Interactions of Candida albicans with Epithelial Cells

Summary

The fungus, Candida albicans interacts with epithelial cells in the human host both as a normal commensal and as an invasive pathogen. It has evolved multiple complementary mechanisms to adhere to epithelial cells. Adherent C. albicans cells can invade epithelial surfaces both by penetrating into individual epithelial cells, and by degrading inter-epithelial cell junctions and passing between epithelial cells. Invasion into epithelial cells is mediated by both induced endocytosis and active penetration, whereas degradation of epithelial cell junction proteins, such as E-cadherin occurs mainly via proteolysis by secreted aspartyl proteinases. C. albicans invasion of epithelial cells results in significant epithelial cell damage, which is probably induced by lytic enzymes, such as proteases and phospholipase secreted by the organism. Future challenges include identifying the epithelial cell targets of adhesins and invasins, and determining the mechanisms by which C. albicans actively penetrates epithelial cells and induces epithelial cell damage.

Introduction

Candida albicans and limited number of other Candida species are part of the normal commensal flora of mucosal surfaces (Kleinegger et al., 1996). These organisms can be isolated from the oropharynx, gastrointestinal tract, and vagina of healthy individuals. However, when either local or systemic host defense mechanisms are impaired, Candida spp. can cause oropharyngeal, esophageal, or vulvovaginal candidiasis. Also, in susceptible hosts, the organisms can penetrate the gastrointestinal mucosa and enter the bloodstream, thereby causing hematogenously disseminated candidiasis.

During both mucosal colonization and induction of disease, Candida spp. interact with epithelial cells. Because the outcomes of these interactions are important in determining whether disease develops, they are the subject of intense investigation by multiple laboratories around the world. Candida albicans is the cause of the majority of cases of mucosal disease (Vazquez et al., 2006; Richter et al., 2005; Willis et al., 1999; Ramirez-Amador et al., 1997) and is thus the most widely studied species of Candida. The interactions of C. albicans with epithelial cells include adherence, invasion, and induction of epithelial cell damage. In turn, epithelial cells respond to candidal infection by secreting pro-inflammatory cytokines and producing factors that inhibit the growth of the organism (Steubesand et al., 2009; Feng et al., 2005; Villar et al., 2005; Yano et al., 2005; Dongari-Bagtzoglou et al., 2004). Moreover, the response of epithelial cells to C. albicans is influenced by the presence of neutrophils (Weindl et al., 2007). The outcomes of these interactions are important in determining whether the organism can colonize a mucosal surface and subsequently cause disease. In this review, we will summarize some of the recent discoveries about how C. albicans adheres to, invades, and damages epithelial cells.

Adherence

Adherence of C. albicans to host epithelial cells is a critical first step in the infection process (Fig. 1). It is essential for both colonization and subsequent induction of mucosal disease. Moreover, colonization of mucosal surfaces is a known risk factor for disseminated candidiasis (Takesue et al., 2004; Marr et al., 2000). Because adherence is essential for C. albicans to persist on mucosal surfaces, it is not surprising that this organism expresses multiple different surface structures that mediate adherence to epithelial cells. These various adhesins frequently exhibit differential expression on yeast versus hyphae, and mediate adherence by different mechanisms.

Fig. 1.

Diagram of three major interactions of C. albicans with epithelial cells. The organism first adheres to epithelial cells. Adherence is mediated by multiple different adhesins that are present on the fungal cell surface. Some adhesins are expressed only by hyphae, whereas others are expressed by both hyphae and yeast-phase organisms. Next, the adherent organism can invade both into and between epithelial cells. Invasion into an epithelial cell can occur by induced endocytosis, whereby Als3 and other invasins on the fungal cell surface bind to E-cadherin and other target proteins on the epithelial cell surface. Binding to these epithelial cell proteins induces the epithelial cell to produce pseudopods that engulf the organism and pull it into the cell. C. albicans actively penetrates into epithelial cells by a mechanism that is currently poorly understood. It actively penetrates between epithelial cells by secreting aspartyl proteases that degrade E-cadherin and other inter-epithelial cell junctional proteins. Invasion into and between epithelial cells is a prerequisite for induction of epithelial cell damage. This damage is induced at least in part by lytic enzymes, such as aspartyl proteases secreted by the organism.

