Persistence within dendritic cells marks an antifungal evasion and dissemination strategy of Aspergillus terreus

Aspergillus terreus is an airborne human fungal pathogen causing life-threatening invasive aspergillosis in immunocompromised patients. In contrast to Aspergillus fumigatus, A. terreus infections are associated with high dissemination rates and poor response to antifungal treatment. Here, we compared the interaction of conidia from both fungal species with MUTZ-3-derived dendritic cells (DCs). After phagocytosis, A. fumigatus conidia rapidly escaped from DCs, whereas A. terreus conidia remained persisting with long-term survival. Escape from DCs was independent from DHN-melanin, as A. terreus conidia expressing wA showed no increased intracellular germination. Within DCs A. terreus conidia were protected from antifungals, whereas A. fumigatus conidia were efficiently cleared. Furthermore, while A. fumigatus conidia triggered expression of DC activation markers such as CD80, CD83, CD54, MHCII and CCR7, persistent A. terreus conidia were significantly less immunogenic. Moreover, DCs confronted with A. terreus conidia neither produced pro-inflammatory nor T-cell stimulating cytokines. However, TNF-α addition resulted in activation of DCs and provoked the expression of migration markers without inactivating intracellular A. terreus conidia. Therefore, persistence within DCs and possibly within other immune cells might contribute to the low response of A. terreus infections to antifungal treatment and could be responsible for its high dissemination rates.

in conidia interactions of A. fumigatus and A. terreus with immune cells have been observed 10 . Once phagocytosed by macrophages, the dihydroxynapthalene-melanin (DHN-melanin) of A. fumigatus conidia inhibits phagolysosome acidification and allows conidia germination and escape from macrophages 10, 11 . In contrast, A. terreus conidia produce a different type of melanin, the so-called Asp-melanin, which protects conidia from UV radiation and reduces phagocytosis by amoeba 12 , but does not inhibit acidification of macrophage phagolysosomes 10 . Consequently, A. terreus conidia reside in the acidic environment of phagolysosomes, in which they show long-term persistence without inactivation. This long-term persistence was further confirmed by in vivo studies using a murine inhalation infection model 13 . Mice treated with high doses of corticosteroids were susceptible not only to A. fumigatus, but also to A. terreus infections. Bioluminescence imaging revealed that all mice infected with A. terreus conidia developed initial invasive disease, but only about 50% of mice succumbed to infection. Survivors showed no clinical signs of invasive disease at day 14 post infection. However, lung sections indicated the presence of conidia and plating of lung homogenates revealed that these conidia were still viable 13 . Furthermore, when immunocompetent mice were infected with A. terreus conidia, only a transient pro-inflammatory response was detected, which implied that infection was rapidly cleared. However, histopathology at days 1, 3 and 5 after infection revealed significant amounts of intracellular conidia and determination of colony forming units showed that these conidia were still viable 10 . This indicates that persistence within immune cells is not only observed under immunocompromised conditions, but also in the immunocompetent situation. However, the subtypes of immune cells in which A. terreus conidia are able to persist have not been investigated yet.
Besides macrophages, dendritic cells (DCs) are another type of professional phagocytic cells that migrate to regional and draining lymph nodes where they present fungal antigens to T-cells and prime T-cell immune responses 14,15 . Furthermore, some fungal pathogens, such as Histoplasma capsulatum are more efficiently restricted in the conidia to yeast transformation by DCs compared to macrophages, indicating an additional role of DCs in the control of fungal infections 16 . A. fumigatus conidia are efficiently phagocytosed by DCs via mannan receptors and C-type lectins of galactomannan specificity 17 . While a short-term interaction with DCs does not inactivate A. fumigatus conidia, three days after infection pulmonary DCs had transported conidia to draining lymph nodes and spleen without signs for viable fungal cells. This indicates the general ability of DCs to inactivate and degrade Aspergillus conidia for efficient antigen presentation 17 . Furthermore, it has been shown that DC recruitment in neutropenic mice increases during A. fumigatus infection and a depletion of DCs under this setting further impairs fungal clearance 18 . However, effective maturation of dendritic cells requires the interaction with neutrophils that stimulate DC maturation and migration to draining lymph nodes 19 . It has also been shown that both A. fumigatus and A. terreus infections trigger Th17/Th1 immune responses 20 . However, while A fumigatus also elicited Th2 activating cytokine production after infection, a Th2 immune response with A. terreus conidia has not been observed 20 . Th2 cytokines are promoted by IL-4 and IL-10 that are secreted by DCs mainly in response to hyphal growth, but expression of these cytokines is not promoted in A. terreus infections 10 . This finding suggests that A. terreus conidia may persist in DCs without germination and escape as observed in previous studies on macrophages 10 .
