LDOC1 Suppresses Microbe-Induced Production of IL-1β in Human Normal and Cancerous Oral Cells through the PI3K/Akt/GSK-3β Axis

Simple Summary Oral microbes often proliferate due to poor oral hygiene (POH). POH is associated with OSCC (oral squamous cell carcinoma). We investigated the role of LDOC1 in the production of IL-1β, an oncogenic proinflammatory cytokine in OSCC, induced by microorganisms in human oral cells. Candida albicans (CA) was detected in OSCC tissues. CA and the oral bacterium Fusobacterium nucleatum stimulate higher levels of IL-1β production in LDOC1-deficient OSCC cells than in LDOC1-expressing oral cells. CA SC5314 increased OSCC incidence in carcinogen-treated mice. Loss and gain of LDOC1 function resulted in increased and decreased, respectively, CA SC5314-induced IL-1β production. LDOC1 deficiency increased active pAktS473 upon SC5314 stimulation and inactive pGSK-3βS9 phosphorylated by pAktS473. PI3K and Akt inhibitors and expression of constitutively active mutant GSK-3βS9A reduced the SC5314-stimulated IL-1β production in LDOC1-deficient cells. These results indicate that the PI3K/Akt/pGSK-3β signaling contributes to LDOC1-mediated inhibition of microbe-induced IL-1β production, suggesting LDOC1 may determine the role of oral microbes in POH-associated OSCC. Abstract Poor oral hygiene (POH) is associated with oral squamous cell carcinoma (OSCC). Oral microbes often proliferate due to POH. Array data show that LDOC1 plays a role in immunity against pathogens. We investigated whether LDOC1 regulates the production of oral microbe-induced IL-1β, an oncogenic proinflammatory cytokine in OSCC. We demonstrated the presence of Candida albicans (CA) in 11.3% of OSCC tissues (n = 80). CA and the oral bacterium Fusobacterium nucleatum stimulate higher levels of IL-1β secretion by LDOC1-deficient OSCC cells than by LDOC1-expressing oral cells. CA SC5314 increased OSCC incidence in 4-NQO (a synthetic tobacco carcinogen) and arecoline-cotreated mice. Loss and gain of LDOC1 function significantly increased and decreased, respectively, CA SC5314-induced IL-1β production in oral and OSCC cell lines. Mechanistic studies showed that LDOC1 deficiency increased active phosphorylated Akt upon CA SC5314 stimulation and subsequent inhibitory phosphorylation of GSK-3βS9 by activated Akt. PI3K and Akt inhibitors and expression of the constitutively active mutant GSK-3βS9A significantly reduced the CA SC5314-stimulated IL-1β production in LDOC1-deficient cells. These results indicate that the PI3K/Akt/pGSK-3β signaling pathway contributes to LDOC1-mediated inhibition of oral microbe-induced IL-1β production, suggesting that LDOC1 may determine the pathogenic role of oral microbes in POH-associated OSCC.


Introduction
Approximately 350,000 new cases of oral cancer are diagnosed annually worldwide [1]. Oral squamous cell carcinomas (OSCCs) account for over 90% of oral cancers [2,3]. The risk of OSCC is substantially increased by alcohol consumption, betel quid chewing, and cigarette smoking (collectively termed "ABC") [4,5]. However, most consumers of alcohol, betel quid, and tobacco will not develop OSCC in their lifetime, indicating that other risk factors are involved. The American Society of Clinical Oncology stated that factors other than ABC, including gender, poor oral hygiene (POH), and a weak immune system, can increase a person's risk of developing oral cancer [6]. Personal oral hygiene and immunity have profound impact on oral microbiota, which may be an important factor in the OSCC tumor microenvironment. The relationship between oral microbes and OSCC has been noted [7][8][9]. Over 700 species of microorganisms have been identified in the oral cavity of humans [10]. Dysbiosis of oral microbial flora induces oral inflammation and contributes to numerous human diseases [11], including OSCC. Oral microbial imbalance dominated by Candida albicans (CA) was identified in OSCC patients [7][8][9]. Analysis of oral microbiota performed in healthy control subjects and patients with OSCC revealed that the species richness and diversity were significantly lower in patients with OSCC. CA was the most abundant microorganisms identified in the oral cavities of patients with OSCC [8]. CA can cause various oral mucosal lesions, including chronic hyperplastic candidiasis (also known as candidal leukoplakia) [9,12], which may increase the risk of OSCC. Furthermore, CA may promote oral malignant transformation from precancer lesions, such as leukoplakia [13]. CA is recognized as an unharmful commensal fungal species in healthy individuals. It can, however, become pathogenic in an immunocompromised person. Consistently, several studies have reported the development of oral carcinoma in immunocompromised patients with chronic mucocutaneous candidiasis [14][15][16]. Besides CA, an epidemiologic study identified that three species of periodontal bacteria, Prevotella tannerae, Fusobacterium nucleatum (FN), and Prevotella intermedia, had a positive association with OSCC risk, and this relationship may be influenced by lifestyle and genetic factors [17]. The oncogenic role of FN has been noted in colorectal cancer [18,19]. Previous investigation showed that FN can induce immune suppression in gut mucosa by suppressing the functions of immune cells [18]. Although the association between CA and FN and OSCC has been noted, it is not clear whether these microorganisms are protumor factors of POH-associated OSCC, and what may be involved in oral microbial carcinogenesis.
Numerous studies have investigated salivary transcriptome and proteome to identify molecules with diagnostic value for OSCC. The results have revealed the potential of specific deregulated cytokines, such as IL-1β, IL-6, and IL-8, as predictive salivary biomarkers for the early diagnosis of oral cancer [20][21][22][23]. Not only a OSCC biomarker, IL-1β also promotes the development and progression of OSCC in cell-based assays and mouse models by substantially inducing IL-6 and IL-8 expression to create a pro-oncogenic microenvironment through autocrine signaling [24]. Moreover, IL-1β transactivates the epidermal growth factor receptor through the CXCL1-CXCR2 axis in oral Gel views of PCR products obtained with DNA extracted from OSCC biopsies. M, molecular weight marker (100-bp ladder); −, negative control; +, PCR carried out with DNA isolated from CA SC5314 as positive control. The size of the CA PCR product is indicated by an arrow (108 bp). Human β-actin PCR product used as internal control for each OSCC samples. (c) CA invade cultured human untransformed CGHNK2 and OSCC TW2.6 cell lines. Immunofluorescence confocal micrograph of CGHNK2 and TW2.6 cells after coculture with CA strain SC5314 expressing GFP (SC5314-GFP, MOI of 100) for 1 h. Invading intracellular CA (green) were confirmed by the single optical section through the host cells and indicated by white arrows. Fixed CGHNK2 and TW2.6 cells were stained with phalloidin-TRITC (red) to show filamentous actin. The nuclear counter stain is DAPI (4′,6-diamidino-2-phenylindole, blue). Bar, 10 μm.

