Histopathologic and immunologic effects of the itraconazole treatment in a murine model of chronic pulmonary paracoccidioidomycosis

Original article

Histopathologic and immunologic effects of the itraconazole treatmentin a murine model of chronic pulmonary paracoccidioidomycosis

Tonny W. Naranjo a,b,*,1, Damaris E. Lopera a,1, Lucy R. Diaz-Granados b, Jhon J. Duque c,Angela Restrepo a, Luz E. Cano a,d

aMedical and Experimental Mycology Group, Corporacion para Investigaciones Biologicas (CIB), Medellın, Colombia

b School of Health Sciences, University Pontificia Bolivariana, Medellın, Colombiac School of Medicine, University of Antioquia, Medellın, Colombia

dSchool of Microbiology, University of Antioquia, Medellın, Colombia

Received 26 March 2010; accepted 25 July 2010

Available online 5 August 2010

Abstract

A comparative study, based on histopathologic findings (inflammation, cellularity, and fibrosis) and immunologic parameters (pro-

inflammatory and anti-inflammatory cytokines), was carried out in order to evaluate the effects of itraconazole (ITC) treatment and its starting

time in a BALB/c murine model of chronic pulmonary paracoccidioidomycosis (PCM), induced by intranasal inoculation of Paracoccidioides

brasiliensis (Pb) conidia. Two different groups of mice were exposed to ITC therapy beginning at the 4th or 8th week after Pb infection,

respectively. ITC was administered daily, via gavage, for a period of sixty days. At weeks 0, 4, 8, 12 and 16 the animals were sacrificed and their

lungs removed for histology staining with hematoxylin and eosin (H&E), Masson’s trichromic and GomorieGrocott; pulmonary levels of IL-1b,

TNF-a, IFN-g, IL-13 and TGF-b were also measured by ELISA. The development or absence of the principal pulmonary PCM sequela, lung

fibrosis, was directly related to the therapy’s starting time. This and other histopathologic findings were related to the behavior of cytokine levels.

! 2010 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved.

Keywords: Paracoccidioidomycosis; Cytokines; Itraconazole; Lung fibrosis

1. Introduction

Invasive fungal infections are one of the major causes of

morbidity and mortality, representing a serious and growing

public health problem, not only for immunocompetent but also

for immunocompromised patients, including those treated

with corticosteroids, receiving organ transplants, or diagnosed

with HIV [1e3].

Paracoccidioidomycosis (PCM) is one of the most impor-

tant and prevalent human endemic and systemic fungal

diseases in Latin America, mainly in Brazil, Colombia and

Venezuela [4], where an estimated 10 million people are

infected with the fungus. PCM is a progressive and chronic

disorder initiated by the inhalation of infectious airborne

propagules (conidia), which are produced by the mycelial

form of the thermo-dimorphic fungus Paracoccidioides

brasiliensis. These propagules change into the pathogenic

yeast form when they reach body temperature [2,4].

PCM has three major clinical forms, an asymptomatic form

(infection) observed in healthy individuals infected with Pb,

and two progressive disease forms, the acute or sub-acute

juvenile type, and the chronic or adult type. The chronic form

represents 90% of all cases and is observed mostly in adult

males in whom the disease may take months or years to fully

develop [5,6].

The primary infection takes place in the lungs and subse-

quently disseminates to other organs and tissues, where it is

recognized by the appearance of secondary lesions in mucous

* Corresponding author. Corporacion para Investigaciones Biologicas

(CIB), Carrera 72a No. 78B-141, Medellın, Colombia. Tel.: þ57 4 4410855;

fax: þ57 4 4415514.

E-mail address: [email protected] (T.W. Naranjo).1 These authors contributed equally to this work.

1286-4579/$ - see front matter ! 2010 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved.

doi:10.1016/j.micinf.2010.07.013

Microbes and Infection 12 (2010) 1153e1162www.elsevier.com/locate/micinf

membranes, skin, lymph nodes, adrenal glands, and other

tissues [2,5,6]. In the lungs, Pb causes chronic parenchymal

damage leading to fibrosis, which severely restricts respiratory

function; such pathology is observed in as many as 60% of the

patients with this disease [2,7].

A number of studies have shown that the acute or juvenile

form of the disease correlates highly with a Th2 type immune

response, with high levels of cytokines IL-10, IL-4 and IL-5,

as well as peripheral blood eosinophilia and low levels of Th1

cytokines. In patients with the chronic adult form, there is

a similar tendency towards diminishing Th1 immune response,

but without polarization towards a Th2 immune response

[8e12].

Soares et al. [8] showed that in patients with the chronic

form, lung lesions had high levels of both TNF-a and TGF-b;

after 20 days of antifungal treatment, TNF-a levels decreased

while those of TGF-b remained high, and these findings

appeared to be associated with the presence of fibrotic areas in

the lungs. It was also found that in healthy infected controls,

the levels of some Th1 cytokines were high, including IL-12,

IL-2, TNF-a and IFN-g [8,9].

It is well known that cytokine profiles strongly influence the

functioning of cells of the monocyteemacrophage lineage,

which are the first to interact with both inhaled conidia and

recently transformed yeast cells [10,12e15].

As it is impossible to detect the precise moment when the

infection occurs in humans, because the fungal niche is

unknown [4], our group developed a murine model of PCM

two decades ago. In this model, BALB/c mice are inoculated

intranasally with Pb conidia, imitating the natural occurrence

in humans, with the objective to know and understand the total

course of the developing disease [7,16e18].

Using this experimental model, we observed that during

the early stages of infection, i.e., the first 72 h, an acute

inflammatory process takes place, involving mainly poly-

morphonuclear (PMN) cells and macrophages. A decrease in

PMN cells starts in the following weeks, preceding the

formation of granulomas composed of macrophages, epithe-

lioid cells, multinucleated giant cells, and lymphocytes, all of

which surround the yeast cells of the fungus. At the end of the

granulomatous process, the accumulation of connective tissue

and the production of collagen lead to the establishment of

fibrosis, generating structural and functional alterations to the

lungs [7,18,19].