Taking into account that A. terreus infections are often refractory to antifungal treatment 4 , show long term persistence in infected murine lungs 10, 13 and display high dissemination rates in human patient populations 6 , we compared the interaction of A. terreus and A. fumigatus conidia with a human dendritic cell line. Our results suggest that A. terreus might hitchhike DCs for dissemination and gets protected from antifungal treatment.

Selection of dendritic cells for interaction studies.
To study the interaction of A. fumigatus and A.
terreus conidia with dendritic cells we tested the suitability of the CD34 + human acute myeloid leukaemia cell line MUTZ-3 21 as an alternative to monocyte-derived human dendritic cells (moDCs) as MUTZ-3 cells have been shown to display the full range of functional antigen processing and presentation 21 . Similar to the differentiation of CD1a-positive and CD14-negative moDCs from peripheral blood monocytes after stimulation with human recombinant GM-CSF and IL-4 ( Fig. S1A), MUTZ-3 cells were reproducibly differentiated into immature DCs (iDCs) when incubated in the presence of human recombinant GM-CSF and IL-4 22 (Fig. S1B). Therefore, subsequent studies were performed with MUTZ-3 cells differentiated into iDCs and relevant data were confirmed by the use or primary moDCs.

Phagocytosis of resting A. fumigatus and A. terreus conidia by iDCs. Previous in vitro studies
showed that A. terreus conidia were rapidly phagocytosed by macrophages and remained resting in acidified phagolysosomes, whereas A. fumigatus conidia suppressed acidification and escaped by the formation of germ tubes 10 . When confronted with MUTZ-3 iDCs, 70% of A. fumigatus conidia were phagocytosed within 3 h and this rate declined with prolonged incubation time (55% at 6 h; Fig. 1A). This appeared mainly due to rapid germ tube formation and escape from DCs (Fig. 1B) and was in line with phagocytosis rates observed with murine fetal skin-derived DCs that showed about 70% internalization of A. fumigatus conidia after two hours with a decreasing rate after 4 h of confrontation 17 . On the contrary, despite higher galactomannan and β-glucan surface presentation of A. terreus conidia compared to A. fumigatus 10 , phagocytosis rate of A. terreus conidia was slightly delayed with 48% internalisation at 3 h. However, phagocytosis rate increased to about 75% at 6 h without significant germ tube formation of intracellular conidia (Fig. 1A). Microscopic analyses revealed some germination of A. terreus conidia after 9 h, but this germination was mainly restricted to conidia that had not been internalised by DCs (Fig. 1B). Intracellular conidia at this 9 h time point were either resting or showed some initial swelling. This implied that A. terreus conidia, once phagocytosed by iDCs do not tend to rapidly escape. These obvious differences in phagocytosis and fungal escape tempted us to determine growth rates of A. fumigatus and A. terreus conidia in the presence and absence of dendritic cells. During co-incubation with iDCs A. fumigatus showed a delay in the increase of optical density by only about 2 h compared to conidia grown in the absence of ScIeNtIfIc REPoRtS | 7: 10590 | DOI:10.1038/s41598-017-10914-w iDCs (Fig. 1C), which is in agreement with the rapid germ tube formation observed by microscopic analyses. In contrast, while A. terreus conidia generally tend to germinate slower than those from A. fumigatus, germination in the presence of DCs was further delayed by at least 5 h (Fig. 1D) and most likely derived from conidia not phagocytosed by DCs. This delay was strictly dependent on the presence of viable DCs, as A. terreus germination speed was not delayed when conidia were incubated in the presence of fixed DCs (Fig. 1D).

