The Journal of Immunology
Candida albicans Triggers Activation of Distinct Signaling Pathways to Establish a Proinflammatory Gene Expression Program in Primary Human Endothelial Cells1 Verena Mu¨ller,* Dorothee Viemann,† Marc Schmidt,* Nicole Endres,* Stephan Ludwig,‡ Martin Leverkus,§ Johannes Roth,† and Matthias Goebeler2* Endothelial cells (EC) actively participate in the innate defense against microbial pathogens. Under unfavorable conditions, defense reactions can turn life threatening resulting in sepsis. We therefore studied the so far largely unknown EC reaction patterns to the fungal pathogen Candida albicans, which is a major cause of lethality in septic patients. Using oligonucleotide microarray analysis, we identified 56 genes that were transcriptionally up-regulated and 69 genes that were suppressed upon exposure of ECs to C. albicans. The most significantly up-regulated transcripts were found in gene ontology groups comprising the following categories: chemotaxis/migration; cell death and proliferation; signaling; transcriptional regulation; and cell-cell contacts/intercellular signaling. Further examination of candidate signaling cascades established a central role of the proinflammatory NF-B pathway in the regulation of the Candida-modulated transcriptome of ECs. As a second major regulatory pathway we identified the stress-activated p38 MAPK pathway, which critically contributes to the regulation of selected Candida target genes such as CXCL8/IL-8. Depletion of MyD88 and IL-1R-associated kinase-1 by RNA interference demonstrates that Candida-induced NF-B activation is mediated by pattern recognition receptor signaling. Additional experiments suggest that C. albicans-induced CXCL8/IL-8 expression is mediated by TLR3 rather than TLR2 and TLR4, which previously have been implicated with MyD88/ IB kinase-2/NF-B activation by this fungus in other systems. Our study provides the first comprehensive analysis of endothelial gene responses to C. albicans and presents novel insights into the complex signaling patterns triggered by this important pathogen. The Journal of Immunology, 2007, 179: 8435– 8445.
I
nvasive infections by pathogenic Candida species are a common cause of nosocomial bloodstream infections (1). Although dimorphic fungi like Candida albicans colonize distinct microanatomical regions such as mucous membranes as a commensal, occasionally they can also cause severe and often lifethreatening disease, especially in patients with compromised immunity. Manifestations may then not only include oropharyngeal, vulvovaginal, or intertriginous candidiasis but also candidemia and invasive fungal infection when Candida gains access to the bloodstream. In such situations, Candida gets in close contact to the vascular endothelium that is strategically located at the border between blood and tissue compartments. In disseminated candidiasis, Candida adheres to and penetrates the endothelium resulting in cellular activation, injury, and eventually cell death (2–5). Increas-
*Department of Dermatology, University Medical Center Mannheim, University of Heidelberg, Mannheim, Germany; †Institute of Experimental Dermatology, Interdisciplinary Center of Clinical Research, and Department of Pediatrics, and ‡Institute of Molecular Virology, University of Mu¨nster, Mu¨nster, Germany; and §Laboratory for Experimental Dermatology, Department of Dermatology and Venerology, Otto-vonGuericke-University Magdeburg, Magdeburg, Germany Received for publication April 2, 2007. Accepted for publication September 26, 2007. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by Grants Go 811/1-5 (to M.G.), SFB293, A16 (to J.R.), and A17 (to S.L.) from the Deutsche Forschungsgemeinschaft, Grant Fo¨2/005/06 from the Interdisciplinary Clinical Research Centre of the University of Mu¨nster (to D.V.), Grant 95.064.3 from the Wilhelm Sander-Stiftung (to M.G.), and by Grant N2 OGU, TP6 from the Exzellenzfo¨rderung Sachsen-Anhalt (to M.L.). 2 Address correspondence and reprint requests to Dr. Matthias Goebeler, Department of Dermatology, University Hospital Medical Center, University of Heidelberg, Theodor-Kutzer-Ufer 1-3, D-68135 Mannheim, Germany, E-mail address:
[email protected]
www.jimmunol.org
ing evidence points to a complex cooperation of different pattern recognition receptors (PRRs)3 in the host innate immune response to fungal pathogens that not only depend on species and morphotype but also on the route of infection (6, 7). Recently, the mechanism of C. albicans recognition by murine macrophages has been elucidated. In those cells, several sensory systems appear to work in parallel and include recognition of O-linked mannosyl residues on the outer layer of the C. albicans cell wall by TLR4, N-linked mannosyl residues by the mannose receptor, and -glucans located at the inner layer of the C. albicans cell wall by the dectin-1/TLR2 receptor complex (8). However, the role of dectin-1 for C. albicans infection is still controversial (9, 10). Interaction and response patterns of human endothelium to C. albicans have not been studied in detail yet. In this study, we used oligonucleotide microarray analysis to elucidate the endothelial gene expression program elicited by infection with Candida. We found that genes involved in the major categories of “chemotaxis/ migration,” “cell death and proliferation,” “signaling,” “transcriptional regulation,” and “cell-cell contacts/intercellular signaling” were over-represented among the Candida-regulated transcripts, indicative for a broad inflammatory response induced by Candida in endothelial cells (ECs). Most striking in this respect was the strong induction of multiple neutrophil-attracting chemokines, which have not been identified in the context of Candida infection
3 Abbreviations used in this paper: PRR, pattern recognition receptor; IKK, IB kinase; EC, endothelial cell; RNAi, RNA interference; shRNA, small hairpin RNA; siRNA, small interfering RNA; IRAK, IL-1R-associated kinase; poly(I:C), polyinosinic-polycytidylic acid; MOI, multiplicity of infection.
Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00
8436 before. We furthermore demonstrate that the well-known proinflammatory IB kinase (IKK)2/IB/NF-B signaling pathway plays a pivotal role in establishing the inflammatory response toward Candida and that activation of the p38 MAPK cascade is crucial for Candida-induced expression of selected genes in ECs such as CXCL8/IL-8. Finally, we provide evidence that in the endothelial system the proinflammatory gene response to Candida is a receptor-mediated event depending on the adapter molecule MyD88 and on IL-1R-associated kinase-1 (IRAK1) without apparent involvement of the common microbial PRRs TLR2 and TLR4. Rather, RNA interference (RNAi) experiments suggest TLR3 as mediator of Candida-induced proinflammatory gene expression in ECs.
Materials and Methods Cytokines and reagents Human recombinant TNF-␣, IL-1, and IL-1R antagonist were obtained from R&D Systems. The pharmacological p38 inhibitor SB202190 was purchased from Calbiochem, the TLR2/TLR6 agonist macrophage-activating lipopeptide-2 (MALP-2) from Alexis, and the TLR agonists Pam3CSK4, polyinosinic-polycytidylic acid (poly(I:C)), and fibroblaststimulating lipopeptide-1 (FSL-1) from InvivoGen, respectively. All other agents were obtained from Sigma-Aldrich, unless specified otherwise. Soluble TNFR2 fusion protein (TNFR2-Fc) was obtained from Wyeth Pharmaceuticals.
Culture of ECs and C. albicans Primary HUVEC were purchased from Clonetics (via Cambrex Biosciences). Cells were used between passages 3 and 5 in all experiments and cultured in a 1:2 mixture of endothelial growth medium (Clonetics) and M199 medium (PAA) supplemented with 6.7% FBS, gentamicin, amphotericin, and Liquemin N5000 (Roche). Twenty-four hours before addition of C. albicans, HUVEC were washed and placed in antibiotic-free medium. HEK293 cells stably expressing human TLR1 and TLR2 (293-hTLR1/ TLR2) or human TLR4, MD-2, and CD14 (293-hTLR4/MD2-CD14) were purchased from InvivoGen and cultured in DMEM supplemented with 10% FBS. The C. albicans strain SC5314 was provided by O. Kurzai (University of Wu¨rzburg, Wu¨rzburg Germany) and cultured overnight in M199 medium (pH 4.0) at 28°C to generate blastospores. Blastospores were then added to wild-type, transfected, or retrovirally infected HUVEC cultures at a multiplicity of infection (MOI) of 1 unless stated otherwise. Thereafter, supernatants or lysates of cultures were processed for ELISA, promoter reporter gene studies, Western blot analysis, or immune complex kinase assays as specified.