The C. albicans ALS gene family

One group of C. albicans adhesins is encoded by the ALS (agglutinin-like sequence) gene family. This family encodes eight glycosylphosphatidylinositol- (GPI) linked cell surface proteins that mediate binding to diverse host substrates (Braun et al., 2005; Zhao et al., 2005; Sheppard et al., 2004; Zhao et al., 2004; Hoyer, 2001; Gaur and Klotz, 1997). Each Als protein has three domains. The N-terminal domain contains the substrate binding region (Rauceo et al., 2006; Loza et al., 2004; Sheppard et al., 2004). The central domain consists of a variable number of tandem repeat sequences. The C-terminal domain is rich in serine and threonine, and contains a GPI anchorage sequence that is predicted to be cleaved as the protein is exported to the cell surface (Hoyer, 2001). Computer-assisted modeling of the N-termini of Als proteins predicts the presence of anti-parallel β sheets, indicating that these proteins are members of the immunoglobulin superfamily. The presence of anti-parallel β sheets in the N-terminus of Als1 and Als5 has been verified by circular dichroism and Fourier transform infrared spectroscopy (Rauceo et al., 2006; Sheppard et al., 2004; Hoyer and Hecht, 2001). Interestingly, the three-dimensional structures of the N-termini of most Als proteins are predicted to be similar to the three-dimensional structure of bacterial adhesins, including invasin of Yersina pseudotuberculosis, and collagen-binding protein and clumping factor A of Staphylococcus aureus (Sheppard et al., 2004).

Two complementary approaches have been used to analyze the role of different Als proteins in adherence: 1) heterologous expression of a C. albicans ALS gene in the normally non-adherent yeast, Saccharomyces cerevisiae (Rauceo et al., 2006; Fu et al., 1998; Gaur and Klotz, 1997), and 2) deletion of both alleles of a given ALS gene in C. albicans (Zhao et al., 2007a; Zhao et al., 2007b; Zhao et al., 2005; Zhao et al., 2004; Fu et al., 2002). Heterologous expression studies of ALS1, ALS3, ALS5, ALS6, ALS7, and ALS9 have determined that different Als proteins have distinct yet overlapping profiles of adherence to diverse host substrates. Als1, Als3 and Als5 mediate adherence to multiple host constituents including oral epithelial cells, whereas Als6 and Als9 bind to a much more limited range of host substrates and do not mediate adherence to epithelial cells. Als7 does not bind to any host substrates tested to date (Sheppard et al., 2004). The adherence function of Als2 and Als4 has not been investigated by this approach.

Studies of C. albicans deletion mutants suggest that Als2, Als3, and Als4 mediate adherence to epithelial cells (Zhao et al., 2005; Zhao et al., 2004). Als1 has been found to mediate adherence to mouse tongues in an ex vivo assay (Kamai et al., 2002), but not to exfoliated human buccal epithelial cells (Zhao et al., 2004). Unexpectedly, deletion of ALS5, ALS6, or ALS7 has been reported to result in increased epithelial cell adherence (Zhao et al., 2007b). These results are difficult to reconcile with findings that heterologous expression of ALS5 and ALS6 in S. cerevisiae results increased adherence (Sheppard et al., 2004). One possible explanation for the increased adherence of C. albicans als5Δ/Δ, als6Δ/Δ, and als7Δ/Δ mutants is that deletion of these genes causes a change in the cell surface of the organism, perhaps due to changes in the cell wall structure or the compensatory up-regulation of other adhesin genes (Zhao et al., 2005).

BLAST searches using the C. albicans Als3 sequence indicate that orthologs of ALS genes are present in the genomes of Candida dubliniensis, Candida tropicalis, Candida parapsilosis, Candida lusitaniae, and Candida guilliermondii. Whether the products of these orthologs mediate adherence to host constituents has not yet been determined.