Survival of A. terreus and A. fumigatus in confrontation with
DCs. Due to the low rate of intracellular germination of A. terreus conidia inside DCs, viability was determined by analysing the number of colony forming units (CFU) after lysis of DCs. Although A. fumigatus escaped DCs and displayed 100% viability after one hour of confrontation with MUTZ-3 iDCs, a significant reduction in viable counts were observed at three and six hours (Fig. 1E). However, due to the formation of dense mycelium from extracellular hyphae, no CFU determination at later time points was possible. In contrast to A. fumigatus, conidia of A. terreus were more resistant against DC-mediated killing. Survival rate of A. terreus conidia remained at about 100% after 3 h and only slightly dropped down to about 70-80% at 9 h (Fig. 1E). When experiments were repeated with human monocyte-derived moDCs comparable results as for MUTZ-3 iDCs were obtained for both fungi (Fig. 1F). This confirms the suitability of the MUTZ-3 cell line in analysing interactions of Aspergillus conidia with DCs and furthermore reveals that A. terreus conidia remain persistent and viable within DCs. Analysis of persistence versus escape strategies. Macrophages acidify phagolysosomes by an active proton pumping mechanism to effectively degrade foreign antigens. A. fumigatus inhibits this acidification by DHN-melanin that coats conidia 11 and contributes to survival within and rapid escape from macrophages. On the contrary, A. terreus conidia produce Asp-melanin 12 , which does not inhibit phagolysosme acidification and, consequently, conidia reside within acidified phagolysosomes 10 . However, it has been shown that the escape of A. terreus conidia from macrophages can be stimulated by either inhibiting the V-H + -ATPase with Bafilomycin or by recombinant expression of the Aspergillus nidulans naphthopyrone synthase gene wA, which is responsible for production of a precursor molecule of DHN-melanin 10 . However, when we confronted A. terreus conidia expressing the wA gene with iDCs, germination showed the same delay ( Fig. S2) as wild-type A. terreus conidia (Fig. 1D). This indicates that the different type of melanin formed by A. terreus conidia is not the direct cause for persistence in DCs. Furthermore, lysotracker staining of DCs after phagocytosis of either A. fumigatus or A. terreus conidia did not indicate acidified phagolysosomes with either species (Fig. S3). This is in agreement with an active NOX2 dependent alkalinisation mechanism in the phagolysome membrane of DCs and is important for antigen cross-presentation and leads to a phagolysosome pH of about 8 23 . This confirmed that an acidic environment is not the cause of delayed germination of A. terreus conidia. However, when we investigated the effect of environmental pH on germination of conidia, only a minor delay of germination was observed with A. fumigatus conidia at pH 8 compared to pH 6 and 7 ( Fig. 2A), whereas a delay of about three hours at a shift from pH 6 to 7 and a 6 h delay at a shift to pH 8 was observed with A. terreus conidia (Fig. 2B). Therefore, an alkaline environment might be one reason for trapping A. terreus conidia in DCs. However, an additional effect on inhibition of germination might derive from nutrient restriction within phagolysosomes. To test this assumption, we studied the germination of A. fumigatus and A. terreus conidia in serial dilutions of RPMI/glucose medium. Indeed, A. fumigatus conidia efficiently germinated without delay when RPMI/glucose medium was diluted by at least 64-fold (Fig. 2C), whereas germination speed of A. terreus conidia was strictly dependent on the concentration of nutrients (Fig. 2D). A subsequent investigation on A. terreus conidia germination in dependence on the limitation of specific nutrient sources indicated that either glucose or nitrogen limitation caused a delay in germination and limitation of both showed a cumulative effect (Fig. S4). As A. terreus and A. fumigatus both grow on minimal media, the overall nutritional requirements for both fungi appear to be similar. However, due to the unrestricted fast germination of A. fumigatus even at very low nutrient concentrations, the required nutrient threshold triggering germination seems much lower for A. fumigatus than for A. terreus. In conclusion, alkalinisation of phagolysosomes accompanied by a restricted nutrient supply traps A. terreus conidia in a resting state, whereas A. fumigatus efficiently germinates at pH 8 and shows rapid germination as soon as nutrients exceed a certain threshold value.
