Laser Scanning Confocal Microscopy Was Used to Validate the Presence of Burkholderia pseudomallei or B. mallei in Formalin-Fixed Paraffin Embedded Tissues

Burkholderia pseudomallei and B. mallei are Gram-negative, facultative intracellular bacteria that cause melioidosis and glanders, respectively. Currently, there are no vaccines for these two diseases. Animal models have been developed to evaluate vaccines and therapeutics. Tissues from infected animals, however, must be fixed in formalin and embedded in paraffin (FFPE) before analysis. A brownish staining material in infected tissues that represents the exopolysaccharide of the pathogen was seen by bright field microscopy but not the actual microorganism. Because of these results, FFPE tissue was examined by laser scanning confocal microscopy (LSCM) in an attempt to see the microorganism. Archival FFPE tissues were examined from ten mice, and five nonhuman primates after exposure to B. pseudomallei or B. mallei by LSCM. Additionally, a historical spleen biopsy from a human suspected of exposure to B. mallei was examined. B. pseudomallei was seen in many of the infected tissues from mice. Four out of five nonhuman primates were positive for the pathogen. In the human sample, B. mallei was seen in pyogranulomas in the spleen biopsy. Thus, the presence of the pathogen was validated by LSCM in murine, nonhuman primate, and human FFPE tissues.


Introduction
Melioidosis is caused by the Gram-negative, facultative intracellular pathogen Burkholderia pseudomallei. It is endemic in Southeast Asia and northern Australia and appears to be much more widely distributed than originally reported [1]. It can be found in soil, wet lands, and water, and the incidence of melioidosis appears to increase during the raining season [2]. Exposure to B. pseudomallei can be through cutaneous inoculation, ingestion of contaminated water or food, or by inhalation. Infection by B. pseudomallei is the third leading cause of disease in Southeast Asia behind tuberculosis and AIDs, and pneumonia is the most common clinical presentation of melioidosis. Host risk factors,

Materials and Methods
2.1. Bacterial Strains, B. pseudomallei K96243 Antibody, and Human Tissue B. mallei GB18-3 was obtained from the Bacteriology Division culture collection at the U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID), Fort Detrick, Frederick, Maryland, and it had been passed through hamsters 3 times [13]. Single use stock cultures of B. pseudomallei K96243 were obtained from the Unified Culture Collection (UCC) at USAMRIID. A rabbit antibody preparation made against an extract of irradiated, whole-B. pseudomallei K96243 (IRBpK) cells was a kind gift from Robert Ulrich (USAMRIID). Human tissue from a patient suspected of exposure to B. mallei was obtained from the Joint Pathology Center (Silver Spring, MD, USA).

Growth of Bacterial Strains and Antigen Preparation
The following procedure describes the general growth conditions and preparation of a whole-cell, bacterial antigen of B. mallei [13] or B. pseudomallei. All procedures were performed under biosafety level 3 (BSL3) conditions and culture/cell manipulations were carried out in a biosafety hood. Two hundred milliliters of 4% glycerol tryptone broth (GTB) (Difco, ThermoFisher Scientific, Walthman, MA, USA) in a 1L flask was inoculated with 5 µl of a stock culture of the organism, and the culture incubated with shaking at 200 rpm overnight at 37 • C (16-18 h). After overnight growth, the culture was placed into 50 mL conical tubes (not more than half full), and the tubes were centrifuged for 25 min at 3700 rpm in a swinging bucket rotor at 4 • C. The culture supernatants were discarded and cell pellets were suspended in 1.0 mL of Hanks Balanced Salt Solution (HBSS with calcium and magnesium, ThermoFisher Scientific). The cell suspensions were combined and formaldehyde (ThermoFisher Scientific) was added to a final concentration of 4%. The cells were left in the formalin solution for 24 h at 4 • C, and then the cells were washed twice with cold HBSS. Ten percent of the total volume was used to test for sterility on sheep blood agar plates that were incubated at 37 • C for 3 days. After validation of the sterility of the formalin-fixed cells, the cells were dialyzed for 3 days against 1 L of water with daily changes (Spectra 3000 MW-cut off, ThermoFisher Scientific) at 4 • C. The dialyzed cells were centrifuged and cell pellets were suspended in sterile water. The absorbance of the cells suspension was compared to a standard curve to calculate the protein concentration, and the cells were stored in aliquots at −70 • C.