Hwp1

Hwp1 (hyphal wall protein 1) is a GPI protein that is expressed on the surface of C. albicans hyphae and mediates adherence to oral epithelial cells by a unique mechanism. The N-terminal region of Hwp1 functions as a substrate for epithelial cell-associated transglutaminases that covalently link it to other proteins on the epithelial cell surface. The usual substrates for these transglutaminases are small proline-rich proteins. Thus, Hwp1 mediates epithelial cell adherence by functionally mimicking host cell proteins (Ponniah et al., 2007; Staab et al., 2004; Staab et al., 1999). An hwp1Δ/Δ mutant has significantly reduced virulence in a mouse model of oropharyngeal candidiasis, indicating that Hwp1 is essential for maximal virulence of C. albicans during this mucosal infection (Sundstrom et al., 2002).

Recently, it has been discovered that Hwp1 also binds to Als1 and Als3, and thereby mediates the adherence of one C. albicans hypha to another. This self adherence is important for biofilm formation, and possibly mating (Nobile et al., 2008a).

Eap1

Eap1 (enhanced adherence to polystyrene) is an adhesin that was discovered by heterologous expression studies in S. cerevisiae (Li and Palecek, 2003). The overall structure of Eap1 is similar to that of the Als proteins in that the substrate binding domain is located in the N-terminus, there are tandem repeats in the central region, and the C-terminus contains a GPI anchorage sequence (Li and Palecek, 2008). Studies with both eap1Δ/Δ null mutants of C. albicans, and expression of C. albicans EAP1 in S. cerevisiae indicate that Eap1 mediates adherence to a kidney epithelial cell line and polystyrene (Li et al., 2007; Li and Palecek, 2003). It is also important for biofilm formation, both in vitro and in vivo (Li et al., 2007). The host cell targets of Eap1 have not yet been identified and it is currently unknown whether this adhesin is required for virulence during mucosal or disseminated candidiasis.

Iff4

Iff4 is a member of a 12 protein family, the members of which are related to Hyr1 due to the presence of a conserved N-terminal domain (Bates et al., 2007). Overexpression of IFF4 in C. albicans results in increased adherence to an oral epithelial cell line and increased virulence in the mouse model of candidal vaginitis (Fu et al., 2008). The C. albicans IFF4 overexpression strain also has increased susceptibility to neutrophil killing, and therefore has attenuated virulence during hematogenously disseminated infection in immunocompetent, but not neutropenic mice (Fu et al., 2008). An iff4Δ/Δ mutant has reduced adherence to plastic in vitro, although its adherence to epithelial cells has not been determined (Kempf et al., 2007). Interestingly the iff4Δ/Δ mutant also has reduced virulence in the immunocompetent mouse model of disseminated candidiasis. These findings indicate that a correct level of IFF4 expression is required for maximal virulence in this model; either overexpression or underexpression of IFF4 results in attenuated virulence.

All of the Iff protein family members, except for Iff11, contain GPI anchorage sequences and are likely to be expressed on the cell surface (Bates et al., 2007). Thus, it is possible that other members of the Iff family may also function as adhesins. However, overexpression of IFF2, IFF3, and HYR1 in C. albicans does not alter adherence to oral epithelial cells in vitro (Fu et al., 2008), suggesting that if the products of these genes are adhesins, they do not mediate binding to epithelial cells.

Proteins with direct versus indirect effects on adherence

Some C. albicans genes specify GPI proteins that influence cell wall structure. When deletion of one these gene results in decreased adherence, it can be difficult to determine whether the product of the gene mediates adherence directly or indirectly. For example, Mp65 is a cell surface GPI protein that is predicted to have glucanase activity. An mp65Δ/Δ deletion mutant has reduced adherence to plastic as well as severely impaired hyphal formation (Sandini et al., 2007). There are two non-exclusive explanations for the decreased adherence of this mutant. The first is that Mp65 may directly mediate the attachment of C. albicans to various substrates. The second is that the glucanase activity of Mp65 modifies the structure of the cell wall to enable the proper expression or function of other adhesins. Importantly, single chain, domain specific antibodies directed against Mp65 are able to block the adherence of wild-type C. albicans yeast cells to vaginal epithelial cells in an ex vivo assay (De Bernardis et al., 2007). This result suggests that Mp65 does indeed function as an adhesin. However, it remains theoretically possible that the anti-Mp65 antibodies also interfer with the glucanase function of this protein.