Efficacy of antifungals on persisting intracellular conidia. Alkaline pH and nutrient limitation appear as major reasons for trapping A. terreus but not A. fumigatus conidia in a resting intracellular state. From this result two questions arose: (i) are conidia persisting in DCs protected from environmental stressors such as antifungals and (ii) how long can conidia survive within DCs? To address these questions we selected the antimycotics amphotericin B (AMB) and voriconazole (VOR) that were used in a concentration of 10 µg/ml, which is far above the MIC values of both drugs for the selected strains with about 0.5 µg/ml (AMB) and 0.25 µg/ml (VOR) for A fumigatus and 1 µg/ml (AMB) and 0.25 µg/ml (VOR) for A. terreus as determined by broth dilution tests. Therefore, we expected that all extracellular fungi that had not been phagocytosed or escaped from DCs were inactivated by the antifungals. Indeed, when conidia were incubated without iDCs in the presence of these antimycotics and analysed for CFUs after 24 h, about 100% of A. fumigatus conidia and about 99.5% of A. terreus conidia were inactivated with either drug (Fig. 3A), showing that the selected concentrations of both antimycotics were highly effective. Subsequently, A. fumigatus and A. terreus conidia were first pre-incubated for six hours in the presence of iDCs to allow for phagocytosis of extracellular conidia before antimycotics were added and CFUs were analysed after 24 h (Fig. 3A). Only about 1% of A. fumigatus conidia remained viable after this procedure, indicating that conidia that escaped DCs were inactivated by extracellular drugs and conidia that failed to escape were inactivated by DCs. In contrast, in the presence of DCs about 60% of A. terreus conidia survived AMB and 40% the VOR treatment. This implies that intracellular conidia were well protected from antifungal treatment and not efficiently inactivated by DCs. To test for a long-term persistence of A. terreus within DCs the experiment was repeated, but analysis of CFUs was performed after three days of incubation (Fig. 3B). While about 99.9% of free extracellular conidia were inactivated by either drug, 30% viable conidia were still re-isolated from DCs under either treatment. This experiment confirmed both assumptions from above: A. terreus conidia show an intracellular long-term persistence within DCs, which is accompanied by a protection from antifungal treatment.

DC activation in the presence of A. fumigatus and A. terreus conidia. Activation of DCs is required
to trigger an adaptive immune response against Aspergillus infections 24 . In this respect, activation of the transmigration marker CCR7 is of special importance as it allows migration of DCs to lymph nodes for T-cell interaction and activation during Aspergillus infection 24 . Since A. terreus conidia showed a persistent phenotype within DCs, we wondered whether these conidia trigger activation and maturation of iDCs. Therefore, we first determined expression of the activation markers CCR7, CD83, CD80 and MHCII after short term incubation (6 h) with viable or UV-inactivated A. fumigatus and A. terreus conidia. Only the confrontation with viable A. fumigatus conidia resulted in a significant induction of CCR7 expression accompanied by induction of the activation marker CD83 and a similar trend was observed by either using MUTZ-3 DCs or moDCs (Fig. S5). UV-inactivated A. fumigatus conidia as well as viable or UV-inactivated A. terreus conidia did not induce CCR7 expression, but slightly induced expression of CD83 (Fig. S5C). Since maturation of DCs may take longer than 6 h especially when no other co-stimulatory cytokines or interactions with immune cells such as neutrophils are present 19 , we aimed in the characterisation of DC activation at later time points. To avoid complete destruction of DCs by extracellular hyphae, we added AMB after 6 h of confrontation, but lowered AMB concentration to 2 µg/ml to avoid perturbance of dendritic cell function. Such treatment has previously been shown not to grossly affect response of human DCs in the interaction with A. fumigatus conidia 24 . Activation marker expression was studied after a further incubation of 18 h and was normalised to cells treated with AMB but not confronted with conidia. As expected 24 , CCR7 and to a minor extend CD62L (not shown) as well as CD83, CD80, CD54 and MHCII were strongly expressed on DCs confronted with viable A. fumigatus conidia (Fig. 4). This indicates that these DCs matured and became activated for transmigration to lymph nodes through expression of CCR7 for stimulation of naïve T-cells. In contrast, while CD83, CD80, CD54 and MHCII, though to a lower extent, were also upregulated by UV-inactivated A. fumigatus and both, viable and UV-inactivated A. terreus conidia, no significant induction of CCR7 surface expression was observed for those pathogen preparations (Fig. 4). These partially activated DCs are assumed not to produce T-cell activating cytokines or stimulate immunogenic T-cells 25 . Cytokine production of DCs in response to A. fumigatus and A. terreus conidia. From analyses of activation markers we assumed that only DCs confronted with viable A. fumigatus conidia are stimulated to produce cytokines that are critical to elicit innate and adaptive immune responses. Therefore, we analysed Th1, Th2, immunosuppressive and pro-inflammatory cytokine production of DCs in response to viable and UV-inactivated A. fumigatus and A. terreus conidia after 24 h of confrontation (Fig. 5). When incubated with viable A. fumigatus conidia DCs produced pro-inflammatory cytokines, such as IL-1β and TNF-α as well as T-cell stimulating IL-12 (Th1) and IL-4 (Th2) cytokines. In contrast, neither confrontation with UV-inactivated A. fumigatus nor viable or UV-inactivated A. terreus conidia resulted in a significant cytokine response of DCs (Fig. 5A-D), whereby the lack of cytokine production was not mediated by production of the immunosuppressive cytokines TGF-β and IL-10 ( Fig. 5E-F). Therefore, we conclude that persisting A. terreus conidia do not induce a strong pro-inflammatory or immune regulatory response in DCs.