In clinical practice, systemic fungal infections are

commonly treated with azolic derivates such as ITC, which is

frequently used and constitutes the treatment of choice

for PCM [20,21]. Its mechanism of action inhibits the fungal

C-14a demethylase, a compound that is required for the

biosynthesis of ergosterol, an essential component of the

fungal cell wall membrane [22,23]. ITC effectively controls

active PCM by restricting fungal proliferation. However,

Tobon et al. [24] showed that over 30% of PCM patients

treated with ITC had lung fibrosis at diagnosis and had not

cleared at the end of the treatment period; during a follow-up

done after therapy, the fibrosis even developed de novo in

some patients [24].

So far, there are only few published studies using in vivo

experimental models to evaluate ITC therapy, and most of

them focused mainly on the antifungal effect rather than the

fibrotic process. Furthermore, in those studies the experiments

were performed using inoculum of yeast cells, which are not

the naturally infecting P. brasiliensis particles [25e28].

In order to better understand the influence of azolic treat-

ment during the establishment of pulmonary PCM and the

development of lung fibrosis, we evaluated the effect of ITC

treatment on immune responses (pro- and anti-inflammatory

cytokines) and pulmonary histopathology (inflammation,

cellularity, and fibrosis) in our animal model, starting at two

different post-infection times.

2. Materials and methods

2.1. Animals

A total of 250 isogenic BALB/c mice, 6e7 weeks old and

18e20 g in weight, were obtained from the Corporacion para

Investigaciones Biologicas (CIB) breeding colony and used in

all experiments. Mice were kept under controlled environ-

mental conditions, temperatures of 24 "C and 12-h light/dark

cycle; they were supplied, ad libitum, with sterile food and

acidified water (pH 2.5e3.0) in sterilized bottles, as well as

clean bedding. The Animal Experimentation Ethics

Committee of our institution approved and verified the

fulfillment of the policies concerning handling and caring of

animals.

2.2. Trial design

The mice were divided into 5 groups of 50 animals to be

sacrificed at 0, 4, 8, 12 and 16 weeks post-challenge respec-

tively. Each group was subdivided by treatment options, as

follows: 10 negative controls, 10 P. brasiliensis infected mice

(positive controls), 10 P. brasiliensis infected mice treated

with ITC starting at the 4th week post-infection, 10 P. brasi-

liensis infected mice treated with ITC starting at the 8th week

post-infection and 10 uninfected mice treated with ITC. Each

of these subgroups was divided equally (5 þ 5 animals) for

immunologic and histopathologic analysis.

2.3. Fungal culture, conidia production and inoculum

preparation

P. brasiliensis American Type Culture Collection isolate

60855, known to produce abundant conidia, was used in all

experiments [29]. For conidia production, the mycelial phase

of the fungus was grown on water-agar plates at 18 "C for 8

weeks. For each experiment, the infectious conidia were

collected as follows: The plates on which P. brasiliensis

cultures had been grown in the mycelial form were flooded

with physiologic saline containing 0.85% Tween 20. The

growth was scraped from the agar surface with a gauged loop.

The suspension was transferred to a capped Erlenmeyer flask

containing three layers of 6 mm glass beads. The flask was

1154 T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

shaken for 30 min at 250 rpm to free the conidia. The fungal

slurry was poured through a sterile syringe packed with glass

wool. The conidial suspension that passed through the glass

wool was concentrated by centrifugation and the supernatant

was decanted. The sediment, which was rich in conidia, was

suspended in 50 ml of physiologic saline and shaken vigor-

ously to suspend the propagules. The quantity and viability

were determined using ethidium bromide staining. The

viability of the conidia was consistently higher than 95% of

the total of conidia counted. The inoculum was adjusted so

that 60 ml contained approximately 3 # 106 viable conidia.

2.4. Experimental infection

Animals were anesthetized by intramuscular injection

(50 ml) of a solution of ketamine 100 mg and xylazine 20 mg.

When deep anesthesia was obtained, 3 # 106 conidia sus-

pended in a 60 ml of the inoculum, divided in two portions,

were instilled intranasally during a 10 min period. Control and

uninfected treated mice received an intranasal inoculum of

60 ml of PBS. In order to determine any changes in collagen

and cytokines during aging, each group of infected mice was

studied together with age-matched uninfected controls that

had been housed under the same conditions as the infected

mice.

2.5. Antifungal and dosing

Itraconazole oral solution was obtained from Ortho-

McNeil-Janssen Pharmaceuticals. The drug dose was 1 mg/

day/mouse. This dose was required to achieve serum levels of

1 mg/ml lower serum levels have been associated with anti-

fungal therapy failures [30]. Itraconazole treatments (100 ml at

10 mg/ml) were given daily for 8 weeks by gavage, and were

started at either 4 or 8 weeks post-infection.

2.6. Sacrifice and sample processing

At the scheduled periods, 10 animals from each treatment

group were sacrificed by intraperitoneal injection of 1.0 ml of

2.5% sodium pentothal (Sandoz Laboratories, Kundl, Austria)

[24]. Upon death, the thoracic cavity was opened, and the

whole lungs were removed and partitioned for histopathologic

(5 mice) and for immunologic studies (another 5 mice).