Plasmids A 6xNF-B.luc construct containing six copies of a NF-B-binding motif cloned upstream of a minimal promoter driving a luciferase reporter gene as well as an ubiquitin-dependent Renilla luciferase construct were obtained from B. Baumann (University of Ulm, Ulm, Germany). The pCFG5IEGZ retroviral vector containing an insert of kinase-dead IKK2 (IKK2KD) coupled to expression of a GFP-zeocin resistance fusion gene through an internal ribosomal entry site has been described earlier (11). A dominant negative mutant of human MyD88 (aa 152–296) was provided by Tularik Inc. (South San Francisco, CA) and cloned into the pCFG5-IEGZ retroviral vector. For RNAi studies, HUVEC were infected with retroviral expression vectors encoding either small hairpin RNA (shRNA) against human IRAK1 (pRS-IRAK1) with targeting sequence 5⬘-AGCACAGGCTTCACAGTGATTT-3⬘, or an empty pRS retrovirus, respectively.
DNA microarray and statistical analysis Total RNA (5 g/experiment) from medium-stimulated controls or HUVEC exposed to C. albicans was collected from four independent experiments and individually processed for microarray hybridization using Affymetrix Human Genome 133A Gene Chip arrays, according to the manufacturer’s instructions as described (12). Microarray data were analyzed using MicroArray Suite (MAS software 5.0; Affymetrix). First, background-adjusted raw intensity accounting for nonspecific binding by removing probe sets with insignificant differences between single perfectmatching and mismatching probes was created. Single raw values were calculated for each probe set from the median of perfect-matching/mismatching discrimination values. Fold change (log ratio change) and change in p value based on a signed rank test were determined for each experiment.
C. albicans-REGULATED GENE EXPRESSION IN ECs Only genes with a change in p ⱕ 0.05 for up-regulated or a change in p ⱖ 0.95 for down-regulated genes in at least three of four experiments and mean log ratio change (calculated as mean of log ratio changes of all experiments with significant change in p values) of at least 2.5 or ⫺2.5 were considered. To identify functional categories of genes that were over-represented in the data sets of regulated genes, we first assigned Gene Ontology (GO) annotations to every probe set spotted on the Affymetrix 133A Gene Chip as well as to every Candida-regulated gene using Expressionist Suite software (GeneData). We then applied Fisher’s exact test to calculate statistical significances with respect to the representation of GO annotations in the groups of up-regulated or down-regulated genes in relation to the universe group of GO annotations represented on the entire array. In case of genes that are represented by two or more probe sets on the array or in the group of Candida-regulated genes, only one transcript was taken into account to avoid potential bias. Microarray data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession number GSE7355.
Quantitative real-time RT-PCR cDNA was synthesized from 4 g of total RNA using SuperScript II RNase H reverse transcriptase (Invitrogen Life Technologies). Specific primers for each gene (sequences available upon request) were designed using Primer Express software package (Applied Biosystems) and obtained from MWG Biotec. Quantitative RT-PCR was performed using the QuantiTect SYBR Green PCR kit (Qiagen) as described (13) and data acquired with the ABI PRISM 7900 (Applied Biosystems). Gene expression was normalized to the endogenous housekeeping control gene GAPDH, and relative expression of respective genes was calculated using the comparative threshold cycle method as described (14).
Immunoprecipitation, immune complex kinase assay, and Western blotting ECs were lysed in Triton lysis buffer (20 mM Tris (pH 7.4), 137 mM NaCl, 10% glycerol, 1% Triton X-100, 2 mM EDTA, 50 mM sodium -glycerophosphate, 20 mM sodium pyrophosphate, 1 mM Pefabloc (Merck), 5 g/ml aprotinin, 5 g/ml leupeptin, and 5 mM benzamidine). Immune complex kinase assays for p38 and IKK2 were essentially performed as described (15, 16) using recombinant ATF-2 or GST-IB␣ as substrates for p38 or IKK2, respectively. Western blot analysis using polyclonal antisera against p38 (C-20) or IKK2 (H-470; Santa Cruz Biotechnology) was performed to confirm equal loading of p38 and IKK2 proteins. For analysis of IB␣ expression and degradation cells were lysed in E1A lysis buffer (150 mM sodium chloride, 50 mM HEPES (pH 7.5), 5 mM EDTA and 0.1% Nonidet P-40) freshly supplemented with 20 mM sodium -glycerophosphate, 0.5 mM sodium orthovanadate, and a commercial protease inhibitor cocktail (complete; Roche Molecular Biochemicals). Western blots were then performed as previously described (17) using polyclonal rabbit antisera against IB␣ (C-21, sc-371; Santa Cruz Biotechnology). IRAK1 and CXCL8/IL-8 were detected using a rabbit antiserum against IRAK1 (06-872; Upstate Biotechnology) or a mouse mAb against CXCL8/IL-8 (G265-8; BD Biosciences), respectively. Equal loading was controlled by labeling with a rabbit Ab against Erk2 (C-14, sc-154; Santa Cruz Biotechnology). Bands were visualized by ECL.
ELISA Supernatants of HUVEC were collected after exposure to C. albicans or cytokines, centrifuged at 2600 ⫻ g to remove cellular debris and analyzed for synthesis of CXCL8/IL-8 using a commercial ELISA system (BD Biosciences) according to the manufacturer’s instructions. For CCL20/MIP3␣, HUVEC were lysed in Triton lysis buffer, and protein levels of CCL20/ MIP-3␣ in the lysates were determined using an ELISA detection kit for CCL20/MIP-3␣ (R&D Systems).
Retroviral infections and retrovirus-mediated expression of shRNA For overexpression of dominant negative versions of IKK2, retroviral infections of HUVEC were essentially performed as described (11). Briefly, the amphotrophic producer cell line NX was stably transfected with the indicated retroviral vectors by calcium phosphate precipitation or Lipofectamine 2000 (Invitrogen Life Technologies) and supernatants from these cells were used to infect HUVEC in two consecutive rounds. After selection for zeocin resistance conferred by a gene expressed from the retroviral backbone using 250 g/ml zeocin, stably transduced HUVEC
The Journal of Immunology were incubated in conventional endothelial growth medium without antibiotics and antimycotics. Cells were subsequently exposed to C. albicans and processed for further analysis. Retroviral infection efficiency of HUVEC was routinely monitored by flow cytometric determination of GFP positivity due to marker gene expression from the retroviral constructs and consistently reached values of ⬎85%. For RNAi studies, HUVEC infected with the pRS retrovirus were selected by treatment with 1 g/ml puromycin for at least 48 h. Cells were further incubated for additional 18 h in puromycin-free medium before stimulation in the respective experiments. Knockdown efficiencies in the selected cell population were routinely controlled by Western blot analysis.
Transient transfection of HUVEC and reporter gene assays ECs were cultured to 50 – 60% confluence before transient transfection according to a DEAE-dextran protocol (18). Briefly, cells were incubated with 4 g/2 ⫻ 106cells 6xNF-B.luc reporter construct, 133 ng/2 ⫻ 106 cells ubiquitin-dependent Renilla luciferase and 1 mg DEAE-dextran (Sigma-Aldrich) in 4 ml of 1 mM HEPES/PBS for 30 min at 37°C. Cells were then incubated with 0.15 mM chloroquine in M199 medium for 3 h and, after removal of medium, exposed to 10% DMSO in M199 medium for another 2.5 min. After 16 h, cells were transferred to 96-well plates and stimulated with C. albicans or other reagents as described. Luciferase activity was measured using the Dual-Glo Luciferase Assay System (Promega). Values were each normalized to the luminescence generated by the cotransfected Renilla luciferase control reporter. HUVEC cultured on 6-well plates were transfected with validated small interfering RNA (siRNA) against human MyD88 (catalog number SI00300909), TLR3 (SI02655156), or TLR4 (SI00151004), all from Qiagen or a scrambled siRNA at a final concentration of 200 nmol/L using Oligofectamine (Invitrogen Life Technologies) according to the manufacturer’s protocol. Cells were exposed to C. albicans or other stimuli 48 h later, and chemokine synthesis determined by Western blot or ELISA analyses.