Another example of a GPI protein that influences both adherence and cell walls structure is Ecm33. This protein is present in the cell wall of C. albicans (Martinez-Lopez et al., 2004). An ecm33Δ/Δ mutant has reduced capacity to adhere to and invade an oral epithelial cell line (Martinez-Lopez et al., 2006). Moreover, this mutant also has an abnormal cell wall and aberrant localization of Als1 on its cell surface (Martinez-Lopez et al., 2006). Heterologous expression of C. albicans ECM33 in S. cerevisiae does not alter the adherence of this yeast, even though it is able to complement the cell wall defects of a S. cerevisiae ecm33Δ mutant (Martinez-Lopez et al., 2006; Martinez-Lopez et al., 2004). Therefore, it is probable that Ecm33 does not function as an adhesin itself, but that it is required for the normal function and localization of other adhesins.

Utr2 is a glycosidase that is expressed on the cell surface. C. albicans utr2Δ/Δ mutants have both reduced adherence to an oral epithelial cell line and abnormal cell wall composition (Pardini et al., 2006; Alberti-Segui et al., 2004). Whether Utr2 mediates adherence directly or indirectly has not yet been reported.

Other adhesins

C. albicans expresses numerous other adhesins that have been incompletely characterized in terms of their primary structure and/or capacity to mediate attachment to epithelial cells. These adhesins include fimbriae (Yu et al., 1994), Csh1 (Singleton et al., 2005; Singleton and Hazen, 2004), Ywp1 (Granger et al., 2005), Pra1 (Soloviev et al., 2007; Sentandreu et al., 1998), and the secreted aspartyl proteases (Saps) (Watts et al., 1998). These potential adhesins are reviewed in (Chaffin, 2008).

Differential adhesin expression and cooperative interactions

Given the large number of C. albicans adhesins that mediate binding to epithelial cells, it is not surprising that deletion of both copies of a single adhesin gene typically results in only a partial reduction in adherence (Li et al., 2007; Zhao et al., 2004; Fu et al., 2002; Staab et al., 1999). Some adhesins are preferentially expressed by specific morphologic forms of C. albicans. For example, ALS3 and HWP1 are expressed by hyphae, but not yeast-phase organisms (Sharkey et al., 1999; Hoyer et al., 1998). In contrast, ALS1 is expressed by yeast cells under some conditions and for only a short time after hyphal formation is initiated (Zhao et al., 2004; Fu et al., 2002). EAP1 is expressed by both yeast and hyphae (Li and Palecek, 2003). Thus, different adhesins mediate the binding of different forms of C. albicans to epithelial cells.

In addition, adhesins can interact with each other to form multimers. For instance, binding of either S. cerevisiae expressing C. albicans Als5 or wild-type C. albicans to fibronectin-coated beads triggers a phase change in the proteins over the entire surface of the fungal cell (Rauceo et al., 2004). This phase change results in increased fluorescence of 8-anilino-1-naphthalene-sulfonic acid and enhanced birefringence, and can be inhibited by congo red (Rauceo et al., 2004). It is also associated with the formation of large clumps of fungal cells. These results suggest that the phase change causes the formation of amyloid-like aggregates on the fungal cell surface. Consistent with these results, Als1, Als3, and Als5 contain a highly conserved heptapeptide sequence that is predicted to have amyloid-forming potential (Otoo et al., 2008). Also, fragments of Als5 containing this region form amyloid-like fibers in solution (Otoo et al., 2008). Collectively, these findings indicate that C. albicans Als proteins interact cooperatively such that the initial attachment of a fungal cell to a host ligand can cause different Als proteins to form amyloid-like aggregates along the cell surface. The formation of these aggregates spreads to adjacent fungal cells, causing them to bind to each other and clump. Whether adhesins other than the Als proteins also participate in this process is not yet known.

