INTRODUCTION
Aspergillus fumigatus is a saprophyte but also an opportunistic human fungal pathogen. It propagates through conidia that are airborne and are constantly inhaled (
1). To establish an invasive infection, conidia have to cross a respiratory barrier that includes epithelial and mucous layers in the upper respiratory tract. Conidia reaching the distal part (lung alveoli) of the respiratory system have to further confront both cellular and humoral immune barriers. Cellular immunity is provided by resident alveolar macrophages and recruited neutrophils. The humoral immune system consists of the complement proteins, collectin, antimicrobial peptides, acute-phase proteins, and immunoglobulins. Among these, the complement system has been speculated to play an important role against
A. fumigatus conidia (
2,
3).
The activation of the complement system consists of a cascade of reactions through classical, lectin, and alternative pathways (
4) that differ according to the activation complexes formed but converge in C3b formation. With
A. fumigatus, the main effect of the complement system is executed through opsonization by C3b, which has been shown to bind to the
A. fumigatus conidial surface (
5–7). It was shown previously that
A. fumigatus conidia activate the alternative pathway, whereas swollen conidia and mycelial morphotypes activate the classical and lectin pathways (
7).
Aspergillus fumigatus conidia are covered by a cell wall (CW), consisting of a proteinaceous rodlet layer and a melanin pigment layer, and an inner CW, composed of different polysaccharides, including β-(1,3)-glucan (BG), α-(1,3)-glucan, chitin, and galactomannan (GM) (
1,
8,
9). The identities of the conidial cell wall ligands associated with the activation of different complement pathways remain to be elucidated. Moreover, the complement activation would be expected to result in the formation of a membrane attack complex (MAC), damaging the pathogen membrane and causing lysis of the pathogens. Nevertheless, the presence of a thick CW in fungi has been hypothesized to prevent lysis of the fungal cell (
10); however, experimental evidence is lacking.
Our study was aimed at identifying the complement components interacting with A. fumigatus conidia, determining the role of conidial CW components in activating complement pathways, and studying the role of the humoral immune system against A. fumigatus. We show that among the proteins interacting with the conidial surface, complement protein C3 is the prominent component. Assays using individual conidial CW components indicated that RodAp, BG, and GM are the main components involved in C3 activation. We observed that C3 opsonization facilitates conidial aggregation and phagocytosis and that complement receptors are mainly involved in conidial phagocytosis. Being airborne, conidia interact first with the alveolar environment; therefore, we compared conidial opsonization with human serum and bronchoalveolar lavage fluid (BALF). Although conidial opsonization with serum or BALF confirmed C3 to be the major complement component binding to the conidial surface, there were significant differences in the interaction of other complement proteins and the cytokines secreted upon phagocytosis of these opsonized conidia with human monocyte-derived macrophages (hMDM), indicating the importance of the source of humoral immune components in the immune response.
DISCUSSION
In the present study, we identified (i) complement proteins interacting with A. fumigatus conidia; (ii) conidial cell wall ligands interacting with C3, a central protein of the complement system; (iii) complement pathways activated by those cell wall components; and (iv) the biological importance of conidial opsonization (conidial aggregation, recognition, phagocytosis, ROS production, and killing). Being airborne, A. fumigatus conidia enter the alveoli via the breath; therefore, inhaled conidia first come into contact with the alveolar environment. Nonetheless, studies using human BALF were lacking. In our study, we used both human serum and BALF and compared complement proteins interacting with conidia. Although C3 is the major complement component interacting with the conidial surface irrespective of the source (serum or BALF), there were substantial differences in the abundance as well as the nature of other complement proteins interacting with conidia. Furthermore, interactions of serum- and BALF-opsonized conidia with hMDM resulted in distinct cytokine profiles, indicating that the immune stimulation pathways elicited by conidia differ with opsonization source.
