Host mechanisms enabling discrimination between commensal and pathogenic organisms are critical in mucosal immune defense and homeostasis. Of particular interest are opportunistic microbes, like C. albicans
, that can act as either commensals or pathogens under suitably predisposing conditions. We and others have previously shown that hypha formation is crucial for C. albicans
pathogenicity and induction of proinflammatory responses at mucosal surfaces (Zakikhany et al., 2007; Korting et al., 2003
), which results in PMN-dependent TLR4-mediated protection against subsequent fungal infection (Weindl et al., 2007
). Given that the key event initiating this protective process appears to be the C. albicans
hypha-induced proinflammatory response, we sought to identify the epithelial mechanisms that discriminate between the yeast and hyphal forms of C. albicans
. We report that specific MAPK-based recognition of the filamentous form appears to initiate a danger response mechanism constituting MKP1 and c-Fos activation via a process dependent on fungal burdens. Activation of this mechanism may inform the host of when this normally commensal fungus has become invasive (pathogenic), which may have important implications for mucosal fungal infections in vivo.
Infection of oral epithelial cells with C. albicans
results in the activation of NF-κB and a biphasic MAPK signaling response, which leads to the induction of a proinflammatory response. Although the activation of NF-κB and MAPK signaling is not unexpected, as these pathways are activated in myeloid cells (Roeder et al., 2004
), our study delineates a biphasic MAPK response in epithelial cells that is morphology and burden dependent. Activation of NF-κB and the first MAPK phase appears to be independent of the morphological form and is likely due to the recognition of general fungal cell wall structures. These include chitin, mannans, and β-glucans, as all three moieties activated NF-κB and the first MAPK phase, but not the second MAPK phase. Also, activation of the first MAPK phase does not necessarily result in cytokine induction, as this was readily activated by killed C. albicans
yeast and hyphae, fungal cell wall agonists, and S. cerevisiae
, all of which were nonimmunostimulatory. This is in contrast to myeloid cells in which strong inflammatory responses are induced by killed C. albicans
and fungal agonists (Netea et al., 2006, 2008
), indicating that different (or more regulatory) innate response mechanisms exist in epithelial cells. We propose that activation of NF-κB and the first MAPK phase is independent of morphology and constitutes a general epithelial response to inform the host of the presence of fungus.
The second MAPK response involves two key proteins, the MAPK phosphatase MKP1 and the transcription factor c-Fos, and is a specific response to the presence of pathogenic fungus (i.e., C. albicans). MKP1 activation is mediated via the MEK1/2-ERK1/2 pathway, is associated with filamentation, and is contact dependent but is independent of viability and avidity of adherence. The hypha-associated moieties that induce MKP1 activation are unknown, but cell wall structural components (mannans, chitin, and β-glucans) are unlikely to be the activating factors, as these only activate NF-κB and the first MAPK phase. We propose that MKP1 is activated via cell surface moieties present on filamentous forms of C. albicans.
Activation of c-Fos is mediated via the p38 pathway and is also hypha associated, but c-Fos is activated via two independent mechanisms. The first mechanism does not require fungal viability and is probably activated via hypha-associated surface moieties, much like MKP1. The second mechanism is associated with the induction of inflammatory mediators, as c-Fos knockdown significantly reduced cytokine release after C
infection, and is likely associated with hyphal invasion/penetration and induction of cell damage (as viability promoted c-Fos activation). Our data contrast with that of myeloid cells in which c-Jun rather than c-Fos is thought to play the dominant role in mediating responses to C. albicans
(Roeder et al., 2004
) and further support the notion that epithelial response mechanisms to fungal pathogens are distinct to those of myeloid cells.
Our data suggest that different hypha-associated moieties are required to induce (1) the epithelial MAPK/MKP-1/c-Fos pathway and (2) the cytokine response. This may explain the differences we observed between the two filamentous mutants, in that Δtup1
might possess one set of surface moieties that activate the MAPK/MKP-1/c-Fos response but lacks another set that results in poor adhesion, lack of invasion, and weak cytokine production. On the other hand, Δnrg1
may possess both sets of surface moieties, similar to the SC5314 wild-type, and is therefore capable of inducing both the MAPK/MKP-1/c-Fos and the cytokine response. Tup1- and Nrg1-regulated genes have been partially identified (Murad et al., 2001a
), and recently, a set of genes was proposed as mediating hyphal virulence independently of morphology (Nobile et al., 2008
). Those studies provide several interesting candidates with regard to determining which hyphal moieties might induce the MAPK/MKP-1/c-Fos and cytokine response in epithelial cells.
