Candida species are fungi that have a major role in human pathology. In healthy individuals, Candida species are commensal in nature and colonize mucous membranes and the skin1. However, they can cause severe invasive disease when tissue homeostasis is disrupted — for example, in patients with neutropenia, pancreatitis or renal insufficiency — or following treatment with glucocorticosteroids, systemic antibiotics, indwelling medical devices, total parenteral nutrition or major abdominal surgery1,2,3. In patients with invasive candidiasis, treatment with antifungal drugs has shown only partial success in improving prognosis, and it is believed that only adjunctive immunotherapy could further improve the outcome of these infections.

In this Review, we provide a detailed account of the interaction of Candida species with the immune system. We describe the various pattern recognition receptors (PRRs) that are involved in sensing Candida species and explain how both innate and adaptive immune cells, as well as non-immune cells, contribute to the antifungal response. In addition, we discuss the emerging data suggesting that the host may develop innate (or 'trained') memory in addition to the well-known adaptive memory responses to Candida species.

Most studies to date have investigated host defences against Candida albicans, which is the most abundant Candida species in humans; for this reason, we focus on C. albicans here. Finally, we explain the consequences of the Candida–host immune system interaction and discuss the future challenges for the field.

Innate sensing of Candida species

The first step in the development of an immune response to Candida species is the recognition of invading fungi via PRRs. Although there are differences in the ways in which innate immune cell populations recognize Candida species, the general framework is identical and involves the recognition of conserved pathogen- associated molecular patterns (PAMPs) by several families of PRRs, including the Toll-like receptors (TLRs), C-type lectin receptors (CLRs), NOD-like receptors (NLRs) and RIG-I-like receptors (RLRs) (Fig. 1).

Figure 1: Recognition of Candida species by innate immune cells.
figure 1

Ligand binding to extracellular Toll-like receptors (TLRs), such as TLR2 and TLR4, leads to the production of pro-inflammatory cytokines during Candida infections. The intracellular TLRs that recognize nucleic acids — namely, TLR3 and TLR9 — might also have a role in anti-Candida host responses. Chitin from Candida albicans has been proposed to induce the production of interleukin-10 (IL-10) via a nucleotide-binding oligomerization domain-containing protein 2 (NOD2)-dependent mechanism and in this way may contribute to dampening pro-inflammatory host responses during fungal infection. The pattern recognition receptors (PRRs) dectin 1, dectin 2 and dectin 3, and Fc receptors for IgG (FcγRs), induce responses in a spleen tyrosine kinase (SYK)-dependent manner, whereas the signalling pathways engaged by the mannose receptor remain unknown. Dectin 1 can interact with TLR2 and can induce intracellular signalling via SYK- and RAF1-dependent pathways. Complement receptor 3 (CR3) is important for the recognition of unopsonized Candida, whereas FcγRs are important for recognition of opsonized Candida by neutrophils. Dendritic cell (DC)-specific -ICAM3-grabbing non-integrin (DC-SIGN) recognizes N-linked mannans of Candida and has a role in inducing T helper (TH) cell responses. There is no known Candida-derived ligand that triggers the C-type lectin receptor MINCLE, whereas β-mannans from Candida are recognized by galectin 3. Although a role for melanoma differentiation-associated protein 5 (MDA5) in anti-Candida host responses has been described, it remains to be determined what ligand induces MDA5 activation. Together, these signalling pathways induce the secretion of cytokines and chemokines and initiate phagocytosis to clear Candida infections. CARD9, caspase activation and recruitment domain-containing 9; C. glabrata, Candida glabrata; NF-κB, nuclear factor-κB; PAMP, pathogen-associated molecular pattern; PKCδ, protein kinase Cδ; ROS, reactive oxygen species.

PowerPoint slide

In the C. albicans cell wall, two layers can be distinguished: the outer layer is mainly composed of O- and N-linked glycoproteins that consist of 80–90% mannose, whereas the inner cell wall contains the skeletal polysaccharides chitin, β-1,3-glucan and β-1,6-glucan, which confer strength and shape to the cells. These polysaccharide structures, which have been reported to differ between yeasts and hyphae4,5,6, represent the main PAMPs recognized by host PRRs during an encounter with the fungus7, as we discuss below.