Oral Microbes May Contribute to High Levels of IL-1β in OSCC with LDOC1 Downregulation
We then examined whether CA and FN can stimulate human oral cells to secrete IL-1β. After cells were cocultured with preparations of either live (L) or heat-killed (HK) CA SC5314 (MOI of 0.5) for 24 h, IL-1β levels in the conditioned medium of all cells tested were significantly elevated, with LDOC1-deficient TW2.6 cells exhibiting a larger increase compared with LDOC1-expressing CGHNK2 cells. In both cell lines tested, L CA more strongly induced IL-1β than HK CA (Figure 2a). In addition to CA, live FN (OD600 = 0.08) can also stimulate CGHNK2 and TW2.6 cell lines to secrete IL-1β during coculture for 6 h (Figure 2b). Similarly, the IL-1β levels induced by FN were significantly higher in LDOC1-deficient TW2.6 than those of LDOC1-expressing CGHNK2 (Figure 2b). These results suggested that the presence of oral microbes, such as CA and FN, may contribute to high levels of IL-1β in OSCC tumors, especially in OSCC cells with LDOC1 downregulation. To assess the correlation between LDOC1 and IL-1β, we explored the mRNA expression of LDOC1 and IL-1β in OSCC specimens by surveying public microarray data deposited on the Oncomine website. In Peng's study [34], mRNA expression of LDOC1 was significantly downregulated in OSCC tumors compared with normal oral tissues (n = 79; Figure 2c). However, oral tumors displayed an obvious increase in IL-1β mRNA expression compared with noncancerous oral tissues (Figure 2c). An inverse correlation between mRNA expression of LDOC1 and IL-1β in OSCC and peritumoral oral tissues supports a potential role of LDOC1 in the production of microbe-induced IL-1β. Bar, 100 µm. (b) Gel views of PCR products obtained with DNA extracted from OSCC biopsies. M, molecular weight marker (100-bp ladder); −, negative control; +, PCR carried out with DNA isolated from CA SC5314 as positive control. The size of the CA PCR product is indicated by an arrow (108 bp). Human β-actin PCR product used as internal control for each OSCC samples. (c) CA invade cultured human untransformed CGHNK2 and OSCC TW2.6 cell lines. Immunofluorescence confocal micrograph of CGHNK2 and TW2.6 cells after coculture with CA strain SC5314 expressing GFP (SC5314-GFP, MOI of 100) for 1 h. Invading intracellular CA (green) were confirmed by the single optical section through the host cells and indicated by white arrows. Fixed CGHNK2 and TW2.6 cells were stained with phalloidin-TRITC (red) to show filamentous actin. The nuclear counter stain is DAPI (4 ,6-diamidino-2-phenylindole, blue). Bar, 10 µm.

Oral Microbes May Contribute to High Levels of IL-1β in OSCC with LDOC1 Downregulation
We then examined whether CA and FN can stimulate human oral cells to secrete IL-1β. After cells were cocultured with preparations of either live (L) or heat-killed (HK) CA SC5314 (MOI of 0.5) for 24 h, IL-1β levels in the conditioned medium of all cells tested were significantly elevated, with LDOC1-deficient TW2.6 cells exhibiting a larger increase compared with LDOC1-expressing CGHNK2 cells. In both cell lines tested, L CA more strongly induced IL-1β than HK CA (Figure 2a). In addition to CA, live FN (OD 600 = 0.08) can also stimulate CGHNK2 and TW2.6 cell lines to secrete IL-1β during coculture for 6 h (Figure 2b). Similarly, the IL-1β levels induced by FN were significantly higher in LDOC1-deficient TW2.6 than those of LDOC1-expressing CGHNK2 (Figure 2b). These results suggested that the presence of oral microbes, such as CA and FN, may contribute to high levels of IL-1β in OSCC tumors, especially in OSCC cells with LDOC1 downregulation. To assess the correlation between LDOC1 and IL-1β, we explored the mRNA expression of LDOC1 and IL-1β in OSCC specimens by surveying public microarray data deposited on the Oncomine website. In Peng's study [34], mRNA expression of LDOC1 was significantly downregulated in OSCC tumors compared with normal oral tissues (n = 79; Figure 2c). However, oral tumors displayed an obvious increase in IL-1β mRNA expression compared with noncancerous oral tissues (Figure 2c). An inverse correlation Cancers 2020, 12, 3148 5 of 20 between mRNA expression of LDOC1 and IL-1β in OSCC and peritumoral oral tissues supports a potential role of LDOC1 in the production of microbe-induced IL-1β. LDOC1-deficient TW2.6 cells secrete more microbe-induced IL-1β as compared with LDOC1expressing CGHNK2 cells. Cytometric bead array (CBA) analysis of IL-1β in conditioned medium of CGHNK2 and TW2.6 cell lines without microbes or with either live (L) or heat-killed (HK) CA SC5314 (MOI 0.5) (a) and L FN (OD600 = 0.08) (b) for 24 and 6 h, respectively. Results are presented as concentrations (pg/mL). Data are representative of at least three independent experiments (mean + S.D.), analyzed using the ANOVA test. * p < 0.05 and ** p < 0.01. (c) The mRNA expression profiles of LDOC1 and IL-1B in OSCC tumors (n = 57) and normal oral tissues (n = 22) were obtained from publicly available microarray data sets (GSE25099) [34] in Oncomine (https://www.oncomine.com).

CA Infection Promoted the Development of OSCC in the Synthetic Tobacco-Related Carcinogen 4-NQO and the Areca Nut Alkaloid Arecoline-Cotreated Mice
To assess the role of CA in oral malignant transformation, we used an OSCC mouse model, established through co-induction of 4-NQO and arecoline (NA), and studied whether the interplay between OSCC-associated carcinogens and CA accelerated the development of OSCC [24]. Figure 3a is a schematic representation of the animal study design. Except for the control mice group was not fed with NA (C0 group), all mice were given water containing NA for 8 or 10 weeks (C8 and C10 groups, respectively). Oral lesions were created by brushing the tongue to mimic the trauma generated by betel quid chewing. Starting from week 8, once a week for 8 weeks mice were either painted on the site of tongue lesions with a brush adhered with live CA SC5314 (1 × 10 6 in 20 μL of 0.9% NaCl) (F+) or painted with normal saline as a vehicle control (F−). We collected oral swabs at week 8, 16, and 28 and tested for CA infection by analysis using CHROMagar Candida plates as described in Materials and Methods. CA infection was determined with the oral swabs obtained at week 28. As shown in Figure 3b, CA SC5314 infection was not observed in mice without treatment of NA (C0 group). For the C8F+ and C10F+ groups, the infection rate was correlated with the period for combined treatment of NA. After sacrifice, GMS staining was carried out with the sections of tongue specimen to cross-validate for CA infection. The representative GSM staining images of the CApositive samples from C8F+ and C10F+ groups are shown (the right panels of Figure 3b). These data suggest a correlation between NA exposure and CA infection. All mice were euthanized at the end of week 28, and the histological slides of tongue tissues were subjected to H&E staining for pathological analysis. Summary statistics based on the incidence of specific pathological features in each group are illustrated in Figure 3c and Table 1. The development of OSCC was observed in all NA-cotreated groups, namely the C8 group (40% and 30% for F-and F+ groups, respectively) and the C10 group (30% and 90% for F− and F+ groups, respectively), but was not observed in the C0 group. These results suggest that the presence of CA alone was not sufficient to induce oral carcinogenesis, The mRNA expression profiles of LDOC1 and IL-1B in OSCC tumors (n = 57) and normal oral tissues (n = 22) were obtained from publicly available microarray data sets (GSE25099) [34] in Oncomine (https://www.oncomine.com).

