Human Monocyte/Macrophage Fungicidal Activity of GM-CSF Against Paracoccidioides brasiliensis Depends on ROS

Paracoccidioidomycosis (PCM) is the major systemic mycosis in Latin America. Its etiological agent is the fungus Paracoccidioides brasiliensis, a microorganism with thermal dimorphism, behaving as yeast at body temperature [1,2]. This fungus causes a natural infection by inhalation of conidia or mycelial elements which are converted into the parasitic yeast form in lungs [3]. This disease shows multiple shapes, ranging from benign and localized to severe and disseminated ones, depending on many factors, such as the host cell immunity and strain virulence of the fungus [4].


Introduction
Paracoccidioidomycosis (PCM) is the major systemic mycosis in Latin America. Its etiological agent is the fungus Paracoccidioides brasiliensis, a microorganism with thermal dimorphism, behaving as yeast at body temperature [1,2]. This fungus causes a natural infection by inhalation of conidia or mycelial elements which are converted into the parasitic yeast form in lungs [3]. This disease shows multiple shapes, ranging from benign and localized to severe and disseminated ones, depending on many factors, such as the host cell immunity and strain virulence of the fungus [4].
Among the immunological mechanisms reported to this infection, innate immunity monocytes (MO)/macrophages appear to play a fundamental role, acting as the first defense line in the organism [3], depending on their state of activation [5]. Ingested conidia or yeast forms of P. brasiliensis readily multiply inside murine alveolar or peritoneal macrophages; however, when cells are activated by cytokines, such as IFN-γ, the multiplication is limited and conidia or yeast cells may be killed [2,3,[6][7][8].
With regards to murine cells, some studies have shown that IFN-γ activation promotes P. brasiliensis killing through the L-arginine/NO pathway [9]. However, works in our laboratory have demonstrated that IFN-γ activation is not enough for the fungicidal activity of human cells against virulent P. brasiliensis strain (Pb18) [10]. This process is effective after cells preactivation with TNF-α or IFN-γ plus TNF-α. Moreover, these studies provided strong evidence of H 2 O 2 participation as an effector mechanism, since catalase, a H 2 O 2 scavenger, inhibited the intracellular killing by TNF-α or TNF-α plus IFN-γ-activated MO [11].
The role of other cytokines, besides IFN-γ and TNF-α concerning human MO/macrophage-P. brasiliensis interaction is still unclear. Broad evidence has indicated that GM-CSF not only promotes proliferation and differentiation of hematopoietic precursor cells, but also induces various aspects of macrophage activation [12][13][14][15][16], e.g., respiratory burst activity [15]. MO release Reactive Oxygen Species (ROS, such as H 2 O 2 , O 2 -and OH), which are cytotoxic to microorganisms and tumor cells.
The effects of GM-CSF on MO/macrophage function against microbial pathogens have been studied in uncontrolled trials in humans and in vitro and in vivo experiments [12,13,16].
Brain-Heart Infusion (BHI) agar medium (Difco Laboratories, Detroit, MI, USA), used for culture plating, contained gentamicin 0.5% (Neoquímica, Anápolis, GO, Brazil), 4% horse normal serum and 5% P. brasiliensis strain 192 culture filtrates (v/v), the latter being the source of growth-promoting factor [27]. 96-well flatbottomed plates were purchased from Nunc, Life Tech. Inc., Maryland, MA, USA. Horseradish peroxidase (type II) was obtained from Sigma Chemical, San Diego, CA, USA. P. brasiliensis strain 18 (Pb 18) was maintained in the yeast-form cells at 35 °C in GPY culture medium for six days [28]. Yeast viability was determined by phase contrast microscopy and bright yeast cells were counted as viable, while dark ones were considered nonviable. Fungal suspensions containing more than 90% viable cells were used for the experiments.

Donors
MO was isolated from volunteer healthy blood donors, after informed consent from the University Hospital of the Botucatu Medical School (FMB), São Paulo State University (UNESP, Brazil). The Hospital Ethics Committee approved this study.

Isolation of peripheral blood mononuclear cells (PBMC):
PBMC were isolated from heparinized venous blood by density gradient. Briefly, 10 ml of heparinized blood were mixed with an equal volume of complete tissue culture medium (CTCM). Samples were layered over 5ml of Histopaque in a 15-ml conical plastic centrifuge tube. After centrifugation at 300×g for 30' at room temperature, the interface layer of PBMC was harvested and washed twice with PBS-EDTA and once with CTCM. Cell viability as determined by 0.2% trypan blue exclusion was >95% in all experiments. The MO were stained with neutral red (0.02%) and the concentrations were adjusted to 2×10 6 MO/ml in CTCM. More than 90% of the cells were considered as MO by morphological examination, neutral red uptake, and staining for unspecific esterase [29].