Results C. albicans modulates expression of multiple genes in primary ECs ECs do not only represent a physical barrier between different tissue compartments but are also capable of regulating defense mechanisms against a multitude of microbial pathogens. To investigate the response of human primary ECs to C. albicans, we performed gene profiling using Affymetrix U133A Gene Chip mi-
8437 croarrays, which cover at least 18,400 transcripts, including 14,500 well-characterized human genes. To this end, C. albicans blastospores (strain SC5314) were cocultured with HUVEC for 5 h, and total RNA from four independent experiments was individually processed for DNA microarray hybridization. A relatively early time point was chosen to minimize indirect autocrine or paracrine effects. In initial experiments we excluded a potential contamination of C. albicans preparations by LPS that might confound endothelial gene expression: neither Limulus amebocyte assays (data not shown) nor experiments with the LPS inhibitor polymyxin B sulfate provided evidence for a potential contamination of our Candida preparations with LPS (data not shown). To increase the reliability of our data sets, we further applied rigorous inclusion criteria in our microarray experiments and only considered transcripts as significantly regulated that showed at least a 2.5-fold alteration in gene expression levels. According to these criteria, we overall identified 56 genes that were up-regulated, whereas 69 were down-regulated upon exposure of ECs to C. albicans. Genes induced by the yeast could be assigned to the following GO categories: chemotaxis/migration, cell death and proliferation, signaling, transcriptional regulation, cell-cell contacts/intercellular signaling, and inflammatory response/cytokines (Fig. 1). Most of the genes have not been described as Candida-regulated genes before. Remarkably, a considerable number of up-regulated genes were chemokines that primarily attract neutrophils, suggesting that ECs contribute to the recruitment of leukocytes to sites of the infectious challenge. Among the down-regulated endothelial genes many belonged to the GO group “integral to plasma membrane” (data not shown). A subset of genes suppressed by Candida is related to glutamate-gated channels and associated signaling pathways (i.e., GRIK2, GRID2, GAD2). A complete list of C. albicans-regulated genes has been deposited in the National Center for Biotechnology Information GEO under accession number GSE7355 (and is available at http://www.ncbi.nlm.nih.gov/projects/geo/). To verify the gene expression data obtained in the microarray analysis by an independent method, selected genes were studied by quantitative
FIGURE 1. Assignment of Candida-regulated genes to functional categories. To identify functional categories of genes statistically over-represented among the Candida-regulated transcripts, Gene Ontology (GO) annotations were assigned to every probe set spotted on the Affymetrix 133A Gene Chip. The representation of single GO annotations within the group of up-regulated genes in relation to the totally represented GO annotations on the chip was calculated using Fisher’s exact test. The diagram displays all statistically relevant GO groups among the up-regulated genes. p ⱕ 0.01. Single GO groups are further assigned to the indicated categories (distinguished by color) with the most over-represented GO groups represented by the longest histogram corresponding to the lowest calculated p value.
8438
C. albicans-REGULATED GENE EXPRESSION IN ECs
real time RT-PCR. Highlighting the reliability of our stringent statistical analysis, up-regulation of these exemplary genes that included chemokines ICAM-1 and VCAM-1 could clearly be confirmed in all cases (Fig. 2A). Our data thus demonstrate that ECs exhibit an impressively strong gene expression response to C. albicans. For additional experiments that aimed to identify intracellular signal transduction pathways influenced by C. albicans, we focused on CCL20/MIP-3␣ and CXCL8/IL-8 as read-out molecules. First, we studied protein expression of these chemokines in response to increasing MOI of Candida blastospores by ELISA and Western blot. Robust induction of CXCL8/IL-8 and CCL20/ MIP-3␣ was achieved at an MOI ⱖ 0.5. In line with the mRNA data, a clear induction of chemokines was also detected at the protein level within 8 h upon infection with MOI ⱖ 0.5 of C. albicans (Fig. 2, B and C). Thus, an MOI of 1 at which a solid induction was found in all assays was chosen for subsequent experiments. Only viable C. albicans blastospores were capable of inducing chemokine expression, whereas heat-inactivated yeast did not elicit such a response even at MOI values of up to 4 (Fig. 2D). C. albicans activates the IKK2/IB␣/NF-B signaling pathway Many of the detected Candida-regulated genes are known targets of the proinflammatory transcription factor NF-B, which is activated via the IKK signalosome (12, 19, 20). To explore whether C. albicans induces activation of the NF-B signaling pathway, we performed immune complex kinase assays to study the activity of the IB␣-phosphorylating kinase IKK2. In time course experiments, we observed an induction of IB␣ phosphorylation within 2 to 4 h after infection with C. albicans (Fig. 3A, left). Consistent with our previous data (21), TNF-␣ stimulation resulted in a considerably faster and transient IB␣ phosphorylation with maximal levels at 20 min (Fig. 3A, right). The kinetics of C. albicansinduced IKK2 activity was then compared with the kinetics of IB␣ degradation for both stimuli. In each case, the timing of IB␣ degradation strictly correlated with the activity of IKK2 (data not shown). Because both stimuli triggered IB␣ degradation, we next determined the transcriptional activity using a Bpromoter-dependent luciferase reporter gene construct (Fig. 3B). We detected a 3- to 4-fold induction of NF-B-dependent transcriptional activity with C. albicans, whereas stimulation with TNF-␣ resulted in a 15- to 20-fold induction of the reporter gene (Fig. 3B). Because the kinetics of C. albicans-induced NF-B activation appeared to be delayed as compared with TNF-␣ we wondered whether C. albicans might activate ECs indirectly via autocrine or paracrine action of IL-1 or TNF-␣, two cytokines that are strong activators of NF-B. However, neither the presence of IL-1R antagonist nor addition of a soluble TNFR2 fusion protein (TNFR2-Fc) could prevent Candida-induced CXCL8/IL-8 induction, whereas effects of IL-1 and TNF-␣ were completely blocked upon exposure of cells to their respective antagonists (Fig. 3C). Furthermore, supernatants of Candida-exposed ECs (i.e., conditioned medium) were not capable of inducing CXCL8/IL-8 expression, excluding the involvement of other soluble factors (Fig. 3D). Altogether, these data indicate that C. albicans is most likely a direct inducer of NF-B activation in primary human ECs, albeit its stimulation capacity is weaker and delayed when compared with TNF-␣. C. albicans-mediated chemokine expression critically depends on the NF-B signaling pathway To investigate whether NF-B is critical for C. albicans-induced chemokine expression, we next studied whether stable expression
FIGURE 2. C. albicans induces expression of multiple chemokines in primary human ECs. A, ECs were incubated with C. albicans blastospores (strain SC5314) for 5 h at a MOI of 1. Total cellular RNA was isolated and mRNA expression of chemokines was studied by quantitative real-time RT-PCR. Gene expression of respective genes was calculated using the comparative threshold cycle method as described in Materials and Methods using the endogenous housekeeping control gene GAPDH for normalization. Coxsackie and adenovirus receptor (CAR) was used as additional control and represents an alternative nonregulated gene. Data are presented as fold stimulation ⫾ SEM and are derived from four independent experiments. B and C, HUVEC were cocultured with C. albicans at different MOI values for 8 h. CXCL8/IL-8 levels in culture supernatants (B, top) or CCL20/MIP-3␣ protein expression in total lysates (C) was determined by ELISA. Additionally, cell lysates were analyzed for CXCL8/IL-8 expression by Western blot (B, bottom). Equal loading was controlled by immunoblotting for ERK1/2. ELISA data are shown as mean ⫾ SEM of four (B) or two (C) independent experiments, respectively. The Western blot in B is representative for one of two experiments. D, HUVEC were exposed to viable or heat-inactivated C. albicans blastospores (95°C, 60 min) at the indicated MOI for 8 h and subsequently studied for CXCL8/IL-8 production by Western blot (top). Equal loading was controlled by staining for ERK1/2 (bottom). One of two independent experiments is shown.