Although MKP1 and c-Fos activation constitute the second MAPK response, they do not appear to be directly linked, as MKP1 is activated by ERK1/2 and c-Fos by p38. However, the function of MKP1 is to dephosphorylate p38 (and JNK) (Zhao et al., 2006; Chi and Flavell, 2008
) in order to stabilize and regulate the overall MAPK network. This is highlighted by an increase in cytokine release (G-CSF, GM-CSF) after MKP1 knockdown or inhibition, as there will be reduced dephosphorylation of p38 (and JNK), which results in an increased effector response (as cytokine induction is primarily mediated via p38 and JNK). In the context of a fungal infection, MKP1 activation will ensure that any immune response to C. albicans
hyphae is both optimal and tightly controlled while concurrently ensuring a rapid deactivation of potential deleterious p38- and JNK-mediated inflammatory responses when no longer being driven by recognition of hyphae. Given that phosphorylation and stabilization of MKP1 are mediated by ERK1/2 activity, our data imply that ERK1/2 plays an important role in the control and resolution of inflammatory responses by regulating the level of cytokine secretion due to p38 and JNK signaling, rather than by direct effects on cytokine production.
This MAPK/MKP1/c-Fos pathway may be important in vivo because a C. albicans
strain (529L) that does not form hyphae or activate MKP1, c-Fos, or cytokine production is able to colonize a murine model, whereas a C. albicans
strain (SC5314) that readily forms hyphae and strongly activates MKP1, c-Fos, and cytokine production is efficiently cleared (Rahman et al., 2007
). Likewise, increased levels of MKP1 and c-Fos expression are evident in organotypic models and human biopsies taken from Candida
-infected individuals. Although we were unable to utilize primary normal oral keratinocytes in our studies, the in vitro data using cell lines appear to be directly applicable to the in vivo situation and suggest that avoidance or activation of this MAPK-based mechanism may be crucial in determining whether a fungal strain colonizes, infects, or is cleared by the host.
Recognition of C. albicans
by monocytes/macrophages is mediated by several different PRRs, including mannose receptor, TLR2, TLR4, and dectin-1 (Netea et al., 2006
). Likewise, in dendritic cells, the initiation of an ERK1/2-c-Fos response is the result of TLR2 and dectin-1 signaling (Agrawal et al., 2003; Dillon et al., 2004, 2006
). We show that the MAPK/MKP1/c-Fos signaling mechanism in epithelial cells is unlikely to utilize the same receptors, as knockdown of TLR2 and TLR4 did not affect MKP1 or c-Fos activation or the cytokine profile. Likewise, SYK (the adaptor utilized by dectin-1) was not activated, and β-glucans (the ligand for dectin-1) did not induce cytokine responses. It has been reported that the surface receptor CDw17 may mediate epithelial responses to Candida
(Li and Dongari-Bagtzoglou, 2009
); however, this was against C. glabrata
, which does not form hyphae. Therefore, it is likely that the MAPK/MKP1/c-Fos system that mediates yeast/hyphal discrimination is through as yet unidentified epithelial PRR(s).
Of major importance is that MKP1 and c-Fos activation is dependent not only on filamentation, but also on fungal burdens, which suggests that a threshold level of stimulation is required prior to full activation of the epithelial response. This may provide a mechanism by which mucosal tissues can remain quiescent in the presence of low fungal burdens (even if hyphae are present) while responding specifically and strongly to damage-inducing hyphae as burdens increase. Therefore, in addition to identifying the switch from the yeast to hyphal form in C. albicans
, the MAPK/MKP1/c-Fos pathway may also constitute a danger response mechanism informing the host of potentially dangerous levels of this fungal pathogen. We propose that, at low burdens, C. albicans
avoids activating the MAPK-based danger response pathway, as a threshold level of activation is not reached. As a result, the fungus is regarded as “nondangerous,” thereby permitting successful colonization of human mucosal surfaces without host challenge. One could regard this as C. albicans
being in the “commensal” state. However, under conditions that permit fungal proliferation, increased fungal burdens together with associated hypha formation activate the MAPK/MKP1/c-Fos pathway to warn the host of the presence of a “dangerous” pathogen. This results in immune activation and secretion of proinflammatory cytokines, which in turn (as we have previously shown), results in PMN recruitment and fungal clearance or reduction of fungal burdens back down below the threshold level of activation and thus a return to the nonactivatory colonization commensal state (Weindl et al., 2007
). In the clinical setting, fungal proliferation and immune activation may translate into signs and symptoms of infection, and the cycle of immune activation (based on fungal burdens and hypha formation) followed by subsequent clearance may represent what occurs in patient groups that experience acute C. albicans