Antifungal Effect of All-trans Retinoic Acid against Aspergillus fumigatus In Vitro and in a Pulmonary Aspergillosis In Vivo Model

Aspergillus fumigatus is the most common opportunistic fungal pathogen and causes invasive pulmonary aspergillosis (IPA), with high mortality among immunosuppressed patients. The fungistatic activity of all-trans retinoic acid (ATRA) has been recently described in vitro.

subgenera (Aspergillus, Circumdati, Fumigati, and Nidulantes) and 20 sections, including 339 recognized species (4). Nevertheless, only a few well-known species (including Aspergillus fumigatus and Aspergillus flavus) are considered important opportunistic pathogens in both humans and animals (5). In humans, Aspergillus fumigatus is the most common and life-threatening airborne opportunistic fungal pathogen, with particular significance among immunosuppressed hosts (5). Aspergillus fumigatus causes invasive pulmonary aspergillosis (IPA), which is associated with severe morbidity and mortality (5). Despite the high IPA-related mortality, the most used three classes of antifungal drugs (polyenes, azoles, and echinocandins) display limited clinical efficacy and are often unsuccessful, despite the noteworthy side effects associated with their long-term use, the critical pharmacokinetic profile, and, lastly, the drug-drug interactions (6). An important challenge for the future is to discover novel drugs that can also interact with specific fungal targets or modify the existing ones by optimizing their antifungal action by reducing dosages and consequently side effects.
In the last decades, the role of vitamin A in the regulation of the immune response has been confirmed by several studies, and its deficiency is associated with increased susceptibility to severe infectious diseases (7). Vitamin A is a nutrient obtained through the diet either as provitamin A (carotenoids) or as preformed vitamin A (retinol and retinyl esters), and liver dehydrogenases convert retinol into retinoic acid, its biologically active metabolite (7). Retinoids are key modulators of cell growth, differentiation, and apoptosis through their receptor-regulated signaling pathways (8). In particular, retinoic acid controls the normal immune system development as a modulator of both innate and adaptive immune responses (9). We recently documented that the topical use of the retinoid derivative tazarotene favors rapid healing of fungal nail infections (10); moreover, all-trans retinoic acid (ATRA) demonstrated fungistatic activity in vitro (11). Based on this evidence, we sought to evaluate ATRA synergistic interaction with conventional antifungal drugs such as amphotericin B and posaconazole in in vitro studies and its therapeutic efficacy in a preclinical model of IPA.

RESULTS
ATRA displays fungistatic activity on Aspergillus fumigatus germination in vitro. At 0.5 to 1 mM concentration, ATRA exerted a strong inhibitory effect on swelling and germination of Aspergillus conidia in vitro ( Fig. 1A and B) as previously reported (11). After 18 h of treatment with lower ATRA concentrations (0.25 and 0.125 mM), Aspergillus conidia appeared swollen ( Fig. 1C and D) and similar to control, but without the development of short germ tubes typical of control cultures (Fig. 1E). The inhibitory effect on conidial germination was maintained until 30 h ( Fig. 1F and G). At that time, ATRA at concentrations lower than 0.5 mM did not interfere with Aspergillus conidia germination, although the development of germ tubes appeared delayed ( Fig. 1H and I). Only control Aspergillus cultures (Fig. 1L) displayed elongated hyphal structures. The fungistatic effect of ATRA was reversible since conidia were able to germinate and to grow in the hyphal form 7 days after ATRA removal (data not shown).