Invasion

Transmission electron microscopic imaging of biopsy specimens from humans with oropharyngeal, vaginal, and cutaneous candidiasis shows the presence of intra-epithelial cell organisms, demonstrating that epithelial invasion occurs during these diseases. (Rajasingham et al., 1989; Garcia-Tamayo et al., 1982; Scherwitz, 1982; Cawson and Rajasingham, 1972; Montes and Wilborn, 1968). Hyphae appear to be the invasive form of the organisms, as the majority of intracellular organisms are hyphae, whereas yeast are typically located either between or on the surface of epithelial cells (Ray and Payne, 1988; Scherwitz, 1982). Epithelial cell invasion is important for the pathogenesis of mucosal candidiasis, because mutants of C. albicans with reduced capacity to invade epithelial cells in vitro usually have reduced virulence in experimental animal models of mucosal candidiasis. These mutants include efg1Δ/Δ, tpk2Δ/Δ, cka2Δ/Δ, and rim101Δ/Δ strains, as well as mutants that lack various SAP genes (Nobile et al., 2008b; Chiang et al., 2007; Park et al., 2005; De Bernardis et al., 1999).

The mechanisms by which C. albicans invades epithelial cells have been investigated using in vitro models. The results of these studies suggest that that C. albicans can invade epithelial cells by two distinct mechanisms (Fig. 1). One mechanism is the induction of epithelial cell endocytosis by the organism. Endocytosis is induced by invasin-like proteins that are expressed on the surface of a C. albicans hypha. These proteins bind to epithelial cell surface proteins and induce the epithelial cell to produce pseudopods that engulf the organism and pull it inside the cell. Another mechanism of invasion is the active penetration of a hypha either into or between epithelial cells. This process requires fungal viability. C. albicans invades epithelial cells from different anatomic sites via different mechanisms. For example, the organism invades oral epithelial cell lines by both induced endocytosis and active penetration, whereas it invades a gastrointestinal epithelial cell line only by active penetration (Dalle et al., 2009; Phan et al., 2007; Park et al., 2005).

Invasion by induced epithelial cell endocytosis

C. albicans induces its own endocytosis by multiple epithelial cell lines, including HeLa cells, HET1-A esophageal cells, FaDu pharyngeal cells, OKF6/TERT-2 oral epithelial cells, and reconstituted human epithelia, which is a three-dimensional model of oral or vaginal epithelium (Zakikhany et al., 2007; Park et al., 2005; Drago et al., 2000; Enache et al., 1996). This process is passive on the part of the organism because killed hyphae are endocytosed similarly to live hyphae (Park et al., 2005). Recently, it has been discovered that C. albicans Als3 not only mediates epithelial cell adherence, but also functions as an invasin that induces epithelial cell endocytosis. An als3Δ/Δ null mutant of C. albicans has markedly impaired capacity to invade epithelial cells (Phan et al., 2007). Also, latex beads coated with the recombinant N-terminal portion of Als3 are avidly endocytosed by epithelial cells (Phan et al., 2007; Sheppard et al., 2004). Similar studies suggest that Als1 can also induce epithelial cell endocytosis, although with lower efficiency than Als3 (Phan et al., 2007; Sheppard et al., 2004).

Als3 induces endocytosis by binding to E-cadherin and other proteins on the epithelial cell surface. Binding of Als3 to E-cadherin is sufficient to induce endocytosis because latex beads coated with recombinant Als3 are internalized efficiently by Chinese hamster ovary cells expressing human E-cadherin (Phan et al., 2007). The interaction of Als3 with E-cadherin activates the clathrin-dependent endocytosis pathway. siRNA knockdown of components of this pathway, including clathrin, dynamin, and cortactin inhibits the endocytosis of C. albicans by approximately 60% (Moreno-Ruiz et al., 2009).

The fact that siRNA knockdown of the clathrin pathway results in incomplete inhibition of endocytosis suggests that additional signaling pathways also govern this process. These alternative signaling pathways are likely activated by receptors other than E-cadherin. In support of this hypothesis, wild-type C. albicans hyphae bind to multiple epithelial cell surface proteins in addition to E-cadherin (Phan et al., 2007). Furthermore, the als3Δ/Δ mutant fails to bind to several of these surface proteins (Phan et al., 2007), suggesting that Als3 has more than one epithelial cell target protein. It is also virtually certain that C. albicans proteins other than Als1 and Als3 can induce epithelial cell endocytosis. For example, an als1Δ/Δ als3Δ/Δ double mutant of C. albicans has only modestly reduced virulence in the mouse model of oropharyngeal candidiasis (Q.T. Phan, N.V. Solis, A.P. Mitchell, and S.G. Filler, unpublished data), indicating that other invasin-like proteins or other mechanisms of invasion can compensate for the absence of Als1 and Als3. Determining how C. albicans invades epithelial cells independently of Als1 and Als3 is an important future challenge.