CA Infection Promoted the Development of OSCC in the Synthetic Tobacco-Related Carcinogen 4-NQO and the Areca Nut Alkaloid Arecoline-Cotreated Mice
To assess the role of CA in oral malignant transformation, we used an OSCC mouse model, established through co-induction of 4-NQO and arecoline (NA), and studied whether the interplay between OSCC-associated carcinogens and CA accelerated the development of OSCC [24]. Figure 3a is a schematic representation of the animal study design. Except for the control mice group was not fed with NA (C0 group), all mice were given water containing NA for 8 or 10 weeks (C8 and C10 groups, respectively). Oral lesions were created by brushing the tongue to mimic the trauma generated by betel quid chewing. Starting from week 8, once a week for 8 weeks mice were either painted on the site of tongue lesions with a brush adhered with live CA SC5314 (1 × 10 6 in 20 µL of 0.9% NaCl) (F+) or painted with normal saline as a vehicle control (F−). We collected oral swabs at week 8, 16, and 28 and tested for CA infection by analysis using CHROMagar Candida plates as described in Materials and Methods. CA infection was determined with the oral swabs obtained at week 28. As shown in Figure 3b, CA SC5314 infection was not observed in mice without treatment of NA (C0 group). For the C8F+ and C10F+ groups, the infection rate was correlated with the period for combined treatment of NA. After sacrifice, GMS staining was carried out with the sections of tongue specimen to cross-validate for CA infection. The representative GSM staining images of the CA-positive samples from C8F+ and C10F+ groups are shown (the right panels of Figure 3b). These data suggest a correlation between NA exposure and CA infection. All mice were euthanized at the end of week 28, and the histological slides of tongue tissues were subjected to H&E staining for pathological analysis. Summary statistics based on the incidence of specific pathological features in each group are illustrated in Figure 3c and Table 1. The development of OSCC was observed in all NA-cotreated groups, namely the C8 group (40% and 30% for F-and F+ groups, respectively) and the C10 group (30% and 90% for F− and F+ groups, respectively), but was not observed in the C0 group. These results suggest that the presence of CA alone was not sufficient to induce oral carcinogenesis, but oral infection of CA acts as a promoting factor of OSCC in mouse with long-term NA exposure. Interestingly, mice from the C0F+ group did not show CA colonization but developed either moderate (n = 3) or severe (n = 2) dysplasia, implying that oral lesions repeatedly exposed to CA may be harmful, even if they are not infected (Figure 3b). Representative images of H&E staining for tongue biopsies in all pathological categories are presented in Figure 3d. but oral infection of CA acts as a promoting factor of OSCC in mouse with long-term NA exposure. Interestingly, mice from the C0F+ group did not show CA colonization but developed either moderate (n = 3) or severe (n = 2) dysplasia, implying that oral lesions repeatedly exposed to CA may be harmful, even if they are not infected (Figure 3b). Representative images of H&E staining for tongue biopsies in all pathological categories are presented in Figure 3d.