of a dominant negative mutant of IKK2 (IKK2KD) in ECs can abrogate induction of selected chemokines after exposure to C. albicans. Indeed, up-regulation of various transcripts including
The Journal of Immunology
FIGURE 3. C. albicans activates the NF-B signaling cascade in ECs. A, HUVEC were stimulated for the indicated time intervals with C. albicans at a MOI of 1 (left) or with 2 ng/ml TNF-␣ (right), respectively. After cell lysis, IKK2 activation was studied by immune complex kinase assay using GST-IB␣ as a substrate (top). Equal immunoprecipitation of IKK2 was confirmed by Western blot using an Ab against IKK2 (bottom). One of two representative experiments is shown. B, C. albicans induces NF-Bdependent transcription. HUVEC were transiently transfected with a NFB-dependent reporter gene construct as outlined in Materials and Methods. At day 2 after transfection, cells were stimulated with C. albicans (MOI ⫽ 1) or TNF-␣ (2 ng/ml) for 8 h and subsequently analyzed for luciferase activity. Data are presented as fold stimulation compared with nonstimulated controls and are normalized to the luminescence generated by cotransfection of a plasmid driving constitutive expression of an ubiquitin-dependent Renilla luciferase reporter. Results are shown as mean ⫾ SEM and were calculated from three independent experiments. C, C. albicans directly activates ECs without autocrine action of TNF-␣ or IL-1. HUVEC were preincubated with IL-1R antagonist (IL-1RA, 200 ng/ml), a soluble TNFR2 protein (TNFR2-Fc, 10 g/ml), or medium with IgG (10 g/ml) as control. Thereafter, cells were exposed to C. albicans (MOI ⫽ 1), TNF-␣ (2 ng/ml), or IL-1 (75 U/ml) for 8 h. CXCL8/IL-8 levels in the culture supernatants were determined. Induced levels obtained after incubation with the indicated stimuli in nonsupplemented culture medium were arbitrarily set 100%. Data are derived from three independent experiments. D, Conditioned medium from C. albicans-exposed ECs is not sufficient to induce CXCL8/IL-8. Conditioned medium was collected from HUVEC previously exposed to C. albicans (MOI ⫽ 1) for 4 or 8 h and after sterile filtration added to independent EC cultures for 8 h. As positive control, cells were directly exposed to C. albicans. ELISA data are shown as fold stimulation ⫾ SEM and are derived from three independent experiments. IL-8 protein present in the conditioned medium was each subtracted to facilitate comparability.
CXCL8/IL-8 and CCL20/MIP-3␣ mRNA (Fig. 4A) was completely inhibited in cells expressing a kinase-dead IKK2 (IKK2KD). Consistently, C. albicans-induced CXCL8/IL-8 and CCL20/MIP-3␣ protein synthesis was also abrogated (Fig. 4B),
8439
FIGURE 4. C. albicans-induced chemokine expression critically depends on the NF-B signaling pathway. A, HUVEC were retrovirally transduced to stably express a dominant negative version of IKK2 (IKK2KD) (f), or a control vector (pCFG5-IEGZ) (u) as described in Materials and Methods. Subsequently, cells were stimulated with C. albicans for 5 h, and mRNA levels of the indicated genes were determined by quantitative realtime RT-PCR. Data are shown as fold stimulation ⫾ SEM and are derived from six independent experiments. B, HUVEC were infected as described in A and exposed to C. albicans (MOI ⫽ 1, left panels) or TNF-␣ (2 ng/ml, right panels) for 8 h as indicated. Expression of the IKK2 mutant transgene was controlled by Western blot (top). More than 85% of the cells expressed the transgene as determined by flow cytometric analysis of GFP expression conferred by the integrated retroviral construct (data not shown). After stimulation, supernatants were obtained and assessed for CXCL8/IL-8 levels by ELISA (middle). Cell lysates were studied for their CCL20/MIP-3␣ content (bottom). ELISA data are presented as fold induction ⫾ SEM calculated from four to six independent experiments.
indicating that C. albicans-induced expression of both chemokines critically requires activation of the IKK2/NF-B pathway. C. albicans activates the p38 MAPK pathway Efficient induction of CXCL8/IL-8 and other chemokines may additionally require activation of the p38 MAPK (12, 22). To determine whether C. albicans also activates the p38 pathway, we used immune complex kinase assays using ATF-2 as a substrate. Fig. 5A indictaes that C. albicans activated endothelial p38 with similar kinetics as observed for IKK2 (after 2– 4 h), whereas TNF-␣ induced a transient p38 activation with a maximum at 20 min followed by a steady decline of activity. To confirm the functional relevance of p38 for C. albicans–induced chemokine induction,
8440
FIGURE 5. C. albicans induces significant but delayed activation of the p38 MAPK. A, HUVEC were stimulated with C. albicans (MOI ⫽ 1, left) or TNF-␣ (2 ng/ml, right) for the indicated time intervals. Total lysates were analyzed for p38 activity by immune complex kinase assay using recombinant ATF-2 as a substrate. TNF-␣ induced a rapid activation of p38, whereas C. albicans-induced p38 activation showed a delayed kinetics (top). Equal immunoprecipitation of p38 was controlled by immunoblot analysis of p38 (bottom). B, HUVEC were preincubated with increasing concentrations of the pharmacological p38 inhibitor SB202190 for 30 min and then stimulated with C. albicans (MOI ⫽ 1, left) or TNF-␣ (2 ng/ml, right) for 8 h in the absence or presence of the inhibitor. Subsequently, cell lysates were studied for CXCL8/IL-8 expression by Western blot. Equal loading of the lanes was controlled by probing the same membrane with an antiserum against ERK1/2. One representative experiment of four is shown. C and D, HUVEC were treated as in B and analyzed for synthesis of the indicated chemokines by ELISA. Data are depicted as mean ⫾ SEM and are derived from three independent experiments. Although Candidainduced CXCL8/IL-8 synthesis was completely blocked (C), CCL20/ MIP-3␣ expression was only marginally affected (D).
ECs were preincubated with the pharmacological p38 inhibitor SB202190. Treatment with the inhibitor resulted in a concentration-dependent repression of C. albicans-induced CXCL8/IL-8 induction with an ID50 of ⬃0.05 M (Fig. 5, B and C). To rule out that exposure of C. albicans with SB202190 will alter the function and virulence of the fungus, ECs were stimulated with C. albicans blastospores that had been cultured for 16 h in the presence of 1 M SB202190 before coculture with HUVEC. Such a treatment did not impair the capability of the yeast to induce chemokines in ECs (data not shown), thus excluding a direct impairing effect of the inhibitor on Candida. Interestingly, although CXCL8/IL-8 clearly was a p38-dependent target gene, C. albicans-induced CCL20/MIP-3␣ expression was not significantly blocked by SB202190, even at concentrations of 10 M. These data suggest distinct signaling requirements for C. albicans-induced expression of different target genes. Taken together, our observations imply an important role of the p38 MAPK for C. albicans-induced activation of CXCL8/IL-8 in ECs, whereas CCL20/MIP-3␣ induction is not considerably affected.