ATRA exerts a synergistic fungistatic activity with amphotericin B and posaconazole in vitro. In order to investigate the synergistic fungistatic activity of ATRA with standard antifungal drugs, we determined a subinhibitory concentration of amphotericin B (AmB). As shown in Fig. 2A, 3.15 mg/ml was the minimal AmB concentration capable of inhibiting Aspergillus conidia germination, while at 1.57 mg/ml concentration, AmB did not interfere with germination (Fig. 2B), and only a delayed hyphal growth was observed. Interestingly, the association of low and noneffective doses of ATRA (0.25 or 0.125 mM; Fig. 2C and D) with that subinhibitory dose of AmB (1.57 mg/ ml) displayed a synergistic efficacy against Aspergillus conidial swelling and germination after 24 h ( Fig. 2E and F), To confirm the synergism of ATRA in combination with AmB or posaconazole (POS), currently used as primary therapy against aspergillosis, antifungal susceptibility tests were also performed in vitro. As shown in Table 1, the lowest concentration of ATRA that caused prominent growth inhibition (MIC $ 50) of the fungus was 75 mg/ml (0.25 mM), while MIC values of AmB and POS were 3.15 mg/ml and 0.0018 mg/ml, respectively. Of note, both the antifungal drugs showed a synergistic inhibitory effect toward A. fumigatus in combination with a low ineffective dose of 35 mg/ml (0.125 mM) of ATRA. Notably, this synergistic interaction was more remarkable with POS. Indeed, the MIC values of AmB and POS combined with ATRA were one dilution and three dilutions lower, respectively, than the MICs of the drugs alone (MIC of AmB plus ATRA, 1.5 mg/ml [versus MIC of AmB alone, 3.15 mg/ml]; MIC of POS plus ATRA, 0.00022 mg/ml [versus MIC of POS alone, 0.0018 mg/ml]). These results suggest a promising interaction in vivo and the therapeutic possibility of reducing classical antifungal drug dosage.
ATRA reduced mortality in a rat model of invasive pulmonary aspergillosis. We investigated the efficacy of ATRA in a rat model of IPA, also in comparison with the antifungal drug posaconazole, commonly used in clinical practice against IPA. Thirty rats were divided into three experimental groups (n = 10), including control, ATRA treated, and posaconazole treated. ATRA dosage and administration protocol (2 mg/kg intraperitoneally), starting 8 days before the infection, were established in a preliminary study (data not shown). Posaconazole treatment (4 mg/kg per os) started concurrently with the infection (12). As shown in Fig. 3, both drugs impacted IPA progression compared to the control group. In particular, 100% of ATRA-treated rats were alive 6 days after the infection. During the second week, ATRA and posaconazole displayed a very similar effect on survival, with 60% of rats alive at the end of the study (P , 0.05 versus control). Those in vivo findings further supported ATRA efficacy as promising for IPA

therapy.
ATRA enhances macrophage phagocytosis of Aspergillus conidia in vitro. Macrophages play a major role in controlling Aspergillus infection (13). In the present study, we also evaluated the capability of ATRA of modulating macrophage phagocytic activity in vitro. As reported in Fig. 4A, ATRA significantly increased conidia phagocytosis by U-937 human macrophages compared to control (at 5 or 10 mM, around 20 and 40%, respectively). Image analysis of macrophage phagocytosis (Fig. 4B) confirmed the increased phagocytosis and adherent macrophage-conidia complex after ATRA treatment (5 and 10 mM). Those data suggested ATRA improves macrophage phagocytosis that likely contributes to antifungal activity against IPA in vivo.
ATRA inhibits Aspergillus Hsp90 expression and Hsp90-related genes in vitro. To confirm whether ATRA is an inhibitor of heat shock protein 90 (Hsp90), Aspergillus conidia were cultured in ATRA-supplemented medium for 24h. Immunoblot analysis showed that after 24 and 48 h of ATRA (1 mM) treatment, Hsp90 protein expression was almost suppressed ( Fig. 5A and B). In the presence of the reduced density and growth rate of Aspergillus conidia, quantitative real-time reverse transcription-PCR (qRT-PCR) documented, in the ATRA-treated group, a strong decrease (about 70%; P , 0.05) of Hsp90 transcripts compared to untreated Aspergillus conidia (Fig. 5C). It was previously reported that the expression of the conidiation-specific genes AbaA and WetA is reduced in the presence of Hsp90 inhibitor (14). To confirm that fungistatic activity of ATRA is Hsp90 related, we also quantified the expression of AbaA and WetA; qRT-PCR analysis showed that ATRAinduced and ATRA-reduced conidia formation associates with the inhibition of Hsp90 expression and decreased AbaA (about 60%) and WetA (about 60%) transcripts in Aspergillus cultures (Fig. 5D). Similar results were also observed at 62 mM ATRA concentration (data not shown). Hsp90 is also reported to interact with the catalytic subunit of calcineurin (15), necessary for correct hyphal growth (16). We investigated ATRA-induced inhibition of the calcineurin pathway and quantified the expression of CrzA, a major signaling component of the zinc finger transcription factor. CrzA expression was strongly reduced (about 40%) in the presence of ATRA (Fig.  5D). Those data demonstrated that ATRA is a potent inhibitor of fungal Hsp90 expression and Hsp90-related genes and that its inhibition likely contributes to the fungistatic activity of ATRA against Aspergillus in vitro.