Fungal sensing by CLRs

Sensing of β-glucans. CLRs are the most important family of innate receptors for the recognition of Candida species. Crucial components of the Candida cell wall recognized by CLRs are the β-1,3- and β-1,6-glucans. Interestingly, structural differences between the β-glucans found in yeasts and those found in hyphae have been reported — with hyphal β-glucans reported to have a 'closed shape' structure — leading to differences in their immunological properties5. β-glucans are shielded from immune recognition by the mannoproteins in C. albicans yeast but become exposed on the budding yeast and in Candida hyphae8. The most well-studied β-glucan receptor is dectin 1 (also known as CLEC7A), a CLR expressed mainly on monocytes and macrophages that induces cytokine production, as well as internalization of the fungus by formation of a 'phagocytic synapse' (Refs 9,10,11). Dectin 1 induces intracellular signals through a pathway mediated by spleen tyrosine kinase (SYK), caspase activation and recruitment domain- containing 9 (CARD9) and protein kinase Cδ12,13,14,15,16, as well as through the RAF1 kinase signalling pathway17. In addition to inducing direct cellular activation, engagement of dectin 1 amplifies responses to TLR2 and TLR4 ligation; these TLRs recognize mannan-containing structures of C. albicans cell wall, as discussed below18,19,20. Recently, another important biological function of dectin 1 has been described. Signalling via dectin 1 is reported to prevent the uncontrolled release of neutrophil extracellular traps (NETs) during fungal infection. Importantly, this prevents extensive tissue damage from occurring during immune responses to fungi21. Interestingly though, not all C. albicans strains are recognized by dectin 1, most likely owing to subtle strain-related differences in the structure of the β-glucan components of the cell wall, which may explain the different susceptibility of dectin 1-deficient hosts to different Candida strains22.

Polymorphisms in dectin 1 are associated with colonization of the genitourinary tract by Candida species, recurrent vulvovaginal candidiasis in humans and other human fungal infections such as aspergillosis23,24,25,26. However, one study found that they are not associated with candidemia (that is, systemic fungal infections) in humans, although more studies are needed24. By contrast, the downstream adaptor molecule CARD9 is essential for systemic antifungal host defence. CARD9-deficient mice are more susceptible than wild-type mice to invasive candidiasis14, and patients with loss-of-function mutations in CARD9 also show increased susceptibility to invasive candidiasis, particularly candidal meningitis15,27. The association of CARD9 deficiency with a much more severe phenotype than dectin 1 deficiency is probably due to the fact that CARD9 also mediates signalling downstream of CLRs other than dectin 1 (Ref. 28).

β-glucans are also recognized by complement receptor 3 (CR3); this receptor is mainly involved in the recognition of β-glucans by neutrophils29. CR3 is used by neutrophils to phagocytose and kill unopsonized C. albicans, whereas Fc receptors for IgG (FcγRs) facilitate the killing of opsonized fungi30. Although both of these pathways of Candida killing by neutrophils require SYK activation, they are independent of dectin 1, although CARD9 is essential for neutrophil killing of unopsonized fungi30. Interestingly, the expression of interleukin-1 (IL-1) receptor antagonist (IL-1RA), which is an anti-inflammatory member of the IL-1 cytokine family, by β-glucans is independent of both dectin 1 and CR3; an unidentified pathway or receptor that specifically activates AKT kinase is expected to be involved31.

Recognition of mannans and mannoproteins by CLRs. Mannans and mannoproteins are important components of the Candida cell wall. Virulence and immune recognition of C. albicans are dependent on modulation of the mannoprotein fibril length by the mannosyltransferase MNN2, and hence, deficiencies of this enzyme have an impact on virulence32. Mannans and mannoproteins are recognized by several CLRs, including mannose receptor, dectin 2 (also known as CLEC6A), dendritic cell (DC)-specific ICAM3-grabbing non-integrin (DC-SIGN; also known as CD209) and MINCLE (also known as CLEC4E)33,34,35,36. The mannose receptor is primarily expressed on macrophages and recognizes Candida N-mannan37; signalling via this pathway has an important role in the expression of pro-inflammatory cytokines, especially IL-17 (Ref. 35). Dectin 2, which is mainly expressed on dendritic cells (DCs), macrophages and neutrophils, recognizes Candida α-mannan34. In addition to its role in modulating T helper 17 (TH17) cell responses, dectin 2 has been associated with the production of reactive oxygen species (ROS), and with phagocytosis and killing of Candida glabrata, which is the second most common strain of Candida38. Dectin 2 has been reported to form heterodimers with dectin 3, leading to pro-inflammatory responses, such as the production of tumour necrosis factor (TNF), IL-1β and IL-6, during C. albicans infection39. Galectin 3 on macrophages recognizes β-mannans40 and induces a protective antifungal response in mouse macrophages through secretion of TNF. Hence, mice deficient in galectin 3 are more susceptible to disseminated candidiasis41. MINCLE is a CLR expressed on both monocytes and neutrophils, and it is responsible for inducing protective responses against C. albicans, mainly by initiating TNF production42. DC-SIGN is present on DCs as well as macrophages and recognizes Candida N-linked mannan33; activation of DC-SIGN promotes adaptive immune responses by inducing the expression of cytokines that drive the activation and differentiation of TH cells. Finally, mannose-binding lectin (MBL; also known as MBPC) is a soluble CLR that also binds mannan and modulates the recruitment of phagocytes and pro-inflammatory responses against Candida species43. The scavenger receptors CD36 (also known as platelet glycoprotein 4) and SCARF1 have also been reported to recognize Candida species44.