C. albicans-REGULATED GENE EXPRESSION IN ECs C. albicans-induced endothelial NF-B-dependent gene expression requires IRAK1 and MyD88 but occurs independently of TLR2 and TLR4 Having shown the importance of NF-B- and p38-dependent signal transduction pathways for C. albicans-induced gene expression, we next focused on critical upstream signaling components. Because TLR2 and TLR4 have been implicated with pattern recognition of C. albicans in the murine system (8), we first studied the role of both receptors for Candida-dependent signaling in our coculture model. LPS, a known TLR4 agonist, strongly induced CXCL8/IL-8 synthesis (Fig. 6A, left) and NF-B-driven transcriptional activity in HUVEC (data not shown). To our surprise the TLR2 agonist Pam3CSK4 neither stimulated endothelial CXCL8/ IL-8 production (Fig. 6A) nor 6xNF-B-dependent luciferase activity. Failure of Pam3CSK4 activation of HUVEC was not due to the agonist itself because the same preparation of Pam3CSK4 strongly induced CXCL8/IL-8 expression in the TLR2-positive monocytic cell line THP-1 (Fig. 6A, right). Moreover, agonists of TLR2/TLR6 (MALP-2 and FSL-1) did not elicit chemokine expression in ECs while they strongly up-regulated CXCL8/IL-8 in THP-1 cells. These observations suggest that, at least under our culture conditions, HUVEC do not contain functional TLR2 receptors. To further assess the role of TLR2 and TLR4 for Candidainduced signaling, we performed experiments with HEK293 cells that do not express significant levels of endogenous TLR2 and TLR4 (data not shown). Alternatively, transgenic HEK293 cell lines stably expressing either TLR1/TLR2 or TLR4/CD14-MD2 were used. Although parental HEK293 cells did not respond to stimulation with Pam3CSK4 or LPS (Fig. 6, B and C, left), HEK293-hTLR1/TLR2 cells produced high levels of CXCL8/IL-8 and significant 6xNF-B-dependent luciferase activity upon stimulation with Pam3CSK4 but not LPS (Fig. 6, B and C, middle). Conversely, HEK293-hTLR4/CD14-MD2 showed increased CXCL8/IL-8 synthesis and B-driven luciferase activity in response to LPS but not to Pam3CSK4 (Fig. 6, B and C, right). Importantly, stimulation of none of the diverse HEK293 cell lines with C. albicans blastospores resulted in CXCL8/IL-8 production or increased 6xNF-B-dependent luciferase activity. In line with these observations, knockdown of endogenous TLR4 in HUVEC by RNAi using validated duplex siRNA did not block C. albicansinduced expression of CXCL8/IL-8, whereas the response induced by the TLR4 agonist LPS was considerably inhibited (Fig. 6D). We therefore have no evidence that TLR2 or TLR4 are critically involved in C. albicans-induced NF-B-dependent gene expression in HUVEC or HEK293 cells. Studying expression of potential receptors for C. albicans in ECs, we also failed to detect significant mRNA levels of TLR2 and TLR6 and other potential receptors for yeast recognition such as dectin-1 and mannose receptor-1 under our culture conditions (data not shown). Signaling via many PRRs requires recruitment of the adaptor protein MyD88 and subsequent activation of IRAK1. MyD88 and IRAK1 are thus functionally located at the bottleneck of PRRdependent signaling. We therefore decided to knockdown endothelial IRAK1 and MyD88 by RNAi. The shRNA expression against IRAK1 resulted in a complete depletion of IRAK1 protein in HUVEC (Fig. 7A). Expectedly, ablation of IRAK1 expression had no significant effect on TNF-␣-mediated expression of CXCL8/IL-8, whereas IL-1-induced expression of the chemokine was largely blocked. Likewise, Candida-induced synthesis of CXCL8/IL-8 and CCL20/MIP-3␣ was almost completely abrogated in HUVEC stably expressing shRNA against IRAK1 (Fig. 7, A and B). Similar results were obtained with a dominant negative mutant of IRAK1 (IRAK83–544) (23) that was stably expressed in
The Journal of Immunology
FIGURE 6. C. albicans-induced expression of chemokines is mediated independently of TLR2 and TLR4. A, Agonists of TLR4 but not TLR2 induce CXCL8/IL-8 expression by HUVEC. HUVEC and the monocyte cell line THP-1 were exposed for 8 h to LPS (100 ng/ml) that activates cells via TLR4, to Pam3CSK4 (200 ng/ml), which is a TLR1/TLR2 agonist, or to ligands of TLR2/TLR6 (MALP-2 or FSL-1, 100 ng/ml). Data are presented as mean ⫾ SEM and are derived from three independent experiments. B, HEK293 cells or HEK293 cells expressing either human TLR1 and TLR2 (HEK293-hTLR1/2) or human TLR4, MD2 and CD14 (HEK293-hTLR4) were exposed to C. albicans, LPS, Pam3CSK4, or medium as control for 8 h. Thereafter, CXCL8/IL-8 synthesis was determined by ELISA. Data are presented as mean ⫾ SEM and are derived from three independent experiments. C, HEK293, HEK293-hTLR1/2, or HEK293hTLR4/MD2-CD14 cells were transiently transfected with a NF-B-dependent reporter gene construct. At day 2 after transfection, cells were stimulated with the indicated stimuli for 8 h and subsequently analyzed for luciferase activity. Data were calculated from three independent experiments and are presented as fold stimulation ⫾ SEM compared with nonstimulated controls and are normalized to the luminescence generated by cotransfection of a plasmid driving constitutive expression of an ubiquitindependent Renilla luciferase reporter. D, C. albicans-induced expression of CXCL8/IL-8 occurs independently of TLR4. HUVEC were transfected with scrambled siRNA or siRNA against TLR4 and stimulated as indicated. ELISA data for CXCL8/IL-8 are from three independent experiments and are ⫾ SEM.
HUVEC (data not shown). Similar to IRAK1 depletion, transfection of ECs with validated siRNA against human MyD88 efficiently blocked C. albicans- and IL-1-induced expression of CXCL8/IL-8 and CCL20/MIP-3␣ (Fig. 7, C and D). Consistently, expression of a dominant negative mutant of MyD88 largely inhibited C. albicans-induced production of CXCL8/IL-8 (Fig. 7E).
8441
FIGURE 7. C. albicans induces endothelial chemokine expression by a signaling pathway involving IRAK1 and MyD88. A, HUVEC were infected with a retroviral construct expressing shRNA against IRAK1 or an empty control vector. Subsequently, cells were exposed to C. albicans (MOI ⫽ 1), TNF-␣ (2 ng/ml), or IL-1 (10 U/ml) for 8 h. Cell lysates were analyzed for efficiency of IRAK1 depletion (upper) and CXCL8/IL-8 synthesis (middle) by Western blot. Equal loading of the lanes was controlled by staining for ERK1/2 (lower). One representative experiment of six independent experiments is shown. B, HUVEC stably expressing shRNA against IRAK1 or an empty control vector were stimulated for 8 h as indicated. CXCL8/IL-8 and CCL20/MIP-3␣ production was determined by ELISA. C, HUVEC were transfected with validated siRNA against MyD88 or scrambled siRNA and subsequently exposed for 8 h to C. albicans and other stimuli as indicated. CXCL8/IL-8 synthesis was detected by Western blot (top) and equal loading controlled by immunolabeling for ERK1/2 (bottom). D, HUVEC transfected with MyD88 siRNA or a scrambled siRNA were stimulated for 8 h as indicated and CXCL8/IL-8 and CCL20/ MIP-3␣ production determined by ELISA. E, HUVEC stably expressing a dominant negative mutant of MyD88 (MyD88C152–296, dnMyD88) were exposed to C. albicans at a MOI of 1 or to IL-1 (10 U/ml) for 8 h. Thereafter, CXCL8/IL-8 production was determined by ELISA. Expression of the transgene was controlled by immunolabeling of a coexpressed myc tag by Western blot (data not shown). D and E are presented as fold stimulation ⫾ SEM of two or three independent experiments, respectively.
8442
C. albicans-REGULATED GENE EXPRESSION IN ECs
These observations unequivocally demonstrate that C. albicans induces NF-B-dependent gene expression via a MyD88/IRAK1dependent signaling pathway. Because our data thus far suggested that neither TLR2 nor TLR4 or autocrine factors (compare Fig. 3D) can sufficiently account for Candida-induced NF-B activation, we considered the involvement of other TLRs. Recently, Morris et al. (24) provided evidence that TLR3 can trigger proinflammatory gene expression in ECs in response to poly(I:C), an established TLR3 agonist. In search for other candidate TLRs, we thus tested whether functional TLR3 is also present in HUVEC under our culture conditions. Indeed, treatment of HUVEC with poly(I:C) resulted in a concentration-dependent expression of CXCL8/IL-8, demonstrating the presence of functional TLR3 in our cells (Fig. 8A). As controversial data have been published concerning the dependency of TLR3 signaling on MyD88 (25), we next analyzed whether siRNA mediated depletion of MyD88 can affect poly(I:C)-induced CXCL8/IL-8 expression. Fig. 8B illustrates that MyD88 depletion reduces poly(I:C)-mediated CXCL8/ IL-8 induction, suggesting that this response is to a considerable degree dependent on MyD88. Finally, to prove that TLR3 is involved in Candida-induced proinflammatory responses, we depleted TLR3 by siRNA. Although TLR3 siRNA transfection had no effect on TNF-␣-induced CXCL8/IL-8 expression, both Candida- and poly(I:C)-mediated CXCL8/IL8 induction was clearly reduced (Fig. 8C). Taken together, these data suggest that the Candida-induced proinflammatory response is at least partially dependent on TLR3.