Heat shock protein 90 ATP-binding site is a target of ATRA. Hsp90 is a chaperone protein required for the activation and stabilization of a wide variety of client proteins, many of them involved in crucial cellular pathways (17). In the absence of A. fumigatus Hsp90 X-ray structure, the crystal structure of Hsp90 from Saccharomyces cerevisiae, with PDB ID 2CG9 (18), whose sequence shares around 78% of identity and 95% of query cover with A. fumigatus, has been used as a template to generate the A. fumigatus Hsp90 model structure (Fig. 6A). A structural bioinformatic analysis was performed to investigate whether Hsp90 may be an ATRA-specific pharmacological target against Aspergillus growth. The ATRA structure, as observed with the ATP molecule, completely filled the Hsp90 ATP-binding site in the N-terminal domain of the Aspergillus fumigatus model structure ( Fig. 6B and C). In particular, the trimethylcyclohexene moiety was located at the bottom of the ATP-binding site, fully stabilized by hydrophobic contacts with the residues (Fig. 6B and C) (19), while the carboxyl-terminal group was involved in hydrogen bonds with residue Arg98 located at the external border of the ATP-binding site. The interaction energy, evaluated by molecular docking simulation, was around 28.9 kcal/mol, a value only slightly below that obtained using the same procedure with the ATP molecule (29.7 kcal/mol). These results strongly supported that ATRA molecule may be considered a competitive inhibitor of the Hsp90 ATP-binding site, in particular at high concentrations, and further confirm Hsp90 as a target of ATRA-induced fungistatic activity.

DISCUSSION
Over the last decades, the incidence of IFIs constantly increased. Moreover, the use of fungicidal drugs in the treatment of fungal diseases led to the development of resistance burden (20). Here, we documented that ATRA displayed a strong fungistatic activity in Aspergillus cultures and reduced mortality in a model of IPA in cortisone acetateimmunosuppressed rats. Our preliminary data also showed that the therapeutic benefit of ATRA is dose dependent and, most importantly, the outcome comparable to that obtained with posaconazole, used as the current standard care for the treatment of IPA. Our in vitro and in vivo findings help to explain some observations previously reported in the clinical practice. In leukemic patients treated with ATRA and other drugs, a lower occurrence of overall fungaemic events has been observed (21,22). Thus, the low clinical incidence of fungal infections is likely linked to the ability of ATRA to rapidly differentiate leukemic promyelocytes (23) but, more importantly, to its direct antifungal and immunomodulatory properties. In this light, we also documented a synergistic therapeutic activity of suboptimal doses of ATRA combined with the subinhibitory concentration of the antifungal drug amphotericin B against Aspergillus conidia germination. This promising interaction could provide an important potential advantage with the use of ATRA in clinical practice, allowing for reduction of the dosage of classic antifungal drugs with a greater control of their side effects as well as a potential better control of the development of fungal drug resistance. However, further experimental in vivo studies are required to establish an optimal therapeutic protocol for the use of ATRA in association with antifungal drugs against IPA and systemic Aspergillus infections.