LDOC1 Is a Negative Regulator of IL-1β Production Induced by CA SC5314
In the groups of mice with long-term NA exposure (C10 groups), oral infection of CA increased the incidence of OSCC, while in the NA-untreated mice groups (C0 groups), neither CA infection nor OSCC was observed in the oral lesions repeatedly treated with CA (Figure 3b,c and Table 1). These results suggest that exposure to NA may be a prerequisite for CA to promote the development of OSCC. Our previous work demonstrated that LDOC1 can be epigenetically silenced by exposure to cigarette smoke condensates [30]. In addition, an inverse relationship between LDOC1 expression and IL-1β induced by CA SC5314 or FN was observed in the cell-based assays (Figure 2a,b). We hypothesized that long-term exposure to NA may cause LDOC1 silence, thereby increased microbe-induced IL-1β promotes oral carcinogenesis. Because most samples from this animal study had poor RNA quality, it was difficult to analyze mRNA expressions of LDOC1 and IL-1β in these mouse tissues. To investigate the role of LDOC1 in IL-1β secretion induced by microbes, we generated cell lines either with LDOC1 knockdown or with ectopic LDOC1 expression from CGHNK2 ( Figure S2a) or TW2.6 cell lines (Figure 4a), respectively. Their vector control cell lines (CGHNK2-shCtrl and TW2.6-GFP), were also established. We examined the effect of LDOC1 on the production of IL-1β induced by CA SC5314. The medium levels of IL-1β were measured using a Cytometric Bead Array (CBA) after cell coculture with live CA SC5314 (MOI 0.5) for 8, 16, and 24 h. The results indicated that CA SC5314 stimulated IL-1β secretion during the period of coculture in all cells tested (Figure 4b,c). The IL-1β secretion in LDOC1-deficient cell lines, CGHNK2-shLDOC1, and TW2.6-GFP significantly increased over time. Notably, the levels of CA SC5314-induced IL-1β produced by CGHNK2-shLDOC1 cells were much higher than in CGHNK2-shCtrl cells (Figure 4b). The CA SC5314-induced IL-1β produced by TW2.6-LDOC1-GFP cells was significantly reduced as compared with TW2.6-GFP cell lines ( Figure 4c). These results indicated that LDOC1 has an inhibitory effect on IL-1β induced by CA in CGHNK2 and TW2.6 cell lines. In the groups of mice with long-term NA exposure (C10 groups), oral infection of CA increased the incidence of OSCC, while in the NA-untreated mice groups (C0 groups), neither CA infection nor OSCC was observed in the oral lesions repeatedly treated with CA (Figure 3b,c and Table 1). These results suggest that exposure to NA may be a prerequisite for CA to promote the development of OSCC. Our previous work demonstrated that LDOC1 can be epigenetically silenced by exposure to cigarette smoke condensates [30]. In addition, an inverse relationship between LDOC1 expression and IL-1β induced by CA SC5314 or FN was observed in the cell-based assays (Figure 2a,b). We hypothesized that long-term exposure to NA may cause LDOC1 silence, thereby increased microbeinduced IL-1β promotes oral carcinogenesis. Because most samples from this animal study had poor RNA quality, it was difficult to analyze mRNA expressions of LDOC1 and IL-1β in these mouse tissues. To investigate the role of LDOC1 in IL-1β secretion induced by microbes, we generated cell lines either with LDOC1 knockdown or with ectopic LDOC1 expression from CGHNK2 ( Figure S2a) or TW2.6 cell lines (Figure 4a), respectively. Their vector control cell lines (CGHNK2-shCtrl and TW2.6-GFP), were also established. We examined the effect of LDOC1 on the production of IL-1β induced by CA SC5314. The medium levels of IL-1β were measured using a Cytometric Bead Array (CBA) after cell coculture with live CA SC5314 (MOI 0.5) for 8, 16, and 24 h. The results indicated that CA SC5314 stimulated IL-1β secretion during the period of coculture in all cells tested (Figure 4b,c). The IL-1β secretion in LDOC1-deficient cell lines, CGHNK2-shLDOC1, and TW2.6-GFP significantly increased over time. Notably, the levels of CA SC5314-induced IL-1β produced by CGHNK2-shLDOC1 cells were much higher than in CGHNK2-shCtrl cells (Figure 4b). The CA SC5314-induced IL-1β produced by TW2.6-LDOC1-GFP cells was significantly reduced as compared with TW2.6-GFP cell lines (Figure 4c). These results indicated that LDOC1 has an inhibitory effect on IL-1β induced by CA in CGHNK2 and TW2.6 cell lines.  2.5. LDOC1 Deficiency Induces Inhibitory Phosphorylation of GSK-3β S9 through Akt/PI3K Activation, and Reduced GSK-3β Activity Controls the Clonogenicity of CGHNK2-shLDOC1 Cells The human phospho-kinase array was used to clarify the underlying mechanisms by identifying key molecules involved in the LDOC1-mediated suppression of IL-1β production in human oral cells. We determined that glycogen synthase kinase-3β (GSK-3β) and protein kinase B (Akt) accounted for two of the top four kinases with enhanced phosphorylation in CGHNK2-shLDOC1 cells compared with CGHNK2-shCtrl cells (Figure 5a). GSK-3β reportedly plays a role in OSCC [35]. This enzyme has also been demonstrated to regulate the inflammatory response caused by microbial pathogens [36,37]. Activated Akt (phospho-Akt S473 , pAkt S473 ) is the major kinase responsible for inhibitory phosphorylation of GSK-3β S9 . Accordingly, we proposed that the Akt/GSK-3β axis may be involved in the LDOC1-mediated suppression of IL-1β production and anti-cancer effects in OSCC. The substantial increase in pAkt S473 accompanied by increased phosphorylated GSK-3β (pGSK-3β S9 , an inactive form of GSK-3β) in CGHNK2-shLDOC1 was validated by Western blotting analysis (Figure 5b). Akt can be activated by phosphatidylinositol 3,4,5-trisphosphate (PIP 3 ) produced by phosphoinositide 3-kinase (PI3K). Specific inhibitors of Akt (SH-6) and PI3K (LY294002) were applied to determine whether activated Akt and PI3K were responsible for the increased pGSK-3β S9 in LDOC1-deficient CGHNK2-shLDOC1 and TW2.6-GFP cell lines. As illustrated in Figure 5c, treatment with SH-6 (5 µM) or LY294002 (10 µM) abolished the phosphorylation of Akt and PI3K, reduced pGSK-3β S9 , and increased GSK-3β in CGHNK2-shLDOC1 and TW2.6-GFP cell lines. These results indicated that activated Akt and PI3K contributed to the increased pGSK-3β S9 and decreased active GSK-3β in LDOC1-deficient CGHNK2-shLDOC1 and TW2.6-GFP cell lines. To examine whether GSK-3β is a critical downstream effector of LDOC1 in OSCC, we examined the effect of ectopic expression of GSK-3β on the acquired clonogenicity of CGHNK2 cells by LDOC1 knockdown. As illustrated in Figure 5d, LDOC1 knockdown resulted in acquired anchorage-independent growth (clonogenicity) of CGHNK2-shLDOC1 cells, and ectopic GSK-3β expression (CGHNK2-shLDOC1-GSK-3β) significantly suppressed the clonogenicity acquired after LDOC1 knockdown, suggesting that the function of GSK-3β is crucial for the anti-cancer activities of LDOC1. phosphokinase array analysis using CGHNK2-shLDOC1 and CGHNK2-shCtrl cells. The bar charts represent the relative protein phosphorylation levels for the top four kinases and the transcription factor STAT3 exhibiting different phosphorylation states in CGHNK2-shLDOC1 and CGHNK2-shCtrl cells (the array contained some non-kinase phosphoprotein spots). Experiments were performed according to the instructions provided by the manufacturer, as described in Materials and Methods. (b) Western blotting analysis of pGSK-3β S9 , GSK-3β, pAkt S473 , and Akt in CGHNK2-shLDOC1 and CGHNK2-shCtrl cells. GAPDH was used as a loading control for each protein sample. (c) The serine kinase activity of Akt is required for the phosphorylation of GSK-3β S9 in CGHNK2-shLDOC1 and TW2.6-GFP cells. Cellular protein lysate was harvested after a 6 h treatment with inhibitors of Akt (SH-6, 5 µM) or PI3K (LY294002, 10 µM). Western blotting analyses of pGSK-3β S9 , GSK-3β, pAkt S473 , Akt, and GAPDH were then performed. Cells treated with PBS (0.1%) were used as controls. The pixel intensity ratio for each band was shown. There are technical duplicates for the human phosphokinase array analysis; Western blotting analysis under each condition was performed at least twice. (d) Ectopic expression of GSK-3β inhibits the acquired clonogenicity of CGHNK2-shLDOC1 cells. Cells were subjected to a colony-forming assay. The bar chart represents the mean values of triplicate tests (mean ± SD). Representative images are presented.