Effect of TNF-α stimulation on DCs confronted with A. terreus conidia. Our analyses revealed that
A. terreus conidia persist within DCs, but neither trigger a significant cytokine response (Fig. 5) nor stimulate the expression of the transmigration markers CCR7 and CD62L (Fig. 4). However, due to their ability to transmigrate for antigen presentation, DCs appear to depict ideal candidates for transporting A. terreus to different body sites. This is of special interest as clinical reports have indicated that aspergillosis caused by A. terreus is frequently associated with high dissemination rates to secondary organs 6 . Previous studies showed that corticosteroid treated mice infected with high doses of A. terreus conidia display elevated levels of TNF-α in lung tissues, which might results from the germination of extracellular conidia 13 . Therefore, we stimulated DCs containing persisting A. terreus conidia with 100 ng/ml TNF-α and inhibited extracellular growth of A. terreus conidia by amphotericin B (Fig. 6). Indeed, TNF-α not only significantly stimulated the expression of the activation markers CD83, CD80, CD54 and MHCII, but also of the transmigration marker CCR7, indicating that TNF-α led to a full maturation of iDCs (Fig. 6A-E). However, when the intracellular survival of A. terreus conidia was investigated after 24 h, about 60% of conidia remained viable and this viability was undistinguishable from iDCs that were not stimulated by TNF-α (Fig. 6F). Therefore, TNF-α stimulation induces migration of DCs harbouring persisting A. terreus conidia and, therefore, might support fungal dissemination.

Discussion
Invasive fungal infections remain a great challenge in critically ill patients and especially patients suffering from haematological malignancies are highly susceptible for invasive aspergillosis 2 . As A. fumigatus is most frequently re-isolated from patients with invasive aspergillosis, studies on immune interactions have strongly focused on this species. However, we have previously shown that macrophage interaction of A. terreus is significantly different from that observed with A. fumigatus 10 . While A. fumigatus inhibits phagolysosome acidification by its DHN-melanin layer and rapidly germinates and pierces the phagocytic cell, A. terreus has been shown to reside in acidified phagolysosomes 10 . Persistence within macrophages is at least partially due to the different type of melanin produced by A. terreus 12 , as expression of the naphtopyrone synthase wA increased A. terreus germination rates. However, as DCs do not tend to acidify their phagolysosomes 23 , it is well conceivable that the type of melanin does not influence persistence and escape in this type of immune cells and, in agreement, the expression of wA in A. terreus conidia did not trigger the escape from DCs.
Our analyses on the reasons why A. terreus conidia remain dormant within DCs indicated that both, the alkaline pH in DC phagolysosomes as well as restriction from nutrient supply significantly delays germination of A. terreus conidia. Strikingly, germination rates of A. fumigatus conidia were quite similar in a range from pH 6 to 8 and even at a very limited nutrient supply germination was initiated without any significant delay. Therefore, while the nutritional supply in DCs seems to exceed the required threshold value for germination of A. fumigatus conidia, the phagolysosomal environment forms a natural barrier for A. terreus germination and indicates that DCs do not need to produce specific growth inhibitory factors to control A. terreus germination. However, since A. terreus conidia can be stored for several weeks in nutrient limited solutions such as phosphate-buffered saline or water without significant loss in viability, they easily survive the nutrient limited environment within phagolysosomes. In addition, as conidia do not tend to germinate under these conditions, they may be treated as inactive antigen particles that induce expression of activation markers on dendritic cells without stimulating maturation by expression of the transmigration marker CCR7 24 . Unfortunately, these activated but not matured DCs may act as negative regulators of T-cell immune response. Interaction of T-cells with these so-called steady-state DCs inhibits their ability to develop into immunogenic effector T-cells, but stimulates development of regulatory T-cells that dampen the immune response 26 . In agreement with the formation of steady-state DCs after phagocytosis of A. terreus conidia neither significant induction of the pro-inflammatory cytokine TNF-α, nor other cytokines was observed. This is in line with observations on A. fumigatus, in which β-glucan exposure from germinating conidia was required to trigger TNF-α production in bone marrow and alveolar macrophages as well as dendritic cells. In contrast, heat-killed A. fumigatus conidia were not immune stimulatory 27

and in our experiments UV-inactivated A. fumigatus conidia behaved comparable in DC interactions as viable or UV-inactivated A. terreus conidia.