Discussion The course of Candida infections is determined by both pathogenspecific and host-dependent factors (26, 27). Prominent target cells of Candida infection are ECs, which play a pivotal role in the early immune response during candidemia and invasive fungal infection as well as during sepsis. In such situations, Candida adheres to and invades the endothelium resulting in cellular activation and injury (2– 4, 6). Using microarray analysis of human primary ECs, we identified 56 genes that were significantly up-regulated and 69 endothelial genes that were significantly suppressed upon coculture of HUVEC with C. albicans blastospores. By bioinformatical approaches we could identify those groups of genes that were significantly over-represented among the regulated genes. Most prominent were GO groups that cover chemotaxis and migration (Fig. 1). Overall, we identified six chemokines that are up-regulated by C. albicans. Remarkably, most of those represent factors that usually recruit neutrophils to sites of pathogen challenge. Because neutrophils play a pivotal role in the resolution of fungal infections (28) with neutropenia being a major indicator of poor prognosis, these findings suggest that ECs are a fully integrated player in an orchestrated defense network against fungal pathogens. Endothelial chemokines elicited by C. albicans include CXCL8/IL-8, CXCL1/Gro␣, CXCL2/Gro, CXCL3/Gro␥ as well as CXCL5/ENA78 and CXCL6/GCP-2, which were detected by quantitative RT-PCR. Induction of those chemokines was a direct effect of Candida exposure and not a consequence of autocrine IL-1 or TNF-␣ secretion because addition of soluble TNFR (TNFR2-Fc) or IL-1R antagonist to the coculture system failed to block their expression. Furthermore, conditioned medium was unable to elicit a CXCL8/IL-8 response excluding the involvement of other soluble factors. Thus, ECs most likely initiate their inflammatory response to Candida independently of host cell- or fungusderived soluble factors. CCL20/MIP-3␣, another previously unknown Candida-regulated endothelial chemokine, belongs to a subgroup of chemokines with defensin-like antimicrobial activity (29) and is a potent che-
FIGURE 8. TLR3 is involved in C. albicans-induced expression of CXCL8/IL-8. A, HUVEC express functional TLR3. ECs were stimulated with increasing concentrations of the TLR3 agonist poly(I:C) or IL-1 (100 U/ml) as indicated. Thereafter, protein expression of CXCL8/IL-8 was assessed by Western blot. B, Poly(I:C)-induced TLR3-mediated expression of CXCL8/IL-8 is partially dependent on MyD88. HUVEC transfected with MyD88 siRNA or a scrambled siRNA were exposed to C. albicans (MOI ⫽ 1), poly(I:C) (1 g/ml), or LPS (100 ng/ml, positive control) for 8 h. CXCL8/IL-8 protein expression was analyzed by ELISA (top) or Western blot (bottom). ELISA data are presented as fold stimulation ⫾ SEM of three independent experiments. C, C. albicans-induced expression of endothelial CXCL8/IL-8 is partially depending on TLR3. HUVEC transfected with validated TLR3 siRNA or scrambled siRNA were exposed to C. albicans (MOI ⫽ 1), poly(I:C) as positive control, or TNF-␣ as negative control. Thereafter, CXCL8/IL-8 expression was detected by ELISA (top). Data are shown as fold stimulation ⫾ SEM of three independent experiments. CXCL8/IL-8 expression was also detected by Western blot (bottom). While TNF-␣-induced expression of the chemokine remains unaffected, induction by C. albicans and poly(I:C) is considerably blocked.
moattractant for dendritic cell precursors. CCL20/MIP-3␣ thus has the properties to act as a danger signal, thereby linking the innate to the adaptive immune system (30), which both determine the course of Candida infection in hosts (31). Besides cytokines a number of cell surface receptors were likewise induced upon C. albicans infection. These include tissue factor, the initiator of the extrinsic coagulation cascade, adhesion molecules such as ICAM-1 and VCAM-1 that are required for leukocyte recruitment, CD24, a receptor for P-selectin (32), and the chemokine receptor CXCR7 (chemokine orphan receptor 1). Furthermore, CD74, an
The Journal of Immunology accessory signaling receptor resulting in NF-B activation and cell survival (33), and the activation-associated receptor CD69 are upregulated upon infection with C. albicans. Other over-represented functional groups of Candida-induced genes can be classified into the main categories cell death and proliferation, signaling, transcriptional regulation, and cell-cell contacts and intracellular signaling. A significant fraction of genes annotated to “transcriptional regulation” were zinc finger proteins that included Kru¨ppel-like transcription factor KLF4, ZFP36 (also known as tristetraprolin), ZNF151, and Lin28. Although ZFP36 is up-regulated during inflammatory responses, it fulfils anti-inflammatory tasks required for the resolution of inflammatory reactions via mRNA destabilization of cytokines including TNF-␣ (34). Among the up-regulated genes many have been reported to act as negative regulators of cell proliferation, indicating that EC proliferation ceases after Candida infection. Systematic analysis of the Candida-regulated gene repertoire revealed that many genes are under control of the transcription factor NF-B, which orchestrates proinflammatory as well as antiapoptotic responses to various stimuli such as IL-1, TNF-␣, or pathogens in multiple cell types (35). In this study, we demonstrate for the first time that Candida activates IKK2 resulting in phosphorylation and subsequent degradation of IB␣ as well as NFB-dependent transcription. However, in sharp contrast to the prototypical activator of endothelial chemokine expression TNF-␣, the kinetics of IKK2 activation by Candida is markedly delayed, which is in line with reports on other stress-like endothelial stimuli such as PMA, LPS, or nickel chloride (21, 36, 37). Delayed and persistent activation imply a different mode of activation. Therefore, it was mandatory to further dissect the intracellular requirements of Candida-induced gene expression. Upon stable expression of a dominant negative mutant of IKK2, which prevents activation of NF-B, expression of various Candida target genes was blocked (see Fig. 4). This result clearly demonstrates NF-B activation as a critical requirement for the expression of many Candida-regulated genes. IKK2/NF-B-dependent gene expression is often modulated by coregulatory signals generated by kinases of the MAPK family (38). In particular the p38 MAPK pathway is known to be involved in the regulation of several proinflammatory mediators in ECs such as CXCL8/IL-8 and CCL2/MCP-1, which exert functional roles in host defense responses (17, 18, 39). Albeit the mechanism of p38dependent gene regulation is not completely understood, increasing evidence suggests an involvement of this pathway in specific posttranscriptional events that result in stabilization or destabilization of mRNA (22, 40). Furthermore, p38 may modulate gene expression at the level of transactivation by interacting with components of the enhanceosome, such as the transcriptional coactivator CBP/p300 (21). Highlighting the regulatory complexity, mitogen- and stress-activated protein kinase-1 (MSK1), which acts downstream of p38, has been identified to phosphorylate the NFB/Rel family member p65 as well as components of the chromatin environment, thereby interfering with NF-B-dependent gene transcription (41, 42). Our data demonstrate that p38, like NF-B, is activated in a delayed kinetic upon exposure to C. albicans. This finding is in agreement with a recent report studying HeLa cells (43). The kinetics also matches findings from previous studies using LPS (39) and nickel compounds (21) as endothelial activators. We additionally identified differences regarding the requirement of a coordinated activation of p38 and NF-B signaling pathways; while CCL20/MIP-3␣ was largely independent of p38, CXCL8/ IL-8 expression was clearly suppressed by a pharmacological inhibitor of p38. These data indicate that activation of NF-B and p38 is critical for IL-8/CXCL8 induction, whereas p38 is dispens-
8443 able for Candida-induced CCL20/MIP-3␣ expression. This result supports our previous observations (using other proinflammatory stimuli for ECs) that some NF-B target genes are coregulated by p38, whereas others obviously do not depend on p38 activation (12). The ability of Candida to grow in different forms as a parasite or a commensal in vivo might have resulted in a complex repertoire of antifungal host response patterns. In systemic candidiasis, primarily cells of the innate immune system are required for an effective host defense. They need to reliably recognize microbial pathogens, which is ensured by evolutionarily ancient PRRs including TLRs, C-type lectin receptors, NOD proteins, and others (9, 10, 44, 45). Upon pathogen recognition, cells appear to activate intracellular signal transduction pathways, which ultimately results in gene expression and cellular activation. Recognition of C. albicans by mouse macrophages could recently be attributed to sensing of O-linked mannosyl residues of the outer layer of the C. albicans cell wall by TLR4, of N-linked mannosyl residues by the mannose receptor, and of -glucans localized in the inner layer of the yeast cell wall by the dectin-1/TLR2 receptor complex (8, 9). Our data demonstrate that at least in primary human ECs alternative sensory mechanisms must exist that mediate responses to Candida. While ECs express substantial levels of TLR4 (46), we failed to detect significant amounts of functional TLR2 in HUVEC under our culture conditions (data not shown). In line with this result, ECs responded to neither the TLR1/TLR2 agonist Pam3CSK4 nor to the TLR2/TLR6 ligands MALP-2 and FSL-1 at concentrations that maximally induced CXCL8/IL-8 expression in TLR2-positive THP-1 cells. Candida-infection was further unable to trigger NF-B activation and CXCL8/IL-8 expression in either parental HEK293 cells lacking TLR4 or HEK293 cells stably expressing TLR4 (together with MD-2/CD14); however, the latter cells were responsive to the TLR4 agonist LPS. It therefore seems unlikely that TLR4 is critical for signal perception of the yeast even more because siRNA against TLR4 failed to abolish Candida-induced CXCL8/IL-8 expression in ECs. According to our expression analysis (data not shown) and in line with an earlier report (47), dectin-1, the myeloid cell-expressed major receptor for fungal -glucan, is not expressed by ECs thus excluding a role in Candidainduced endothelial signaling. Furthermore, the mannose receptor, another C-type lectin receptor, is absent in HUVEC (48). Thus, neither C-type lectin receptors nor TLR2 and TLR4 appear to be responsible for C. albicans-induced gene expression in our EC model. Remarkably, RNAi-mediated depletion of the adapter molecule MyD88 as well as of IRAK1, a kinase downstream of TLRs, abolished C. albicans-induced expression of NF-B target genes such as CXCL8/IL-8. This observation clearly indicates that endothelial activation by Candida is a receptor-mediated process. Dependency on MyD88 and IRAK1 suggests sensing of C. albicans by a TLR. Our siRNA data indicate that most likely TLR3 represents this receptor. TLR3 was initially reported as intracellular receptor primarily implicated with the perception of viral dsRNA and endogenous RNA species released upon cell damage (49 –51). These data are compatible with a scenario in which viable yeasts enter a cell and cause a certain degree of EC injury as suggested by Filler et al. (5). Such damaging events may be associated with development of putatively endogenous dsRNA species that successively activate endothelial defense mechanisms via TLR3. Although the exact mechanism of Candida-induced TLR3 activation remains to be elucidated, a model of Candida-induced cell damage might also account for the observed delayed kinetics of Candida-induced signaling and provide a reasonable explanation why only viable C. albicans is capable of inducing NF-B-dependent signaling in ECs
8444 (2) (Fig. 2D). Furthermore, TLR3 is well compatible with our observed MyD88 dependency because Alexopoulou et al. (52) initially described TLR3 as MyD88-dependent receptor. We confirm these data by showing that CXCL8/IL-8 induction by the TLR3 agonist poly(I:C) at least partially depends on MyD88. Although we cannot exclude a contribution by TLR3-independent mechanisms, our data demonstrate that TLR3 is involved in mediation of the proinflammatory response induced by the yeast. In conclusion, our data underscore the importance of primary human ECs as an integrated constituent of the innate immune system in the defense of fungal pathogens and provide novel insights into the molecular mechanisms by which C. albicans elicits excessive immune responses that under unfavorable conditions can lead to life-threatening sepsis.