The PI3K/Akt/GSK-3β Signaling Pathway Is Involved in the CA-Stimulated IL-1β Production by LDOC1 Deficienct Oral Cells.
To investigate the involvement of the PI3K/Akt/pGSK-3β signaling pathway in the CA-induced IL-1β production by LDOC1-deficient oral cells, we examined whether SC5314 can activate Akt. As illustrated in Figure 6a,b, coculture with CA SC5314 increased pAkt S473 in all cells tested, with a peak more rapidly appearing in TW2.6-derived cell lines (20 min) than in CGHNK2-derived cell lines (120 min). In CGHNK2-and TW2.6-derived cell lines, LDOC1 expression significantly attenuated the induction effect of SC5314 on pAkt S473 . Furthermore, treatment with inhibitors of Akt (SH-6 (2.5 μM and 5 μM for CGHNK2-and TW2.6-derived cell lines, respectively)) and PI3K (LY294002 (10 μM)) markedly reduced the CA SC5314-induced production of IL-1β in LDOC1-deficient CGHNK2-shLDOC1 ( Figure 6c) and TW2.6-GFP cells (Figure 6d). These results indicated that the enhanced PI3K/Akt signaling caused by LDOC1 downregulation resulted in the augmented IL-1β production stimulated by microorganisms such as CA.  The PI3K/Akt signaling pathway involved in CA-induced IL-1β production in LDOC1-deficient oral cells. (a,b) Effect of CA SC5314 on the phosphorylation of Akt S473 in LDOC1-deficient or LDOC1-expressing oral cell lines. Cells cultured in a normal growth medium were treated with CA SC5314 (MOI = 0.5) for the times indicated. Equal amounts of whole-cell lysates were subjected to Western blotting analysis with antibodies specific for pAKT S473 , Akt, and GAPDH. A representative blot is presented in the upper panels. As presented in the bottom panels, densitometric analyses were performed to quantify the fold change (FC) in the intensity of the pAKT S473 blots with untreated controls (time 0) set as 1. Data are expressed as mean ± SD (n = 2, biological duplicates). * p < 0.05, ** p < 0.01 vs. basal activation. (c,d) Levels of IL-1β induced by CA SC5314 were suppressed by inhibitors of Akt and PI3K (SH-6 and LY294002, respectively) in LDOC1-deficient CGHNK2-shLDOC1 (c) and TW2.6-GFP (d) cell lines. Concentrations of IL-1β in the conditioned medium of cells were measured by CBA after coculture with live CA SC5314 for 16 h, with or without pretreatment of SH-6 (2.5 µM and 5 µM for CGHNK2-and TW2.6-derived cell lines, respectively) or LY294002 (10 µM) for 6 h. Data are presented as mean ± SD (n = 3), analyzed using the ANOVA test. * p < 0.05 and ** p < 0.01.
To determine whether inactivated GSK-3β contributes to the dysregulated IL-1β response stimulated by CA in LDOC1-deficient cells, we increased the abundance of active GSK-3β by inducing the expression of a constitutively active mutant GSK-3β S9A in CGHNK2-shLDOC1 cells. By treatment of CGHNK2-shLDOC1-V cells with CA SC5314, pGSK-3β S9 expression was observed at only 90 min after coculture with CA SC5314 (Figure 7a, left panel). GSK-3β phosphorylation at S9 leads to its inactivation through proteasomal degradation [38,39], which may explain the absence of pGSK-3β S9 in the CGHNK2-shLDOC1-V cells at most time points in the experiment. In the CGHNK2-shLDOC1-GSK-3β S9A cells stimulated with CA SC5314, the protein expression of pGSK-3β S9 decreased at 60 min and then increased significantly at 90 min, peaking at 150 min (Figure 7a, center panel). Ninety minutes after CA SC5314 stimulation of CGHNK2-shLDOC1-V cells, the expression of active GSK-3β decreased gradually, and only one-tenth of the original amount remained at 150 min (Figure 7a, right panel). In the CGHNK2-shLDOC1-GSK-3β S9A cells stimulated with CA SC5314, the total protein expression of GSK-3β and GSK-3β S9A , which accounted for the enzyme activity, remained stable without apparent reduction. These results indicated that expressing GSK-3β S9A can compensate for the reduced active form of GSK-3β in response to CA SC5314 treatment (Figure 7a). Figure 7b illustrated that CA SC5314 (MOI 0.5) can also induce CGHNK2-shLDOC1-GSK3β S9A cells to secrete IL-1β after coculture for 24 h. As compared to those secreted by the corresponding control cell line CGHNK2-shLDOC1-V, the levels of CA SC5314-induced IL-1β produced by CGHNK2-shLDOC1-GSK3β S9A significantly reduced. Altogether, these results demonstrated that the PI3K/Akt/GSK-3β axis is involved in the augmented CA-induced IL-1β production caused by LDOC1 deficiency in oral cells. The immune actions affected by GSK-3β are largely attributable to the numerous crucial immune-related transcription factors (TFs) that it regulates, such as NF-kB, CREB, AP-1, STAT1-3, and β-catenin [37,40]. Results from TF DNA-binding profiling assay performed with CGHNK2-shLDOC1 and CGHNK2-shCtrl cells (Figure 7c) demonstrated that up to 96% (23/24) of the total 24 LDOC1-mediated TFs identified (5 and 19 had decreased and increased DNA-binding ability by LDOC knockdown, respectively) were relevant to immune development and inflammatory responses. These results support that GSK-3β is a crucial acting point for the immunomodulating function of LDOC1 in human oral cells.  Figure 4; three independent experiments were performed for IL-1β, each with three duplicates. Data are presented as mean ± SD (n = 3), analyzed using the ANOVA test. ** p < 0.01; (c) LDOC1 modulates immune-related TFs. Nuclear proteins isolated from CGHNK2-shCtrl and CGHNK2-shLDOC1 cells were subjected to a TF DNA-binding profiling assay. The fold change (FC) of intensities derived from TFII in CGHNK2-shLDOC1 versus CGHNK2-shCtrl was used for normalization. Experiments were performed according to the instructions provided by the manufacturer, as described in Materials and Methods. The bar charts represent the average FC values of two independent experiments.

Discussion
IL-1β has been recognized as a salivary biomarker for OSCC and can promote oral malignant transformation [22,24]. In general, the pathogenesis of OSCC differs between regions and ethnicities. Some studies indicated that CA and FN are promoting factors in the development of OSCC [7,8,17]. However, published studies on the association between CA and OSCC require further investigation of cases from Taiwan. Two specific reasons motivated our experiment, as illustrated in Figure 1. First, LDOC1 is a tumor suppressor gene that was identified in OSCC specimens in Taiwan, and second, we primarily used OSCC cell lines generated from Taiwan OSCC tissues to conduct the subsequent cell-based functional analysis of LDOC1 in OSCC. Specifically, we decided to initially test the proportion of CA-positive oral specimens in OSCC cases from Taiwan (Figure 1a,b), and, for consistency, we also verified that CA SC5314 invaded TW2.6 and CGHNK2 cell lines, which were two cell lines established from the tissues of Taiwanese patients with OSCC (Figure 1c). Therefore, the results in Figure 1 are novel in that they elucidate the roles of LDOC1 and CA in OSCC in Taiwanese patients. In this study, we demonstrated that CA and FN are potent inducers of IL-1β in GSK-3β participates in the enhanced IL-1β production in LDOC1-deficient CGHNK2-shLDOC1 cells after coculture with CA SC5314. (a) Western blotting analyses of the quantities of active GSK-3β, GSK-3β S9A , and inactive pGSK-3β S9 proteins in CGHNK2-shLDOC1-V (left panel) and CGHNK2-shLDOC1-GSK-3β S9A (center panel) cell lines. Densitometric analyses were performed to quantify the FC in the intensity of the active non-phosphorylated GSK-3β blots with untreated controls (time 0) set as 1 (right panel). Data are expressed as mean ± SD (n = 2, biological duplicates). ** p < 0.01 versus basal activation. (b) Concentrations of IL-1β in the conditioned medium were measured as described in the caption to Figure 4; three independent experiments were performed for IL-1β, each with three duplicates. Data are presented as mean ± SD (n = 3), analyzed using the ANOVA test. ** p < 0.01; (c) LDOC1 modulates immune-related TFs. Nuclear proteins isolated from CGHNK2-shCtrl and CGHNK2-shLDOC1 cells were subjected to a TF DNA-binding profiling assay. The fold change (FC) of intensities derived from TFII in CGHNK2-shLDOC1 versus CGHNK2-shCtrl was used for normalization. Experiments were performed according to the instructions provided by the manufacturer, as described in Materials and Methods. The bar charts represent the average FC values of two independent experiments.