Fungal recognition by TLRs and RLRs

TLRs have been extensively studied in the context of fungal infection. In addition to membrane-bound TLRs such as TLR2, TLR4 and TLR6, which mainly recognize mannoprotein constituents of the fungal cell wall37, it has become apparent that the intracellular receptors that recognize cytoplasmic nucleic acids — namely, TLR3 and TLR9 — may also have a role in anti-Candida host defence. TLR3 has a protective role against Candida species, as expression of the mutated L412F variant of TLR3 leads to reduced activation of nuclear factor-κB (NF-κB) and decreased levels of interferon-γ (IFNγ), resulting in increased susceptibility to cutaneous candidiasis in humans45. Furthermore, chitin is recognized by TLR9 resulting in the anti-inflammatory cytokine production that is needed to maintain a balanced immune response46.

RLRs are another family of PRRs that recognize cytoplasmic nucleic acids, and they are important for viral recognition47. A recent study has described the involvement of the RLR melanoma differentiation-associated protein 5 (MDA5; also known as IFIH1), an intracellular RNA sensor, in the recognition of C. albicans, with polymorphisms in this receptor influencing susceptibility to disseminated candidiasis in humans48.

Fungal sensing by NLRs

The role of NLRs in pattern recognition. NLRs are cytoplasmic receptors that fulfil several important biological functions, including the recognition of bacterial peptidoglycans, antigen processing and presentation, and activation of the inflammasomes. The main recognition function of NLRs in the context of Candida infection is represented by the recognition and mediation of chitin-mediated responses, especially the production of IL-10 in the context of interaction between nucleotide-binding oligomerization domain-containing protein 2 (NOD2), mannose receptor and TLR9 (Ref. 46).

The role of NLRs in inflammasome activation and IL-1β processing. Another important biological effect of NLRs is their function as components of the inflammasome. The best-characterized inflammasome is the NOD-, LRR- and pyrin domain-containing 3 (NLRP3) inflammasome, which activates caspase 1 and processes pro-IL-1β and pro-IL-18 into their biologically active forms49. Activation of the NLRP3 inflammasome by Candida hyphae, but not yeasts, has been reported in several studies50,51,52, and this has been proposed to be important for enabling the host to discriminate between Candida colonization and invasion53. The unmasking of β-glucan in hyphae due to ineffective mannan fibril formation leads to the dectin 1- dependent activation of the inflammasome54 (Fig. 2). Other possible activators of the inflammasome are the secreted aspartic proteases Sap2 and Sap6 (Ref. 55). Mice deficient in NLRP3 are unable to activate caspase 1, and hence their macrophages release less bioactive IL-1β; these mice are therefore more susceptible to Candida infections50,51,56. Recent studies have reported that in addition to caspase 1, a non-canonical caspase 8 inflammasome is important for the processing of pro-IL-1β in response to C. albicans57. Another study reported that caspase 8 activation is in turn important for the NLRP3-dependent maturation of IL-1β maturation following recognition of Candida species58. In addition, activation of the NLRP3 inflammasome by C. albicans has been described to trigger pyroptosis in macrophages59,60. This represents a second mechanism of macrophage damage by the fungus, in addition to hyphae-induced mechanical disruption of membrane integrity. Finally, it was recently proposed that the NOD-, LRR- and CARD-containing 4 (NLRC4) inflammasome has a role in mucosal defence against C. albicans through its induction of pro-inflammatory cytokines and antimicrobial peptides61.

Figure 2: Interleukin-1β processing during Candida infection.
figure 2

Activation of NOD-, LRR- and pyrin domain (PYD)-containing protein 3 (NLRP3) by Candida hyphae or secreted aspartic protease (Sap) proteins triggers the assembly of the inflammasome with subsequent activation of caspase 1, which cleaves pro-interleukin-1β (pro-IL-1β) and pro-IL-18 into bioactive cytokines. In addition, caspase 8 has also been demonstrated to cleave pro-IL-1β into active IL-1β in response to Candida. When pro-IL-1β is released in the inflammatory environment where neutrophils are present, it can also be cleaved into bioactive IL-1β by neutrophil-derived serine proteases, such as proteinase 3. Candida itself can also contribute to IL-1β activity as it can release proteases that cleave pro-IL-1β. CARD, caspase activation and recruitment domain; NBD, nucleotide-binding domain.

PowerPoint slide

However, it is important to note that inflammasome activation is not the only mechanism through which pro-IL-1β can be processed and activated (Fig. 2). Primary human monocytes express constitutively active caspase 1 (Ref. 62), and hence inflammasome activation in these cells is not needed to process pro-IL-1β63. Neutrophil-derived serine proteases such as proteinase 3 can also cleave pro-IL-1β64,65. As neutrophils are the main cellular component in inflammatory infiltrates in the tissues during disseminated candidiasis66, these enzymes may account for the generation of active IL-1β during the initial stages of Candida infection67. Indeed, neutrophil-derived proteinase 3, rather than caspase 1, is likely to have an important role in protecting mice against disseminated candidiasis68. Finally, during infection with C. albicans, a fungus-derived protease can lead to the processing and activation of host-derived pro-IL-1β and thus can activate of the immune system69 (Fig. 2).