2.6.1. Cytokine assays

After the lungs were removed, they were individually

homogenized using a tissue grinder (Tissue Tearor, model

985-370, Biospec Products) in 2 ml of a cocktail solution

containing protease inhibitors (pepsin 0.1 mM, leupeptin

0.1 mM, phenylmethyl sulfonide fluoride 1 mM, N-tosyl-L-

phenylalanine chloromethyl ketone 0.2 mM, (a)-p-methyl-L-

lysine chloromethyl ketone 0.1 mM from Sigma chemical,

plus ethylene diamine tetra-acetic acid (EDTA) 1 nM from

Merck Germany). The homogenates were kept in ice and then

centrifuged at 3000 rpm during 15 min at 4 "C, keeping the

supernatants frozen at $70 "C until assayed. The lung

homogenate’s supernatants were thawed immediately prior to

processing with commercial mouse cytokine ELISA assays

(mouse IL-13 and mouse IL-1b, Biosource California USA,

mouse IFN-g and mouse TNF, OptEIA Set Pharmingen, San

Diego, CA, USA, and TGF-b1 Immunoassay kit, Promega) in

order to measure the pulmonary levels of these cytokines. The

assays were performed according to the manufacturers’

instructions. Samples were run in duplicate and the results

were read at 450 nm in an automated microplate reader (Bio-

Rad model 680XR, Bio-Rad Laboratories, Australia).

2.6.2. Histopathology

Before removing the lungs, 10 ml of PBS and 10 ml of 10%

phosphate-buffered formalin were injected directly in the heart

to insufflate the lungs and remove any remaining blood. The

lungs were fixed in buffered formalin and embedded in

paraffin to be sectioned at 5 mm thickness; the sections were

stained with Grocott’s methenamineesilver nitrate (Grocott)

for identification of fungi, hematoxylin and eosin (H&E) to

characterize the inflammatory response, and Gomori’s retic-

ulin stained sections were used to determine the changes

occurring in the organization of reticulin fibers (collagen III),

both thick and thin; Masson’s trichrome stained sections were

used to identify collagen I, thick and thin fibers. Two pathol-

ogists who were blinded to the experimental conditions

examined the stained sections independently.

2.7. Criteria used for grading lung granulomatous

inflammation, fibrosis and fungal load

To determine the severity of lung involvement in granulo-

matous inflammation and fibrosis, we used the same parame-

ters previously reported by Franco et al. [17] and Cock et al.

[18], with slight modifications. The inflammatory reaction was

reported as percentage of the lung occupied by granulomatous

inflammation. The appearance (thin or thick) of collagen and

reticulin fibers was recorded and assigned a score according to

their relative increase compared to control mice: 0: no

apparent increase, 1: mild increase, 2: moderate increase and

3: strong increase. In order to evaluate the fungal load, five

Grocott’s stained slides for each experimental group were

scanned, and the Aperio positive pixel count algorithm was

used to quantify the amount of Grocott pixels present on each

scanned image. The results were expressed as the sum of

strong, medium and weak Grocott pixels in the total pulmo-

nary area, measured using ImageScope software.

2.8. Statistical analysis

Statistical analyses between groups were performed by

the Student t test and time changes were analyzed by

ANOVA, using Prism version 5.0 (Graph-Pad, USA). Results

were considered significant when p < 0.05. To evaluate the

pathologists’ agreement, the Kappa coefficient of agreement

was used.

1155T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

3. Results

3.1. Effect of itraconazole therapy on fungal load in lung

In P. brasiliensis infected mice, the number of Grocott

positive pixels (GPP) was measured; at the 4th week post-

infection these mice had an average of 1 # 106 GPP, and the

value increased to 7.2 # 106, 2.6 # 106 and 3.4 # 106 GPP for

the 8th, 12th and 16th week post-infection, respectively

(Fig. 1). Although the number of GPP began to decrease once

the antifungal treatment was started, a rapid reduction was

observed 4 weeks after treatment, but only in those mice

treated at 4 weeks post-infection (Fig. 1A). By contrast, when

the treatment was started at 8 weeks post-infection, twice the

time (8 weeks) was needed to observe a significant reduction

in the number of GPP (Fig. 1B).

3.2. Effect of itraconazole therapy on pulmonary

histopathology

3.2.1. Inflammatory process

At 4 weeks post-infection period, granulomatous inflam-

mation was present in all infected mice, with an average of

10% of their lung parenchyma affected. At weeks 8, 12, and

16 post-infection, the percentages of affected lung paren-

chyma were 22.5%, 20.3%, and 15%, respectively.

In the mice that received ITC treatment at 4 or 8 weeks post-

challenge, a clear reduction in the pulmonary area with

inflammatory reaction was observed immediately after the

treatment started. This decrease was more pronounced (and

statistically significant, p< 0.05) inmice beginning treatment at

the 4th week after infection than in those starting treatment at

the 8th week post-infection (Fig. 2). Controls and uninfected

ITC-treated mice did not show histopathologic pulmonary

abnormalities.

3.2.2. Cellularity

During the granulomatous inflammatory period, the cellular

infiltrate was composed mainly of mononuclear and PMN

cells. In the absence of ITC treatment, infected mice had an

average of 40% PMN and 60% mononuclear cells during the

infectious process (Fig. 3A). In mice receiving treatment,

these cellular proportions changed depending on the time of

therapy initiation. For mice that started treatment at the 4th

week post-infection, the proportion of PMN cells diminished

and became statistically significant ( p < 0.01) at 8 weeks after

Fig. 1. Effects of itraconazole on fungal load in the lung. BALB/c mice were infected intranasally with 3 # 106 P. brasiliensis conidia, treated with ITC starting at

the 4th or 8th week post-infection (PI) and the amount of Grocott positive pixels (GPP) as marker of fungal load was measured. Filled bars represent infected

untreated mice and open bars represent infected ITC-treated mice. A: Infected mice starting ITC therapy at 4th week. B: Infected mice starting ITC therapy at 8th

week PI. C: Grocott stained lung section (10#) from an infected untreated mouse. D: Representative Grocott stained lung section (10#) from infected mouse

treated with ITC. The results are expressed as mean % SEM of GPP. Five Grocott stained slides for each experimental group were scanned and the Aperio positive

pixel count algorithm in free ImageScope software was used to measure the GPP. Symbols * and *** denote significant difference between infected untreated mice

and infected ITC-treated mice ( p < 0.05 and p < 0.001 respectively). ND, no significant difference.

