Candida albicans β-Glucan-Containing Particles Increase HO-1 Expression in Oral Keratinocytes via a Reactive Oxygen Species/p38 Mitogen-Activated Protein Kinase/Nrf2 Pathway

ABSTRACT

Oral keratinocytes provide the first line of host defense against oral candidiasis. We speculated that interactions of fungal cell wall components with oral keratinocytes regulate the stress response against Candida infection and examined the expression of genes induced by heat-killed Candida albicans in oral immortalized keratinocytes using a cDNA microarray technique. Of 24,000 genes revealed by that analysis, we focused on HO-1, a stress-inducible gene, as its expression was increased by both heat-killed and live C. albicans. In histological findings, HO-1 expression in the superficial layers of the oral epithelium following Candida infection was elevated compared to that in healthy epithelium. We then investigated fungal cell wall components involved in induction of HO-1 expression and found that β-glucan-containing particles (β-GPs) increased its expression. Furthermore, β-glucan was observed on the surface of both heat-killed C. albicans and Candida cells that had invaded the oral epithelium. Fungal β-GPs also promoted induction of intracellular reactive oxygen species (ROS), NADPH oxidase activation, and p38 mitogen-activated protein kinase (MAPK) phosphorylation, while those specific inhibitors inhibited the HO-1 expression induced by fungal β-GPs. Moreover, fungal β-GPs induced Nrf2 translocation into nuclei via p38 MAPK signaling, while the HO-1 expression induced by fungal β-GPs was inhibited by Nrf2-specific small interfering RNA (siRNA). Finally, knockdown of cells by HO-1- and Nrf2-specific siRNAs resulted in increased β-GP-mediated ROS production compared to that in the control cells. Our results show that the HO-1 induced by fungal β-GPs via ROS/p38 MAPK/Nrf2 from oral keratinocytes may have important roles in host defense against the stress caused by Candida infection in the oral epithelium.

INTRODUCTION

Oral candidiasis is a superficial mucosal infection caused by Candida species, most commonly, Candida albicans (1, 2). Following adherence to oral mucosa, C. albicans penetrates the epithelial surface at microscopic wound sites (3) and invades the oral epithelium (4). Oral keratinocytes provide the first line of host defense against C. albicans infection (5) and actively respond to live C. albicans organisms by producing inflammatory mediators (6, 7). In an in vitro model, heat-killed C. albicans did not enhance immune responses in the oral epithelium, whereas the contact of live C. albicans organisms with the epithelium was shown to increase the expression of proinflammatory cytokines, such as interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α) (6). In contrast, heat-killed C. albicans and Candida cell wall fractions have been reported to increase the expression of inflammatory mediators, such as IL-8 and granulocyte-macrophage colony-stimulating factor, in oral keratinocytes (8). Therefore, interactions of fungal cell wall components with oral keratinocytes may regulate the stress response against C. albicans infection.

C. albicans and the budding yeast Saccharomyces cerevisiae share similarities in regard to their cell wall structures, in both of which the cell walls are composed of an inner layer of β-glucan covalently linked to a variety of cell surface mannoproteins (9–11). β-Glucan has been shown to induce phagocytosis, cytotoxic activities, and proinflammatory cytokine production in mouse macrophages (12). Furthermore, β-glucan has been observed on the surface of biofilms formed by C. albicans in mice with oropharyngeal candidiasis showing invasion of the tongue mucosa (13). However, it is unknown whether fungal cell wall components, such as β-glucan, participate in the activation of stress-mediated immune responses by oral keratinocytes.

Heme oxygenase 1 (HO-1) is an enzyme that catalyzes the first rate-limiting step in the degradation of free heme to produce carbon monoxide, ferrous iron, and biliverdin (BV) (14). Furthermore, HO-1 is also thought to be a stress-inducible enzyme that mediates antioxidative and cytoprotective effects to maintain cellular redox homeostasis and provide protection against oxidative stress (14). This enzyme is induced by an oxidative stressor, such as hydrogen peroxide, and its inhibition increases hydrogen peroxide-induced oxidative damage (15–17). On the other hand, following its induction by some bacterial components, HO-1 enhances host defense and oxidative signaling in response to bacterial infection. The Gram-negative bacterial outer membrane component lipopolysaccharide (LPS) has been shown to increase HO-1 expression in immune cells, such as macrophages and monocytes (18, 19), while HO-1 was also shown to be increased by the Gram-positive bacterial cell wall component lipoteichoic acid (LTA) in human tracheal smooth muscle cells (20). Although the inducer and signaling events involved in HO-1 expression in oral keratinocytes have not been completely elucidated, the HO-1 induced by microbial components in oral keratinocytes may play a role in protective intercellular stress against oral microorganism infection.

We speculated that cell wall components of C. albicans participate in mediation of the stress responses against C. albicans infection in the oral epithelium. Therefore, we investigated the expression profiles of genes induced by heat-killed C. albicans in oral immortalized (RT7) keratinocytes using a cDNA microarray technique and focused on the HO-1 expression induced by C. albicans and the fungal cell wall component involved in its increase. Furthermore, we examined the mechanisms of the intercellular signaling pathway and antioxidative stress functions involved in induction of HO-1 expression by β-glucan-containing particles (β-GPs), the fungal cell wall components.

RESULTS

Differences in gene expression between heat-killed C. albicans-exposed cells and nontreated cells.

Scatter plot analysis of the microarray signals was performed (Fig. 1). To determine the top 10 genes with a known biological function among the various genes regulated by heat-killed C. albicans, we initially set the cutoff value for gene selection at an 8-fold change in the level of expression by cells treated with heat-killed C. albicans in comparison with their level of expression by nontreated cells. Among the 24,000 genes detected by the cDNA microarray, 33 genes were upregulated greater than 8-fold in heat-killed C. albicans-exposed cells in comparison with their levels of expression in nontreated cells. Pseudogenes were excluded, and the known biological functions of 9 of the 33 genes are shown in Table 1. On the other hand, 2 genes were found to be downregulated greater than 8-fold in heat-killed C. albicans-exposed cells compared to their levels of expression in nontreated cells (Table 1). We confirmed the alterations in expression of the upregulated and downregulated genes in RT7 keratinocytes when exposed to live or heat-killed C. albicans using quantitative reverse transcription (RT-PCR) analysis (Fig. 2). Of the 9 upregulated genes, the expression of 7 was increased by both live and heat-killed C. albicans, while 2 genes downregulated by exposure to heat-killed C. albicans were upregulated in cells exposed to live C. albicans. Among these genes, we focused on HO-1, as it is known to be a stress-inducible gene whose expression is increased by both heat-killed C. albicans and live C. albicans organisms. Furthermore, HO-1 expression was rather well induced by the heat-killed organism, implying its regulation in response to cell wall components.

FIG 1.

Scatter plot analysis revealed several genes that were up- or downregulated by heat-killed C. albicans (HKCA) in RT7 keratinocytes. Cells were incubated with heat-killed C. albicans (10⁸/ml) for 12 h, and then gene expression profiling was performed using a cDNA microarray. The upper 33 circles show genes, including HO-1, in cells exposed to heat-killed C. albicans that were upregulated greater than 8-fold compared to their levels of expression in nontreated cells. The lower 2 circles show genes in cells exposed to heat-killed C. albicans that were downregulated greater than 8-fold compared to their levels of expression in nontreated cells.

TABLE 1.