Recognition of yeasts versus hyphae

The capacity to undergo reversible yeast to hyphae morphogenesis is linked to the virulence of C. albicans53,70. Studies of differences in the stimulation of cytokines by yeasts and hyphae have shown that hyphae are incapable of inducing IL-12 expression in DCs, resulting in a tolerant phenotype71, and this was attributed to the lack of TLR4-mediated recognition of the hyphae72. In addition, the differential exposure of β-glucans on the surface of yeasts and hyphae has been proposed to account for differences in stimulation of cytokines, with β-glucans being exposed on the bud scars of yeasts but not on hyphae73. However, other studies suggest that recognition of hyphal β-glucans can still be mediated by dectin 1, perhaps due to the shorter and less abundant mannan fibrils on the surface of hyphae54. The loss of shielding by mannans and thus the exposure of other immunostimulatory PAMPs could account for the difference in inflammasome activation by hyphae and yeast in tissue macrophages. Strikingly, inflammasome activation is induced by Candida hyphae but not yeasts. This reflects an important mechanism of discrimination between colonization and invasion53. Finally, differential recognition of mannans from hyphae and yeasts by dectin 2 has been proposed34, but not fully confirmed, in additional studies33. Newer data suggest that the chemistry of cell wall polysaccharides differs significantly between C. albicans yeasts and hyphae, and this may add an additional mode of discrimination5.

When considering the importance of the dimorphism between yeast and hyphae for the virulence of Candida species, it should be mentioned that C. glabrata is an exception, as it does not germinate into hyphae but yet is virulent. Much remains to be learned about the particularities of the infection by C. glabrata and about the immune response to this pathogen (for a review of this topic see Ref. 74).

Effector mechanisms of defence

After the initial recognition of fungal PAMPs by the various families of PRRs, a chain of effector mechanisms is initiated that ultimately leads to the clearance of the invading fungi (Fig. 3). Below, we summarize the immune and non-immune cell populations that contribute to the antifungal response.

Figure 3: Effector mechanisms for the clearance of Candida.
figure 3

The epithelium provides a mechanical barrier against invading Candida. However, Candida can invade the tissue by inducing endocytosis or actively penetrating the epithelial cells. Epithelial cells produce cytokines via a mitogen-activated protein kinase 1 (MAPK1)- and FOS-dependent pathway when Candida hyphae are present, which leads to the recruitment of phagocytic immune cells. Epithelial cells can also respond by producing β-defensins that have direct anti-Candida activity. After Candida has penetrated the tissue, it will first encounter the tissue resident-macrophages, which will phagocytose and clear Candida. Upon invasion, neutrophils are recruited by pro-inflammatory cytokines produced by macrophages and epithelial cells. In addition to their capacity to kill Candida by phagocytosis and the production of reactive oxygen species (ROS), neutrophils can also release neutrophil extracellular traps (NETs) that capture Candida conidia and hyphae and contain antimicrobial proteins such as calprotectin that inhibit fungal growth. Inflammatory monocytes are also recruited to the site of infection via CC-chemokine ligand 2 (CCL2) and will contribute to clearing Candida. Dendritic cells (DCs) can migrate to the lymph nodes and contribute by shaping adaptive T helper (TH) cell responses. TH17 cell responses have an important role in mucosal host defences against Candida by producing interleukin-17 (IL-17) that recruits and activates neutrophils, and IL-22 that induces the secretion of β-defensins. TH1 cell-derived interferon-γ (IFNγ) strongly activates phagocytic cells. Innate lymphoid cells also have the capacity to produce pro-inflammatory cytokines and contribute to mucosal antifungal defence. The last step of Candida invasion is breach of the endothelium, allowing access to the bloodstream. In addition to neutrophils, systemic Candida can activate platelets that can produce CCL5 and platelet factor 4 (PF4), which both have anti-Candida activity. CCR2, CC-chemokine receptor 2; IL-22R, IL-22 receptor; NF-κB, nuclear factor-κB; NK, natural killer; TNF, tumour necrosis factor.

PowerPoint slide

Epithelial cells. An intact epithelium and endothelium constitute important mechanical barriers against tissue invasion by fungi. Tissue invasion is associated with morphological changes from yeast into hyphae, with Candida hyphae penetrating the epithelial cells through two distinct mechanisms: induced endocytosis and active penetration. Epithelial cells respond to Candida species colonization through a TLR4-dependent mechanism75. This leads to the activation of NF-κB and JUN (also known as AP-1), and this response is independent of fungal morphology. By contrast, when Candida germinates and forms hyphae it also activates mitogen-activated protein kinase 1 (MAPK1) and FOS signalling in epithelial cells, a mechanism that permits sensing of tissue invasion76,77. It remains to be determined which Candida cell wall components and which host receptors are important for mediating these different epithelial cell responses to yeasts and hyphae. Activation of the second phase of the response is accompanied by the release of pro-inflammatory cytokines and initiation of a host response. In addition, epithelial cells contribute to controlling the commensal state of Candida by producing β-defensins, which have potent antifungal activity, in response to the IL-22 that is released by TH17 cells or innate lymphoid cells (ILCs)78,79.

