Structural and Topographic Dynamics of Pulmonary Histopathology and Local Cytokine Profiles in Paracoccidioides brasiliensis Conidia-Infected Mice

Background Paracoccidioidomycosis (PCM), an endemic systemic mycosis caused by the fungus Paracoccidioides brasiliensis (Pb), usually results in severe lung damage in patients. Methods and Findings Considering the difficulties to sequentially study the infection in humans, this work was done in mice inoculated intranasally with infective Pb-conidia. Lungs of control and Pb-infected mice were studied after 2-hours, 4, 8, 12 and 16-weeks post-infection (p.i) in order to define histopathologic patterns of pulmonary lesions, multiplex-cytokine profiles and their dynamics during the course of this mycosis. Besides the nodular/granulomatous lesions previously informed, results revealed additional non-formerly described lung abnormalities, such as periarterial sheath inflammation and pseudotumoral masses. The following chronologic stages occurring during the course of the experimental infection were defined: Stage one (2-hours p.i): mild septal infiltration composed by neutrophils and macrophages accompanied by an intense “cytokine burst” represented by significant increases in IL-1α, IL-1β, IL-4, IL-5, IL-6, IL-10, IL12p70, IL-13, IL-17, Eotaxin, G-CSF, MCP1, MIP1α, GM-CSF, IFN-γ, MIP1β and TNFα levels. Stage two (4-weeks p.i): presence of nodules, evidence of incipient periarterial- and intense but disperse parenchymal- inflammation, abnormalities that continued to be accompanied by hyper-secretion of those cytokines and chemokines mentioned in the first stage of infection. Stages three and four (8 and 12-weeks p.i.): fungal proliferation, inflammation and collagenesis reached their highest intensity with particular involvement of the periarterial space. Paradoxically, lung cytokines and chemokines were down-regulated with significant decreases in IL-2,IL-3,IL-5,IL-9,IL-13,IL-15,GM-CSF,IFN-γ,MIP1β and TNFα. Stage five (16-weeks p.i.): inflammation decreased becoming limited to the pseudotumoral masses and was accompanied by a “silent” cytokine response, except for PDGF, MIG, RANTES and IL12p40 which remained up-regulated for the duration of the experiment. Conclusions Results of this study identified both classic and novel patterns corresponding to histopathologic and immunologic responses occurring during the course of experimental PCM.


Introduction
The chronic adult form of PCM progresses slowly with minor symptoms and may take months to years to become clinically manifested; consequently, when the patients are finally diagnosed and receive medical treatment, pulmonary lesions are already well established, presenting different stages of development [1][2][3].
The primary infection takes place in the lungs where the fungus may persist inactive for years. Tuder et al. (1985), in twelve patients with chronic PCM, described five main pathologic aspects: (a) pneumonic reaction; (b) early granulomatous formation; (c) mature and healed granulomas; (d) mixed pattern (early and mature granuloma in the same pulmonary area); (e) pulmonary fibrosis, with the latter being present in all cases [4]. In addition, in chronic human PCM, the lesions may be either diffuse or circumscribed [3]. Human studies, however, do not allow the histopathologic sequential analyses required to build up the sequence of events. Therefore, several experimental models have been applied to compensate this limitation, aiming to better define the pathogenic mechanisms in PCM [5][6][7][8][9][10][11]. Despite important differences in the experimental designs due to infective forms (conidia, yeasts) and fungal isolates used, inoculum concentration, route of inoculation and host genetic background, overall results have contributed to determine some stages of the inflammatory dynamics, delineate cellular and extracellular matrix composition of the granulomas and related aspects [5][6][7][8][9][10][11]. Nonetheless, a detailed description of the topographic location of the lesions in the lung compartments and the characteristics of other inflammatory patterns, besides the granulomatous reaction, had been incomplete and superficially referred to in the literature. Additionally, the immunological studies reported have been focused on specific cell populations or tested a restricted number of cytokines produced in vivo or in vitro [12][13][14]. Therefore, the aim of this study was to sequentially determine histopathologic patterns of lesions and multiplex-determined cytokine profiles in an experimental model of PCM.
A deeper knowledge of the pattern of lesions and cytokines profiles induced by the fungus and their changes according to the time of infection could contribute to prognosis prediction.