Genes upregulated and downregulated 8-fold in heat-killed Candida albicans

Gene	Function	GenBank accession no.	Fold change in expression
Upregulated genes
UBE2NL	Enzyme activity	NM001012989	8.08
TRPC	Stress response	NM016113	8.94
HO-1	Stress response	NM002133	9.04
KAP10-2	Physiological	NM198693	9.19
HCATD26	Cell proliferation	NM018687	10.37
PPIAL4	Enzyme activity	NM178230	10.51
RPS4Y2	Enzyme activity	NM001039567	11.47
LCE3A	Cell differentiation	NM178431	12.30
ADAM	Inflammatory response	NM005099	14.23
Downregulated genes
CYP450F24SA	Enzyme activity	NM000782	0.09
FGF-1	Cell growth	NM000800	0.10

FIG 2.

Alterations of expression of 11 genes identified by microarray analysis following exposure of RT7 keratinocytes to live C. albicans (LCA) and heat-killed C. albicans (HKCA). Cells were incubated with live C. albicans cells (10⁵/ml) or heat-killed C. albicans cells (10⁸/ml) for 12 h. Gene mRNA levels relative to the level of the β-actin gene are shown. Values are shown as the fold increase in expression compared to the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. *, significantly different from nontreated cells (paired t test, P < 0.05).

HO-1 expression in healthy and Candida-invaded oral epithelium.

Using immunocytochemistry, we examined the expression of HO-1 in oral keratinocytes, which has not been widely reported, and detected the expression of the gene in the cytoplasm of those cells (Fig. 3A). Immunohistochemical staining showed that HO-1 expression was faint in the basal cell layer of healthy oral epithelium as well as Candida-invaded epithelium (Fig. 3B). On the other hand, HO-1 expression in Candida-invaded oral epithelium in the superficial layer near the site of invasion was significantly higher than that in the superficial layer of healthy oral epithelium (Fig. 3B and C).

FIG 3.

(A) Localization of HO-1 expression in RT7 keratinocytes. Cells were stained with anti-HO-1 or rabbit IgG as a negative control along with Alexa Fluor 568-conjugated rabbit IgG. HO-1 was observed in the cytoplasm of cells, as indicated by red staining. Nuclei were counterstained with DAPI (blue). The experiments were performed at least 3 times, and representative results are shown. (B) HO-1 expression in healthy and Candida-invaded oral epithelium. (a) Healthy epithelium; hematoxylin-eosin (H&E) staining; magnification, ×100. (b) Immunostaining for HO-1; magnification, ×100. (c) Candida-invaded epithelium; H&E staining; magnification, ×100. (d) Candida-invaded epithelium; immunostaining for HO-1; magnification, ×100. (e, f) High-magnification views of Candida-invaded epithelium showing HO-1-positive cells in the superficial layer close to the site of invasion of Candida. (e) PAS staining for Candida hyphae; magnification, ×200. (f) Immunostaining for HO-1; magnification, ×200. The experiments were performed at least 3 times, and representative results are shown. (C) HO-1 expression in the superficial layer of healthy (n = 4) and Candida-invaded (n = 9) oral epithelium. The HO-1-positive rate was determined on the basis of the ratio of the number of HO-1-positive cells to the total number of cells examined. *, P < 0.05, Mann-Whitney U test.

Effects of heat-killed C. albicans, non-albicans Candida species, microbial components, and inflammatory cytokines on HO-1 expression in oral keratinocytes.

We examined the effects of live and heat-killed C. albicans on HO-1 mRNA expression in primary oral keratinocytes. Both live and heat-killed C. albicans organisms increased HO-1 mRNA levels in primary oral keratinocytes obtained from 3 different donors as well as in immortalized cells (see Fig. S2A in the supplemental material). Next, the effects of 2 different heat-killed C. albicans strains on HO-1 mRNA expression in RT7 keratinocytes were examined over time, and both strains caused an increase in a time-dependent manner (Fig. 4A). We also investigated the effects of heat-killed non-albicans Candida species, as well as those of Staphylococcus aureus and Escherichia coli, on HO-1 mRNA expression and found that heat-killed C. glabrata and C. tropicalis as well as C. albicans increased HO-1 mRNA levels in RT7 keratinocytes (Fig. 4B), whereas heat-killed S. aureus and E. coli had no effect (Fig. 4B). Furthermore, we examined the effects of UV-killed C. albicans and found that the increase in HO-1 expression was lower after exposure to UV-killed C. albicans than after exposure to heat-killed C. albicans at the same concentrations (Fig. S2B). Some investigators found that inflammatory cytokines and bacterial components, such as LPS, induce an increase in HO-1 mRNA levels in various types of cells (19, 21). However, the Toll-like receptor (TLR) agonists LPS and Pam3CSK4 and the Th1 cytokines gamma interferon (IFN-γ) and TNF-α had no effects on HO-1 mRNA expression in our experiments (Fig. 4C).

FIG 4.

(A) Effects of heat-killed C. albicans (HKCA) on HO-1 mRNA expression over time. RT7 keratinocytes were incubated with heat-killed C. albicans IFO1385 or IFM4009 (10⁸/ml) for the indicated times, and then RT-PCR assays were performed for HO-1 and β-actin. The experiments were performed at least 3 times, and representative results are shown. (B, C) Effects of heat-killed Candida species, other bacteria, microbial components, and inflammatory cytokines on HO-1 mRNA expression. RT7 keratinocytes were incubated with heat-killed non-albicans Candida species, S. aureus, or E. coli (10⁸/ml), LPS (10 μg/ml), Pam3CSK4 (1 μg/ml), IFN-γ (10 ng/ml), or TNF-α (10 ng/ml) for 12 h. The gene mRNA levels shown are relative to those of the β-actin mRNA gene. Values are shown as the fold increase in expression compared to the levels of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. *, significantly different from the nontreated control (paired t test, P < 0.05). (D) Effects of fungal cell wall component on HO-1 mRNA expression. RT7 keratinocytes were incubated with C. albicans β-glucan-containing particles (CA β-GPs; 200 μg/ml), the alkali-soluble fraction (200 μg/ml), the mannan fraction (200 μg/ml), S. cerevisiae β-glucan-containing particles (SC β-GPs) (200 μg/ml), and water-soluble β-glucan (200 μg/ml) for 12 h. The gene mRNA levels shown are relative to those of the β-actin mRNA gene. Values are shown as the fold increase in expression compared with the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. *, significantly different from the nontreated control (paired t test, P < 0.05). (E) Effects of fungal cell wall component on HO-1 protein expression. RT7 keratinocytes were incubated with heat-killed C. albicans cells (10⁸/ml), C. albicans β-GPs (200 μg/ml), or S. cerevisiae β-GPs (200 μg/ml) for 24 h. Cell extracts were subjected to SDS-PAGE and Western blotting with antibodies against HO-1. The experiments were performed at least 3 times, and representative results are shown.

Effects of β-glucan-containing particles derived from C. albicans and S. cerevisiae on HO-1 expression.