While persisting A. terreus conidia may cause a down-regulation of the immune response, they may also explain the difficulties to treat A. terreus infections with antifungals despite the fact that, except for AMB, antifungal susceptibility of A. terreus is largely comparable to that of A. fumigatus 4 . It has been observed that significantly higher drug concentrations are required in therapy of A. terreus infections than assumed from in vitro susceptibility analyses and it virtually affects treatment with all common antimycotics 4 . This may, at least partially, be due to a fungistatic effect of drugs on A. terreus, which results in large differences between minimal inhibitory and minimum fungicidal drug concentration 28 . However, at 10 µg/ml of AMB or VOR both drugs showed high fungicidal activity on A. terreus, which was completely different in the presence of DCs. Within DCs between 40 and 60% of A. terreus conidia survived this drug treatment for the first 24 h and about 30% viable conidia were still re-isolated after three days. Therefore, persistence within DCs may not only tune down immune activation, but also supports evasion of A. terreus from antifungal treatment.
For using DCs as a vehicle for dissemination as frequently observed with A. terreus infections 6 , full maturation and expression of transmigration markers is required. As observed in moribund corticosteroid treated mice, A. terreus can indeed elicit a strong TNF-α response in infected lungs once hyphal elements are formed 13 . Figure 6. Effect of extracellular TNF-α on DCs containing persisting A. terreus conidia. MUTZ-3-derived iDCs were pre-incubated for 6 h with A. terreus conidia before amphotericin B (2 μg/ml) was added. Medium was either supplement with 100 ng/ml TNF-α or left untreated (control). Cells were incubated for a further 18 h before expression of (A) CCR7, (B) CD54, (C) CD83, (D) CD80 and (E) MHC II was analysed by flow cytometry. Data were normalised to expression levels from uninfected DCs without TNF-α supplementation (dashed line). (F) Survival of persisting A. terreus conidia under TNF-α treatment. Cells were lysed after 24 h and CFUs were determined. Statistics were performed by using the two-tailed Student's t-test. All data represent mean values + SD from three independent experiments (***p < 0.005).
ScIeNtIfIc REPoRtS | 7: 10590 | DOI:10.1038/s41598-017-10914-w Alternatively, TNF-α production may be caused by bacterial (co-)infection as common in critically ill patients 29 . We have shown that external addition of TNF-α results in complete maturation and expression of CCR7 on DCs containing persisting A. terreus conidia, without resulting in increased conidia inactivation rates. TNF-α and IFN-γ activate production of reactive oxygen species in macrophages, which has been described to support clearance of intracellular fungal pathogens 8,9 . However, in case of A. terreus infections, this appears of minor importance. A transient increase in IFN -γ in immunocompetent mice infected with A. terreus conidia has been observed 10 , which is in line with the Th1 immune response described for A. terreus infections 20 . However, as viable conidia remain persistent in infected lung tissues 10 , this clearance mechanism seems not effective for A. terreus conidia. Even more, the activation of DCs by these external cytokines may cause the initiation of transmigration to draining lymph nodes and spleen, transporting viable A. terreus conidia to organs outside the pulmonary tract. However, it needs to be mentioned that providing a definite role for the observed persistence of A. terreus conidia in dendritic cells during pathogenesis will require the in vivo detection of DCs carrying viable fungal elements.
In conclusion, MUTZ-3 cells are well suited for the investigation of Aspergillus conidia interactions with dendritic cells. Parallel studies on the survival of A. fumigatus and A. terreus conidia with MUTZ-3 DCs and moDCs showed identical results and our analyses on A. fumigatus conidia matched results previously obtained from primary or monocyte-derived DCs. As previously observed for macrophages, A. terreus follows a sit and wait strategy in DCs and may use these cells for dissemination once an external trigger stimulates DCs maturation. Most importantly, A. terreus was protected from antifungal treatment inside DCs, which could explain treatment failure despite in vitro susceptibility of A. terreus against antifungals 4 . Finally, we speculate that long-term persistence of A. terreus conidia within immune effector cells such as macrophages or dendritic cells may also occur in immunocompetent persons as previously observed in immunocompetent mice 10 . Thus, once an immunosuppressive regimen is applied these conidia may be released from immune cells and may cause invasive aspergillosis with high dissemination rates.

