). Even so, mycobacteriosis in most teleosts is characterized by caseating granulomas (Noga, 2010). By contrast, the lesions in the reproductive organs of both cases of the sharks in our study showed granulomatous inflammation composed of loose aggregates of macrophages without central necrosis, capsules composed of fibrous tissue or epithelioid tissue and infiltrations of giant cells. Unlike the typical granulomas seen in teleosts, the histopathological features observed in the reproductive organs of the current cases might make the histopathological diagnosis of mycobacteriosis in an elasmobranch quite difficult.
In teleost fishes, the ovary and testis are not major organs involved in mycobacterial infection (Novotny et al., 2010), though have been described in rare cases, including spontaneous and experimental infections of Mycobacterium spp. in association with granulomatous inflammation (Broussard & Ennis, 2007;Whipps et al., 2008).
In elasmobranchs, although clinical reproductive abnormalities (Anderson et al., 2012) and yolk coelomitis (Tuxbury et al., 2017) were reported in guitarfishes with mycobacterial infection, the involvement of some mycobacterial species in these abnormalities of the reproductive system was not determined, because no AFB were detected in the lesions of those cases. The presence of intralesional M. marinum in the reproductive organs in the current cases clarified that mycobacteria could cause gonadal lesions in elasmobranchs as in teleosts.
In conclusion, our study indicates that mycobacteriosis can be suspected in elasmobranchs even without typical caseating granulomas and it should be added to potential zoonotic diseases in aquaria housing sharks.

ACK N OWLED G EM ENT
The authors wish to thank the aquarist team of Shimane Aquarium for their work and commitment. This project was funded by Grantin-Aid for JSPS Research Fellow from the Japan Society for the Promotion of Science (JSPS; grant no. 20 J22583).

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated during and/or analysed during the current study are available from the corresponding author upon reasonable request.