Discussion
IL-1β has been recognized as a salivary biomarker for OSCC and can promote oral malignant transformation [22,24]. In general, the pathogenesis of OSCC differs between regions and ethnicities. Some studies indicated that CA and FN are promoting factors in the development of OSCC [7,8,17]. However, published studies on the association between CA and OSCC require further investigation of cases from Taiwan. Two specific reasons motivated our experiment, as illustrated in Figure 1. First, LDOC1 is a tumor suppressor gene that was identified in OSCC specimens in Taiwan, and second, we primarily used OSCC cell lines generated from Taiwan OSCC tissues to conduct the subsequent cell-based functional analysis of LDOC1 in OSCC. Specifically, we decided to initially test the proportion of CA-positive oral specimens in OSCC cases from Taiwan (Figure 1a,b), and, for consistency, we also verified that CA SC5314 invaded TW2.6 and CGHNK2 cell lines, which were two cell lines established from the tissues of Taiwanese patients with OSCC (Figure 1c). Therefore, the results in Figure 1 are novel in that they elucidate the roles of LDOC1 and CA in OSCC in Taiwanese patients. In this study, we demonstrated that CA and FN are potent inducers of IL-1β in human normal and cancerous oral cells, especially for cells deficient in LDOC1 expression (Figure 2a,b). We analyzed IL-1β in the conditioned medium of tested cells. We did so in a direct contact coculture system where the human oral cells were cultured in a growth medium containing either L or HK CA preparations. Therefore, even if HK CA cannot invade oral cells, HK CA still has direct contact with the cells. Previous studies have shown that fungal β-glucan (glucose polymers found in the cell walls of fungi) is key to immune recognition in macrophages; it stimulates the production of inflammatory cytokines, mainly through Dectin-1, which is a β-glucan-specific membrane receptor. Recent reports have shown that surface mannoproteins can mask the underlying β-glucan from the receptors. CA cells expose their masked β-glucan after being heat killed. In doing so, heat-killing treatment facilitates the interaction between Dectin-1 and the exposed β-glucan, and, in turn, internalization of β-glucans mediates the production of proinflammatory cytokines such as IL-1β and IL-8. Dectin-1 expresses on the cell surface of phagocytes as well as in normal and cancerous oral cells (data not shown). Our results were consistent in indicating that the HK CA-induced IL-1β production was nearly abolished by pretreatment with Dectin-1-neutralizing antibodies (data not shown). LDOC1 as a negative regulator of CA-induced IL-1β production (Figure 4b,c) and this inhibitory effect involved PI3K/Akt/GSK-3β signaling (Figures 6 and 7). This stimulation effect can be nearly abolished or markedly reduced by LDOC1 expression. Inhibitors of PI3K and Akt (Figure 6c,d), and increased GSK-3β activity (Figure 7b), attenuated the augmented production of CA-stimulated IL-1β in LDOC1-deficient oral cells, whereas this suppressing effect was lower in LDOC1-expressing cells, suggesting that the PI3K/Akt/GSK-3β signaling pathway is involved in LDOC1-mediated microbe-induced IL-1β production. We proposed a model based on our data for the role of LDOC1 in OSCC (Figure 8). LDOC1 being silenced or downregulated by promoter hypermethylation in smokers causes the activation of PI3K and Akt, which reduces the activities of GSK-3β and thus results in IL-1β response aberrations. In this setting, human oral cells secrete high levels of IL-1β to promote malignant transformation of oral lesions in the presence of microbes, such as CA and FN.
receptors. CA cells expose their masked β-glucan after being heat killed. In doing so, heat-killing treatment facilitates the interaction between Dectin-1 and the exposed β-glucan, and, in turn, internalization of β-glucans mediates the production of proinflammatory cytokines such as IL-1β and IL-8. Dectin-1 expresses on the cell surface of phagocytes as well as in normal and cancerous oral cells (data not shown). Our results were consistent in indicating that the HK CA-induced IL-1β production was nearly abolished by pretreatment with Dectin-1-neutralizing antibodies (data not shown). LDOC1 as a negative regulator of CA-induced IL-1β production (Figure 4b,c) and this inhibitory effect involved PI3K/Akt/GSK-3β signaling (Figures 6 and 7). This stimulation effect can be nearly abolished or markedly reduced by LDOC1 expression. Inhibitors of PI3K and Akt (Figure 6c,d), and increased GSK-3β activity (Figure 7b), attenuated the augmented production of CA-stimulated IL-1β in LDOC1-deficient oral cells, whereas this suppressing effect was lower in LDOC1-expressing cells, suggesting that the PI3K/Akt/GSK-3β signaling pathway is involved in LDOC1-mediated microbeinduced IL-1β production. We proposed a model based on our data for the role of LDOC1 in OSCC (Figure 8). LDOC1 being silenced or downregulated by promoter hypermethylation in smokers causes the activation of PI3K and Akt, which reduces the activities of GSK-3β and thus results in IL-1β response aberrations. In this setting, human oral cells secrete high levels of IL-1β to promote malignant transformation of oral lesions in the presence of microbes, such as CA and FN.
Several lines of evidence have indicated that numerous tumor suppressor genes, such as p53, PTEN, RB1, and ARF, play important roles in the immune system [41]. Therefore, tumor suppressor genes may act not only as guardians of genomic integrity but also as critical factors in the homeostatic regulation of immune response. Findings from this study revealed that LDOC1 is crucial for the IL-1β response to microbial exposure. Cigarette smoke modulates both the innate and adaptive immune systems. In addition to activating macrophage and dendritic cell activity, tobacco smoke exposure increases the production of numerous proinflammatory cytokines, such as IL-1, IL-6, IL-8, TNF-α, and GM-CSF, and reduces the levels of anti-inflammatory cytokines, such as IL-10 [42]. LDOC1 being epigenetically silenced by tobacco exposure [30] may explain the effect of smoking on immunity and inflammation [42]. Altered immunity can cause commensal microorganisms to become pathogenic and induce high levels of oncogenic IL-1β for oral cancer. This assumption is indirectly supported by observations from epidemiologic studies that have indicated that poor oral hygiene, which may cause proliferation of oral microorganisms, is an independent risk factor for OSCC, particularly among cigarette smokers and alcohol users [17,26].  Several lines of evidence have indicated that numerous tumor suppressor genes, such as p53, PTEN, RB1, and ARF, play important roles in the immune system [41]. Therefore, tumor suppressor genes may act not only as guardians of genomic integrity but also as critical factors in the homeostatic regulation of immune response. Findings from this study revealed that LDOC1 is crucial for the IL-1β response to microbial exposure. Cigarette smoke modulates both the innate and adaptive immune systems. In addition to activating macrophage and dendritic cell activity, tobacco smoke exposure increases the production of numerous proinflammatory cytokines, such as IL-1, IL-6, IL-8, TNF-α, and GM-CSF, and reduces the levels of anti-inflammatory cytokines, such as IL-10 [42]. LDOC1 being epigenetically silenced by tobacco exposure [30] may explain the effect of smoking on immunity and inflammation [42]. Altered immunity can cause commensal microorganisms to become pathogenic and induce high levels of oncogenic IL-1β for oral cancer. This assumption is indirectly supported by observations from epidemiologic studies that have indicated that poor oral hygiene, which may cause proliferation of oral microorganisms, is an independent risk factor for OSCC, particularly among cigarette smokers and alcohol users [17,26].
The mechanisms that control GSK-3β activity are not yet fully understood. However, GSK-3β activity is thought to be controlled, in large part, by phosphorylation. GSK-3β is constitutively active under basal cellular conditions. Inhibitory phosphorylation of GSK-3β S9 can be triggered by cellular stimuli, such as those received through growth factor receptors, Toll-like receptors, T cell receptors, and interleukin receptors. Our results revealed that the activated Akt and PI3K are two key molecules controlling the activities of GSK-3β (Figure 5c), and that PI3K/Akt/GSK-3β signaling significantly affects IL-1β production in LDOC1-deficient cells (Figure 6c,d and Figure 7b), indicating that LDOC1 at least partly regulates immune actions by modulating the activities of GSK-3β via Akt and PI3K. The activation of Akt and PI3K are tightly regulated by ubiquitination. Ubiquitination is a post-translational modification by the covalent attachment of the small polypeptide ubiquitin that alters the properties of its target proteins in a variety of manners and serves as a valuable signal in the host immune response to pathogenic infection. In eukaryotes, the conjugation of ubiquitin to target proteins involves the consecutive actions of three enzymes, the ubiquitin-activating enzyme E1, the ubiquitin-conjugating enzyme E2, and the ubiquitin ligase E3 [43]. Ubiquitination is a reversible modification. Ubiquitin can be removed from its targets by deubiquitinating enzymes (DUBs). Lin's research group discovered that Akt also undergoes K63-linked ubiquitination by the E3 ligase TNF receptor-associated factor (TRAF) 6 upon cellular stimulation with growth factors or cytokines inducing Akt activation [44]. This K63-linked ubiquitination of Akt does not alter the protein stability of Akt, but it is crucial for membrane recruitment and subsequent activation of Akt. Furthermore, pAkt associates with the E3 ligases TTC3 and MULAN to facilitate ubiquitination and degradation within the nucleus [45,46]. These findings reflect the profound influence of ubiquitination on Akt activity. PI3K also undergoes ubiquitination [47,48]. We identified two E3 ligases as potential LDOC1-interacting partners by searching the STRING database (http://string-db.org). Therefore, LDOC1 may negatively regulate the activation of Akt and PIK3 through ubiquitination. Our data support the assumption that LDOC1-deficient and LDOC1-expressing cells exhibited different kinetic patterns of Akt phosphorylation in oral premalignant and OSCC cell lines in response to SC5314 stimulation (Figure 6a,b).
Activation of NF-kB and the assembly of inflammasomes are two essential signals for the production and secretion of IL-1β [49]. LDOC1 may also control the production of IL-1β through the ubiquitination of NF-kB and inflammasome. Inflammasome is composed of three major components, pattern recognition receptors (PRRs), an adaptor protein (PYCARD or ASC), and procaspase-1 or procaspase-5 [49]. When PRRs sense pathogen-associated or danger-associated molecular pattern signals, the PRRs interact with PYCARD, which recruits and activates procaspase-1 or procaspase-5. NLRP3 inflammasome plays an essential role in host defense against bacterial infection, and its regulation has been extensively studied. Increasing evidence indicates a central role for ubiquitination in NLRP3 inflammasome regulation [50]. The Alnemri group demonstrated that the deubiquitylation of NLRP3 by DUB BRCC3 was required for NLRP3 inflammasome activation [51]. Furthermore, David Brough et al. identified that DUB USP7 and USP47 regulate NLRP3 inflammasome activation by promoting ASC oligomerization and speck formation [52]. Regulation of the NF-κB activation pathway by the ubiquitin proteolytic system is well documented [53]. PRRs can activate the canonical NF-κB pathway. The primary mechanism for canonical NF-kB activation is the inducible degradation of IκBα, triggered by site-specific phosphorylation by the IκB kinase (IKK) complex. The IKK comprises two catalytic subunits, IKKα and IKKβ, and a regulatory subunit, NF-κB essential modulator (NEMO). IKK is activated by stimulation with various agents, such as inflammatory cytokines, mitogens, growth factors, and microbial components. Upon activation, IKK phosphorylates IκBα, which is then conjugated with K48-linked polyubiquitin chains by the E3 ligase SCF β-TRCP and degraded by the proteasome, resulting in the release of NF-κB from IκBs and translocation to the nucleus to induce the transcription of target genes. These findings emphasize the central role of ubiquitination in NF-kB and inflammation regulation. Inhibition of PI3K/Akt/pGSK-3β signaling reduced the production of microbe-induced IL-1β but failed to fully compensate for the augmented secretion of IL-1β caused by LDOC1 loss (Figure 6c,d and Figure 7b). These results suggested that, in addition to the regulation of PI3K/Akt/pGSK3β signaling, LDOC1 may suppress the biosynthesis of IL-1β by modulating the ubiquitination efficiency for components of NF-kB and inflammasome. Future studies should investigate the effects of LDOC1 on the activities of these critical molecular modules through ubiquitination.