Material and Methods
Fungal strains, cultivation and preparation of conidia suspensions. A. fumigatus CBS 144.89 (CBS, Utrecht, Netherlands), A. terreus SBUG844 (HKI strain collection, Jena, Germany 30 ) and A. terreus SBUG844 expressing the A. nidulans wA gene 10 were used in this study. For conidia preparation, fungi were cultivated at 37 °C on solid 50 mM glucose containing Aspergillus minimal media with nitrate as nitrogen source and 2% agar 30 . A. fumigatus conidia were harvested after 3-4 and A. terreus conidia after 5-6 days of incubation. Conidia were harvested by overlaying plates with 10 ml sterile phosphate-buffered saline (PBS) and filtered through 40 μm cell-strainers. Conidia suspensions were washed twice in PBS and counted by using a haemocytometer.
Labelling and UV-inactivation of conidia. For phagocytosis assays conidia were pre-labelled with fluoresceine isothiocyanate (FITC) as described previously 10 . In brief, 1-2 × 10 8 conidia were suspended in 10 ml of 0.1 M sodium carbonate solution containing 0.1 mg/ml FITC and incubated at 37 °C for 1 hour. Conidia were washed twice in sterile PBS and diluted to selected concentrations for phagocytosis assay. For UV inactivation 1 ml of conidia suspension containing 2-3 × 10 8 conidia were poured in small petri dish (5 cm in diameter). UV treatment was performed with the sterilisation programme of a GS Gene Linker UV chamber (BioRad Laboratories). Complete inactivation of conidia (>99.9%) was confirmed by spreading aliquots on malt extract agar plates.

Differentiation of DCs from human monocytes.
Human monocyte-derived DCs were generated as previously described 31 . In brief, human peripheral blood mononuclear cells (PBMCs) were harvested by density gradient centrifugation using Biocoll (Biochrom AG). CD14 + primary human monocytes were further purified from PBMCs by CD14 + microbeads (Miltenyi Biotech) using a magnetic cell sorting system. Primary monocytes were differentiated into monocyte-derived dendritic cells (moDCs) by treatment with 800 U/ml human recombinant GM-CSF and 1,000 U/ml human recombinant IL-4 (Miltenyi Biotech). At day six cell surface marker expression (Fig. S1A) and phenotype of immature moDCs was monitored prior to performing conidia interaction studies.
Phagocytosis assays. iDCs (3 × 10 5 cells) were transferred to wells of a 24-well plate containing polylysine-coated coverslips (12 mm diameter) and FITC-labelled conidia of A. fumigatus or A. terreus were added at a multiplicity of infection (MOI) of 0.5. At selected time points plates were centrifuged at 200 × g to pellet cells and fixed over night at 4 °C in the presence of 4% of paraformaldehyde. After washing with PBS, cells were blocked for 15 min at room temperature (RT) with Fc receptor blocking solution (Human TruStain FcX TM , BioLegend). Plates were centrifuged again for 10 min at 200 × g before blocking solution was removed. To stain conidia not phagocytosed by DCs, a rabbit anti-conidia antiserum (kindly provided by Frank Ebel, Munich) diluted 1:100 in 1% bovine serum albumin (BSA, sigma) was added and incubated for 1 hour at RT. After washing with PBS, Cy3-labelled anti-rabbit IgG antibody diluted 1:300 in 1% BSA was added and incubated for 1 hour at RT in the dark. Coverslips were washed three times with PBS, mounted in ProLong Gold mounting solution (Invitrogen) and analysed by fluorescence microscopy. Conidia co-localising with DCs were counted. Phagocytosis rates were calculated from FITC + , Cy3 − conidia divided by the total number of conidia counted resulting in the percentage of phagocytosis. At least 900 conidia were analysed in each experimental group.
ScIeNtIfIc REPoRtS | 7: 10590 | DOI:10.1038/s41598-017-10914-w Analysis of conidia germination. A. terreus and A fumigatus conidia (10 5 /well) were cultivated at 37 °C in a CO 2 incubator in 96-well plates in the absence or presence of DCs (MOI = 1). Optical density of cultures (OD 600 ) was analysed at 600 nm every hour. To investigate the effect of pH or nutrient limitation on germination of Aspergillus conidia, RPMI 1640 (Sigma) with 2% glucose was adjusted to pH 6-8 or MOPS-buffered medium (pH 6.5) was serially diluted with PBS. Cells were incubated at 37 °C in a Multiskan Go microplate spectrophotometer (Thermo) and OD 600 was taken every 15 min. Similarly, Aspergillus minimal medium with ammonium chloride as nitrogen source was serially diluted and analysed for conidia germination. To analyse the effect of carbon or nitrogen starvation, either glucose or ammonium chloride were serially diluted, but all other components were kept constant.