Monocytes and macrophages. Tissue-resident macrophages are key effector cells that function in antifungal defence. These macrophages produce inflammatory cytokines and chemokines that recruit and activate other immune cells at the site of infection (Fig. 3). The relevance of the macrophage lineage for anti-Candida host defence was demonstrated by early in vivo studies in macrophage-depleted mice; these animals showed accelerated fungal proliferation in tissues and increased mortality80. By contrast, neutrophil depletion in the blood did not reverse inhibition of Candida growth in the blood81, despite the important role that neutrophils have in the elimination of the fungus in organs (see below). Therefore, tissue-resident macrophages exert potent candidacidal activity81,82.

Blood monocytes are also recruited to the infected tissue where they differentiate into inflammatory macrophages. In mice deficient in CX3C-chemokine receptor 1 (CX3CR1), impaired accumulation of monocyte-derived macrophages in the kidney leads to renal failure and death82. Moreover, patients with a polymorphism leading to a decreased CX3CR1 function are more susceptible to disseminated candidiasis82, whereas their susceptibility to mucosal Candida infections is not affected83. Furthermore, both human blood classical CD14hiCD16 monocytes and non-classical CD14+CD16low monocytes have candidacidal activity84, and deficiency of CC-chemokine receptor 2 (CCR2; which is essential for monocyte recruitment to inflamed tissues) leads to increased susceptibility to disseminated candidiasis81. These data highlight the essential role of monocytes and macrophages in controlling disseminated fungal infection.

Neutrophils. Neutrophils are of major importance in host defence against Candida infections. Activated epithelial cells and tissue-resident macrophages release chemokines that recruit neutrophils to the site of fungal infection37. Neutrophil activation is essential for clearance of Candida, with neutropenia being a major risk factor for invasive fungal infections85,86. Neutrophils are the most potent killers of Candida and are the only host cells that are capable of successfully inhibiting the germination of Candida yeasts into hyphae87. Neutropenic mouse models have clearly demonstrated the crucial role of neutrophils in disseminated fungal infection66.

Neutrophils use both oxidative and non-oxidative effector mechanisms to kill Candida88. Although both mouse and human granulocytes that are deficient in either NADPH oxidase or myeloperoxidase (MPO) are incapable of efficient Candida killing in vitro89,90, NADPH oxidase deficiency in patients with chronic granulomatous disease is associated with significantly increased susceptibility to invasive mould infection, but it has little effect on susceptibility to Candida infection. This suggests that alternative mechanisms in vivo can compensate for a defect in NADPH oxidase-dependent killing mechanisms. Similarly, MPO deficiency in humans does not predispose to Candida infection, unless there are concomitant risk factors (such as diabetes)91.

In addition to ROS, neutrophils also use non-oxidative mechanisms to kill Candida; they produce antimicrobial factors such as lysozyme, lactoferrin, elastase, β-defensins, gelatinases and cathepsin G92. Neutrophil elastase and cathepsin B have been described to have antifungal activity, and elastase was found to contribute to the release of NETs that form large DNA-containing fibril structures93. These fibrils bind and neutralize pathogens, providing a mechanism to combat hyphae that are too big to be phagocytosed94. Calprotectin was also identified as an important component of NETs95, and NET formation induces the release of antimicrobial substances (for example, MPO, lactoferrin, azurocidin and cathelicidin) from the granules of the neutrophils. Proteinase 3 derived from these cells cleaves cathelicidin into the antimicrobial peptide LL-37 (also known as CAMP)96, which has a number of antimicrobial effects: it promotes disruption of the fungal cell membrane97,98, inhibition of biofilm formation and fungal adhesion99, enhanced chemotaxis, production of ROS and inhibition of neutrophil apoptosis100,101. Very recently, it has been shown that simultaneous exposure of Candida to oxidative and cationic stresses synergizes with the fungicidal capacity of human neutrophils102. With regard to the role of autophagy, a process by which cells can clear unwanted cytosolic content, there is some controversy. Although an initial study reported that mice deficient in autophagy are more susceptible to Candida infection103, recent studies have shown that autophagy is most likely to be redundant in host defence against Candida104,105.

The receptor pathways that lead to Candida killing by neutrophils have recently been elucidated. The ROS-dependent mechanisms that are needed for clearance of opsonized Candida depend on FcγRs and protein kinase C, whereas the ROS-independent pathway — which is important for killing unopsonized Candida — relies on CR3 ligation and CARD9 recruitment30. Interestingly, dectin 1 is dispensable for both of these mechanisms. A recent study demonstrated that glycosylation of proteins is important for the fungicidal function of neutrophils: defects in the Jagunal homologue 1 protein result in impaired responses to C. albicans and an increased susceptibility to infection106.

Natural killer cells. In addition to neutrophils, monocytes and macrophages, natural killer (NK) cells contribute to the rapid innate immune response against invading pathogens. Depletion of NK cells had variable results in mouse models of Candida, with either a lack of effect107 or an increase in susceptibility to disseminated candidiasis12. The additional depletion of NK cells in lymphocyte-deficient severe combined immunodeficiency (SCID) mice increased the susceptibility to systemic candidiasis, whereas such depletion had no effect in immunocompetent mice108. Although NK cells do not inhibit hyphal growth of Candida, perforin-dependent antifungal activity by NK cells has been reported109.