It was shown previously that
A. fumigatus conidia predominantly activate the alternative pathway (
5,
7). Confirming that conidia activate C3 through the alternative pathway, we observed that the depletion of Ca
2+-Mg
2+ during interaction between C3 and RodAp, the protein that coats the conidial surface, had no effect on C3 activation. C1q and MBL were also shown to bind to dormant conidia of
A. fumigatus, activating the classical and lectin pathways (
13–15). Ficolin, the other lectin responsible for activating the lectin pathway, was also shown to bind to
A. fumigatus (
15–17). However, these observations were based on the interaction of conidia with serum. In contrast to studies with serum-opsonized conidia, MBL and the mannose-binding serine proteases (MASP1 and MASP2) were not identified in our proteomic analysis of the BALF-opsonized conidia, suggesting the absence of the lectin pathway upon conidial opsonization with BALF. Although C1q was identified in both serum- and BALF-opsonized conidial protein extracts, C2, one of the complement components required for the activation of the classical pathway, was absent in the BALF-opsonized conidial extract. The importance of C1q binding to the conidial surface is obscure with these data, as C1q is a component of the classical pathway. The structure of C1q resembles that of collectins, the soluble pattern recognition receptors (PRR) (
18); the collectin receptors share binding sites with C1q (
19). Therefore, C1q may function as a conidium-recognizing PRR, although this hypothesis needs to be validated. Of note, it has been reported that C1q knockout mice are susceptible to invasive aspergillosis (
20). Pentraxin 3 (PTX3) is involved in the complement activation via classical pathway by recruiting C1q to the microbial surfaces (
21). We could identify PTX3 in the pooled BALF sample after subjecting it to in-solution digestion followed by proteomic analysis but not in the protein extract from BALF-opsonized conidia. This could be due to the overwhelming C3 recruitment on the conidial surface through the activation of the alternative pathway or due to the technical issues (in-gel digestion) in identifying PTX3 if its PSM score in the extracted sample is significantly low.
To date, studies of complement activation by
A. fumigatus have always utilized intact fungus. As a result, the role of conidial CW, the first fungal component to interact with the host immune system and the one responsible for complement activation, remained unclear. The
A. fumigatus CW is a dynamic and immunomodulating component whose composition changes across fungal morphologies (
22). Understanding the interaction between the complement system and the CW components is essential for elucidating how the fungal pathogen is eliminated in healthy hosts, as phagocytosis of inhaled conidia may not be an immediate process (
23). We found that RodAp and the CW polysaccharides BG and GM could activate the complement system, in contrast to the rest of the CW components. It has been shown that there is stage-specific exposure of BG during conidial germination, and an atomic force microscopy analysis using a ConA-ligated probe indicated at least 5 to 7% mannan positivity on the conidial surface (
24), suggesting that BG and GM could play roles in the partial activation of the complement system.
The melanin pigments present in the
Aspergillus niger and
Cryptococcus neoformans CWs are of dihydroxyphenylalanine (DOPA) origin, and were shown to activate the alternative pathway and to bind activated C3 (
25). However, melanin pigment in the CW of
A. fumigatus is dihydroxynaphthalene (DHN) derived (
26), and in our assay it failed to activate the complement system. In
A. fumigatus conidia, melanin pigments form a layer underneath the rodlet layer, and at places it is exposed on the conidial surface (
12). Upon disruption of a major enzyme in the
A. fumigatus DHN melanin biosynthesis, the deposition of C3 was enhanced, which suggested that the intact DHN melanin indeed impairs C3 binding to the conidial surface (
27). On the other hand, chitin, a polymer of
N-acetylglucosamine, should be a ligand for ficolin, a lectin containing a collagenous domain and a fibrinogen domain that recognizes
N-acetylated compounds (
28). Ficolin has been shown to bind to chitin and activate the complement system (
17) as well as to activate the alternative pathway in human plasma (
29). However, in the present work, although it was significant compared to the control, we did not observe a major C3b deposition on chitin upon opsonization with human serum. This discrepancy could be explained by the fact that in the earlier studies, crustacean chitin (extracted from crab or shrimp) was used, while we used chitin isolated directly from
A. fumigatus conidial cell wall.
In our view, the role of the complement system in host defense against pathogenic fungi has not been given enough attention. This could be due to the presumed resistance of the thick fungal CW to the complement-based MAC. In accordance, we did not see MAC formation on the
A. fumigatus conidial surface. However, opsonization is known to render its effector function, causing microbial agglutination (
19). We observed conidial aggregation upon opsonization with serum as well as BALF. Moreover, the complement system plays a critical role in host elimination of the fungal pathogens through opsonin-mediated phagocytosis and facilitation of the inflammatory response (
10,
11,
30). In agreement with this, we observed that conidial opsonization resulted in a significant increase in phagocytosis of conidia as well as their killing.