We also examined the ability of C. albicans cell wall components to increase HO-1 expression. Mannan and alkali-soluble fractions extracted from C. albicans had no effect, whereas C. albicans β-glucan-containing particles (β-GPs) dramatically increased the expression of HO-1 mRNA (Fig. 4D), with the same results being found for primary oral keratinocytes obtained from 3 different donors (Fig. S2A). Furthermore, we used β-GPs derived from the yeast S. cerevisiae (purchased from InvivoGen) through a series of alkaline and acid extractions (22). β-GPs from S. cerevisiae increased the level of HO-1 mRNA expression (Fig. 4D). Similarly, fungal β-GPs enhanced the amount of HO-1 protein at a molecular size of 32 kDa (Fig. 4E and S2C). To investigate whether the β-glucan binding receptor is involved in the regulation of HO-1 in oral keratinocytes, the effects of knockdown of Dectin-1, TLR2, and TLR4, well-known β-glucan receptors, on fungal β-GP-induced HO-1 were examined using their specific small interfering RNAs (siRNAs) and neutralizing antibodies, though no effects were seen (Fig. S3). On the other hand, water-soluble β-glucan binds to Dectin-1, and that has an ability to block fungal particulate β-glucan-induced signaling (23). We examined the effect of water-soluble β-glucan derived from S. cerevisiae on fungal β-GP-induced HO-1 expression, though no effects were seen (Fig. S4).

β-Glucan exposure in oral keratinocytes by heat-killed C. albicans or the oral epithelium following invasion by Candida.

The β-glucan in the inner layer of the cell wall in C. albicans is masked in the outer mannan layer (24). To examine the possibility that contact with the β-glucan in oral keratinocytes induces an immune response against Candida infection, we investigated β-glucan exposure on the surface of heat-killed C. albicans and Candida-invaded oral epithelium using a β-glucan-specific antibody. Heat-killed C. albicans organisms were observed as yeast-phase cells, and β-glucan was found to be exposed on the cell surface of heat-killed C. albicans cells (Fig. 5A). Furthermore, β-glucan was found on the surface of Candida-invaded oral epithelium, in accordance with the site of invasion (Fig. 5B).

FIG 5.

(A) β-Glucan exposure in heat-killed C. albicans (HKCA). (Top) Phase-contrast images; (bottom) confocal fluorescence images. Heat-killed C. albicans organisms were stained with anti-β-glucan or mouse IgG as a negative control along with Alexa Fluor 568-conjugated mouse IgG. Red staining indicates exposure of β-glucan in heat-killed C. albicans. The experiments were performed at least 3 times, and representative results are shown. (B) β-Glucan exposure in Candida-invaded oral epithelium. Candida-invaded epithelium was stained with PAS for Candida hyphae or anti-C. albicans along with Alexa Fluor 568-conjugated rabbit IgG. Red staining indicates C. albicans. Candida-infected oral epithelium was stained with anti-β-glucan along with Alexa Fluor 488-conjugated mouse IgG, and green staining indicates exposed β-glucan. Normal healthy epithelium was stained with anti-β-glucan with Alexa Fluor 488-conjugated mouse IgG and used as a negative control. Nuclei were counterstained with DAPI (blue). The experiments were performed at least 3 times, and representative results are shown. Magnifications, ×200.

Effects of fungal β-GPs on induction of intercellular ROS.

Several investigators have reported that oxidative stress contributes to the induction of HO-1 in various cell types (25); thus, we examined the effects of fungal β-GPs on induction of intercellular reactive oxygen species (ROS) in oral keratinocytes. Addition of heat-killed C. albicans or fungal β-GPs dramatically promoted intercellular ROS production within 1 h (Fig. 6A), while the levels of ROS were then decreased with additional exposure to either of those (data not shown). In addition, the ROS production and HO-1 expression induced by C. albicans β-GPs were attenuated by pretreatment with N-acetyl-l-cysteine (NAC), an ROS scavenger, in a dose-dependent manner (Fig. 6B and C).

FIG 6.

(A) Effects of heat-killed C. albicans (HKCA) and fungal β-GPs on induction of intercellular ROS induction. RT7 keratinocytes were labeled with DCF-DA for 30 min and then stimulated with heat-killed C. albicans cells (10⁸/ml), C. albicans β-glucan-containing particles (CA β-GPs) (200 μg/ml), or S. cerevisiae β-glucan-containing particles (SC β-GPs) for 1 h. The fluorescence intensity was then measured. Values are shown as the fold increase in expression compared with the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. *, significantly different from the nontreated controls (paired t test, P < 0.05). (B) Effects of an ROS inhibitor on C. albicans β-GP-induced ROS. Cells were labeled with DCF-DA, incubated with NAC for 1 h, and then stimulated with C. albicans β-GPs (200 μg/ml) Values are shown as the fold increase in expression compared with the level of expression by nontreated cells, and values are presented as the mean ± SD from 3 independent experiments. #, significantly different from C. albicans β-GPs alone (paired t test, P < 0.05). (C) Effects of an ROS inhibitor on C. albicans β-GP-induced HO-1 expression. Cells were incubated with NAC for 1 h and then stimulated with C. albicans β-GPs (200 μg/ml) Values are shown as the fold increase in expression compared with the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. #, significantly different from C. albicans β-GPs alone (paired t test, P < 0.05). (D) Effects of NADPH oxidase inhibitor on heat-killed C. albicans and fungal C. albicans β-GP-induced HO-1 mRNA expression. RT7 keratinocytes were preincubated with apocynin (1 μM) for 1 h and then exposed to heat-killed C. albicans (10⁸/ml) or fungal C. albicans β-GPs (200 μg/ml) for 12 h. The HO-1 mRNA levels shown are relative to the levels of the β-actin mRNA gene. Values are shown as the fold increase in expression compared with the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. #, significantly different from heat-killed C. albicans or fungal C. albicans β-GPs alone (paired t test, P < 0.05). (E) Effects of an NADPH inhibitor on HO-1 protein expression induced by C. albicans β-GPs. RT7 keratinocytes were preincubated with APO (1 μM) or DPI (1 μM) for 1 h and then incubated with C. albicans β-GPs (200 μg/ml) for 24 h. Cell extracts were subjected to SDS-PAGE and Western blotting with antibodies against HO-1. The experiments were performed at least 3 times, and representative results are shown. (F) Effects of fungal β-GPs on NADPH oxidase activation. RT7 keratinocytes were stimulated with C. albicans β-GPs (200 μg/ml) or S. cerevisiae β-GPs (200 μg/ml) for the indicated times. Membrane extract (ME) and cytosolic extract (CE) fractions were prepared and then subjected to Western blot analysis with the anti-p47^phox antibody. Gα proteins (GαS) and GAPDH were used as marker proteins for the membrane and cytosolic fractions, respectively. The experiments were performed at least 3 times, and representative results are shown.

Effects of fungal β-GPs on activation of NADPH oxidase.

Since NADPH oxidase is an important enzymatic source for production of ROS (26), we investigated its role in the ROS generation associated with β-GP-induced HO-1 expression. Initially, we found that the NADPH oxidase inhibitor apocynin (APO) decreased the HO-1 mRNA expression induced by heat-killed C. albicans and fungal β-GPs (Fig. 6D). In addition, β-GP-induced HO-1 proteins were inhibited by another NADPH oxidase inhibitor, diphenyleneiodonium (DPI) (Fig. 6E and S5A). Activated NADPH oxidase is a multimeric protein complex consisting of 3 cytosolic subunits, p47^phox, p67^phox, and p40^phox, and organizes their translocation from the cytosol to the membrane (27). Therefore, we next investigated the effect of fungal β-glucan on the translocation of p47^phox in RT7 keratinocytes and found that it was increased within 2 h (Fig. 6F and S5B).