Killing and antifungal sensitivity of conidia in the presence of DCs. DCs (1 × 10 6 cells/well) were co-incubated with Aspergillus conidia at an MOI of 0.1 in 6 well plates. DCs, including all extracellular fungal material, were scraped at 1 h, 3 h, 6 h and 9 h and lysed by the addition of sterile cold ddH 2 O with subsequent centrifugation at 2000 × g to release and collect intracellular conidia. Serial dilutions of cell lysates were plated on malt extra agar plates and incubated for 24-40 h at 37 °C for determination of CFUs. Survival rates were calculated with reference to CFUs from conidia not confronted with DCs. For antifungal sensitivity assays DCs were allowed to phagocytose Aspergillus conidia for 6 h prior to addition of 10 μg/ml AMB or VOR. DCs were collected and lysed as described after a total of either 24 h or 72 h. Serial dilutions of lysates were plated on malt extract agar plates and CFUs and survival rates were determined after 40 h of incubation at 37 °C as described above.

Lysotracker staining. DCs and MH-S cells (3 × 10 5 cells/well) were co-incubated with FITC-labelled
Aspergillus conidia at an MOI of 0.5 on polylysine coated coverslips (12 mm) in 24 well plates. After 3 h plates were centrifuged at 200 × g for 10 min and medium was removed. Pre-warmed fresh medium containing 200 nM lysotracker (DND-99, Invitrogen) was added and incubation was continued for 1 h to allow for staining of phagolysosomes. After a brief centrifugation step, cells were fixed for 15 minutes with 4% paraformaldehyde. After washing with PBS, coverslips were mounted with ProLong Gold mounting solution and analysed by fluorescence microscopy.
Analysis of DC activation markers and cytokine production. DCs (10 6 /well) were co-incubated with viable or UV-inactivated Aspergillus conidia (MOI = 2) in 6 well plates. For short-term activation studies cells were analysed after 6 h of co-incubation, whereas in long-term interaction studies amphotericin B (2 μg/ml) was added after 6 h to inhibit extracellular fungal growth. Cells were collected at 24 h and blocked for 10 min at RT with Fc receptor blocking solution (Human TruStain FcX TM ). DCs (2-3 × 10 5 cells) were treated with 1 μg/ml FITC-anti-human CCR7, APC-anti-human CD54, PE-anti-human CD80 or PerCP-anti-human HLA-DR (MHCII) monoclonal antibodies (BioLegend) diluted by FACS staining buffer (dPBS supplemented with 1% heat-inactivated FBS) and incubated for 30 min in the dark at 4 °C. After two washes in FACS staining buffer, cells were fixed in 1% paraformaldehyde and subjected to flow cytometry (LSRII). For preparation of intracellular cytokine staining, Brefeldin A (5 μg/ml, Fluka) was added 6 h before harvest to block intracellular protein transport. Cells were collected at 30 h and treated with Fc receptor blocking solution for 10 min at RT. Cells were fixed with 4% paraformaldehyde for 20 min at 4 °C. To make cell membranes permeable, fixed cells were washed with perm wash buffer (FACS staining buffer containing 0.1% saponin). The cells were stained at 4 °C for 30 min with 1 μg/ml of either PE-anti-human TNF-α, FITC-anti-human IL-1β, PE-anti-human IL-12/IL-23 p40, PE-anti-human IL-4, PE-anti-human TGF-β or APC-anti-human IL-10 monoclonal antibodies (BioLegend) prepared in perm wash buffer. After washing with perm wash buffer, the cells were left in 1% paraformaldehyde at 4 °C in the dark before analysed by flow cytometry. Mean fluorescence intensity (MFI) was analysed by FlowJo software and normalised to mock-infected DCs.

Statistical analyses.
Comparisons between multiple groups were analysed by one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test using GraphPad Prism (GraphPad Software). Statistical differences between two groups were analysed by two-tailed Student's t-test. Ethics statement. Human peripheral blood for generation of moDCs was collected from healthy volunteers after written informed consent. The study was conducted in accordance with the Declaration of Helsinki, all protocols were approved by the Ethics Committee of the University Hospital Jena (permit number: 273-12/09).