The receptors involved in fungal recognition dictate immune responses. It has been shown that Dectin-1 binds to BG to recognize fungi (
31) and that Toll-like receptor-2 (TLR2) is implicated in the cross talk between fungal cell wall α-(1,3)-glucan and human dendritic cells (
32). Although β- and α-(1,3)-glucans are the major components of the conidial cell wall, they are covered by the melanin and rodlet layers (
1,
8). Recently, we showed that the C-type lectin MelLec, expressed by human myeloid immune cells, recognizes the
A. fumigatus conidial surface melanin pigment (
12). Nevertheless, the study of
MELLEC knockout mice and single nucleotide polymorphism analysis in humans indicated that this receptor is implicated in the dissemination of the fungal infection. Moreover, we also showed that Dectin-1 inhibition only partially blocks BG uptake by hMDM (
33). Since immunoglobulins are rich in BALF (
34), we suspected the involvement of FcR-γ (receptors recognizing immunoglobulin G) in the uptake of immunoglobulin-opsonized conidia. However, there was no difference in the uptake of opsonized conidia by hMDM with all three FcR-γ blocked (CD16 [FcR-γIII], CD32 [FcR-γII], and CD64 [FcR-γI] with respective monoclonal antibodies) and opsonized conidial uptake by hMDM, in agreement with the
in vivo data showing that conidial uptake by alveolar neutrophils of FcR-γII knockout mice was comparable to that of wild-type mice (
35). Thus, it was still unclear which receptors were involved in conidial recognition. Complement receptors CR3 and CR4 are the receptors known to recognize activated C3 fragments (
36). In our study, upon blocking these two CRs on hMDM, we observed a significant decrease in phagocytosis of opsonized conidia, suggesting that CR3 and CR4 are the major receptors involved in the recognition of
A. fumigatus conidia, facilitating their phagocytosis. However, there was a significant (∼35%) amount of phagocytosis of unopsonized as well as opsonized conidia even after complement receptor blockage, suggesting that other receptors are involved.
A common practice in performing
in vitro cell culture assays is to supplement medium with serum. However,
A. fumigatus is an airborne pathogen, and its conidia first encounter the alveolar environment. With a panel of five cytokines, we showed here that the levels of induction of cytokines when hMDM encounters conidia opsonized with serum and BALF are different. Levels of TNF-α, IL-6, and IL-8 secretion were higher and the level of secretion of IL-1β was lower with serum-opsonized conidia than BALF-opsonized conidia, with no significant difference in IL-10 secretion. TNF-α, IL-1β, and IL-6 are involved in the proinflammatory response. IL-1β contributes to the augmentation of antimicrobial properties of phagocytes as well as to the differentiation of T cells in to Th1/Th17 cells (
37,
38) and expansion of Th17 cells (
39), and TNF-α maintains a normal innate immune response when an infection is encountered (
40,
41), whereas IL-6 has been shown to induce IL-17 production upon
Aspergillus infection (
42). Thus, a significantly higher secretion of IL-1β than other cytokines by hMDM upon interaction with BALF-opsonized conidia may indicate a beneficial role in the clearance of
A. fumigatus conidia through multiple axes. Classically, inflammasomes are thought to be critical for the release of IL-1β; however, IL-1β release could also be TLR mediated (
38). Also, it has been demonstrated that TLR2 aggregates at the site of conidial phagocytosis (
43). Therefore, even after the blockage of complement receptors, a significant conidial phagocytosis by hMDM leads us to speculate that there might be cross talk between complement proteins and/or other humoral immune components and TLRs.