Effects of fungal β-GPs on p38 MAPK activation.

Members of the mitogen-activated protein kinase (MAPK) family have been reported to mediate HO-1 expression in various cell types (28); thus, we investigated the effects of 3 different MAPK family inhibitors on fungal β-glucan-induced HO-1 expression. Fungal β-GP-induced HO-1 mRNA was inhibited by SB203580, a p38 MAPK inhibitor, but not by PD98059, an extracellular signal-regulated kinase (ERK) inhibitor, or SP600125, a c-Jun N-terminal protein kinase (JNK) inhibitor (Fig. 7A and B and S6A). Furthermore, fungal β-GP treatment increased the level of p38 MAPK phosphorylation (Fig. 7C and S6B), which was attenuated by the NADPH oxidase inhibitor, indicating that fungal β-GP-induced HO-1 expression via NADPH/ROS generation is associated with activation of p38 MAPK (Fig. 7D and S6C).

FIG 7.

(A) Effects of 3 MAPK inhibitors on induction of HO-1 mRNA by fungal β-glucan-containing particles (β-GPs). RT7 keratinocytes were preincubated with SP600125, SB203580, or PD98059 (each 10 μM) for 1 h and then exposed to fungal β-GPs (200 μg/ml) for 12 h. The HO-1 mRNA levels shown are relative to the levels of the β-actin mRNA gene. Values are shown as the fold increase in expression compared with the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. #, significantly different from fungal β-GPs alone (paired t test, P < 0.05). (B) Effects of a p38 MAPK inhibitor on fungal β-GP-induced HO-1 protein. RT7 keratinocytes were preincubated with SB203580 (10 μM) for 1 h and then exposed to fungal β-GPs (200 μg/ml) for 24 h. Cell extracts were subjected to SDS-PAGE and Western blotting with antibodies against HO-1. The experiments were performed at least 3 times, and representative results are shown. (C) Effects of fungal β-GPs on p38 MAPK phosphorylation. RT7 keratinocytes were exposed to fungal β-GPs (200 μg/ml) for the indicated times, after which cell extracts were subjected to SDS-PAGE. p38 MAPK activation was examined by Western blotting with antibodies against phospho-p38 MAPK and total p38 MAPK. The experiments were performed at least 3 times, and representative results are shown. (D) Effects of an NADPH inhibitor on fungal β-GP-induced p38 MAPK phosphorylation. RT7 keratinocytes were preincubated with apocynin (1 μM) for 1 h and then exposed to fungal β-GPs (200 μg/ml) for the indicated times, after which cell extracts were subjected to SDS-PAGE. p38 MAPK activation with antibodies against phospho-p38 MAPK and total p38 MAPK were examined by Western blotting. The experiments were performed at least 3 times, and representative results are shown. P-p38, phosphorylated p38.

Fungal β-GPs induce HO-1 expression via Nrf2 signaling.

Nrf2 is a redox-sensitive basic leucine zipper transcription factor and a member of the NADPH oxidase complex which shows translocation into the nucleus upon activation by oxidative stress and which has been reported to have important effects on antioxidant genes (29, 30). First, we examined the effect of fungal β-GP treatment on Nrf2 expression in nuclear fractions by Western blotting and staining for immunofluorescence. Treatments with heat-killed C. albicans and fungal β-GPs increased the expression of Nrf2 in the nuclear fraction in a time-dependent manner (Fig. 8A and S7A), while Nrf2 expression in the nuclei of cells stimulated with C. albicans β-GPs was also observed by immunofluorescence staining (Fig. 8B). Fungal β-GP-induced Nrf2 expression in the nucleus was inhibited by the p38 MAPK inhibitor SB203580 (Fig. 8C and S7B). Next, to determine whether β-GP-induced HO-1 expression is associated with Nrf2 activation, we performed siRNA-mediated knockdown of Nrf2 (Fig. 8D) and found that siRNA specific for Nrf2 decreased the level of HO-1 protein expression induced by fungal β-GPs (Fig. 8E and S7C). These results indicate that fungal β-GP-induced HO-1 expression is mediated via the p38/Nrf2 pathway.

FIG 8.

(A) Effects of heat-killed C. albicans (HKCA) and fungal β-glucan-containing particles (β-GPs) on Nrf2 expression in nuclei. RT7 keratinocytes were stimulated with heat-killed C. albicans (10⁸/ml) or fungal β-GPs (200 μg/ml) for the indicated times. Whole-cell extracts (WE) and nuclear extracts (NE) were prepared and then subjected to Western blot analysis with the Nrf2 antibody. Lamin B and GAPDH were used as marker proteins for the nuclear and whole-cell fractions, respectively. The experiments were performed at least 3 times, and representative results are shown. (B) Localization of Nrf2 in RT7 keratinocytes treated with C. albicans β-GPs. RT7 keratinocytes were stimulated with C. albicans β-GPs for 4 h and then stained with anti-Nrf2 along with Alexa Fluor 568-conjugated rabbit IgG as the secondary antibody. Red staining indicates Nrf2 and was observed in cell nuclei. The experiments were performed at least 3 times, and representative results are shown. (C) Effects of SB203580 on localization of Nrf2 in nuclei. RT7 keratinocytes were preincubated with SB203580 for 1 h and then stimulated with fungal β-GPs (200 μg/ml) for the indicated times. Nuclear fractions were prepared and subjected to Western blot analysis with the Nrf2 antibody. Lamin B was used as the marker protein for nuclei. The experiments were performed at least 3 times, and representative results are shown. (D) Knockdown of Nrf2 expression by siRNA. Cells were transfected with siRNA specific for Nrf2 (Si Nrf2) for 48 h. The gene mRNA levels shown are relative to the levels of the β-actin mRNA gene. Values are shown as the fold increase in expression compared with the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. #, significantly different from control siRNA (Si Control)-transfected cells (paired t test, P < 0.05). (E) Effects of knockdown of Nrf2 on fungal β-GP-induced HO-1. Cells were transiently transfected with siRNA specific for Nrf2 for 48 h and then exposed to fungal β-GPs (200 μg/ml) for 24 h. Whole-protein extracts of cells were subjected to SDS-PAGE. HO-1 expression was examined by Western blotting with the HO-1 antibody. The experiments were performed at least 3 times, and representative results are shown.

Effects of HO-1 knockdown on C. albicans β-GP-mediated ROS production.

Finally, to examine the direct role of HO-1 in β-GP-mediated intercellular ROS production, RT7 keratinocytes were transfected with HO-1-specific siRNA and then stimulated with heat-killed C. albicans or C. albicans β-GPs for 6 h. Those transfected cells showed decreased levels of HO-1 expression (Fig. 9A). Furthermore, the ROS level found when the cells were stimulated with heat-killed C. albicans or β-GPs was increased compared to that found when the cells were transfected with control siRNA or in wild-type cells (Fig. 9B).

FIG 9.