MO/macrophage monolayers
100µL of MO suspension (2×10 6 MO/ml) was dispensed into 96well flat-bottomed plates. After incubation during 2h at 37 °C in 5% CO 2 , non adherent cells were removed by aspiration and each well was rinsed twice with CTCM. After adherence, MO were cultured in CTCM at 37 °C in 5% CO 2 during 18h, alone (MO) or containing GM- Fungicidal activity After supernatants removal of control and treated monolayers, MO/macrophages were challenged with 100 µl of 2×10 4 viable units/ml of Pb18 (ratio 50:1) in CTCM containing 10% fresh human AB serum, in absence or presence of CAT (20,000U/mL) or L-NMMA (450U/mL). After coculture during 18h (experimental cultures), cells were harvested by aspiration with sterile distilled water to lyse MO. Each culture and well washing was contained in a final volume of 2ml. The number of colony forming units (CFU) of Pb18 per culture was determined by plating 100µl of the 2-ml harvested volume, in triplicate, on BHI containing 4% normal horse serum and 5% P. brasiliensis strain 192 culture filtrates (v/v). A control culture only containing 100µl of yeastform Pb18 (2×10 4 viable units/ml) was submitted to the same procedures used for the experimental cultures. Inoculated plates were incubated at 35 °C in sealed plastic bags to prevent drying. After 10 days the number of CFU per plate was counted and the percentage of fungicidal activity was determined by the formula: mean CFU of control culture

Reactive oxygen intermediates (ROI) determination
ROI production was indirectly measured by assessing H 2 O 2 release from MO and macrophages, according to the method previously described by PICK &KEISARI [30] and adapted by PICK & MIZEL [31]. MO or MO-derived macrophages were obtained as previously described, and cultured in duplicate in 96-well plates for 24 hours, at 37 °C in 5% CO 2 tension, with or without 100µL of recombinant human GM-CSF (rh-GM-CSF) (1, 10, 31.25, 62.5, 125, 250, 500 and 1000U/mL). After this period, culture supernatants were used to nitric oxide (NO) determination and the adherent cells were resuspended to the original volume (0,1mL) in phenol red buffer solution containing: 140mM of NaCl; 10nM of phosphate buffer, pH 7; 5.5mM of dextrosis; 0.56mM of phenol red; 0.01mg/ mL of peroxidase from radish peroxidase type II and, in the presence or absence of 1mg of phorbol myristateacethate (PMA), and were incubated at 37 °C in a dark, humid chamber. After 60 minutes, the reaction was interrupted by the addition of 0.01mL of NaOH 1N. Absorbances were measured at 620nm in an automatic ELISA microreader. Results were expressed in nanomols (nM) of H 2 O 2 /2x10 5 cells, using a standard curve.