1156 T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

therapy initiation. In mice that began their treatment at the 8th

week post-infection, the reduction to essentially the same

proportion of PMN cells took only 4 weeks (Fig. 3A). For

mononuclear cells, therewas a slight increase in their proportion

only when the ITC treatment was started at the 8th week of

infection (Fig. 3B).

In uninfected ITC-treated mice, no differences in lung

cellularity were observed with respect to negative control mice

(data not shown).

3.2.3. Development of lung fibrosis

Thin fibers of collagen and reticulin became visible after the

4th week following conidial inoculation and increased over

time, suggesting the beginning of a fibrotic process. Thick fibers

of both proteins, which are indicators of established fibrosis,

became evident only after the 8th week post-infection (Fig. 4A).

In healthy as well as in uninfected ITC-treated mice, collagen

and reticulin fibers, both thin and thick, showed no histopatho-

logic changes during the period of observation (Fig. 4B).

Fig. 2. Effects of itraconazole on the pulmonary inflammatory response. A: BALB/c mice were infected intranasally with 3 # 106 P. brasiliensis conidia and treated

with ITC starting at the 4th or 8th week post-infection (PI). Symbols * and # denote significant difference ( p < 0.05) between infected untreated mice and infected

ITC-treated mice starting therapy at the 4th and 8th week PI, respectively. B: Control group, mice that received PBS. C: Infected untreated mice (8 weeks of

infection). D: Infected mice (8 weeks of infection) treated with ITC from the 4th week PI. Results are expressed as mean % SEM (n ¼ 5 animals per group). Lung

sections (10#) were stained with H&E.

Fig. 3. Effects of itraconazole on the cellularity of the pulmonary inflammatory response. Mice were infected intranasally with 3 # 106 P. brasiliensis conidia and

treated with ITC starting at the 4th or 8th week post-infection (PI). Symbols * and # denote significant difference ( p < 0.05) between infected untreated mice and

infected ITC-treated mice starting therapy at the 4th week and 8th week PI, respectively. A: Percent of PMN cells. B: Percent of mononuclear cells. Results are

expressed as mean % SEM (n ¼ 5 animals per group). Two pathologists who were blinded to the experimental conditions examined independently one sagittal

frame of total lung from each animal; they showed a Kappa coefficient of 0.59.

1157T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

When infected mice were treated with itraconazole, the thin

reticulin fibers were reduced from 8 weeks after initiation of

treatment, regardless of the treatment starting time (Fig. 5A).

In the case of thin collagen fibers, their reduction occurred

rapidly, in 4 weeks, when the ITC treatment started at the 4th

week post-infection; when the treatment was initiated at the

8th week post-infection, the reduction of these fibers took 8

weeks to be noticed (Fig. 5B).

As for thick fibers of both proteins, collagen and reticulin,

these fibers were reduced only when treatment was started

early at the 4th week post-infection (Fig. 5C and D).

3.3. Pulmonary cytokines in chronic murine PCM

3.3.1. Interleukin 1 beta (IL-1b)

Lung homogenates from infected mice revealed a significant

increase in levels of IL-1b ( p < 0.05) soon (2 h) after intranasal

inoculation as compared to control mice; these levels remained

significantly high ( p < 0.001) until week 8, returning to normal

levels at week 12 post-infection. In the case of ITC treatment in

infectedmice, themedication led to a significant decrease in IL-1b

level ( p< 0.05) when the therapy started early (4thweek), but not

when the therapy started at the 8th week post-infection (Fig. 6A).

3.3.2. Tumor Necrosis Factor alpha (TNF-a)

Similar to IL-1b, TNF-awas found to be elevated early after

2 h of infection with Pb conidia; the corresponding levels were

significantly high ( p < 0.001) throughout all the periods eval-

uated (4, 8, 12 and 16 weeks) in comparison to uninfected mice.

In infected mice that were treated with itraconazole at either the

4th or 8th week post-infection, TNF-a levels were reduced to

levels similar to those found in control mice (Fig. 6B).

3.3.3. Interleukin 13 (IL-13)

In infected mice, IL-13 lung levels were significantly

higher ( p < 0.005) at 2 h, 4 and 8 weeks compared to control

mice; at 12 and 16 weeks, levels were similar to those

observed in uninfected mice. When ITC therapy was given to

infected mice at the 4th week, IL-13 reached normal levels.

When the therapy was given at the 8th week post-infection,

levels of IL-13 were reduced even below normal levels of

control mice (Fig. 6C).

3.3.4. Transforming Growth Factor beta (TGF-b)

TGF-b levels of infected mice showed a rather variable

behavior compared to the control group; TGF-b had

a tendency to go over the normal levels throughout the

observation period (significantly higher at the 8th and 16th

week post-infection, p < 0.01). During ITC treatment, TGF-b

levels in infected mice depended on the therapy starting time:

when the treatment started at the 4th week post-infection,

TGF-b levels tended to decrease, whereas when ITC therapy

began at the 8th week post-infection, they tended to increase

(Fig. 6D).

Fig. 4. Chronic pulmonary paracoccidioidomycosis. BALB/c mice were infected intranasally with 3 # 106 P. brasiliensis conidia and the development of lung

fibrosis was evaluated at 4, 8, 12 and 16 weeks post-infection. A: To reticulin and collagen fibers were assigned a score quantifying their relative increase compared

to uninfected mice, which do not exhibit histopathologic changes (scores ¼ 0.0). B: Lung slide from healthy control mouse showing no changes in collagen I fibers;

similar morphology was found in uninfected ITC-treated mice. C: Increment of thin and thick collagen III (reticulin) fibers forming bands around granulomas.

D: Increment in thin and thick collagen I fibers. Results in (A) are expressed as mean % SEM (n ¼ 5 animals per group). B and D show Masson’s trichrome stained

lung sections; C shows Gomori’s stained lung section (10#).