(A) Knockdown of HO-1 expression by siRNA. Cells were transfected with HO-1-specific siRNA for 48 h. The gene mRNA levels shown are relative to the levels of the β-actin mRNA gene. Values are shown as the fold increase in expression compared with the level of expression by nontreated cells and presented as the mean ± SD from 3 independent experiments. #, significantly different from control siRNA-transfected cells (paired t test, P < 0.05). (B) Effects of knockdown of HO-1 on ROS induction. Cells were transfected with HO-1- or Nrf2-specific siRNA for 48 h, labeled with DCF-DA for 30 min, and then stimulated with heat-killed C. albicans (HKCA) (10⁸/ml) or C. albicans β-glucan-containing particles (CA β-GPs) (200 μg/ml) for 6 h. The fluorescence intensity was then measured. Values are shown as the fold increase in expression compared with the level of expression by nontreated cells stimulated with heat-killed C. albicans or β-GPs and presented as the mean ± SD from 3 independent experiments. *, significantly different from control siRNA-transfected cells stimulated with heat-killed C. albicans (paired t test, P < 0.05); #, significantly different from control siRNA-transfected cells stimulated with C. albicans β-GPs (paired t test, P < 0.05).

DISCUSSION

Oral keratinocytes play a crucial role in protecting the oral epithelium against invasion by C. albicans, as contact of that pathogen with those cells results in proinflammatory cytokine and chemokine production (6, 7). However, some reports have shown that such immune responses were abrogated in oral keratinocytes when heat-killed C. albicans organisms were used (6, 31), though others have reported that higher concentrations of heat-killed C. albicans can modulate immune response factors in oral keratinocytes, such as IL-8 (7, 32). In the present study, we used a cDNA microarray system and identified 9 genes that showed an upregulation of at least 8-fold as well as 2 genes that showed a downregulation of at least 8-fold when the cells were exposed to heat-killed C. albicans. Of the upregulated genes, 7 were upregulated by both live and heat-killed C. albicans. In addition, we found that β-glucan was exposed on the cell surface of heat-killed C. albicans, as previously reported (33). Therefore, contact with cell wall components, such as β-glucan, can trigger and orchestrate an immune response via regulation of differential genes involved in multiple biochemical functions in oral keratinocytes.

Our particular interest in this study was the upregulation of stress response genes, such as HO-1, by heat-killed C. albicans. HO-1 has been identified to be an enzyme that cleaves the toxic heme with carbon monoxide (CO), biliverdin, and iron (27) and has also been shown to have an important role in iron homeostasis, while its constitutive expression has been reported to be low in various tissues and cells and its localization in cytoplasm has been confirmed (14, 25). Our findings showed that HO-1 is constitutively expressed in the cytoplasm of oral keratinocytes, while its expression was faint in the basal cell layer of healthy oral epithelium as well as Candida-infected epithelium. On the other hand, HO-1 expression in Candida-invaded oral epithelium in the superficial layer near the site of invasion was higher than that in the superficial layer of healthy oral epithelium. Indeed, HO-1 is thought to be a stress-inducible enzyme that is upregulated in response to oxidative cellular stress by oxidants, such as hydrogen peroxidase. Furthermore, several other inducers of HO-1, including inflammatory cytokines and bacterial components, have been found in a variety of cell types (25). As for the Th1 cytokines, TNF-α was reported to increase the expression of HO-1 in endothelial and myelomonocytic leukemic cell lines (21, 33) and was also shown to be upregulated by the bacterial component LPS in various immune cells, such as macrophages and monocytes (18, 19). Lipoteichoic acid (LTA), a cell wall component of Gram-positive bacteria, was also found to increase HO-1 expression in human tracheal smooth muscle cells via Toll-like receptor 2 (TLR2) (20). In the present experiments, HO-1 expression in oral keratinocytes was dramatically induced by heat-killed Candida species, whereas TNF-α, LPS, and the TLR2 ligand Pam3CSK4 had no effects. These results indicate that oral keratinocytes specifically induce HO-1 in response to a fungal component rather than a bacterial component or Th1 cytokines and are different from the results presented in other reports.

The pathogenic fungus C. albicans and nonpathogenic fungus S. cerevisiae are similar in regard to their cell wall components (9–11). The walls of those cells are mainly composed of two different types of polysaccharides, mannan in the outer layer and β-glucan in the inner layer, with the major part of β-glucan being a particulate that is essentially insoluble in H₂O or NaOH (34). It has been reported that fungal β-glucan can activate host immune responses in various cell types, while it also promotes direct cellular responses, including phagocytosis in macrophages and dendritic cells (23). Typically, β-glucan in the cell wall is initially masked in live C. albicans cells, while it later becomes exposed under various conditions, such as following drug treatment (24). Dongari-Bagtzoglou et al. reported that the display of β-glucan was found on fungal cell surfaces during Candida biofilm development in both in vivo and in vitro models (13), while our findings of β-glucan in superficial layers in Candida-invaded oral epithelium raises the possibility of exposure of β-glucan on the C. albicans cell surface in the oral epithelium. Therefore, β-glucan exposure in the oral epithelium caused by Candida infection may promote a stress response in oral keratinocytes to increase HO-1 expression.

In our experiments, UV-killed C. albicans induced a lower level of HO-1 expression than heat-killed C. albicans. Although, heat-killed C. albicans showed the greatest cell wall disruption, the general architecture of the inner and outer cell wall layers remained apparent, whereas the cell walls of UV-killed C. albicans showed less visible perturbation of the cell wall structure (35). The significant increase in expression of HO-1 induced by heat-killed C. albicans may be due to disruption of the cell wall and exposure of the underlying β-glucan.

Fungal β-GPs are primarily recognized by the C-type lectin receptor Dectin-1, which is expressed in both immune cells, such as monocytes and macrophages (36, 37), and nonimmune cells, such as human bronchial epithelial and intestinal epithelial cells (38, 39). Dectin-1 recognizes β-glucan, which promotes antifungal effector mechanisms and production of inflammatory mediators (39), and in some cases cooperates with TLR2 and TLR4 to mediate this response (40, 41). Although the function of Dectin-1 and its expression in the oral epithelium are still unknown (42), we detected its expression in RT7 keratinocytes. However, knockdown of Dectin-1, TLR2, and TLR4 in RT7 keratinocytes did not affect the expression of HO-1 induced by fungal β-GPs. On the other hand, soluble β-glucan binds to Dectin-1, which can block fungal β-glucan-induced signaling (23). Our findings also showed that water-soluble β-glucan did not have an effect on fungal β-GP-induced HO-1. Therefore, the mechanism of β-GP-induced HO-1 expression in the oral epithelium appears to be different from the mechanism of induction of the immune response by β-glucan receptors, such as Dectin-1 and TLRs.

ROS are formed in the human oral cavity upon exposure to oxidative stress, which leads to activation of intercellular signals and enhancement of inflammatory reactions (43). Some have reported that fungal β-glucan promotes phagocytosis and fungicidal activity via ROS production in immune cells, such as macrophages and neutrophils (23, 44). NADPH oxidase is also an important enzymatic source for the production of ROS under various pathological conditions, such as through translocation and binding of p47^phox to the membrane complex of NOX2 and p22^phox, which were found to be key events leading to activation of NADPH oxidase and generation of ROS (27). In the present study, the NADPH oxidase inhibitor decreased β-GP-induced HO-1 levels, while fungal β-GPs and heat-killed C. albicans increased NADPH oxidase activation and ROS generation in the early phase. These results show that induction of HO-1 expression is involved in the response to intercellular oxidative stress via the NADPH oxidase activation induced by fungal β-GPs and heat-killed C. albicans. Interestingly, our microarray analysis also revealed that the CYP450F24SA and FGF-1 genes were downregulated by heat-killed C. albicans. HO-1 is involved in the regulation of biotransformation reactions that depend on cytochrome P450 and may directly attenuate the synthesis of cytochrome P450 by metabolizing heme (45). Also, FGF-1 was previously shown to have antagonistic stimulatory and inhibitory effects on NADPH-dependent H₂O₂ generation (46). Together, these results may also implicate the induction of ROS and NADPH activation by heat-killed C. albicans.