Activation of a potent innate immune response by epithelial cells and phagocytic cells, possibly aided by NK cells, is in most cases enough to counteract displacement of Candida species from surface colonization to tissue invasion and to prevent a disseminated infection. However, for situations in which the infection is not promptly controlled by the innate immune mechanisms, adaptive immunity is activated to establish an additional layer of defence.

DCs and activation of T cells. DCs have been shown to be essential for the host response against Candida species through their production of IFNβ via a SYK- and IFN-regulatory factor 5 (IRF5)-dependent pathway110,111,112. Although DCs can also ingest and kill Candida species, they are less efficient than macrophages at fungal killing113. Instead, DCs are important for processing and presenting fungal antigens for the activation of TH cell responses.

Candida-specific TH cells in humans were found to produce a combination of IL-17 and IFNγ114. The importance of TH1 cell responses and IFNγ production for the fungicidal activities of both neutrophils115 and macrophages116 is well established. IFNγ-deficient mice are more susceptible to disseminated candidiasis, as are mice defective in the pro-inflammatory cytokine IL-18, which induces TH1 cell responses117. Treatment of mice with either recombinant IFNγ118 or IL-18 (Ref. 119) protects them against systemic candidiasis. In line with this, a recent proof-of-principle trial in patients with systemic candidiasis demonstrated improvement of immunological parameters after treatment with recombinant IFNγ120.

TH17 cell production of IL-17 and IL-22 is also important for host defence against Candida species. These cytokines induce neutrophil recruitment and activation, and are responsible for the activation of epithelial cells and the release of antifungal β-defensins78, often in a cooperative manner121. Several studies have shown that mice deficient in IL-17 signalling are more susceptible to both systemic candidiasis122,123 and mucosal infections124. However, patients with a deficiency of IL-17 production or signalling owing to mutations in IL-17F, IL17RA (which encodes IL-17 receptor A), DECTIN1, STAT1 (which encodes signal transducer and activator of transcription 1) or STAT3 suffer from mucosal candidiasis but not invasive candidiasis. This suggests that in humans, TH17 cell responses are mainly necessary for mucosal antifungal responses25,125,126,127,128. TH17 cells also seem to be less important for the control of vaginal candidiasis than for the control of other forms of mucosal infection. In mice, the recruitment of neutrophils in response to vaginal epithelial cell-mediated production of S100 alarmins was found to be independent of IL-17 production129, and patients with defects in TH17 cell responses do not suffer from recurrent vulvovaginal candidiasis126,130. Moreover, the importance of IL-17 for the host defence against Candida infections has been underlined by the increased number of infectious complications seen in patients with psoriasis who have been treated with IL-17A-targeted antibodies131.

In contrast to these TH1- and TH17-type immune responses that are important for a protective host response to Candida species, cytokines associated with TH2-type immunity have been shown to have conflicting roles. On the one hand, therapeutic ablation of IL-4 or IL-10 signalling in wild-type mice increased their resistance to systemic candidiasis132, and IL-10-deficient mice showed greater resistance to candidiasis, suggesting a detrimental role for these cytokines133,134. On the other hand, other studies showed that IL-4 is necessary for the development of a protective host response to Candida infection135 and that early IL-10 production contributes to the development of protective TH1 cell responses in IL-12-deficient mice136.

Antifungal roles for ILCs. In addition to the classical TH cell subsets, the recently described group of ILCs may also contribute to immunity to Candida species (Fig. 3). Three groups of ILCs have been proposed: ILC1s that express T-bet and produce IFNγ; ILC2s that express GATA-binding protein 3 (GATA3) and produce IL-5 and IL-13; and ILC3s that express nuclear receptor RORγt and produce IL-17 and IL-22 (Ref. 136). Evidence that ILCs may help to control fungal infection and colonization came from experimental mouse models of mucosal infections, in which fungal control of oral candidiasis was found to be mediated by IL-17-secreting ILCs. Both ILC-depleted Rag1−/− mice (which lack expression of recombination activating gene 1) and Rorc−/− ILC-deficient mice failed to control mucosal Candida infection137. However, it is still unclear how ILCs contribute to the immune response during invasive fungal infections. In addition to the ILCs, other 'innate-like' lymphocyte subpopulations may have a role in host defence against Candida species, especially at the level of skin and mucosa; γδ T cells are one such cell type that releases protective IL-17 (Ref. 138).

Humoral antifungal mechanisms. Humoral immune mechanisms have also been suggested to be involved for host defence against Candida infections, although their contribution to antifungal defence is likely to be more modest than the cellular mechanisms discussed above. Although activated complement cannot kill Candida hyphae during infection, it can contribute to the induction of a proper cytokine response139. Mice with a deficiency of either complement factor C3 or C5 exhibit increased mortality as a consequence of impaired anti-Candida resistance or excessive exuberant infection-driven immunopathology, respectively140,141. The impact of B cells and antibodies on host defence against Candida species is less clear. Although patients with agammaglobulinaemia or hypogammaglobulinaemia do not suffer from an increased susceptibility to fungal infection, eliciting protective antibodies through vaccination has been proposed as a viable strategy for improving resistance to infection142,143,144,145.