1158 T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

3.3.5. Interferon gamma (IFN-g)

During the observation period, there were no differences in

the relative level of IFN-g in Pb infected mice compared to the

control group.

When the ITC treatment started at week 8, the IFN-g levels

in infected mice were reduced significantly even below those

in control mice ( p < 0.05) (Fig. 6E).

4. Discussion

Our results demonstrate, for the first time, the effect of ITC-

antifungal therapy started at two different times post-infection

in a chronic pulmonary paracoccidioidomycosis model induced

by intranasal inoculation of Pb conidia, reproducing the normal

route for human infections. Our study strongly suggests

relations, but does not yet demonstrate direct causalities,

among fungal load, histopathologic findings (inflammation,

cellularity and fibrosis) and the immune responses (pro-

inflammatory and anti-inflammatory cytokines). As expected,

the therapy with itraconazole significantly reduced the fungal

burden present in infected mice (Fig. 1). ITC therapy led to the

following changes, at both start times: i) a significant decrease

in the pulmonary inflammatory response compared with that

observed in untreated infected mice; ii) a change in the cellular

infiltrate composition, with reduction of the proportion of

PMN cells; and iii) a strong reduction of the levels of

pro-inflammatory cytokines (IL-1b and TNF-a). In contrast,

changes related to the fibrotic process depended on the therapy

start time: a reduction in pro-fibrotic cytokine levels (IL-13

and TGF-b), followed by an evident decrease in the fibrotic

sequela, was observed only when the therapy was started

promptly.

Pulmonary PCM is a chronic granulomatous fungal infec-

tion that affects mainly the lungs and is influenced by different

factors such as cellular infiltration, cytokine production,

granuloma formation and, ultimately, development of pulmo-

nary fibrosis [2,7,18,19].

In this study the granulomatous inflammation was observed

after 4 weeks post-infection, when it involved 10% of the lung

parenchyma. It progressed to cover 22.5%, 20.3% and 15% by

weeks 8, 12 and 16, respectively (Fig. 2). In infected mice that

started ITC treatment at the 4th week post-infection, the

proportion of pulmonary granulomatous inflammation was

significantly less than in untreated controls as early as 4 weeks

after start of treatment. However, when the antifungal therapy

was started later (8th week post-infection), a significant

reduction was observed only after a longer period of treatment

(8 weeks), probably because the inflammatory response to

infection was already well established and the fibrotic process

was at a more advanced stage when the therapy began [7,18]

(Fig. 2). In agreement with these findings, when the ITC

treatment was started at the 4th week post-infection, we

Fig. 5. Effects of itraconazole on pulmonary histopathology of BALB/c mice intranasally inoculated with 3 # 106 P. brasiliensis conidia in relation to the treatment

starting time, 4th or 8th week post-infection (PI). Symbols * and # denote significant difference ( p < 0.05) between infected untreated mice and infected ITC-

treated mice starting therapy at the 4th week and 8th week PI, respectively. A: Thin reticulin fibers. B: Thin collagen fibers. C: Thick reticulin fibers. D: Thick

collagen fibers. Results are expressed as mean % SEM (n ¼ 5 animals per group).

1159T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

observed that it caused a prompt diminishing of the relative

number of Grocott positive pixels (GPP), a marker of fungal

load (Fig. 1A). In the case of the treatment started at the 8th

week post-infection, the GPP took more time to diminish

(Fig. 1B).

The histopathologic observations of the pulmonary

inflammatory response after itraconazole treatment corre-

sponded to the decreased fungal load and the observed cyto-

kine levels. It has been shown that the formation of

paracoccidioidal granuloma depends on fungal cell wall

composition, as the fungal components such as chitin, b-1,3

and b-1,6 glucans can modulate the immune response, influ-

encing in this way the extent and morphology of tissue lesions

[18,31,32].

Additionally, it is known that in the early stages of

P. brasiliensis infection (12e96 h) the host’s cellular infiltrates

are composed mainly of PMN cells (over 90% of the inflam-

matory infiltrate); and that the proportion of those cells

diminishes gradually during granuloma formation [7]. In

agreement with this, we observed that in the absence of anti-

fungal treatment, at the 4th week of infection, the cellular

infiltrate consisted of 40% PMN cells and 60% mononuclear

cells, and essentially the same proportions were observed also

at 8, 12 and 16 weeks post-infection (Fig. 3). From the 4th

week of infection, when the granuloma is already forming,

mononuclear cells play a critical role in host defense [19,33]:

they influence several mechanisms of resistance against fungi

[34,35], and, very important, they are responsible for forming

and maintaining the granuloma structure in an attempt to avoid

fungal dissemination [36,37].

In this study, mice treated with ITC (both start times)

showed a drastic change in their pulmonary cell profile, with

marked increase in the number of mononuclear cells (75%)

and concomitant reduction in PMN count (15%) (Fig. 3). In

mice infected with P. brasiliensis, PMN cells are typically

present in larger quantities during the first week of infection;

since PMN cells are considered the first line of defense against

fungi, it is possible that after itraconazole treatment it appears

that they are not so involved in the granuloma formation. It has

been suggested that in BALB/c mice, the mechanisms of

defense against P. brasiliensis infection proceed in two steps:

first, phagocytosis by PMN cells and second, intervention of

cell-mediated immunity and granuloma formation [19].

Granuloma formation depends on the synthesis of several

cytokines that mediate the local immune response by phago-

cytic cells, mainly IL-1b and TNF-a [17,38,39]. In our study

we observed high levels of these pro-inflammatory cytokines

in lung homogenates at the 4th, 8th, 12th and 16th week post-

infection and in agreement with the fungal load observed at

these times. Although IL-1b decreased to normal levels at the

12th week post-infection, TNF-a levels remained elevated

during all periods evaluated, coinciding with a sustained

granulomatous inflammation; the role of TNF-a in maintain-

ing granulomas has been widely recognized in different

diseases [40,41]. Other cytokines such as IFN-g, IL-13 and

TGF-b were also evaluated (Fig. 6). IFN-g, which activates

macrophages, is considered crucial in host defense against

various infectious disease agents [42e44]; nonetheless, no

changes in levels of this cytokine were observed in the

infected mice. In accordance with the findings presented here,

Uran et al. found by RT-PCR that during the chronic stages in

the same PCM model, IFN-g mRNA levels were not different

from those in control mice (unpublished data; M. Uran,

personal communication).