Activation of 3 major MAPK families, extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK), and p38 MAPK, has been reported to play important roles in induction of HO-1 gene expression (28). p38 MAPK activation is required for HO-1 induction in response to various inducers, and hemin-induced HO-1 expression was found to be associated with that activation in rat alveolar macrophages (47), while ERK has also been reported to be associated with HO-1 expression in some cell types (48, 49). The present findings show that the HO-1 expression induced by fungal β-GPs is inhibited by the p38 MAPK inhibitor but not by inhibitors of the JNK and ERK pathways and that β-glucan treatment increased p38 MAPK phosphorylation, which was attenuated by the inhibitor of NADPH oxidase. Therefore, MAPK family signaling related to HO-1 expression may differ depending on the inducer and cell type.

In a study that was performed to screen for proteins that interact with the NF-E2 binding motif, the antioxidant Nrf2 was identified (50). Nrf2 is emerging as a known regulator of cellular resistance to oxidants, as well as of expression of genes involved in the oxidative stress response and cytoprotection (51). It has also been reported that activation of Nrf2 induces its nuclear translocation along with HO-1 expression in association with MAPK family member signaling pathways (52). On the other hand, the effect of p38 MAPK on Nrf2 is controversial, as Nrf2 activation and HO-1 expression have been shown to be attenuated by overexpression of p38 MAPK (53), whereas curcumin-induced HO-1 was reported to be involved in Nrf2 activation via p38 MAPK in renal epithelial cells (54). Also, p38 MAPK activation was found to be associated with cadmium-induced HO-1 expression via Nrf2, whereas neither ERK nor JNK was required for that in breast cancer cells (49). In our study, treatment with heat-killed C. albicans and fungal β-GPs resulted in the translocation of Nrf2 into the nucleus, which was blocked by the p38 MAPK inhibitor. Furthermore, knockdown of Nrf2 decreased fungal β-GP-induced HO-1 expression, indicating that p38 MAPK-dependent Nrf2 activation via ROS generation plays a key role in the HO-1 expression induced by β-GPs.

HO-1 is thought to function as an antioxidant factor as part of the anti-inflammatory defense mechanism in response to various stimuli that trigger induction of stress responses (14). Induction of HO-1 occurs in response to various cellular stress conditions, including alteration of oxygen tension and production of inflammatory cytokines in response to ROS (55), while its inhibition increases H₂O₂-induced injury in various cell types, which is associated with oxidative stress (15–17). Also, knockdown of the HO-1 gene by a specific siRNA was reported to induce intracellular ROS generation (56). Furthermore, siRNA specific for HO-1 was shown to augment ROS production in human bronchial epithelial cells caused by exposure to cigarette smoke extract (57). Another study found that IL-19 increased the expression of HO-1 in human vascular smooth muscle cells, and HO-1-specific siRNA-transfected cells increased ROS production in the presence of IL-19 (58). In the present study, we noted that heat-killed C. albicans and fungal β-glucan increased intercellular ROS production within 1 h, after which ROS levels were decreased with additional exposure. However, knockdown of HO-1 and Nrf2 increased the level of ROS production in comparison with that by control siRNA-transfected cells when stimulated with heat-killed C. albicans and β-GPs for 6 h. Oral keratinocytes are exposed to oxidative stress elicited by inflammatory reactions that result from contact with β-glucan during Candida infection. A continuous accumulation of intercellular ROS shifts the balance toward oxidative stress, which has deleterious effects, such as DNA damage, tissue degeneration, and cell apoptosis (59). Thus, intracellular ROS induction results in activation of downregulation pathways for protection against oxidative stress to restore the balance between cellular oxidants and antioxidants (55). Fungal β-GP-induced HO-1 may maintain the oxidant/antioxidant balance and tissue homeostasis against stress caused by Candida infection.

β-GPs cause a stress response in oral keratinocytes following infection, which induces an increase in intercellular ROS, while ROS signaling promotes the p38 and Nrf2 antioxidative pathways to increase HO-1 levels. Therefore, the HO-1 from oral keratinocytes induced by fungal β-GPs via the ROS/p38 MAPK/Nrf2 pathway may have important roles in the oral epithelium as part of the host defense against the stress caused by Candida infection.

MATERIALS AND METHODS

Reagents.

Diphenyleneiodonium (DPI) and N-acetyl-l-cysteine (NAC) were purchased from Sigma-Aldrich (St. Louis, MO, USA), and apocynin (APO) was purchased from R&D Systems (Minneapolis, MN, USA). The signal inhibitors SP600125, SB203580, and PD98059 were purchased from Cayman Chemical Company (Ann Arbor, MI, USA). The Toll-like receptor (TLR) agonists Pam3CSK4 (synthetic triacylated lipopeptide) and Escherichia coli lipopolysaccharide (LPS) were purchased from Imgenex (San Diego, CA, USA). Yeast whole β-glucan particulates (WGPs) and water-soluble β-glucan extracted from S. cerevisiae were purchased from InvivoGen (San Diego, CA, USA). The Th1 cytokines TNF-α and IFN-γ were purchased from R&D Systems. Antibodies used for immunoblotting, immunohistochemistry, and immunofluorescence included anti-HO-1 (catalog number ADI-SPA-896; Enzo Life Sciences, Farmingdale, NY, USA), anti-P47^phox (catalog number BS3261; Bioworld Technology, Inc., St. Louis Park, MN, USA), anti-p38 MAPK (catalog number 9212; Cell Signaling Technology, Danvers, MA, USA), anti-phospo-p38 MAPK (catalog number 4631; Cell Signaling Technology, Danvers, MA, USA), anti-Nrf2 (catalog number BS1258; Bioworld Technology, Inc., St. Louis Park, MN, USA), anti-lamin B (catalog number E13490; Spring Bioscience, Pleasanton, CA, USA), anti-G α proteins (catalog number ab97629; Abcam, Cambridge, MA, USA), anti-GAPDH (anti-glyceraldehyde-3-phosphate dehydrogenase; catalog number MAB374; Millipore Corporate Headquarters, Billerica, MA, USA), (1-3)-β-glucan antibody (Biosupplies Australia, Parkville, Australia), and anti-C. albicans (catalog number 12-6258-1; American Research Products, Belmont, MA). Secondary antibodies for labeling were a horseradish peroxidase (HRP)-conjugated antibody (Amersham Biosciences, Uppsala, Sweden) and Alexa Fluor 488-conjugated rabbit IgG, Alexa Fluor 488-conjugated mouse IgG, and Alexa Fluor 568-conjugated rabbit IgG (all from Invitrogen, Carlsbad, CA, USA). Neutralizing antibodies were anti-Dectin-1 IgG, anti-TLR2 IgG, and mouse IgG (all from InvivoGen) and anti-TLR3 IgG (catalog number ab12085; Abcam).

Cell lines.