Cell Lines and Transduction
The immortalized normal oral keratinocyte cell line CGHNK2 was a gift from Dr. Cheng, Ann-Joy (Chang Gung University, Taoyuan, Taiwan) and was grown in keratinocyte serum-free medium (KSFM) with supplements (EGF, Human Recombinant/Bovine Pituitary Extract from Life Technologies, Inc., Gibco BRL, Rockville, MD, USA, respectively) as previously described [26,47]. TW2.6 has been established from the surgically resected specimen of an untreated primary squamous cell carcinoma of the buccal mucosa from a patient who was an betel quid chewer and tobacco smoker, and was cultured as previously described [27,29]. In vitro transduction was carried out at one multiplicity of infection (MOI) per cell. To generate stable cell lines, the lentivirus-infected cell pools were selected with either 5 µg/mL puromycin (Sigma-Aldrich, St. Louis, MO, USA) or 20 µg/mL blasticidin S (Invitrogen, Waltham, MA, USA).

Quantitative Real-Time PCR (qPCR)
Total RNA was isolated using RNeasy Mini Kits (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The synthesis of cDNA from total RNA was performed with M-MLV reverse transcriptase (Promega, Madison, WI, USA). The resulting cDNAs were subjected to qPCR using the following primers: 5 -ATGACGACGAAGACGACGA-3 (F) and 5 -GAGGGTCGAGGGCCTAATAA-3 (R) for LDOC1. QPCR was performed with FastStart Universal Probe Master Kit (Roche Applied Science, Madison, WI, USA). The gene expression level was normalized using GAPDH mRNA.