Regarding IL-13 and TGF-b, although these cytokines are

known as anti-inflammatory mediators, their levels tended to

Fig. 6. Effects of itraconazole on lung cytokine levels of BALB/c mice intranasally inoculated with 3 # 106 P. brasiliensis conidia in relation to the treatment

starting time, 4th or 8th week post-infection (PI). A: Interleukin 1b levels. B: Tumor Necrosis Factor alpha levels. C: Interleukin 13 levels. D: Transforming Growth

Factor beta levels. E: Interferon gamma levels. Symbols #, * and f denote significant difference ( p < 0.05) between infected untreated mice and control ($) mice,

infected ITC-treated mice starting at the 4th week PI, and infected ITC-treated mice starting at the 8th week PI, respectively. Results are expressed as mean % SEM

(n ¼ 5 animals per group).

1160 T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

increase during the P. brasiliensis infection. There are two

possible explanations for this tendency. First, it is probable

that these cytokines play a regulatory role during inflamma-

tion, and second, they are essential in the genesis of fibrosis, as

they are the major pro-fibrotic cytokines, due to their capacity

to induce both the fibroblast proliferation and collagen

synthesis [45e47]. The above processes could also explain the

appearance of fibrosis and the concomitant deterioration of

lung function, such as is seen in human with chronic pulmo-

nary PCM [24].

Regarding pro-inflammatory cytokines, measurement of

IL-1b and TNF-a showed a substantial reduction in their

levels immediately after itraconazole treatment started; in fact,

they declined to the levels found in control mice (Fig. 6).

These changes are in agreement with the decrease in fungal

load observed in Fig. 1, and with the histopathologic results

presented in Fig. 2, where the percentage of the lung exhib-

iting granulomatous inflammation decreased when ITC

therapy was started.

Concerning the fibrotic process, the histopathologic anal-

yses from infected untreated mice showed a gradual but clear

increase in reticulin and collagen deposition, progressing from

thin to thick fibers in both proteins as the pulmonary fibrotic

sequela was established (Fig. 4). When the itraconazole

treatment was started at the 4th week post-infection, the fungal

load was rapidly reduced and the collagen and reticulin thin

fibers showed a marked and faster tendency to decrease; also,

thick collagen fibers were never formed, in contrast to the

observations made when ITC therapy was started at the 8th

week post-infection (Fig. 5AeD).

In view of the above findings and in order to establish

a correlation with the immune response, TGF-b was moni-

tored. This cytokine, which is considered to be the principal

pro-fibrotic cytokine [48,49], was found to be reduced to

normal levels only when the ITC treatment was started

“early”, at the 4th week after infection. When ITC therapy was

started “later”, at the 8th week post-infection, TGF-b levels

increased significantly and INF-g levels diminished signifi-

cantly, below the levels observed in control mice (Fig. 6). It is

important to clarify that in the context of the fibrotic process,

INF-g is considered as one of the few molecules with anti-

fibrotic activity. These results could be explained by the

observation that thick collagen fibers were already present

when antifungal therapy was started, as is also the case in

humans [24,50].

Although at present several medications can be used for the

treatment of paracoccidioidomycosis, the principal sequela of

this disease in humans, pulmonary fibrosis, persists even in the

presence of well-conducted antifungal therapy. In the study

presented here, we demonstrated the importance of starting

itraconazole treatment early in an animal model. In the normal

course of lung PCM in human, it is difficult to detect the

precise moment of P. brasiliensis infection, and for this reason

it is extremely important to implement antifungal treatment as

soon as possible, or even consider using combined therapies

designed not only to eliminate the fungus but also to prevent

the development of fibrosis.

Acknowledgments

This work was supported by COLCIENCIAS (project No.

2213-04-16439), Bogota, Colombia, by the Corporacion para

Investigaciones Biologicas (CIB) and by the University of

Antioquia (UdeA), Medellın, Colombia.

References

[1] M.D. Richardson, Changing patterns and trends in systemic fungal

infections, J. Antimicrob. Chemother. 56 (2005) 5e11.

[2] E. Brummer, E. Castaneda, A. Restrepo, Paracoccidioidomycosis: an

update, Clin. Microbiol. Rev. 6 (1993) 89e117.

[3] S.M. Caramori, R. Teixeira, J. Afonso Jr., R. Carraro, T. Strabelli,

M. Samano, P. Pego-Fernandes, F. Jatene, Bacterial and fungal pneu-

monias after lung transplantation, Transplant. Proc. 40 (2008) 822e824.

[4] D.R. Matute, L.M. Quesada-Ocampo, J.T. Rauscher, J.G. McEwen,

Evidence for positive selection in putative virulence factors within the

Paracoccidioides brasiliensis species complex, PLoS Negl. Trop. Dis. 17

(2008) 296.

[5] M.I. Borges-Walmsley, D. Chen, X. Shu, A.R. Walmsley, The pathobi-

ology of Paracoccidioides brasiliensis, Trends Microbiol. 10 (2002)

80e87.

[6] M.M. Franco, R. Montenegro, R.P. Mendes, S.A. Marques, N.L. Dillon,

N.G. Mota, Paracoccidioidomycosis: a recently proposed classification of

its clinical forms, Rev. Soc. Bras. Med. Trop. 20 (1987) 129e132.