It is well-known that conidial germination represents a crucial step in the pathogenesis and progression of IPA. Alveolar macrophages are the first important line of the host's immune defense against Aspergillus conidia. In healthy individuals, lung resident macrophages are able to rapidly engulf and kill inhaled conidia and are not capable of internalizing hyphal cells (24). In immunocompromised patients, nonphagocytosed conidia produce germ tubes that elongate into hyphae in the lung tissue, causing a systemic infection for their capacity to invade blood vessels. Here, we demonstrated that ATRA increases, in a dose-dependent manner, the number of adherent and internalized resting Aspergillus conidia by macrophages, suggesting its modulatory action on innate immunity. Our results were in line with previous studies showing that ATRA is able to enhance macrophage phagocytosis against Candida yeasts and to favor . The hydrogen bonds have been indicated with green dashed lines between the interaction partners. The residues composing the active sites that are in proximity of the molecules are shown. Images have been produced using the LigPlot1 software. (C) Best molecular docking complexes between Aspergillus fumigatus Hsp90 with ATP (left) and ATRA molecules (right). The a-helices and loops are shown as light blue spirals and wires, respectively. The molecular surface is shown, and the ligand molecules, hosted in the Hsp90 ATP-binding site, are depicted by sphere representations. Panels A and C have been produced using the Chimera program. healing of pneumonia by Pneumocystis carinii in an affected patient (23,25). ATRA, by suppressing the conidial germination and hyphal growth, could exert a crucial role in facilitating the conidia uptake and their elimination by phagocytic cells, preventing the spread of infection as well as host tissue damage caused by hyphal invasion. Considering the crucial role of macrophages in Aspergillus conidia clearance and that the phagocytic activity is impaired in severely immunocompromised patients, the impact of ATRA on phagocytic activity of macrophages is quite relevant. Recent studies also support the hypothesis that efficacy of vitamin A and ATRA alone or in combination with other drugs is related to the capacity to stimulate the monocyte-mediated immune response (26).
Recent works indicate that fungal Hsps are implicated in several processes, including pathogenicity, phase transition, and antifungal drug resistance (27). The function of Hsp90 in Aspergillus has been also investigated, and genetic and pharmacologic studies indicate an important role of Hsps in conidiation and in cell wall stress-compensatory mechanisms (27). Here, we report a reduced Hsp90 expression in ATRA-treated Aspergillus conidia cultures. Thus, Hsp90 could represent an important target for inhibiting replication and controlling fungal drug resistance. It was demonstrated that Hsp90 itself is subjected to posttranslational modification in a cell cycle-dependent manner. In S. cerevisiae, during the S phase of the cell cycle, the tyrosine kinase Swe1 phosphorylates a residue in the N terminus of Hsp90 that facilitates chaperone interaction with numerous client proteins (28). A role for Hsp90 in Candida albicans cell cycle progression has been also established and a molecular link between Hsp90, cell cycle regulation, and morphogenesis defined, identifying Hsp90 as an important player in cell cycle progression (29). Besides Hsp90, our results showed that ATRA downregulates transcription of other genes, in particular, AbaA and WetA. Their inhibition is reported in association with decreased conidia formation (30). We documented the downregulation of CrzA expression in ATRA-treated Aspergillus cultures. CrzA is involved in the calcineurin pathway and hyphal growth (31).
In order to better elucidate how ATRA targets the expression of Hsp90, molecular docking demonstrated that ATRA can fit the Hsp90 ATP-binding site in the N-terminal domain of the Aspergillus fumigatus model structure. Hsp90 is an ATP-dependent molecular chaperone which is essential in eukaryotes. Hsp90 is required for the activation and stabilization of a wide variety of client proteins, and many of them are involved in important cellular pathways (32). Structurally, Hsp90 is a flexible dimeric protein composed of three different domains which adopt structurally distinct conformations. ATP binding triggers directionality in these conformational changes and leads to a more compact state. Posttranslational modifications of Hsp90, such as phosphorylation and acetylation, may provide another level of its regulation. Phosphorylation and acetylation influence the conformational cycle, cochaperone interaction, and interdomain communications (33). Targeting Hsp90 to reduce stress response compensatory pathways with fungal-specific inhibitors of ATRA may represent an attractive novel antifungal strategy (34).