Ethics statement
All animals were handled according to the national (Law 84 of 1989, Res No. 8430 of 1993) and international (Council of European Communities and Canadian Council of Animal Care, 1998) guidelines for animal research and the experimental protocols were approved by Corporación para Investigaciones Biológicas (CIB) research ethics committee.

Animals
BALB/c mice were originally obtained from Taconic Farms, Inc., Quality Laboratory Animals and Services for Research, New York, USA, with the breeding colony being expanded at the (CIB), Medellin, Colombia. Male mice, 6-7 weeks old and approximately 20 g in weight were used in this study. Mice were divided into 2 groups: non-infected control mice (n = 50) and infected mice (n = 50).

Fungus
Paracoccidioides brasiliensis isolate ATCC-60855 from a Colombian patient registered at the American Type Culture Collection (Rockville, MD), and known to produce abundant conidia was used in all experiments [15]. The techniques applied to grow the mycelial forms with subsequent collection and dislodgement of conidia had been previously reported [15][16][17].

Experimental infection
Mice were anesthetized by the intramuscular injection of a solution containing -ketamin hydrochloride (Park, Davis & Company, Berlin, Germany; 100 mg/kg) and xylazine (Bayer, Brazil 10 mg/kg) [18]. When deep anesthesia was obtained, 3610 6 Pb viable conidia (in 0.06 ml of the inoculum) were intranasally instilled. Control mice received an intranasal inoculum of 0.06 ml of PBS.

Image analysis
Location, shape, size, frequency, cellular composition, fungal budding patterns, and collagen deposits features were evaluated for each inflammatory pattern using scanned slides images and the free ImageScope software (http://www.aperio.com/downloadimagescope-viewer.asp). Tissue sections were examined and evaluated blindly by a pathologist.
The total inflammatory area was measured by using one panoramic image of both lungs per mouse. Regions of interest (ROIs), correspondent to the inflammatory regions, were manually drawn and measured. The percentage of pulmonary area with inflammatory reaction was calculated dividing the sum of total ROIs by the total area of lung (excluding the air space).
The Aperio positive pixel count algorithm was used to quantify the amount of a specific stain present in a scan slide image. Red pixels in the PIFG stained slides were measured to quantify collagen, and brown/black pixels in the Grocott's stained slides to quantify yeast cells.
Cellular composition was evaluated using a semi-quantitative approach. A score of 2 to +++ was given according to the degree of infiltration with (+++) for intense, (++) for moderate, (+) for slight, (6) for very slight and (2) for no reaction [24].

Author Summary
Paracoccidioidomycosis (PCM), an endemic fungal infection of pulmonary origin resulting in severe disseminated disease, occurs in rural areas of most South American countries and presents several clinical forms. The infection is acquired by inhalation of specific fungal propagules, called conidia. Considering the difficulties encountered when studying the infection in humans, this work was done in mice infected by inhalation of infective fungal conidia thus mimicking the human natural infection. The lungs of mice were sequentially studied by histopathological and multiplex cytokine methods from 2 h to 16 weeks after infection to verify the course of the disease. The mycosis presented different morphologic aspects during the course of time, affecting several pulmonary compartments. Otherwise and based on the analysis of 30 cytokines, the immune response also showed heterogeneous responses, which were up or down regulated depending on the time of infection. By recognizing the different stages that correspond to the evolution of pulmonary lesions, the severity (benign, chronic or fibrotic) of the disease could be predicted and the probable prognosis of the illness be inferred.
Immunopathologic Profiles in Experimental PCM www.plosntds.org Immunohistology Rabbit immunoglobulin to human fibronectin (DAKO, Cat # A245), known to react with fibronectin-covered Pb yeasts [25], was used to better observe the yeasts into a ''perifungal space''. Indirect immunofluorescence was done on paraffin lung sections as described [25]. All slides were counterstained with Evans blue at a dilution of 1/10.000 and mounted in buffered glycerol containing p-phenylenediamine and analyzed by confocal microscopy (LSM-510-META, Zeiss).
Cytokine levels were determined using a multiplex array reader from Luminex TM Instrumentation System (Bio-Plex Workstation from Bio-Rad Laboratories). The analyte concentration was calculated using software provided by the manufacturer (Bio-Plex Manager Software) and expressed as pg.