Acknowledgments We thank Sybille Schmid, Martina Cimander, and Sigrid Lempert for excellent technical assistance and advice.
Disclosures The authors have no financial conflict of interest.
References 1. Wisplinghoff, H., T. Bischoff, S. M. Tallent, H. Seifert, R. P. Wenzel, and M. B. Edmond. 2004. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin. Infect. Dis. 39: 309 –317. 2. Filler, S. G., A. S. Pfunder, B. J. Spellberg, J. P. Spellberg, and J. E. Edwards, Jr. 1996. Candida albicans stimulates cytokine production and leukocyte adhesion molecule expression by endothelial cells. Infect. Immun. 64: 2609 –2617. 3. Orozco, A. S., X. Zhou, and S. G. Filler. 2000. Mechanisms of proinflammatory response of endothelial cells to Candida albicans infection. Infect. Immun. 68: 1134 –1141. 4. Cannom, R. R. M., S. W. French, D. Johnston, J. E. Edwards, and S. G. Filler. 2002. Candida albicans stimulates local expression of leukocyte adhesion molecules and cytokines in vivo. J. Infect. Dis. 186: 389 –396. 5. Filler, S. G., J. N. Swerdloff, C. Hobbs, and P. M. Luckett. 1995. Penetration and damage of endothelial cells by Candida albicans. Infect. Immun. 63: 976 –983. 6. Filler, S. G., and D. C. Sheppard. 2006. Fungal invasion of normally non-phagocytic host cells. PLoS Pathog. 2: e129. 7. Dostert, C., and J. Tschopp. 2007. DEteCTINg fungal pathogens. Nat. Immunol. 8: 17–18. 8. Netea, M. G., N. A. Gow, C. A. Munro, S. Bates, C. Collins, G. Ferwerda, R. P. Hobson, G. Bertram, H. B. Hughes, T. Jansen, et al. 2006. Immune sensing of Candida albicans requires cooperative recognition of mannans and glucans by lectin and Toll-like receptors. J. Clin. Invest. 116: 1642–1650. 9. Taylor, P. R., S. V. Tsoni, J. A. Willment, K. M. Dennehy, M. Rosas, H. Findon, K. Haynes, C. Steele, M. Botto, S. Gordon, and G. D. Brown. 2007. Dectin-1 is required for beta-glucan recognition and control of fungal infection. Nat. Immunol. 8: 31–38. 10. Saijo, S., N. Fujikado, T. Furuta, S. H. Chung, H. Kotaki, K. Seki, K. Sudo, S. Akira, Y. Adachi, N. Ohno, et al. 2007. Dectin-1 is required for host defense against Pneumocystis carinii but not against Candida albicans. Nat. Immunol. 8: 39 – 46. 11. Denk, A., M. Goebeler, S. Schmid, I. Berberich, O. Ritz, D. Lindemann, S. Ludwig, and T. Wirth. 2001. Activation of NF-B via the IB kinase complex is both essential and sufficient for proinflammatory gene expression in primary endothelial cells. J. Biol. Chem. 276: 28451–28458. 12. Viemann, D., M. Goebeler, S. Schmid, K. Klimmek, C. Sorg, S. Ludwig, and J. Roth. 2004. Transcriptional profiling of IKK2/NF-B- and p38 MAP kinasedependent gene expression in TNF-␣-stimulated primary human endothelial cells. Blood 103: 3365–3373. 13. Rajeevan, M. S., D. G. Ranamukhaarachchi, S. D. Vernon, and E. R. Unger. 2001. Use of real-time quantitative PCR to validate the results of cDNA array and differential display PCR technologies. Methods 25: 443– 451. 14. Liu, W., and D. A. Saint. 2002. A new quantitative method of real time reverse transcription polymerase chain reaction assay based on simulation of polymerase chain reaction kinetics. Anal. Biochem. 302: 52–59. 15. Goebeler, M., A. Trautmann, A. Voss, E. B. Bro¨cker, A. Toksoy, and R. Gillitzer. 2001. Differential and sequential expression of multiple chemokines during elicitation of allergic contact hypersensitivity. Am. J. Pathol. 158: 431– 440. 16. Schmidt, M., M. Goebeler, G. Posern, S. M. Feller, C. S. Seitz, E. B. Bro¨cker, U. R. Rapp, and S. Ludwig. 2000. Ras-independent activation of the Raf/MEK/ Erk pathway upon calcium-induced differentiation of keratinocytes. J. Biol. Chem. 275: 41011– 41017. 17. Viemann, D., M. Schmidt, K. Tenbrock, S. Schmid, V. Mu¨ller, K. Klimmek, S. Ludwig, J. Roth, and M. Goebeler. 2007. The contact allergen nickel triggers a unique inflammatory and proangiogenic gene expression pattern via activation of NF-B and hypoxia-inducible factor-1␣. J. Immunol. 178: 3198 –3207.