In conclusion, we documented the efficacy of ATRA in a rat model of IPA and its synergistic capacity to improve the efficacy of antifungal agents in vitro. ATRA, for its direct fungistatic action targeting the expression of fungal Hsp90, combined with its immunoadjuvant effect against opportunistic fungi, may represent a promising novel candidate in the armamentarium of systemic therapeutic strategies for the treatment and/or prophylaxis of IPA and IFIs, either alone or in association with conventional antifungal drugs.

MATERIALS AND METHODS
Microorganism and fungal cultures and treatments. The Aspergillus fumigatus strain obtained from a clinical isolate, according to routine hospital laboratory procedures, was grown on Sabouraud dextrose agar (SDA) (Difco, Detroit, MI) supplemented with chloramphenicol. Resting conidia were obtained from the SDA-grown fungal culture as previously described (11). For in vitro studies, ATRA (catalog no. R2625; Sigma-Aldrich, Milan, Italy) and amphotericin B (Fungizone) (Sigma-Aldrich, Milan, Italy) were tested in the range of 1 to 0.06 mM (300 to 37.5 mg/ml) and in the range of 50 to 0.09 mg/ml, respectively. To investigate the potential synergistic interaction between ATRA and AMB on the conidial germination, resting conidia were plated at the density of 1 Â 10 5 in 96-well microplate in RPMI 1640 with 10% fetal calf serum (FCS; Euroclone, Milan, Italy), supplemented with penicillin (100 U/ml) and streptomycin (100 mg/ml) and treated with different concentrations of ATRA or AmB, alone or in combination. In the combined treatment, ATRA was given 30 min before amphotericin B. Fungal cultures were maintained at 37°C and monitored during the first 30 h after the treatments.
Evaluation of fungistatic activity. Aspergillus fumigatus development was followed using an optical microscope (Carl Zeiss, UK) with a Â40 magnification objectives lens. Aspergillus conidia vitality was evaluated microscopically by the trypan blue dye exclusion method.
In vitro antifungal susceptibility testing. The synergistic antifungal activity of ATRA in combination with the antifungals AmB or POS (Sigma-Aldrich) against Aspergillus fumigatus was also assessed using the broth microdilution method as described in M38-A2, a document produced by the Clinical and Laboratory Standards Institute (35). RPMI 1640 2% glucose with L-glutamine (Sigma-Aldrich, St. Louis, MO, USA) buffered to pH 7.0 was used as fungal culture medium. A. fumigatus strain was grown on Sabouraud dextrose agar (Difco Laboratories, Detroit, MI, USA) supplemented with chloramphenicol. Resting conidia were harvested by washing the slant cultures with sterile saline as previously described (11). Aspergillus inoculum was set to 0.4 Â 10 4 to 0.5 Â 10 5 CFU/ml, standardized spectrophotometrically. We added 100 ml of this suspension to 96-well flat-bottom plates. Twofold serial dilutions of ATRA (from 150 mg/ml to 37.5 mg/ml), AmB (from 50 mg/ml to 0.09 mg/ml), and POS (4 mg/ml to 0.00005 mg/ml) were prepared in RPMI 1640 medium. Aspergillus conidia were treated with 100 ml of each antifungal drug either alone or in combination with ATRA. Positive (100 ml of RPMI medium with 100 ml of conidia) and negative controls (200 ml of RPMI containing ATRA alone or in association with the antifungals without conidia) were included in all experiments. The plates were incubated with agitation at 30°C for 24 h. The MIC was determined spectrophotometrically at 510 nm with an enzyme-linked immunosorbent assay (ELISA) reader. For each compound, the MIC was defined as the lowest drug concentration that produced a significant inhibition (MIC $ 50) of fungal growth compared with the drug-free control. The MICs were tested in three independent experiments carried out in triplicate.