Statistical analysis
Histopathology results were obtained from five biological replicates at each time point. Multiplex cytokine bioplex assays were analyzed in five additional animals at each time and run in duplicates.
Cytokine data analysis was performed through the use of R language (R Development Core Team, 2010). R: A language and environment for statistical computing. (R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org). Cytokine data were initially analyzed by exploratory data analysis (EDA). The variables were individually plotted with 40% of smoothing by lowess (local regression), aiming to define the behavior of each variable during the infection. Pb-infected and control groups were compared by non-parametric Kruskal-Wallis test with Dunn's multiple comparisons post-test (Prism 5.0 software, Graph Pad, USA) dividing the groups up to four and after eight weeks of infection. p values less than 0.05 were considered statistically significant; p values less than 0.01 were considered statistically highly significant.

Histopathological aspects
None of the animal receiving PBS exhibited histopathologic alterations except transient and slight septal infiltration of neutrophils with some eosinophils at 2 h post-inoculations. In comparison, the lungs of BALB/c mice infected with Pb conidia exhibited, from the 4 weeks onwards, pulmonary lesions with three main histopathologic patterns: the classical and commonly reported nodular or granulomatous pattern, followed by inflammatory peri-arterial sheaths and pseudotumoral masses. A quantitative evaluation of those inflammatory infiltrates is illustrated in the Figure 1. Immunopathologic Profiles in Experimental PCM www.plosntds.org Nodular pattern. It corresponded to multiple sphere-like or oval-shaped parenchymal granulomas adjacent to terminal bronchioles, sometimes isolated, but becoming confluent with the course of infection (Figure 2A). Nodules presented the following frequencies: 10615, 34626, 665 and 767 at 4, 8, 12 and 16 weeks, respectively (p,0.05 at 8 weeks in reference to the other times). The mean diameter of the nodules was 360.26SD 103.7 mm.
Their cellular composition varied according to time of infection. At 4 weeks, the granulomas were composed by a central core of macrophages with fungi and intermixed neutrophils sometimes presenting an apoptotic aspect ( Figure 2B). The most peripheral zone often consisted of non-continuous halo of small lymphocytes, at times forming pseudo-follicular lymphocyte aggregates.
At 8 weeks post-infection (p.i.), the nodules showed an increase in their number and a higher confluence among them. The most significant change in the central zone was the appearance of some xantomatous cells and cholesterol crystals. Additionally, the number of neutrophils significantly decreased and a large number of mature plasma cells appeared at the granuloma periphery.
At 12 weeks p.i., the xantomatous aspect of the macrophages and the number of cholesterol crystals were intensified ( Figures 2C, 2D). Some of the granulomas were confluent and/or in direct contact with the periarterial lesions.
At 16 weeks p.i., nodules presented a significant increase in the number of xantomatous and Touton cells.
Inflammatory periarterial sheath pattern. At 4 weeks, the periarterial spaces were globally preserved but presented lymphatic dilatation and edema. The inflammatory cells consisted in only focal accumulation of small lymphocytes and some macrophages with neutral glycoproteins (PAS positive stain).
At 8 weeks, the involvement of this pulmonary compartment was more diffuse and defined periarterial inflammatory sheaths ( Figures 2E, 2F). The inflammatory infiltrate consisted of a large number of macrophages, fungi, mature plasmocytes, small lymphocytes and eosinophils. The lymphatic vessels were more prominent.
The same aspects were also seen at 12 weeks p.i. except for presenting a higher number of eosinophils, plasmocytes and Mott cells ( Figures 2G, 2H). At 16 weeks p.i., there was a reduction in the periarterial involvement.
Peri-venous inflammation occurred less frequently than arterial involvement with the former being more prominent at 12 weeks p.i. The perivenular cellular composition was similar to that in arteries.
Pseudotumoral pattern. Pseudotumoral masses were 5 to 6 times larger (20006194 mm of length) than the nodules (Figures 2I, 2J). This pattern appeared mainly at 16 weeks p.i. and was the main lesion in 80% of mice at this time. Masses were adjacent to hilar bronchi and their large size produced bronchial compression ( Figure 2K). In one case, the mass projected into a large bronchus, forming an endobronchial polyp ( Figure 2J). These masses exhibited a combination of the main features described in the two previous patterns, and presented also a larger number of plasmocytes with plasmacytic differentiation (plasmocytogenesis) ( Figure 2L) and eosinophils in the peripheral zone.