While the two invasins discovered to date, Als1 and Als3, also have adherence function, it is possible that some invasins may mediate invasion, but not adherence. It also seems probable that some C. albicans surface structures may only mediate adherence, but not invasion. However, relatively few C. albicans adhesins have been evaluated for invasin function; analyzing additional adhesins will help clarify the relationship between adherence and invasion.

One method for identifying additional C. albicans invasins is an overexpression rescue approach. Rim101 is a C. albicans transcription factor that governs response to alkaline pH (Davis et al., 2000; El Barkani et al., 2000). Nobile et al., (Nobile et al., 2008b) determined that a rim101Δ/Δ mutant had reduced capacity to induce its own endocytosis by oral epithelial cells. This mutant also had significantly attenuated virulence in a mouse model of oropharyngeal candidiasis, demonstrating the importance of Rim101 in governing epithelial cell interactions. Next, microarray analysis of the rim101Δ/Δ mutant was performed to identify genes with Rim101-dependent expression. To determine which of these Rim101 target genes were required for C. albicans to be endocytosed by epithelial cells, 10 of them were overexpressed in the rim101Δ/Δ mutant under the control of the strong TDH3 promoter. These overexpression strains were then tested for their capacity to induce their own endocytosis by epithelial cells in vitro. Overexpression of ZRT1, CHT2, PGA7, or ALS3 in the rim101Δ/Δ background consistently rescued the endocytosis defect of the rim101Δ/Δ mutant (Nobile et al., 2008b). ZRT1 specifies a zinc transporter that is located in the plasma membrane, so it may have indirect effects on endocytosis. CHT2 encodes a chitinase that is expressed on the C. albicans cell surface, and PGA7 also specifies a cell surface protein. Whether these proteins act as invasins themselves or influence the function of other invasins is currently under investigation.

Active penetration

C. albicans can also invade epithelial cells by an active process that is independent of endocytosis. Active penetration seems to require hyphal formation and is not inhibited by when the epithelial cells are treated with a microfilament inhibitor, such as cytochalasin D (Dalle et al., 2009). During active penetration, the organism can either invade into an epithelial cell without inducing the formation of epithelial cell pseudopods or pass through the intercellular junction between epithelial cells (Dalle et al., 2009; Villar et al., 2007).

The mechanism by which C. albicans actively penetrates into an epithelial cell is incompletely understood. It is possible that Saps contribute to this process. Pepstatin A, an inhibitor of aspartyl proteinases, has been reported to inhibit the invasion of C. albicans into corneocytes in mice with cutaneous candidiasis (Ray and Payne, 1988). Some, but not all investigators have found that pepstatin A reduces C. albicans-induced damage to reconstituted human epithelia (Lermann and Morschhauser, 2008; Naglik et al., 2008; Schaller et al., 1999) This reduced damage may be due in part to decreased epithelial cell invasion. However, a limitation of the pepstatin A studies is that this protease inhibitor not only blocks C. albicans Sap activity, but it also inhibits the function of aspartyl proteases produced by the epithelial cell. Therefore, pepstatin A has effects on both the organism and the epithelial cell (Perez-Torres et al., 2008). C. albicans mutants containing disruptions of various SAP genes have reduced capacity to damage vaginal and oral epithelial cells in some in vitro models, but not others (Lermann and Morschhauser, 2008; Naglik et al., 2008; Schaller et al., 2003; Schaller et al., 1999). It is possible that Sap activity is required for active penetration into epithelial cells because it alters the surface characteristics of C. albicans rather than degrading host cell proteins. Dalle et al. (Dalle et al., 2009) recently found that pepstatin A inhibits C. albicans invasion into epithelial cells only when the hyphae are pre-incubated with this inhibitor, prior to being added to the epithelial cells. They also found that triple mutants lacking either Sap1–3 or Sap4–6 have reduced capacity to invade epithelial cells. Importantly, these defects in epithelial cell invasion persist even when the mutants are killed. Taken together, these results suggest the possibility that Saps may activate by proteolysis some C. albicans surface proteins that are require for the organism to invade into an epithelial cell.