[7] S. Restrepo, A. Tobon, J. Trujillo, A. Restrepo, Development of

pulmonary fibrosis in mice during infection with Paracoccidioides

brasiliensis conidia, J. Med. Vet. Mycol. 30 (1992) 173e184.

[8] M.R. Parise-Fortes, S.A. Marques, A.M. Soares, C.S. Kurukawa,

M.E. Marques, M.T. Peracoli, Cytokines released from blood monocytes

and expressed in mucocutaneous lesions of patients with para-

coccidioidomycosis evaluated before and during trimethoprime

sulfamethoxazole treatment, Br. J. Dermatol. 154 (2006) 643e650.

[9] S.J. Oliveira, R.L.Mamoni, C.C.Musatti, P.M. Papaiordanou,M.H.Blotta,

Cytokines and lymphocyte proliferation in juvenile and adult forms of

paracoccidioidomycosis. Comparison with infected and non-infected

controls, Microbes Infect. 4 (2002) 139e144.

[10] S.A. Calvi, A.M. Soares, M.T. Peracoli, R.P. Mendes, Effect of cytokines

on the in vitro fungicidal activity of monocytes from para-

coccidioidomycosis patients, Microbes Infect. 5 (2003) 107e113.

[11] L.M. Mello, M.L.S. Vergara, J.V. Rodrigues, Patients with active infec-

tion with Paracoccidioides brasiliensis immune response characterized

by high interleukin 4 and interleukin 5 production, Hum. Immunol. 63

(2002) 149e154.

[12] S.A. Karhawi, A.L. Colombo, R. Salomano, Production of IFN-gamma is

impaired in patients with paracoccidioidomycosis during active disease

and is restored after clinical remission, Med. Mycol. 38 (2002) 225e229.

[13] G. Benard, C.C. Romano, C.R. Cacere, et al., Imbalance of IL-2, IFN-

gamma and IL-10 secretion in the immunosuppression associated with

human paracoccidioidomycosis, Cytokine 13 (2001) 248e252.

[14] M.T. Peracoli, C.S. Kurokawa, S.A. Calvi, et al., Production of pro and

anti-inflammatory cytokines by monocytes from patients with para-

coccidioidomycosis, Microbes Infect. 5 (2003) 8e13.

[15] M.C. Fornazim, A. Balthazar, J.R. Quagliato, et al., Evaluation of

bronchoalveolar cells in pulmonary paracoccidioidomycosis, Eur. Respir.

J. 822 (2003) 895e899.

[16] J.G. McEwen, V. Bedoya, M.M. Patino, M.E. Salazar, A. Restrepo,

Experimental murine paracoccidioidomycosis induced by the inhalation

of conidia, J. Med. Vet. Mycol. 25 (1987) 165e175.

[17] L. Franco, L. Najvar, B. Gomez, S. Restrepo, J.R. Graybill, A. Restrepo,

Experimental pulmonary fibrosis induced by Paracoccidioides brasi-

liensis conidia: measurement of the host local responses, Am. J. Trop.

Med. Hyg. 58 (1998) 424e430.

[18] A.M. Cock, L.E. Cano, D. Velez, B. Aristizabal, J. Trujillo, A. Restrepo,

Fibrotic sequelae in pulmonary paracoccidioidomycosis: histopathological

1161T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162

aspects in BALB/c mice infected with viable and non-viable propagules,

Rev. Inst. Med. Trop. Sao Paulo 42 (2000) 59e66.

[19] A. Gonzalez, A. Restrepo, L.E. Cano, Pulmonary immune responses

induced in BALB/c mice by Paracoccidioides brasiliensis conidia,

Mycopathologia 165 (2008) 313e330.

[20] V.M. Menezes, B. Soares, C. Fontes, Drugs for treating para-

coccidioidomycosis, Cochrane Database Syst. Rev. (Apr 19 2006)

CD004967.

[21] F. Queiroz-Telles, L.Z. Goldani, H.T. Schlamm, J.M. Goodrich,

A. Espinel-Ingroff, M.A. Shikanai-Yasuda, An open-label comparative

pilot study of oral voriconazole and itraconazole for long-term treatment

of paracoccidioidomycosis, Clin. Infect. Dis. 45 (2007) 1462e1469.

[22] K. De Beule, J. Van Gestel, Pharmacology of itraconazole, Drugs 61

(2001) 27e37.

[23] Z.K. Khan, P. Jain, Antifungal agents and immunomodulators in systemic

mycoses, Indian J. Chest Dis. Allied Sci. 242 (2000) 345e355.

[24] A.M. Tobon, C.A. Agudelo, M.L. Osorio, D.L. Alvarez, M. Arango,

L.E. Cano, A. Restrepo, Residual pulmonary abnormalities in adult

patients with chronic paracoccidioidomycosis: prolonged follow-up after

itraconazole therapy, Clin. Infect. Dis. 37 (2003) 898e904.

[25] J.G. McEwen, G.R. Peters, T.F. Blaschke, E. Brummer, A.M. Perlman,

A. Restrepo, D.A. Stevens, Treatment of paracoccidioidomycosis with

itraconazole in a murine model, J. Trop. Med. Hyg. 88 (1985) 295e299.

[26] S. Restrepo, A.M. Tabares, A. Restrepo, Activity of two different tri-

azoles in a murine model of paracoccidioidomycosis, Rev. Inst. Med.

Trop. Sao Paulo 34 (1992) 171e176.

[27] R. Martinez, M.H. Malta, A.V. Verceze, M.R. Arantes, Comparative

efficacy of fluconazole and amphotericin B in the parenteral treatment of

experimental paracoccidioidomycosis in the rat, Mycopathologia 146

(1999) 131e134.

[28] R. Scavone, E. Burger, Paracoccidioidomycosis: reduction in fungal load

and abrogation of delayed-type hypersensitivity anergy in susceptible

inbred mice submitted to therapy with trimethoprimesulfamethoxazole,

Med. Microbiol. Immunol. 193 (2004) 53e59.