Reactive nitrogen intermediates (RNI) determination
NO production was determined based on Griess reaction [32]. Culture supernatants were mixed in with an equal volume of Griess reagent (1% Sulphanilamide, 0.1% NEED, in 5% phosphoric acid) at room temperature for 10min [32]. Sodium nitrite (NaNO 2 -) was used as standard. Absorbances were measured at 540nm in an ELISA microreader. Assays were carried out in quadriplicate. Results were expressed in µmols of NO 2 -/2x10 5 cells, comparing the optical density (OD) with a standard curve of known NO 2 -concentrations. Statistical Analysis: Statistical procedures were performed using Graphpad Instat software (San Diego, California -USA). Significant differences among the various groups were detected by Repeated Measures Analysis of Variance (ANOVA), followed by Tukey Kramer Multiple Correlations. Significance level was set at p<0.05.   As shown in Table 1 Table 1 suggest that the fungicidal activity presented by GM-CSF-activated MO is mediated by H 2 O 2 , while NO seems not to be involved.  Similarly, the data obtained with MO pre incubation, GM-CSF activated macrophages (M∅) displayed a dose-dependent increase in fungicidal activity, from 31.25U/mL (15.5%), to the most significant concentrations of 125 and 250U/mL (29.8% and 27.2%, respectively), when compared to control M∅ (7.2%). It shall be reinforced that, compared to MO (Figure 1; 33.9 and 32.9%), the activity of activated-M∅ was not statistically significant. To detect the effector mechanisms involved in GM-CSFactivated macrophages (M∅) for P. brasiliensis killing, we challenged these cocultures with Pb18 and treated them concomitantly with CAT or L-NMMA. Similarly to MO cultures (MO- Figure 2), CAT inhibited the fungicidal activity induced by GM-CSF, confirming the role of H 2 O 2 in this process, as well with M∅. On the other hand, LNMMA did not change fungicidal activity of GMCSF-activated M∅, like in MO cultures (Figure 4). One may see in Table 2 that the levels of H 2 O 2 were significantly diminished after non activated macrophages challenge with Pb18 (M∅+Pb=0.71±0.23nM), when compared to macrophages alone (M∅=1.24±0.53nM). However, Macrophages preactivated with GM-CSF and challenged with Pb showed similar H 2 O 2 levels (M∅+GM-CSF+Pb=2.17±0.6nM) to those detected in M∅+GM-CSF culture. These data show that, differently of non-activated M∅ challenged with Pb18 (M∅+Pb), the challenge of activated cells (M∅+GMCSF+Pb) did not lead to H 2 O 2 inhibition. In all the cocultures treated with CAT, a significant inhibition in H 2 O 2 levels was obtained, compared to those detected in the absence of this scavenger. In relation to NO 2 -, similarly to H 2 O 2 release and also similarly to MO cultures, there was a significant inhibition in this metabolite production in cocultures supernatants of M∅+Pb (=2.46±0.53µM), compared to M∅ (=3.17±0.98). However, conversely to H 2 O 2 , GM-CSF (3.03±0.91µM) did not stimulate the cells to increase NO2-production, compared to M∅. Moreover, M∅+GM-CSF+Pb (1.89±0.39) did not release higher levels of NO 2 . Once more, as occurred in supernatants of MO cultures, the results showed that NO 2 -levels were very low in all cocultures supernatants, suggesting that they were not correlated with fungicidal activity. Again, these results are associated with the lack of L-NMMA effect on fungicidal activity presented by GM-CSFactivated M∅. Together, the results presented in Table 2 and Figure  S3 & 4 showed that the fungicidal activity presented by GM-CSFactivated M∅ is mediated by H 2 O 2 , similarly to MO experiments. In experiments with M∅, once more, NO seems not to be involved.   (1,10,31,25,62,5,125,250, 500 and 1000U/mL) for 18h. Following, they were challenged with P. brasiliensisduring 4h and assessed for fungicidal activity in vitro. Results are expressed as Mean (M) ± SEM of 8 subjects. Figure 5 shows an increased fungicidal activity of human MO-derived macrophages activated with GM-CSF, after 7 days of treatment in the presence of this cytokine (G-M∅), only with 125 and 250U/mL (32.4% and 31.0%, respectively), in a similar way compared to fungicidal activity presented by GMCSF-activated-MO and M∅. One may observe that the MO differentiation process into macrophages in the presence of GM-CSF, did not change significantly the responsiveness to GM-CSF pre activation, compared to macrophages culture without this cytokine.   As to NO 2 -, similarly to H 2 O 2 release and similarly to MO and M∅ cultures, in cocultures supernatants of M∅+Pb (= 2.46±0.53µM), we detected a significant inhibition in this metabolite production, compared to M∅ (= 3.17±0.98). However, conversely to H 2 O 2 , M∅+GM-CSF (3.03±0.91µM) did not stimulate the cells for increased NO 2 -production, compared to M∅. Moreover, M∅+GM-CSF+Pb (1.89±0.39) did not release higher levels of NO 2 -, which were very low in all cocultures supernatants, suggesting that they were not correlated with fungicidal activity. Once again, these results are associated with the lack of L-NMMA effect on fungicidal activity of MO+GM-CSF cells. Therefore, it follows, from the results shown in Table 3 and Figure 5 & 6, that the fungicidal activity presented by GM-CSF-activated G-M∅ is mediated by H 2 O 2 similarly to MO and M∅. Again, NO seems not to be involved.

Discussion
The aim of this work was to investigate the role of GM-CSF on human mononuclear phagocytes activation and the effector mechanisms developed by these cells for P. brasiliensis killing. The results presented here allow us to consider the existence of an important fungicidal activity of GM-CSF-activated human MO and MO-derived macrophages against P. brasiliensis. Stimulation of MO/macrophage for antimicrobial activity by GM-CSF is reported by a number of studies in vivo and in vitro, evidenciating that this cytokine activates neutrophils, MO and macrophages, and enhances the ability of these cells to kill intracellular parasites, such as Candida albicans [17], Aspergillus fumigates [18,19], Histoplasma capsulatum [20,21], Trypanosoma cruzi [23], Mycobacterium avium [24]; M. lepraemurium [25], and MAC [26]. Since GM-CSF increased the fungicidal activity of human MO and macrophages against P. brasiliensis strain 18 (Pb18), we were interested in clarifying the mechanisms by which the cells would exert this activity.
Our results clearly demonstrated that CAT inhibited the fungicidal activity of the cells tested. Moreover, despite of the Pb18 inhibits H 2 O 2 release, as shown in previous works [33], when the cells are activated by GM-CSF, a compensatory effect on this production was detected, and H 2 O 2 concentrations were enough for a fungicidal effect. Thus, this work provides evidence that the mechanism by which GM-CSF activated cells kill P. brasiliensis is This mechanism has been shown to be efficient in the inhibition of replication and killing of fungi [9,39] and other microorganisms by murine mononuclear phagocytes. Previous results from our group have demonstrated that killing of Pb18 by IFN-γ and TNF-αactivated peritoneal murine macrophages is mediated by NO and H 2 O 2 [40]. Although it is known that NO is abundantly synthesized by phagocytes from mice and rats, its secretion by human mononuclear phagocytes has become a controversial issue [24,37,40,41]. Mac Micking et al. [40] and Albina [42] have reported that human MO/ macrophages express nitric oxide synthase type II (NOS 2 ) as protein synthesis, in response to various stimuli. Furthermore, while it has been found to be a potent antifungal compound of phagocytes in mice, its antifungal role has not been established in human [37][38][39]. Specifically, in this work, NO does not appear to be involved in the fungicidal activity of human phagocytes against Pb18 yeasts [43,44].