Extracellular matrix aspects of the lesions
The aspect and quantity of extracellular matrix were dependent on the type of lesion and time of infection. Collagen fibers gradually increased up to 12 weeks, with further decrease at 16 weeks.
Comparative features between extracellular matrix arrangements in the lesions are presented in the Figure 3.

Fungal morphology
The morphology and quantity of the fungi in the lung tissue also varied depending on the location of the lesions and the time of the infection with a tendency to increase fungal numbers at 8 and 12 weeks post-infection.
In nodules, fungal cells were restricted to the central zone with predominance of large yeasts (10-20 mm). In comparison with the nodular pattern, fungi in perivascular lesions were more dispersed and frequently presenting a perifungal space surrounding a large mother cell with multiple small oval buds, less than 2 mm. These buds formed concentric aggregates around the mother cell and/or were dispersed in or attached to the external border of the perifungal space. Fungal cells were present in high amounts in pseudotumoral masses.
A summary of the main histopathologic aspects for each pulmonary lesion is represented in Table 1.

Immunological aspects
By means of exploratory data analysis, the kinetic of each cytokine was plotted against time of infection using Lowess smoothing (40%). The observed trajectories (data not showed) clearly showed a switch of some cytokine levels up to 4-and from 8-weeks afterwards. To test this statistical hypothesis, each cytokine in the Pb-infected group was compared with those of the control group by the Kruskal-Wallis test using two time periods consisting of two hours up to four weeks and eight to sixteen weeks of infection ( Figure 4). These analyses enabled us to identify five different patterns of inflammatory mediators, as follows:
In the second stage (4 wks p.i.), nodules and incipient periarterial inflammatory reaction, together with disperse and intense parenchymal inflammation became evident. In addition, hyper-secretion of the cytokines and chemokines grouped in patterns 1 and 2 was also recorded.
In the third and fourth stages (8 and 12 wks p.i., respectively) fungal growth, inflammation and collagenesis reached their highest intensity with particular involvement of the periarterial space. Paradoxically at this point, lung cytokines and chemokines were down-regulated as revealed by significant decreases in IL-5, IL-13, GM-CSF, IFN-c, MIP1b, TNFa, IL-2, IL-3, IL-9 and IL-15 (patterns 2 and 3).
Finally, in the fifth stage (16 wks p.i.), the inflammation decreased and became practically restricted to the pseudotumoral masses which were now accompanied by a ''silent'' cytokine response, except for PDGF, and MIG, RANTES and IL12p40, all of which remained upregulated during the whole experimental period. Some associations among morphological changes and cytokine profiles could be attempted, as follows: (A) High levels of IL-1a, IL-1b, IL-6, IL-17, TNFa and chemokines such as eotaxin, MCP1 and MIP1a indicated a highly proinflammatory milieu at 2 h and 4 wks p.i. and would explain the recruitment of monocytes, PMNs and eosinophils [26].  Immunopathologic Profiles in Experimental PCM www.plosntds.org airway hypersensitivity, eosinophilia, abundant plasma cells and pathological changes in the lungs [27]. (C) IFN-c is considered to mediate resistance to PCM [28], and in this experiments high levels of this cytokine were observed only at 2 h and 4 wks p.i. at the same time with occurrence of the pro-inflammatory burst. Later on, however, IFN-c decreased to lower levels than those in controls. This change towards IFN-c down-regulation together with low IL-2 levels has been also regularly observed in patients with chronic PCM [29]. (D) For the first 4 wks p.i., pro-inflammatory cytokines, chemokines and growth factors co-existed with high concentrations of immune-regulators such as IL-10 and IL-12p40. IL-10 hinder full activation of monocytes/ macrophages and T