[29] A. Restrepo, M.E. Salazar, L.E. Cano, M.M. Patino, A technique to

collect and dislodge conidia produced by Paracoccidioides brasiliensis

mycelial form, J. Med. Vet. Mycol. 24 (1986) 247e250.

[30] J. Heykants, A. Van Peer, V. Van de Velde, et al., The clinical pharma-

cokinetics of itraconazole: an overview, Mycoses 32 (1989) 67e87.

[31] L.M. Alves, F. Figueiredo, S.L. Brandao Filho, I. Tincani, C.L. Silva, The

role of fractions from Paracoccidioides brasiliensis in the genesis of

inflammatory response, Mycopathologia 97 (1987) 3e7.

[32] C.L. Silva, L.M. Alves, F. Figueiredo, Involvement of cell wall glucans

in the genesis and persistence of the inflammatory reaction caused by the

fungus Paracoccidioides brasiliensis, Microbiology 140 (1994)

1189e1194.

[33] M. Moscardi-Bacchi, A. Soares, R. Mendes, S. Marques, M. Franco,

In situ localization of T lymphocyte subsets in paracoccidioidomycosis,

J. Med. Vet. Mycol. 27 (1989) 149e158.

[34] A. Gonzalez, W. De Gregori, D. Velez, A. Restrepo, L.E. Cano, Nitric

oxide participation in the fungicidal mechanism of gamma interferon-

activated murine macrophages against Paracoccidioides brasiliensis

conidia, Infect. Immun. 68 (2000) 2546e2552.

[35] A. Gonzalez, B.H. Aristizabal, E. Caro, A. Restrepo, L.E. Cano, TNF-a-

activated macrophages inhibit transition of Paracoccidioides brasiliensis

conidia to yeast cells through a mechanism independent of nitric oxide,

J. Trop. Med. Hyg. 71 (2004) 828e830.

[36] M. Moscardi-Bacchi, E. Brummer, D.A. Stevens, Support of Para-

coccidioides brasiliensis multiplication by human monocytes or macro-

phages: inhibition by activated phagocytes, J. Med. Microbiol. 40 (1994)

159e164.

[37] B.M. Saunders, A.A. Frank, I.M. Orme, Granuloma formation is required

to contain bacillus growth and delay mortality in mice chronically

infected with Mycobacterium tuberculosis, Immunology 98 (1999)

324e328.

[38] V.L. Calich, S.S. Kashino, Cytokines produced by susceptible and

resistant mice in the course of Paracoccidioides brasiliensis infection,

Braz. J. Med. Biol. Res. 31 (1998) 615e623.

[39] N. Kurita, M. Oarada, M. Miyaji, E. Ito, Effect of cytokines on antifungal

activity of human polymorphonuclear leukocytes against yeast cells of

Paracoccidioides brasiliensis, Med. Mycol. 38 (2000) 177e182.

[40] A.G. Bean, D.R. Roach, H. Briscoe, M.P. France, H. Korner,

J.D. Sedgwick, et al., Structural deficiencies in granuloma formation in

TNF gene-targeted mice underlie the heightened susceptibility to aerosol

Mycobacterium tuberculosis infection, which is not compensated for by

lymphotoxin, J. Immunol. 162 (1999) 3504e3511.

[41] K. Phillips, M. Weinblatt, Granulomatous lung disease occurring during

etanercept treatment, Arthritis Care Res. 53 (2005) 618e620.

[42] M.A. Lucchiari, M. Modolell, K. Eichmann, C.A. Pereira, In vivo

depletion of interferon-gamma leads to susceptibility of A/J mice to

mouse hepatitis virus 3 infection, Immunobiology 185 (1992) 475e482.

[43] C.H. Mody, C.L. Tyler, R.G. Sitrin, C. Jackson, G.B. Toews, Interferon-g

activates rat alveolar macrophages for anti-cryptococcal activity, Am. J.

Respir. Cell Mol. Biol. 5 (1991) 19e26.

[44] L.E. Cano, S.S. Kashino, C. Arruda, D. Andre, C.F. Xidieh, L.M. Singer-

Vermes, C.A. Vaz, E. Burger, V.L. Calich, Protective role of gamma

interferon in experimental pulmonary paracoccidioidomycosis, Infect.

Immun. 66 (1998) 800e806.

[45] M. Kaviratne, M. Hesse, M. Leusink, A.W. Cheever, S.J. Davies,

J.H. McKerrow, L.M. Wakefield, J.J. Letterio, T.A. Wynn, IL-13 activates

a mechanism of tissue fibrosis that is completely TGF-beta independent,

J. Immunol. 173 (2004) 4020e4029.

[46] N.K. Malavia, J.D. Mih, C.B. Raub, B.T. Dinh, S.C. George, IL-13

induces a bronchial epithelial phenotype that is profibrotic, Respir. Res. 9

(2008) 27.

[47] T.A. Wynn, Cellular and molecular mechanisms of fibrosis, J. Pathol. 214

(2008) 199e210.

[48] K. Koli, M. Myllarniemi, J. Keski-Oja, V.L. Kinnula, Transforming

growth factor-beta activation in the lung: focus on fibrosis and reactive

oxygen species, Antioxid. Redox Signal. 10 (2008) 333e342.

[49] N. Decologne, M. Kolb, P.J. Margetts, F. Menetrier, Y. Artur, C. Garrido,

J. Gauldie, P. Camus, P. Bonniaud, TGF-beta1 induces progressive pleural

scarring and subpleural fibrosis, J. Immunol. 179 (2007) 6043e6051.

[50] A.M. Tobon, I. Gomez, L. Franco, A. Restrepo, Seguimiento post-terapia

en pacientes com Paracoccidioidomicosis tratados con itraconazol, Rev.

Colomb. Pneumol. 7 (1985) 74e78.

1162 T.W. Naranjo et al. / Microbes and Infection 12 (2010) 1153e1162