C. albicans-REGULATED GENE EXPRESSION IN ECs 18. Goebeler, M., K. Kilian, R. Gillitzer, M. Kunz, T. Yoshimura, E. B. Bro¨cker, U. R. Rapp, and S. Ludwig. 1999. The MKK6/p38 stress kinase cascade is critical for tumor necrosis factor-a-induced expression of monocyte-chemoattractant protein-1 in endothelial cells. Blood 93: 857– 865. 19. Karin, M., and M. Delhase. 2000. The IB kinase (IKK) and NF-B: key elements of proinflammatory signalling. Semin. Immunol. 12: 85–98. 20. Leverkus, M., M. R. Sprick, T. Wachter, A. Denk, E. B. Bro¨cker, H. Walszak, and M. Neumann. 2003. TRAIL-induced apoptosis and gene induction in HaCaT keratinocytes: differential contribution of TRAIL receptors 1 and 2. J. Invest. Dermatol. 121: 149 –155. 21. Goebeler, M., R. Gillitzer, K. Kilian, K. Utzel, E. B. Bro¨cker, U. R. Rapp, and S. Ludwig. 2001. Multiple signaling pathways regulate NF-B-dependent transcription of the monocyte chemoattractant protein-1 (MCP-1) gene in primary endothelial cells. Blood 97: 46 –55. 22. Hoffmann, E., O. Dittrich-Breiholz, H. Holtmann, and M. Kracht. 2002. Multiple control of interleukin-8 gene expression. J. Leukocyte Biol. 72: 847– 855. 23. Cao, Z., W. J. Henzel, and X. Gao. 1996. IRAK: a kinase associated with the interleukin-1 receptor. Science 271: 1128 –1131. 24. Morris, G. E., L. C. Parker, J. R. Ward, E. C. Jones, M. K. Whyte, C. E. Brightling, P. Bradding, S. K. Dower, and I. Sabroe. 2006. Cooperative molecular and cellular networks regulate Toll-like receptor-dependent inflammatory responses. FASEB J. 20: 2153–2155. 25. Akira, S., and K. Takeda. 2004. Toll-like receptor signalling. Nat. Rev. Immunol. 4: 499 –511. 26. Berman, J., and P. E. Sudbery. 2002. Candida albicans: a molecular revolution built on lessons from budding yeast. Nat. Rev. Genet. 3: 918 –930. 27. Roeder, A., C. J. Kirschning, R. A. Rupec, M. Schaller, and H. C. Korting. 2004. Toll-like receptors and innate antifungal responses. Trends Microbiol. 12: 44 – 49. 28. Urban, C. F., S. Lourido, and A. Zychlinsky. 2006. How do microbes evade neutrophil killing? Cell. Microbiol. 8: 1687–1696. 29. Yang, D., Q. Chen, D. M. Hoover, P. Staley, K. D. Tucker, J. Lubkowski, and J. J. Oppenheim. 2003. Many chemokines including CCL20/MIP-3␣ display antimicrobial activity. J. Leukocyte Biol. 74: 448 – 455. 30. Dieu-Nosjean, M. C., C. Massacrier, B. Homey, B. Vanbervliet, J. J. Pin, A. Vicari, S. Lebecque, C. Dezutter-Dambuyant, D. Schmitt, A. Zlotnik, and C. Caux. 2000. Macrophage inflammatory protein 3␣ is expressed at inflamed epithelial surfaces and is the most potent chemokine known in attracting Langerhans cell precursors. J. Exp. Med. 192: 705–718. 31. Bellocchio, S., C. Montagnoli, S. Bozza, R. Gaziano, G. Rossi, S. S. Mambula, A. Vecchi, A. Mantovani, S. M. Levitz, and L. Romani. 2004. The contribution of the Toll-like/IL-1 receptor superfamily to innate and adaptive immunity to fungal pathogens in vivo. J. Immunol. 172: 3059 –3069. 32. Aigner, S., Z. M. Sthoeger, M. Fogel, E. Weber, J. Zarn, M. Ruppert, Y. Zeller, D. Vestweber, R. Stahel, M. Sammar, and P. Altevogt. 1997. CD24, a mucin-type glycoprotein, is a ligand for P-selectin on human tumor cells. Blood 89: 3385–3395. 33. Starlets, D., Y. Gore, I. Binsky, M. Haran, N. Harpaz, L. Shvidel, S. Becker-Herman, A. Berrebi, and I. Shachar. 2006. Cell-surface CD74 initiates a signaling cascade leading to cell proliferation and survival. Blood 107: 4807– 4816. 34. Carballo, E., W. S. Lai, and P. J. Blackshear. 1998. Feedback inhibition of macrophage tumor necrosis factor-␣ production by tristetraprolin. Science 281: 1001–1005. 35. Karin, M., Y. Yamamoto, and Q. M. Wang. 2004. The IKK NF-B system: a treasure trove for drug development. Nat. Rev. Drug Discov. 3: 17–26. 36. Johnson, D. R., I. Douglas, A. Jahnke, S. Ghosh, and J. S. Pober. 1996. A sustained reduction in IB- may contribute to persistent NF-B activation in human endothelial cells. J. Biol. Chem. 271: 16317–16322. 37. Zen, K., A. Karsan, T. Eunson, E. Yee, and J. M. Harlan. 1998. Lipopolysaccharide-induced NF-B activation in human endothelial cells involves degradation of IB␣ but not IB. Exp. Cell Res. 243: 425– 433. 38. Ashwell, J. D. 2006. The many paths to p38 mitogen-activated protein kinase activation in the immune system. Nat. Rev. Immunol. 6: 532–540. 39. Hippenstiel, S., S. Soeth, B. Kellas, O. Fuhrman, J. Seybold, M. Kru¨ll, C. Eichel Streiber, M. Goebeler, S. Ludwig, and N. Suttorp. 2000. Rho proteins and the p38-MAPK pathway are important mediators for LPS-induced interleukin-8 expression in human endothelial cells. Blood 95: 3044 –3051. 40. Frevel, M. A. E., T. Bakheet, A. M. Silva, J. G. Hissong, K. S. A. Khabar, and B. R. G. Williams. 2003. p38 mitogen-activated protein kinase-dependent and -independent signaling of mRNA stability of AU-rich element-containing transcripts. Mol. Cell Biol. 23: 425– 436. 41. Schmitz, M. L., S. Bacher, and M. Kracht. 2001. IB-independent control of NF-B activity by modulatory phosphorylations. Trends Biochem. Sci. 26: 186 –190. 42. Vermeulen, L., G. De Wilde, P. van Damme, W. Vanden Berghe, and G. Haegeman. 2003. Transcriptional activation of the NF-B p65 subunit by mitogen- and stress-activated protein kinase-1 (MSK1). EMBO J. 22: 1313–1324. 43. Deva, R., P. Shankaranarayanan, R. Ciccoli, and S. Nigam. 2003. Candida albicans induces selectively transcriptional activation of cyclooxygenase-2 in HeLa cells: pivotal roles of toll-like receptors, p38 mitogen-activated protein kinase, and NF-B. J. Immunol. 171: 3047–3055. 44. Netea, M. G., G. Ferwerda, C. A. van der Graaf, J. W. Van der Meer, and B. J. Kullberg. 2006. Recognition of fungal pathogens by toll-like receptors. Curr. Pharm. Des. 12: 4195– 4201. 45. Romani, L., F. Bistoni, and P. Puccetti. 2002. Fungi, dendritic cells and receptors: a host perspective of fungal virulence. Trends Microbiol. 10: 508 –514.
The Journal of Immunology 46. Faure, E., O. Equils, P. A. Sieling, L. Thomas, F. X. Zhang, C. J. Kirschning, N. Polentarutti, M. Muzio, and M. Arditi. 2000. Bacterial lipopolysaccharide activates NF-B through toll-like receptor 4 (TLR-4) in cultured human dermal endothelial cells: differential expression of TLR-4 and TLR-2 in endothelial cells. J. Biol. Chem. 275: 11058 –11063. 47. Sobanov, Y., A. Bernreiter, S. Derdak, D. Mechtcheriakova, B. Schweighofer, M. Duchler, F. Kalthoff, and E. Hofer. 2001. A novel cluster of lectin-like receptor genes expressed in monocytic, dendritic and endothelial cells maps close to the NK receptor genes in the human NK gene complex. Eur. J. Immunol. 31: 3493–3503. 48. Gro¨ger, M., W. Holnthoner, D. Maurer, S. Lechleitner, K. Wolff, B. B. Mayr, W. Lubitz, and P. Petzelbauer. 2000. Dermal microvascular endothelial cells
8445
49. 50.
51. 52.
express the 180-kDa macrophage mannose receptor in situ and in vitro. J. Immunol. 165: 5428 –5434. Kariko, K., H. Ni, J. Capodici, M. Lamphier, and D. Weissman. 2004. mRNA is an endogenous ligand for Toll-like receptor 3. J. Biol. Chem. 279: 12542–12550. Brentano, F., O. Schorr, R. E. Gay, S. Gay, and D. Kyburz. 2005. RNA released from necrotic synovial fluid cells activates rheumatoid arthritis synovial fibroblasts via Toll-like receptor 3. Arthritis Rheum. 52: 2656 –2665. Marshak-Rothstein, A. 2006. Toll-like receptors in systemic autoimmune disease. Nat. Rev. Immunol. 6: 823– 835. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognition of double-stranded RNA and activation of NF-B by Toll-like receptor 3. Nature 413: 732–738.