Keratinocytes of the RT7 immortalized human oral keratinocyte cell line were established by transfection of human telomerase reverse transcriptase and E7, as previously described (60), and were then cultured in keratinocyte growth medium (KGM) supplemented with human epidermal growth factor, insulin, hydrocortisone, calcium, bovine pituitary extract, gentamicin sulfate, and amphotericin B (Lonza, Walkersville, MD). We also prepared primary cultures of human oral keratinocytes obtained from healthy gingival tissues excised during extraction of impacted teeth from 3 different patients, as previously reported (61). Informed consent for acquisition of those tissues was obtained according to a protocol approved by the Ethical Committee of Hiroshima University. The cells (primary keratinocytes 1, 2, and 3) were used after 3 to 5 passages.

Tissue samples.

Specimens were obtained from patients with Candida-invaded oral epithelium, determined on the basis of histological findings of invasive hyphal formation in the oral epithelium, while healthy oral mucosal tissues collected during oral surgery were used as control samples. All specimens were obtained after receiving informed consent, using a protocol approved by the Institutional Review Board of Hiroshima University, Hiroshima, Japan, and related facilities. Tissue blocks were cut thinly and then deparaffinized and subjected to hematoxylin-eosin (H&E) and periodic acid-Schiff (PAS) staining for immunohistochemistry and immunofluorescence examinations.

Microorganisms and growth conditions.

C. albicans IFO1385 and IFM48311, Candida glabrata IFM5678, Candida tropicalis IF5797, Staphylococcus aureus 209P, and E. coli HB101 were used in this study.

IFM48311, IFM5678, and IF5797 were obtained from the Chiba University Research Center for Pathogenic Fungi and Microbial Toxicoses, while IFO1385, 209P, and HB101 were generously provided by the Department of Bacteriology, Hiroshima University Graduate School of Biomedical Sciences. All yeast cells were harvested after growth in Sabouraud's broth medium at 37°C overnight, while S. aureus and E. coli were grown in a brain heart infusion medium at 37°C. These cells were washed twice with phosphate-buffered saline (PBS) and heat killed at 60°C for 30 min (7, 62). UV-killed cells were exposed to UV radiation (100 mJ/cm²) in a UV-DNA cross-linker device (catalog number CL-1000; UVP, Upland, CA, USA). Live, heat-killed, and UV-killed cells were separately resuspended in medium before the assays.

β-Glucan-containing particles and mannan extraction from C. albicans.

C. albicans β-glucan-containing particles were obtained using a hot alkali and acid method, as previously described (63). C. albicans IFO1385 was cultured in 1,200 ml of Sabouraud's broth medium and then washed with PBS and suspended in 1% (wt/vol) NaOH at 100°C for 24 h, and the supernatant for the alkali-soluble fraction was collected using the following method. The insoluble residue was treated with 0.5 M acetic acid at 80°C for 24 h and subjected to repeated washings with PBS, and then the pellet was lyophilized and stored at 4°C. We were able to obtain 1.01 g of those β-glucan-containing particles (β-GPs). The total β-glucan concentration in 10 mg of β-GPs from C. albicans was 61.7% ± 4.26%, and that in 10 mg of β-GPs from S. cerevisiae (WGPs; Invivogen) was 67.5% ± 6.07%. These values were determined using a β-glucan assay kit (enzymatic yeast β-glucan assay kit; Megazyme, Wicklow, Ireland). We used the mass of β-GPs noted in those experiments as the amount to be added to cell cultures. The supernatant for the alkali-soluble fraction was neutralized with glacial acetic acid and then concentrated by flash evaporation at 30°C, dialyzed against deionized water for 24 h, and lyophilized before use as the alkali-soluble fraction (64). Mannan from C. albicans was isolated as previously described by Kogan et al. (65). Briefly, C. albicans IFO1385 was suspended in 2% (wt/vol) KOH and heated for 1 h at 100°C, and then insoluble residues were separated by centrifugation and mannan was precipitated from the supernatant using Fehling's reagent. Mannan-copper complex sediment was dissolved in a minimum volume of 3 M HCl and added by drops to methanol-acetic acid at a ratio of 8:1 (vol/vol). The procedures for dissolution and precipitation were repeated twice. Finally, the sediment was separated, dissolved in distilled water, dialyzed for 24 h, and lyophilized.

RNA extractions.

Initially, we examined the effects of C. albicans on cytotoxicity in a preliminary experiment using a cytotoxicity detection kit (Roche Applied Science). Neither live C. albicans cells (10⁵ CFU/ml; ratio of RT7 keratinocytes to live C. albicans cells, approximately 5:1) nor heat-killed C. albicans cells (10⁸ CFU/ml) caused a significant increase in lactate dehydrogenase release by RT7 keratinocytes relative to that from the noninfected control cultures after 12 and 24 h (see Fig. S1 in the supplemental material). Therefore, those concentrations of live and heat-killed C. albicans cells were used in the following examinations. For microarray analysis, RT7 keratinocytes were seeded into 10-cm cell culture plates and cultured until reaching 70 to 80% confluence. Then, after the plates were washed with PBS and the medium was exchanged for antibiotic-free medium, the cells were incubated with heat-killed C. albicans (10⁸ CFU/ml) for 12 h. For RT-PCR analysis, RT7 keratinocyte and primary cell cultures were seeded into 6-well cell culture plates and cultured until reaching 70 to 80% confluence. Then, after they were washed with PBS and the medium was exchanged for antibiotic-free medium, the cultures were incubated with heat-killed C. albicans (10⁸ CFU/ml) or live C. albicans (10⁵ CFU/ml) for 12 h. RNA from those cultures was extracted using an RNeasy minikit (Qiagen, Hilden, Germany). Single-stranded cDNA for use as a PCR template was synthesized using a first-strand cDNA synthesis kit (Amersham Biosciences) and then subjected to microarray analysis, RT-PCR, and real-time PCR assays.

Microarray analysis.

Microarray analysis was performed using a NimbleGen microarray system containing 24,000 genes from 60-mer oligonucleotides. cDNA was cleaned and labeled in accordance with the NimbleGen gene expression analysis protocol (NimbleGen Systems, Inc., Madison, WI, USA). Following hybridization and washing, the array slides were scanned using an Axon GenePix 4000B scanner (Molecular Devices Corporation, Sunnyvale, CA, USA) piloted by GenePix Pro (version 6.0) software (Axon). Scanned images were then imported into NimbleScan software (NimbleGen Systems, Inc., Madison, WI, USA). Expression data were normalized using quantile normalization and the robust multichip average (RMA) algorithm included as part of NimbleScan software.

RT-PCR and real-time PCR assays.

Synthesized cDNA was used for quantitative PCR analysis with the oligonucleotide primers shown in Table 2. Quantitative PCR analysis was performed using a CFX Connect real-time PCR detection system (Bio-Rad Laboratories, Inc.) and SYBR green master mix (Applied Biosystems) for 40 cycles at 95°C for 15 s and then at 60°C for 60 s. Chemokine mRNA levels are shown relative to the levels of β-actin mRNA, the internal control, and are presented as the mean ± standard deviation (SD) from 3 independent experiments. The RT-PCR conditions with the above-described primers prepared from synthesized cDNAs were 1 cycle of 95°C for 15 min; 40 cycles of 95°C for 2 min, 55°C for 30 s, and 72°C for 1 min; and 1 cycle of 72°C for 7 min. The products were analyzed on 2% agarose gels containing SYBR green (Invitrogen). β-Actin was included as an internal control.

TABLE 2.