Platelets. Less well-appreciated players in antifungal host defence are platelets. When C. albicans are injected into mice, the fungi bind and activate platelets in the bloodstream146. Platelets produce immune mediators such as CC-chemokine ligand 5 (CCL5; also known as RANTES) and platelet factor 4, which have antimicrobial activity against Candida species, and platelet-rich plasma inhibits the growth of Candida147. More studies are needed to fully decipher the role of platelets in antifungal host defence.

Compartmentalized response to Candida species. In summary, the host defence response against Candida infections is complex and involves a cascade of mechanisms. The initial recognition of fungi by epithelial cells and tissue macrophages promotes inflammation and innate immune cell activation, and breach of epithelial barriers leads to antigen presentation and the induction of specific antifungal lymphocyte responses. Innate immune mechanisms are crucial for host defence against both mucosal and systemic candidiasis, whereas adaptive host defences are mainly involved in mucosal anti-Candida responses. However, one has to be aware of the often compartmentalized nature of the immune responses: whereas adaptive T cell responses are crucial for anti-Candida defence at the level of the intestinal mucosa, these cells have a much more restricted role in host defence in the vaginal mucosa148. Although the vast majority of the studies have been performed with C. albicans, similar mechanisms have also been described for other Candida species: although this is not the focus of this Review, other very good recent papers have described these mechanisms in detail (for reviews, see Refs 149,150).

Evasion of host defences by Candida

Although the host immune system is generally very efficient in keeping fungal infections at bay, Candida species have developed strategies to escape clearance by the immune system. One mechanism involves the shielding of PAMPs that are capable of eliciting an immune response. C. albicans hyphae are less inflammatory than the yeast forms, and this is thought to be due to the shielding of β-glucans from the surface of hyphae by mannans73. These findings also demonstrate how morphogenetic changes that occur in Candida species may modulate the immune response to the advantage of the fungus72.

Candida species are also capable of inhibiting phagolysosome maturation and the production of nitric oxide by macrophages151, although the precise molecular mechanisms underlying these effects still need to be investigated. Furthermore, C. albicans can enhance its survival by inducing macrophages to switch from a more inflammatory M1 phenotype to a less inflammatory M2 macrophage phenotype152. Interestingly, Candida species can also hijack certain PRR pathways. For example, Candida-mediated activation of TLR2 can induce immunomodulatory signals that promote a tolerogenic DC profile153 and the proliferation of regulatory T cells154. Understanding these evasion strategies that interfere with successful clearance of the fungus could guide the way for the development of novel therapeutic approaches.

Insights from functional genomics

In the past decade, novel technologies such as genomics, proteomics, metabolomics and transcriptomics, as well as advances in computational analyses, have led to a better understanding of the host immune response to Candida infections155.

Comparisons of RNA profiles in the blood of mice infected with C. albicans have shown that certain inflammatory pathways (TLR4–MYD88, inflammasome and STAT3 pathways) are activated throughout all phases of infection, whereas other pathways characterize specific phases of the infection, such as caspase 3- dependent apoptosis in the late phase156. Assessment of transcriptional profiles induced by C. albicans in bone marrow-derived macrophages identified host–pathogen modules, such as MTA2–Hap3, that drive host IL-2 and IL-4 production157. Using an elegant combination of transcriptome and functional validation experiments, the same study also identified pentraxin-related protein PTX3 as an important modulator of antifungal host defence157. In humans, assessment of the transcriptome profile in vascular endothelial cells after stimulation with C. albicans revealed that chemotaxis, angiogenesis and inhibition of apoptosis were among the main processes induced158. By contrast, the interaction between C. albicans and human granulocytes induced genes involved in cell–cell signalling, cell signal transduction and cell growth159. A recent study compared the transcriptome induced by stimulation with C. albicans and different bacterial stimuli in human blood mononuclear cells: surprisingly, a strong type I IFN signature was identified to be specific for Candida stimulation160. This relatively unexpected finding was validated by genetic and functional studies and supported by independent studies in mouse models111,161. Finally, transcriptome studies have also been performed in epithelial cells stimulated with Candida albicans: these studies reported early expression in epithelial cells of genes that are involved in immune pathways. These genes included components of immune signalling pathways, such as the NF-κB, MAPK and phosphoinositide 3-kinase (PI3K)–AKT pathways162, as well as chemokines and adhesion molecules163.