We demonstrated that the
A. fumigatus conidial surface rodlet layer masks conidial recognition by immune cells (
1). In agreement, there was no cytokine production when paraformaldehyde (PFA)-fixed (inactivated) conidia were made to interact with hMDM in a medium supplemented with human serum, suggesting that metabolically active conidia are essential for phagocytes to mount an antifungal defense mechanism. CR3-mediated phagocytosis has been considered a silent mode of entry for pathogens, resulting in limited induction of proinflammatory cytokines (
44,
45). In line with this, even at an hMDM-conidium ratio of 2:1, the rate of secretion of proinflammatory cytokines by hMDM in our study system was not very high (the positive-control lipopolysaccharide [10 ng/well] used in our study resulted in the secretion of 6,792 ± 69 pg/ml of TNF-α, 1,503 ± 58 pg/ml of IL-6, and 6,789 ± 332 pg/ml of IL-8), suggesting that
A. fumigatus conidia may also utilize a complement receptor-mediated route of phagocytosis for a silent entry. On the other hand, it could be the metabolic activeness of conidia and recognition of pathogen-associated molecular patterns in the phagolysosome following conidial swelling that stimulate phagocytes to produce reactive intermediates, a host defense mechanism that results in conidial killing. It should be noted that hMDM could take up a significant number of unopsonized conidia; nevertheless, conidial killing and ROS production were lower under these conditions. Thus, our study suggests that
A. fumigatus conidial phagocytosis and killing are two independent processes, in agreement with the earlier observation (
23).
Altogether, our data indicate that the complement system is activated on the
A. fumigatus conidial surface through the alternative pathway, facilitating conidial opsonization, aggregation, and phagocytosis.
Aspergillus fumigatus conidial interaction with the complement system in the alveolar environment results in the activation of phagocytosis and enhanced antimicrobial properties, facilitating conidial clearance in healthy hosts. However, we could not identify the functional role played by some of the complement proteins interacting with conidia in our study, for example, MBL. This could be due either to competition of MBL with the structurally similar lung collectins SP-A and SP-D, which, as shown here, interact with conidia, or to the absence of MBL in the BALF from uninfected lungs (
46). However, MBL has been reported to be the activator of complement in the presence of low immunoglobulin levels in aspergillosis (
15). Furthermore, genetic polymorphism leading to MBL deficiency has been reported to be associated with chronic pulmonary and severe invasive aspergilloses (
47,
48). A significantly lower level of serum MBL was found in invasive-aspergillosis patients than in controls, suggesting an association between MBL deficiency and invasive aspergillosis (
48). Interestingly, mRNA transcripts related to complement components C3 and CFB in primary human bronchial epithelial cells were downregulated upon
A. fumigatus infection (
49). These observations demand further investigation of the role played by the complement and humoral immune system against
A. fumigatus during infection.
ACKNOWLEDGMENTS
This work was supported by the Centre Franco-Indien pour la Promotion de la Recherche Avancee (CEFIPRA) grant no. 5403-1. S.S.W.W. and I.D. were supported by a CEFIPRA fellowship; S.S.W.W. was also supported by a Pasteur-Roux-Cantarini fellowship.
We thank Thierry Fontaine (Institut Pasteur, Fungal Biology and Pathogenicity Unit) for providing the galactomannan extracted from the A. fumigatus plasma membrane fraction. We also thank Françoise Dromer, Stéphane Bretagne (Institut Pasteur, Molecular Mycology Unit, Paris), and Arvind Sahu (National Centre for Cell Sciences, Pune, Maharashtra, India) for their constructive comments on our study.
V.A., Conceptualization, Project administration, Supervision, Visualization, and Writing (original draft); S.S.W.W., I.D., J.M.J., S.D., R.S., J.B., and V.A., Investigation and Methodologies; J.-P.G., H.G., T.M., and J.I.G., Resources; V.A., J.M.J., P.L., and D.K., Funding acquisition; S.S.W.W., I.D., J.M.J., S.D., J.B., and V.A., Data curation and formal analysis; all authors, Validation, Writing (editing and reviewing the draft manuscript). The final version of the manuscript was reviewed, edited, and approved by all the authors.
We have no conflict of interest to disclose.