Saps may be more important for C. albicans to invade epithelial surfaces by passing between epithelial cells rather than penetrating into them. Several groups have observed that infection of epithelial cells by C. albicans in vitro results in proteolytic degradation of E-cadherin (Rollenhagen et al., 2009; Frank and Hostetter, 2007; Villar et al., 2007). This molecule is concentrated at the intercellular junctions between epithelial cells, and its degradation is associated with loss of integrity of the epithelium. In support of these in vitro findings, E-cadherin antigen is reduced in the oral epithelium of HIV-infected patients who have oropharyngeal candidiasis compared to those without this disease (McNulty et al., 2005). Degradation of E-cadherin is likely mediated at least in part by Saps because it can be blocked by pepstatin A (Frank and Hostetter, 2007; Villar et al., 2007). In addition, a rim101Δ/Δ mutant, which has reduced expression of SAP4, SAP5, and SAP6, has reduced capacity to degrade E-cadherin and disrupt epithelium in a three-dimensional model. These defects are rescued by overexpression of SAP5 in this mutant (Villar et al., 2007). Collectively, these results suggest that the major mechanism of C. albicans invasion between epithelial cells is by proteolysis of intercellular junctions, whereas invasion into individual epithelial cells occurs both by induction of epithelial cell endocytosis and active penetration via a mechanism that is yet to be defined.

Damage

A characteristic finding during oropharyngeal candidiasis is the destruction and loss of the superficial oral epithelium due to fungal invasion (Farah et al., 2000; Eversole et al., 1997). Similarly, when live C. albicans is incubated with epithelial cells in vitro, significant epithelial cell damage occurs. C. albicans must be at least partially endocytosed to cause epithelial cell damage because mutants with defects in inducing epithelial cell endocytosis cause less damage to these cells (Nobile et al., 2008b; Chiang et al., 2007; Phan et al., 2007; Martinez-Lopez et al., 2006; Park et al., 2005) (Fig. 1). Similarly, inhibiting epithelial cell endocytosis of wild-type C. albicans with the microfilament inhibitor, cytochalasin D protects epithelial cells from damage (Park et al., 2005).

However, epithelial cell damage is not a direct consequence of the endocytic process, because killed hyphae are avidly endocytosed, but cause no detectable damage (Park et al., 2005). Also, altering the capacity of a C. albicans mutant to induce endocytosis does not necessarily alter the amount of epithelial cell damage that it causes. For example, a rim101Δ/Δ has impaired capacity to induce epithelial cell endocytosis and cause epithelial cell damage. Overexpression of PGA7 in the rim101Δ/Δ mutant results in increased epithelial cell endocytosis, but has no effect on the extent of epithelial damage this is induced (Nobile et al., 2008b). Thus, epithelial cell endocytosis can be dissociated from epithelial damage.

The mechanism by which C. albicans induces epithelial cell damage is incompletely understood. As mentioned above, Saps likely contribute to this process. However, C. albicans must cause epithelial cell damage by additional mechanisms because in some systems pepstatin A does not protect epithelial cells from damage, and sap mutant strains of C. albicans cause wild-type levels of epithelial cell damage (Lermann and Morschhauser, 2008; Naglik et al., 2008). It is possible that phospholipases secreted by C. albicans may also contribute to epithelial cell damage (Theiss et al., 2006; Naglik et al., 2003; Schofield et al., 2003; Leidich et al., 1998), although this hypothesis has not yet been rigorously tested.

Concluding remarks

The molecular mechanisms by which C. albicans adheres to, invades, and damages epithelial cells are being dissected. This work has been greatly aided by the sequencing and annotation of the C. albicans genome, facile methods for performing gene disruption, and good in vitro and in vivo models for investigating C. albicans-epithelial cell interactions. Future challenges are to identify additional epithelial cell surface proteins that are bound by the various candidal adhesins and invasins, elucidate how C. albicans actively penetrates epithelial cells and causes epithelial cell damage, and to gain a better understanding of the epithelial cell signaling pathways that are activated in response to C. albicans infection.