Production of Rabbit Polyclonal Antibodies Against B. mallei or B. pseudomallei
Two methods were used to produce rabbit polyclonal antibodies against formalin-treated whole-cells. In the first method, formalin-treated B. mallei (fBm) GB18-3 cells were formulated with Ribi TriMix as the adjuvant (Ribi ImmunoChem Research Inc., Hamilton, MT, USA) [13]. In the second method, formalin-treated B. pseudomallei K96243 (fBpK) cells were formulated with Freund's complete adjuvant (FCA) or Freund's incomplete adjuvant (FIA) (Sigma-Alrich, Saint Louis, MO, USA). The general procedure to generate polyclonal antibodies in 2 female NZW rabbits (~2.5 kg) were as follows (Covance Research Products, Denver, PA, USA): prebleed, 21 days before the primary vaccination; primary vaccination, 250 µg of fBpK in FCA; 3 boost (21 days apart) vaccinations starting 21 days after the primary vaccination, 125 µg of fBpK in FIA; terminal bleed, 14 days after the last boost. Antibody (IgG) titers against IRBpK, fBpK, and fBm cells were determined at least twice by ELISA as previously described [14]. See Table A1 in Appendix A for antibody titers of antibodies used in the present study. No new animals were used at USAMRIID for this report.

Immunohistochemistry
Immunohistochemistry (IHC) was performed using the Dako Envision system (Dako Agilent Pathology Solutions, Carpinteria, CA, USA). Briefly, after deparaffinization, peroxidase blocking, and antigen retrieval, sections were covered with a rabbit polyclonal anti-B. mallei or B. pseudomallei antibody (USAMRIID, Frederick, MD, USA) at a dilution of 1:6000 and incubated at room temperature for forty five minutes. They were rinsed, and the peroxidase-labeled polymer (secondary antibody) was applied for thirty minutes. Slides were rinsed and a brown chromogenic substrate 3,3 Diaminobenzidine (DAB) solution (Dako Agilent Pathology Solutions) was applied for eight minutes. The substrate-chromogen solution was rinsed off the slides, and the slides were counterstained with hematoxylin and rinsed. The sections were dehydrated, cleared with Xyless, and then coverslipped. Stained sections were digitized and examined with Aperio Image Scope software (Aperio Technologies, Vista, CA). The specimens were examined with an Olympus BX53 microscope (Olympus America, Center Valley, PA, USA).

Immunofluorescence and Laser Scanning Confocal Microscopy Imaging
Formalin-fixed paraffin embedded (FFPE) tissue sections were deparaffinized using xylene and a series of ethanol washes before staining single sections with H&E. 0.1% Sudan black B (Sigma-Alrich) treatment was used to eliminate the autofluorescence background, and sections were heated in a citrate buffer (pH 6.0) for 15 min to reverse formaldehyde crosslinks. After rinsing with PBS (pH 7.4), the sections were blocked with PBS containing 5% normal goat serum overnight at 4 • C. The sections were incubated with rabbit anti-B. pseudomallei or anti-B. mallei polyclonal antibody (1:1000-1500) for 2 h at room temperature. After rinsing with PBS, the sections were incubated with a secondary Alexa Fluor 488 conjugated goat anti-rabbit antibody for 1 h at room temperature. Sections were cover slipped using the Vectashield mounting medium with or without 4 ,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA, USA) to stain nuclei. Additionally, in some cases, to visualize the presence of macrophages in the nonhuman primate samples, a mouse anti-human/NHP CD68 antibody was used (Dako Agilent Pathology Solutions) or CD45 antibody for lymphocytes (Dako). To visualize B. pseudomallei/B. mallei, z-stacks (multiple slices) were used. Images were captured on a Zeiss LSM 880 confocal system (Carl Zeiss, Oberkochen, Germany) and processed using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

IHC Analysis of FFPE by Bright Field Microscopy
In the animal model studies of melioidosis and glanders, bright field microscopy was used in the IHC analysis of FFPE tissue from exposed animals. An anti-B. pseudomallei K96243 polyclonal antibody was used as the primary antibody to detect the presence of B. pseudomallei in the FFPE tissue. An example is shown of a spleen from a C57BL/6 mouse 47 days post-infection (PI) that was exposed to B. pseudomallei 22 by aerosol ( Figure 1). The exopolysaccharide from B. pseudomallei 22 was seen as a brownish staining material associated with pyogranulomatous lesions and the immediate surrounding cells ( Figure 1C,D). However, the actual pathogen was not seen on closer examination.