Animals, fungal infection, and treatment. Sprague-Dawley rats (Envigo, Huntingdon, UK) were immunosuppressed by cortisone 21-acetate (catalog no. C3130; Sigma-Aldrich, St. Louis, USA) treatment subcutaneously administered at 150 mg/kg body weight 7 days before the day of infection; immunosuppression was maintained with 80 mg/kg of cortisone 21-acetate for 3 days a week until the end of the experiment. As the immunosuppression treatment started and throughout the study, rats received a low-protein (8%) diet, as well as drinking water containing tetracycline hydrochloride antibiotic (260 mg/ liter; catalog no. T3383; Sigma-Aldrich), supplied ad libitum. Rats were then infected with 5 Â 10 5 Aspergillus conidia in 0.1 ml of saline solution by oropharyngeal aspiration. For treatment, ATRA (Sigma-Aldrich) was resuspended in 100% dimethyl sulfoxide (DMSO) at 50 mg/ml and stored at 220°C in the dark. Before using, ATRA was diluted in peanut oil (catalog no. P2144; Sigma-Aldrich) in order to obtain a 2-mg/kg/2 ml suspension ready for in vivo administration. The posaconazole oral suspension was obtained by diluting the clinical drug Noxafil in water at a concentration of 4 mg/10 ml (Merck Sharp & Dohme, Kenilworth, NJ, USA). Posaconazole treatment started concurrently with the infection (day 0), and it was administered for 5 days a week until the end of the study. As a negative control, the vehicle was prepared by adding 2% DMSO in peanut oil and 5% glucose to water. In the prophylaxis schedule optimization experiment, three groups of rats (n = 10) were treated with ATRA by intraperitoneal (i.p.) injection for 7 days before infection (day 0) and then for 5 days a week up to the end of the study.
The mortality rate was recorded daily until the end of the studies. All animal procedures were carried out in accordance with the standards established by the Animal Ethical Committee of Takis.
Phagocytosis assay. For phagocytosis assay (36), human macrophages U-937 (1 Â 10 6 cells/ml) were incubated overnight in 6-ml polypropylene tubes in Dulbecco's modified Eagle's medium (DMEM; Gibco, Milan, Italy) supplemented with 2% (vol/vol) FCS, 2 mM L-glutamine, 100 IU/ml penicillin, and 100 mg/ml of streptomycin with vehicle (DMSO) or different ATRA concentrations (1, 5, and 10 mM). After overnight incubation, the cells were incubated for an additional 1 h at 37°C with unopsonized resting Aspergillus conidia at an effector-to-fungal cells ratio of 1:10 before being evaluated for internalization and visualized by light microscopy. Phagocytic cells were separated from nonphagocytosed A. fumigatus by centrifugation on a fetal serum gradient. A 0.1-ml sample of the harvested phagocytic cells was used for cytospin preparation. After Diff-Quik staining, fungal cell internalization was expressed according to the following formula: percentage of internalization = number of cells containing one or more fungal cells/ 100 cells counted. In parallel, macrophage phagocytosis was also performed in 24-well plates for a realtime analysis of phagocytosis by microscopy. U-937 macrophages cultured at 1 Â 10 6 cells/ml were treated overnight with ATRA and incubated for an additional 1 h at 37°C with Aspergillus conidia as described above.
RNA extraction and real-time reverse transcription-PCR. A sample of frozen biomass was placed into a 2-ml extraction tube containing beads and disrupted using TissueLyser. RNA was isolated using TRI reagent (Sigma-Aldrich) according to the manufacturer's instructions. To verify the general purity of the RNA, a NanoDrop instrument was used. One microgram of total RNA was reverse transcribed by a QuantiTect reverse transcription kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. qRT-PCR analyses were performed using 20 ng cDNA per well in triplicate with the SYBR Green master mix (Applied Biosystems, Foster City, CA USA) (37,38) according to the manufacturer's instructions. Reactions were run on 7900HT system (Applied Biosystems).
Statistical analysis. For in vitro experiments, statistical analysis was performed using an unpaired two-tailed t test between control and experimental groups. For in vivo studies, survival curves were determined by the Kaplan-Meier method, and the log-rank test was used to conduct comparisons between the groups. Test of statistical significance was two-sided; a value of P , 0.05 was considered to be statistically significant.