Primer sequences

Gene	Primer sequence
Ubiquitin-conjugating enzyme E2N-like (UBE2NL)	5′-GCCCGTTATTTTCATGTGGT-3′
	5′-CATCTGGATTGGGAGCATTT-3′
Transient receptor potential channel (TRPC)	5′-CCTCCCATGGAGTCACAGTT-3′
	5′-GTTCAGCACAGCCTTCATCA-3′
Heme oxygenase 1 (HO-1)	5′-TCCGATGGGTCCTTACACTC-3′
	5′-ATTGCCTGGATGTGCTTTTC-3′
Keratin-associated protein 10-2 (KAP10-2)	5′-TGTCTGCTGCAAGTCCATCT-3′
	5′-GAGAGGAGCCAGTGAGCATC-3′
Hepatocellular carcinoma-associated gene TD26 (HCATD26)	5′-ACATACATCGGGTGTGAGCA-3′
	5′-GAATCCACTCTCCCAGGTCA-3′
Peptidylprolyl isomerase A (cyclophilin A)-like 4 (PPIAL4)	5′-AGGGTTCCTGCTTTCACAGA-3′
	5′-GTCTTGGCAGCACAGATGAA-3′
Ribosomal protein S4, Y linked 2 (RPS4Y2)	5′-GGTTGCACCGTAAAAGGAGA-3′
	5′-CGAACCTTGCCATCAATTTT-3′
Late cornified envelope 3A (LCE3A)	5′-CAGCAGAACCAGCAGCAGT-3′
	5′-CTGACCACTTCCCCTGTCAC-3′
ADAM metallopeptidase (ADAM)	5′-AGGCAGTGATGGGTTAGTGG-3′
	5′-TATCACCACCACCCTGGATT-3′
Cytochrome P450, family 24, subfamily A (CYP450F24SA)	5′-GGCAACAGTTCTGGGTGAAT-3′
	5′-TATTTGCGGACAATCCAACA-3′
Fibroblast growth factor 1 (FGF-1)	5′-GTGTGTTAAGGGGTCGGCTA-3′
	5′-CATGGAGATGCCATCCTTCT-3′

Immunofluorescence microscopy.

Cells were seeded into chamber slides (Matsunami Glass, Osaka, Japan) and fixed in 4% paraformaldehyde in PBS for 15 min, followed by permeabilization with 0.2% Triton X-100 in PBS for 5 min and incubation overnight with anti-HO-1 (1:200) and anti-Nrf2 (1:200) diluted in PBS containing 5% bovine serum albumin. Next, the cells were washed and incubated with diluted Alexa Fluor 568-conjugated rabbit IgG (diluted 1:1,000) for 1 h. Vectashield antifade medium containing DAPI (4′,6-diamidino-2-phenylindole; Vector Laboratories, Burlingame, CA, USA) was used to mount the cells. Fluorescent and phase-contrast images were visualized using a BZ-9000 microscope (Keyence, Osaka, Japan).

Immunohistochemistry.

Immunohistochemical staining was performed using an immunoperoxidase technique following antigen retrieval with microwave treatment (500 W, 10 min) in citrate buffer (pH 6.0). Following peroxidase blocking with 3% H₂O₂ in methanol for 10 min, the specimens were blocked with PBS containing 5% normal horse serum (Vector Laboratories, Inc., Burlingame, CA). For staining, anti-HO-1 (diluted 1:200) was used. After 6 h of incubation with the primary antibody at room temperature, the specimens were briefly rinsed with PBS and incubated with an HRP-conjugated antibody as the secondary antibody (diluted 1:1,000) for 1 h at room temperature. The specimens were rinsed with PBS and incubated with diaminobenzidine (Dako, Tokyo, Japan), and then counterstaining was done using hematoxylin solution (Santa Cruz Biotechnology, CA, USA). Specificity was ascertained by substituting normal rabbit IgG for the primary antibody. Next, the proportion of epithelial cell cytoplasm that stained brown in ×200 microscopic fields was calculated for each case by two investigators. We counted at least 1,000 epithelial cells in the superficial layers of Candida-infected epithelium and normal healthy epithelium specimens. The HO-1-positive rate was determined on the basis of the ratio of the number of HO-1-positive cells to the total number of cells counted.

Immunofluorescence staining.

We used an anti-β-glucan antibody (diluted 1:100) and Alexa Fluor 488-conjugated mouse IgG (1:1,000) as the secondary antibody for β-glucan staining, while we used an anti-C. albicans antibody (diluted 1:100) and Alexa Fluor 568-conjugated rabbit IgG (1:1,000) as the secondary antibody for C. albicans staining. Vectashield antifade medium containing DAPI was used to mount the specimens. Fluorescent images were visualized using a BZ-9000 microscope (Keyence, Osaka, Japan).

Preparation of whole-cell extracts and membrane and nuclear fractions.

Cell cultures were washed with ice-cold PBS and subjected to lysis with sodium dodecyl sulfate (SDS) sample buffer using a mammalian cell lysis kit (Sigma-Aldrich) to yield whole-cell extracts. Membrane and cytosolic fractions were extracted from the cell cultures using a transmembrane protein extraction kit (Millipore Corporation, Billerica, MA, USA). Nuclear fractions were extracted from the cell cultures using a nuclear extraction kit (Cayman Chemical Company). In a preliminary study, Western blotting with cell extracts was performed using these kits with Gαs and Lamin B as internal controls, and the results confirmed that the cultured cells could be divided into those fractions (data not shown).

Western blotting.

Protein concentrations were determined using a protein assay kit (Bio-Rad Laboratories), and then each sample was separated on a 10% SDS-polyacrylamide gel and transferred to a polyvinylidene fluoride membrane (Amersham Biosciences). After incubation with the specific antibody, immunoblots were labeled with the HRP-conjugated secondary antibody and developed using an enhanced chemiluminescence Advance Western blotting detection kit (GE Healthcare Life Sciences, Tokyo, Japan). Image data were analyzed using an LAS 4000 mini-imaging system (Fuji Film, Tokyo, Japan). Protein expression was evaluated by comparing the integrated densities of the phosphorylated bands revealed by Western blotting. Densitometric scanning was performed using Kodak Digital Science 1D software (Eastman Kodak, Rochester, NY, USA), and the levels of total proteins were compared with the total GAPDH protein level.

Measurement of intracellular ROS accumulation.

The fluorescent probe 2′,7′-dichlorodihydrofluorescein diacetate (DCF-DA; Sigma-Aldrich) was used to monitor net intracellular accumulation of ROS (66). RT7 keratinocytes were washed with PBS and incubated in cell medium containing 10 μM DCF-DA at 37°C for 45 min, and then the medium was removed and replaced with fresh medium. Next, the cells were incubated with various concentrations of β-glucan and washed 3 times with PBS. The fluorescent intensity was measured at an excitation wavelength of 485 nm and an emission wavelength of 530 nm using a fluorescence microplate reader (Appliskan; Thermo Scientific, Waltham, MA, USA), and the results are presented in relative fluorescence units.

siRNA.

Small interfering RNA (siRNA) specific for Nrf2, HO-1, Dectin-1, TLR2, and TLR4 and a negative-control siRNA were purchased from Invitrogen. RT7 keratinocytes were transiently transfected with various combinations of those siRNAs using the Lipofectamine RNAi max reagent (Invitrogen), according to the manufacturer's recommendations.

Statistical analysis.

Data were analyzed using Student's t test, one-way analysis of variance (ANOVA), and the Mann-Whitney U test, and the values obtained are presented as the mean ± standard deviation.