IHC Analysis of FFPE by Bright Field Microscopy
In the animal model studies of melioidosis and glanders, bright field microscopy was used in the IHC analysis of FFPE tissue from exposed animals. An anti-B. pseudomallei K96243 polyclonal antibody was used as the primary antibody to detect the presence of B. pseudomallei in the FFPE tissue. An example is shown of a spleen from a C57BL/6 mouse 47 days post-infection (PI) that was exposed to B. pseudomallei 22 by aerosol ( Figure 1). The exopolysaccharide from B. pseudomallei 22 was seen as a brownish staining material associated with pyogranulomatous lesions and the immediate surrounding cells ( Figure 1C,D). However, the actual pathogen was not seen on closer examination.

LSCM Analysis of FFPE Tissue
Because of the difference in technology between bright field microscopy and LSCM on how the image is captured, would this method enable us to see B. pseudomallei bacterial cells in FFPE tissues? A polyclonal antibody raised against a formalin-treated B. mallei GB18 was used as the primary antibody to examine FFPE tissue by LSCM from a C57BL/6 mouse exposed to B. pseudomallei 22. The same infected spleen shown in Figure 1 was used for comparison that was examined by bright field microscopy (see Table 1 for tissue source). Unlike the results with bright field microscopy, however, B. pseudomallei 22 cells were seen within the pyogranulomatous lesions in the spleen from the infected C57BL/6 mouse with LSCM ( Figure 2).

LSCM Analysis of FFPE Tissue
Because of the difference in technology between bright field microscopy and LSCM on how the image is captured, would this method enable us to see B. pseudomallei bacterial cells in FFPE tissues? A polyclonal antibody raised against a formalin-treated B. mallei GB18 was used as the primary antibody to examine FFPE tissue by LSCM from a C57BL/6 mouse exposed to B. pseudomallei 22. The same infected spleen shown in Figure 1 was used for comparison that was examined by bright field microscopy (see Table 1 for tissue source). Unlike the results with bright field microscopy, however, B. pseudomallei 22 cells were seen within the pyogranulomatous lesions in the spleen from the infected C57BL/6 mouse with LSCM ( Figure 2).

LSCM Analysis of Murine Tissue from Animal Model Studies
The previous results encouraged us to begin a retrospective study of archival FFPE tissues from our murine melioidosis animal model studies with LSCM to see if B. pseudomallei could be visualized in other animal tissues (see Table 1 for tissue source). FFPE tissues were examined from 10 mice (one to four organs from each mouse) that were exposed to B. pseudomallei by LSCM. Not all mice (2) were positive for B. pseudomallei. Tissues that were positive were also those with local pyogranulomatous inflammation, and areas without pyogranulomatous inflammation were negative for the pathogen (Figure 3). In Figure 3A-C, clusters of B. pseudomallei K96243 cells were seen in the dorsal thoracic region of a BALB/c mouse that showed staining of the outer surface of single and dividing cells of the microorganism. In pyogranulomas present in the spleen, the microorganism was seen primarily within the pyogranuloma, and very few were outside the pyogranuloma ( Figure 3D-E). B. pseudomallei K96243 was seen in the lumbar ( Figure 3G-I) and lung ( Figure 3J-L) of aerosol exposed BALB/c mice. Figure 3M-O showed a negative liver from a C57BL/6 mouse that was exposed to B. pseudomallei K96243. See Figure A1 for LSCM analysis of spleens from naïve BALB/c mice.

LSCM Analysis of Murine Tissue from Animal Model Studies
The previous results encouraged us to begin a retrospective study of archival FFPE tissues from our murine melioidosis animal model studies with LSCM to see if B. pseudomallei could be visualized in other animal tissues (see Table 1 for tissue source). FFPE tissues were examined from 10 mice (one to four organs from each mouse) that were exposed to B. pseudomallei by LSCM. Not all mice (2) were positive for B. pseudomallei. Tissues that were positive were also those with local pyogranulomatous inflammation, and areas without pyogranulomatous inflammation were negative for the pathogen (Figure 3). In Figure 3A-C, clusters of B. pseudomallei K96243 cells were seen in the dorsal thoracic region of a BALB/c mouse that showed staining of the outer surface of single and dividing cells of the microorganism. In pyogranulomas present in the spleen, the microorganism was seen primarily within the pyogranuloma, and very few were outside the pyogranuloma ( Figure 3D-E). B. pseudomallei K96243 was seen in the lumbar ( Figure 3G-I) and lung ( Figure 3J-L) of aerosol exposed BALB/c mice. Figure 3M-O showed a negative liver from a C57BL/6 mouse that was exposed to B. pseudomallei K96243. See Figure A1 for LSCM analysis of spleens from naïve BALB/c mice.