cells decreasing their fungicidal activity [30]. IL-12p40 is a competitive antagonist of IL-12p70 and also acts as a chemoattractant for macrophages [31]. As pro and anti-inflammatory stimuli co-existed, it was not possible to determine the predominant cytokine effect. (E) MIG (CXCL9), RANTES (CCL5) and IL-12p40 were upregulated at all times during the experimental period, including most chronic times (8, 12 and 16 wks) when the remaining cytokines decreased. The molecules referred to above could be considered as markers when discriminating PCM cases from controls and could also be useful when treatment follow-up is considered. Indeed, in a study with human PCM patients, serum levels of MIG/CXCL9 were higher in all untreated cases and decreased progressively with treatment [32]. In addition, MIG, RANTES and IL12p40 appeared to be of help in maintaining cellular recruitment during the chronic periods of the Pb-infection when the remaining chemokines showed low levels. MIG is a chemotactic factor selective for lymphocytes which exerts its greatest influence on activated CD4+ Th1 cells, CD8+ T cells and NK cells [33]. RANTES is a potent chemokine for monocytes, macrophages, T cells, dendritic cells, eosinophils, and basophils [34]. IL12p40 also acts as a chemokine for macrophages [31].
In the course of these studies we have noticed certain histopathological findings that due to their novelty deserve further exploration, as discussed below: (A) Foamy cells were frequent in the granuloma central areas, as were Touton cells and large cholesterol crystals suggesting participation of lipids in Pb-infection. Foamy macrophages have been described in individuals developing secondary or adult type tuberculosis [35], in Mycobacterium avium-infected AIDS patients, in chronic stages of M. tuberculosis infection in mice [36,37] and in Chlamydia infections [38]. However, their presence or significance has not been often discussed in PCM. (B) Another peculiar feature in experimental PCM was the involvement of the periaterial space. Although this compartment has been regularly neglected, its involvement has been reported in several animal models, such as respiratory syncycial virus, bacteria including M. tuberculosis, in some asthma models and in lipopolysaccharide-induced inflammation [39]. The mechanisms and route of recruitment of inflammatory cells into the perivascular space (PVS) remain largely unknown. An important feature in the peri-arterial sheath inflammation was the higher frequency of plasmocytes and eosinophils in comparison to those existing in nodules and granulomas. These cells are associated with a Th2 cytokine profile, as well as with antibody production [40,41]. The participation of B-lymphocytes in paracoccidioidomycosis has not been completely elucidated; however, B-1 cells contribution to susceptibility in experimental disease has been reported [42,43]. In addition, activation of B-cells is associated with a severe disease outcome and is indicative of a progressive disease [44]. Mott cells containing Russell bodies indicate plasma cells indigestion due to a failure to eliminate misfolded or incorrectly assembled proteins [45,46]. (C) We reported pseudotumoral masses in the lungs of Pbinfected mice. Pulmonary masses have been recorded in patients in whom chest x-rays revealed dense large solitary masses with well-defined borders sometimes referred as paracoccidioidomas [47][48][49]. The preference of this lesion for the hilar zone could probably be explained by the characteristics of the pulmonary lymph flow which proceeds in opposite directions: lymphatics in the lung perimeter flow to the pleura and the rest flows towards the hilum [50].
These overall results complement previous work by our group [51], where noninvasive conventional medical X-ray tomography allowed the in vivo follow-up of the three main pathological patterns described above.
In conclusion, this study has revealed different patterns of pulmonary lesions according to time post-infection, as well as