Detection of Fungal Organisms Using Grocott's Methenamine Silver (GMS) Staining
For fungal microorganism examination, deparaffinized and hydrated oral tissue sections were stained with Grocott methenamine silver (GMS) stain kit (ScyTek Laboratories, Inc., West Logan, UT, USA) following manufacturer's instructions. Briefly, GMS is a chromic acid which oxidizes fungal cell wall mucopolysaccharide components to form aldehyde groups. The aldehyde groups present in the fungal cell wall then react with the silver nitrate, reducing it to metallic silver, which appears black. The stained samples were observed under a light microscope (Olympus-Japan BX 41, magnification × 400, Tokyo, Japan). The fungi were turned black, while everything else is stained light green.

Detection of CA DNA Using Real-Time PCR Assay
DNA extracted from tissue sections using QIAamp DNA FFPE tissue kit (Qiagen 56404) following the manufacturer's instructions. The PCR amplicon specific for internal transcribed space of CA rDNA is 108 base pairs in size using the primer sequences and probe as described previously [54]. Real-time PCR was conducted using PowerUp™ SYBR Green Master Mix using an ABI 7500 fast Real-time PCR System (Applied Biosystems, Foster City, CA, USA). Each reaction was run in duplicate and contained 100 ng gDNA template along with 50 nM primers in a final reaction volume of 20 µL. The cycling parameters began with 95 • C for 2 min, followed by 40 cycles of 95 • C for 3 s and 60 • C for 30 s. Samples were also run a 3% agarose gel to confirm specificity.

Mouse OSCC Model with CA-Infected Oral Lesions
The animal experimental protocol was approved by Institutional Animal Care and Use Committee. C57BL/6JNarl mice (8 weeks of age) were purchased from the National Laboratory Animal Center (Taipei, Taiwan). The mouse OSCC was induced with combination of 4-NQO and arecoline (NA) as described previously [24]. Briefly, mice were given water containing NA for 8 (C8) or 10 (C10) weeks. Mice without NA treatment were used as control group (C0). Oral lesions were created by brushing the tongue to mimic the trauma generated by betel quid chewing. Starting from week 8, once a week for 8 weeks mice were either painted with live CA SC5314 at the site of lesions on their tongue (F+) or painted with normal saline as a vehicle control (F−). All mice were euthanized at the end of week 28 (day 196) and the tongue sections were carefully collected immediately after euthanasia. The histological slides of tongue tissues were processed for H&E staining and GSM staining for histopathological and fugus examination, respectively.

Determination of CA Infection in Mouse Oral Lesions
CA infection was examined from oral swabs collected at week 8, 16, and 28. Oral infection of CA was assessed using CHROMagar™ Candida plate (CHROMagar™, Paris, France). Briefly, samples were collected by gently rubbing a sterile cotton swab over the mouse oral cavity and were immediately transferred for microbiological analysis. The collected oral swabs were soaked in PBS and shaken for one minute, and then the PBS was streaked onto CHROMagar plates and incubated at 37 • C for 24 h. CA infection was determined with the oral swabs obtained at week 28. Positive for CA infection was defined as the appearance of creamy and convex pasty colonies with a moldy odor. No colonies observed even after 24 h of incubation was recognized as negative. After sacrifice, the section of mouse tongue specimen was subjected to GMS staining to cross-validate for CA infection. The infection rate is defined as the rate of CA-positive mice to the total number of mice in an experimental group.

Cytometric Bead Array (CBA) Analysis
The production of IL-1β in the culture supernatant was determined by a Human Inflammation CBA Kit (BD Biosciences, CA) following the manufacturer's instructions. Conditioned medium was prepared from 10 5 cells seeded onto a 6 cm dish and cultured in 2 mL of serum-free medium in the presence or absence of preparations of CA SC5314 or FN. The BD FACSCalibur flow cytometer was used to acquire data by setting the red diode laser for exciting the CBA particles and detection of particle emission in the FL4 channel.

Gene Expression Microarray Analysis
Extraction of total RNA, RNA quality evaluation, and cDNA microarray experiments were performed according to Affymetrix standard protocols by the microarray core laboratory at the National Health Research Institutes. The gene chips (Clariom S Assay HT06, Affymetrix Inc, Santa Clara, CA, USA) were scanned with an Affymetrix Gene ChIP Scanner 3000 7G, and the CEL files generated were analyzed using Affymetrix Expression Console Software (version 1.4, Affymetrix, Santa Clara, CA, USA), which normalizes array signals using Signal Space Transformation (SST) and a robust multiarray averaging (RMA) algorithm. Normalized data were analyzed using Transcriptome Analysis Console (TAC) 3.0 software (Affymetrix, Santa Clara, CA, USA). A paired t-test was applied to identify differentially expressed transcript genes between sample pairs and probes, with P-values less than 0.05 and fold-change ≥2 declared significant. All gene level files were imported into QIAGEN's Ingenuity Pathway Analysis (IPA) software for pathway and molecular function analysis. Microarray expression data are available at the U.S. National Center for Biotechnology Information Gene Expression Omnibus (GEO) database under accession number GSE149108.

Human Phospho-Kinase Array Analysis
Total protein concentrations in cell lysates were measured using a NanoDrop ND-1000 spectrophotometer (Wilmington, DE, USA). Four hundred micrograms of cell lysates were hybridized with the Proteome Profiler Human Phospho-Kinase array (ARY003B, R&D systems, Minneapolis, MN, USA) according to the manufacturer's instructions. Chemi Reagent Mix was applied for visualizing the hybridized spots, then spot densities were quantified using ImageJ v1.52.

Transcription Factor (TF) Profiling Array
Experiments were performed as previously described [31]. Briefly, cellular nuclear proteins were harvested using a Nuclear Extraction Kit (Chemicon, Rolling Meadows, IL, USA). Experiments were carried out with The TF Activation Profiling Plate Array system (Signosis, Santa Clara, CA, USA) according to the manufacturer's instructions. The DNA binding activities of TFIID were used for normalization.

Soft Agar Assay
Experiments were performed and quantified as previously described [24].

Statistics
Student tests and ANOVA tests were analyzed using Microsoft Excel (Microsoft, Redmond, WA, USA) and statistical software NCSS (NCSS, Kaysvill, UT, USA), respectively.

Conclusions
This study highlighted that LDOC1 acts as a novel IL-1β modulator to control inflammatory responses to oral microbes, such as CA and FN. Furthermore, we identified molecules, such as PI3K, Akt, and GSK-3β, that acted as key players in the LDOC1-mediated IL-1β response. Chronic inflammation that is characterized by a sustained level of IL-1β may create a protumor microenvironment for OSCC. Accordingly, to prevent OSCC, we recommend smoking cessation to avoid LDOC1 silencing, which can cause immune abnormalities, and good dental hygiene so as to not allow proliferation of microbes, such as CA or FN, in oral cavities.