LSCM Analysis of Nonhuman Primate (NHP) Tissue
FFPE tissues from four NHPs (two African Green Monkeys [AGM], and two Rhesus macaques) were examined by LSCM that were exposed to B. pseudomallei HBPUB10134a by aerosol. Three tissues were examined by LSCM (lung, liver, spleen) for each NHP (see Table 1 for tissue source). Both AGM were infected (all tissues examined), but only one rhesus appeared to be infected (all tissues examined). Generally, more B. pseudomallei HBPUB10134a microorganisms were found in AGM tissues when present than in tissues from Rhesus macaques.

LSCM Analysis of Suspected Human Tissue
Other FFPE tissues were examined to determine if LSCM could identify the presence of B. mallei. A historical spleen biopsy (FFPE) from a human suspected of having been exposed to B. mallei was obtained. Figure 5 shows examples of two areas of pyogranulomatous inflammation in the spleen biopsy. In Figure 5A-C, positive cells were seen within an area of pyogranulomatous inflammation by LSCM that upon closer examination looked like bacterial rods ( Figure 5C). In another area with a small pyogranuloma ( Figure 5D-F), a cluster of positive cells ( Figure 5E) was seen that appeared to consist of rod-shaped bacterial cells at higher magnification ( Figure 5F). Thus, the possible presence of B. mallei in FFPE tissue was demonstrated from a human suspected of being exposed to B. mallei by LSCM.

LSCM Analysis of Suspected Human Tissue
Other FFPE tissues were examined to determine if LSCM could identify the presence of B. mallei. A historical spleen biopsy (FFPE) from a human suspected of having been exposed to B. mallei was obtained. Figure 5 shows examples of two areas of pyogranulomatous inflammation in the spleen biopsy. In Figure 5A-C, positive cells were seen within an area of pyogranulomatous inflammation by LSCM that upon closer examination looked like bacterial rods ( Figure 5C). In another area with a small pyogranuloma ( Figure 5D-F), a cluster of positive cells ( Figure 5E) was seen that appeared to consist of rod-shaped bacterial cells at higher magnification ( Figure 5F). Thus, the possible presence of B. mallei in FFPE tissue was demonstrated from a human suspected of being exposed to B. mallei by LSCM.

Comparison of Polyclonal Antibodies Used to Examine FFPE Tissue by LSCM
A question arose if other types of antibodies would work with LSCM when examining FFPE tissue. In the previous study above with LSCM, an antibody was used that was developed against a formalin-treated B. mallei GB18 (fBm) whole cell. Figure 6A shows an area with pyogranulomatous inflammation in the lung from an AGM exposed to B. mallei FMH that was examined with two other antibody preparations (see Table A1 for antibody ELISA titers). Figure 6B shows the presence of B. mallei FMH in the area of inflammation with the antibody (raised against fBm) that was used in the previous study. In Figure 6C, similar results were seen with an antibody raised against a whole-cell extract of B. pseudomallei K96243 (extBpK). Similarly, the presence of B. mallei was seen in the lung of the AGM with an antibody raised against formalin-treated B. pseudomallei K96243 whole-cells (fBpK).
Therefore, it appears that antibodies raised against three different Burkholderia antigen preparations will work with LSCM of infected FFPE tissue.
tissue. In the previous study above with LSCM, an antibody was used that was developed against a formalin-treated B. mallei GB18 (fBm) whole cell. Figure 6A shows an area with pyogranulomatous inflammation in the lung from an AGM exposed to B. mallei FMH that was examined with two other antibody preparations (see Table A1 for antibody ELISA titers). Figure 6B shows the presence of B. mallei FMH in the area of inflammation with the antibody (raised against fBm) that was used in the previous study. In Figure 6C, similar results were seen with an antibody raised against a whole-cell extract of B. pseudomallei K96243 (extBpK). Similarly, the presence of B. mallei was seen in the lung of the AGM with an antibody raised against formalin-treated B. pseudomallei K96243 whole-cells (fBpK). Therefore, it appears that antibodies raised against three different Burkholderia antigen preparations will work with LSCM of infected FFPE tissue.