Although genome-wide association studies (GWASs) have been a part of mainstream genetic research for almost a decade, comprehensive genetic studies in individuals with Candida infections have been notoriously scarce. Recently, the first GWAS on a fungal infection has been published. This study of patients with candidemia reports an association with three genes that were not previously known to be involved in antifungal defence: CD58, TAGAP (which encodes T cell activation RHO-GTPase-activating protein) and LCE4A (which encodes late cornified envelope 4A)164. CD58 has been shown to be involved in the inhibition of Candida germination, whereas TAGAP influences cytokine production induced by the fungus. In the same cohort of patients, genetic variation in MDA5 was subsequently demonstrated to also influence susceptibility to candidiasis48.

The microbiome–immune response interaction

It is becoming clear that other components of the host microbiota also affect fungal colonization and the antifungal immune response. Metagenomic analysis showed that the human microbiome contains bacterial, fungal and viral communities that vary significantly by body site and across individuals165. A first insight into this complex interaction is highlighted by the skin microbiome of patients with primary immunodeficiencies. The microbiome of patients with chronic mucocutaneous candidiasis showed a significant reduction in the prevalence of species that normally colonize the skin (such as Corynebacterium species) and an increased prevalence of Gram-negative bacteria (such as Acinetobacter species); these changes were associated with a suppressed cytokine response to C. albicans166. By contrast, Pseudomonas aeruginosa and Enterococcus faecalis inhibit hyphal growth of C. albicans167, whereas H2O2 and bacteriocin-like compounds released by lactobacilli inhibit fungal adhesion and growth168. In addition, the interaction between C. albicans and streptococci has been suggested to contribute to the colonization of the oral cavity by C. albicans169. Fungi themselves belong to the 'rare biosphere' (Ref. 170), which functions as a reservoir for potentially pathogenic microorganisms and can expand when the environment is disturbed or when the host is immunocompromised171. In the setting of dectin 1 deficiency, the mycobiome of the gut in mice is significantly altered, and this was associated with increased susceptibility to experimental colitis172. The microbiota can also influence the immune system through the release of metabolites. Tryptophan metabolites produced by lactobacilli in the gut induce IL-22 production from NKp46+NK1.1low cells via the aryl hydrocarbon receptor (AHR)173. IL-22 provides resistance to C. albicans colonization at mucosal surfaces. In addition, lactobacilli can produce short-chain fatty acids that can inhibit fungal growth174 and — via G protein-coupled receptors such as GPR41 and GPR43 — modulate host immune responses175,176. Therefore, changes in the microbiome might have a dramatic impact on Candida-induced host immune responses and disease severity177,178,179.

Trained immunity in the response to Candida

Studies performed more than two decades ago reported that infection of mice with an attenuated C. albicans strain provided protection against lethal invasive candidiasis in a T cell-independent but macrophage-dependent manner180,181. This unexpected 'innate memory' response has been referred to as 'trained immunity', and it has recently been demonstrated to be mediated by epigenetic reprogramming in innate immune cells (see the figure in Box 1)182.

C. albicans and β-glucans induce trained immunity through ligation of dectin 1 and activation of the non-canonical RAF1 signalling pathway, which leads to stable changes in histone methylation and acetylation183. In this way, the capacity of monocytes and macrophages to produce pro-inflammatory cytokines is enhanced183. The induction of trained immunity by β-glucans is dependent on a pathway mediated by AKT, mammalian target of rapamycin (mTOR) and hypoxia-inducible factor 1α (HIF1α), and on a switch of glucose metabolism from oxidative phosphorylation to aerobic glycolysis184. The clinical relevance of trained immunity for antifungal host defence has recently been suggested by the report of defective trained immunity in patients with chronic mucocutaneous candidiasis who are deficient for STAT1 (Ref. 185).

Conclusions and future perspectives

In conclusion, the past decade has witnessed a marked improvement in our understanding of the molecular and immunological mechanisms that lead to the induction of an efficient antifungal immune response. The discovery of PRRs has greatly increased our knowledge of the innate immune response, and the specific roles of TH cell subsets for systemic or mucosal immunity to Candida have become apparent. Moreover, the recent descriptions of patients with dectin 1 or CARD9 deficiency and those with defective TH17-type immune responses have underlined the crucial role of these responses for mucosal immunity to fungal pathogens.

In which areas will the main discoveries occur in the coming years? First of all, additional efforts are being made to characterize the genetic mutations that are responsible for the development of severe forms of Candida infection in patients for which no genetic defects have yet been identified. Second, the ongoing recruitment of large cohorts of patients with both invasive candidiasis and recurrent vulvovaginal candidiasis is expected to lead to a much more detailed and precise understanding of how common genetic variability in PRRs affects susceptibility to disseminated disease. Third, an increased understanding of the immunological, cellular and molecular pathways responsible for activation of host defence will be obtained both by refining our molecular research armamentarium and by improving analyses of 'omics' data through systems biology. With regard to metagenomics, we are only at the very beginning of understanding the complex interplay between bacteria, fungi and the human host, and understanding the implications of this interplay for health and disease. Another exciting area of new research is that of epigenomics. The concept of trained immunity, which is epigenetically determined, will enable us to learn how to modulate the immune response to greater benefit of patients. We should thus expect that further research in the field may lead to novel discoveries and new therapies that will have an important impact on the treatment of fungal infections.