Discussion
The presence of B. pseudomallei or B. mallei bacterial cells in FFPE tissue by LSCM was demonstrated that it was not by bright field microscopy. Burkholderia cells in FFPE tissues were seen from mice and nonhuman primates exposed to B. pseudomallei or B. mallei, and in a historical spleen biopsy from a human suspected of being exposed to B. mallei. The difference in the technology of image formation between bright field microscopy and LSCM made it possible to clearly visualize the pathogen in FFPE tissue. Briefly, in LSCM the excitation (laser) and emission light sources are limited (either both or only emission) by pinhole apertures with the focus on a point(s) within a single plane of the sample. Excitation and emission light derived from above, below, and away from the point or plane of focus are generally excluded by the pinhole apertures that result in a higher resolution image. As the image is scanned, the point of focus stays in the same plane (optical sectioning) [14]. In contrast,

Discussion
The presence of B. pseudomallei or B. mallei bacterial cells in FFPE tissue by LSCM was demonstrated that it was not by bright field microscopy. Burkholderia cells in FFPE tissues were seen from mice and nonhuman primates exposed to B. pseudomallei or B. mallei, and in a historical spleen biopsy from a human suspected of being exposed to B. mallei. The difference in the technology of image formation between bright field microscopy and LSCM made it possible to clearly visualize the pathogen in FFPE tissue. Briefly, in LSCM the excitation (laser) and emission light sources are limited (either both or only emission) by pinhole apertures with the focus on a point(s) within a single plane of the sample. Excitation and emission light derived from above, below, and away from the point or plane of focus are generally excluded by the pinhole apertures that result in a higher resolution image. As the image is scanned, the point of focus stays in the same plane (optical sectioning) [14]. In contrast, in bright field microscopy, the entire field is exposed by the light source, and the resulting image may focus on multiple planes which results in a lower resolution image than obtained with LSCM. This resolution is also true with epi-fluorescence microscopy when compared with LSCM.
In earlier reports on the visualization of B. pseudomallei or B. mallei in FFPE tissue samples, it was reported that the presence of the pathogen was seen after immunohistochemical (IHC) staining of paraffin sections [15,16], which was different than in our present report because the pathogen after IHC staining was not visualized. It was not clear why there was a difference from our study, but in Wong et al. [15], they reported using a Gram stain on their FFPE tissues, while in Glaros et al. [16], they may have used a different imaging system. In other reports, fluorescent in situ hybridization (FISH) was used to identify B. pseudomallei or B. mallei present in FFPE tissue in infected murine [17] or human tissue samples [18]. Furthermore, DNA was extracted from infected murine FFPE tissue samples and polymerase chain reaction (PCR) was run with B. pseudomallei specific primers to determine if the pathogen was present in the samples [19,20]. In two other reports, transmission electron microscopy was used to visualize B. pseudomallei in experimental mouse studies [21] or clinical human melioidosis patients [22] without or with FFPE tissue, respectively.
There are few reports of the use of LSCM to visualize B. pseudomallei or B. mallei in infected tissue. LSCM was used to visualize B. mallei, B. pseudomallei, or B. thailandensis in murine macrophage-like RAW 264.7 cells [23,24], and B. pseudomallei infection of A549 human lung epithelial cells [22,25]. A recent study of B. mallei in an infected mouse FFPE spleen sample that was examined by LSCM was reported [26]. Finally, LSCM has been used to detect the presence of other pathogens, such as Mycobacterium tuberculosis in lung tissue from human patients [27].
One caution to this report is that the figures presented in this report may show more microorganisms than present than seen in human cases of melioidosis or glanders. Animal tissues are easier to recover and manipulate than human tissue, for example the mouse spleen. Burkholderia appear to accumulate in the spleen and cause the formation of pyogranulomas that are in most cases easy to observe upon autopsy. A mouse spleen with a pyogranuloma may contain from 10 4 to 10 9 CFU [9]. Additionally, because these tissues come from experimental melioidosis or glanders animal models, they may be exposed to more CFU than normally encountered by humans that acquired melioidosis or glanders. Thus, they are more likely to be acutely infected. In addition, animals may be chosen that are more susceptible to infection than others, such as BALB/c mice versus C57BL/6 mice, where the latter species is generally more resistant than the former. This could result in higher CFU in the susceptible animal than in the resistant animal [7,8].
This report established the feasibility of using LSCM to validate the presence of B. pseudomallei or B. mallei in FFPE tissue from different animal or human sources. This technology may be diagnostic for melioidosis (or glanders) or complement the diagnosis of the disease with the isolation of the pathogen. In addition, it would be useful to have B. pseudomallei and B. mallei capsule-specific monoclonal antibodies, if there is a question between diagnosis of melioidosis or glanders [28][29][30]. One advantage of LSCM is that it can validate the diagnosis of melioidosis or glanders, but some disadvantages of LSCM are the cost of the system, the expertise to operate the system, and the preparation of the sample would limit its usefulness in an on-site clinical setting where diagnosis is needed. At present, it may have to be part of a core facility that serves a wide area in need of such supporting technology.