Porphyromonas gingivalis promotes tumor progression in esophageal squamous cell carcinoma

Abstract

Purpose

Increasing evidence indicates that the microbiome may influence tumor growth and modulate the tumor microenvironment of gastrointestinal cancers. However, the role of oral bacteria in the development of esophageal squamous cell carcinoma (EsoSCC) has remained unclear. Herein, we investigated the relationship between the periodontal pathogen Porphyromonas gingivalis and EsoSCC.

Methods

To identify bacterial biomarkers associated with EsoSCC, we analyzed microbiomes in oral biofilms. The presence of P. gingivalis in esophageal tissues and relationships of P. gingivalis infection with clinicopathologic characteristics in 156 patients with EsoSCC were assessed using immunohistochemistry. The role of P. gingivalis infection in in vitro and in vivo EsoSCC progression was also assessed.

Results

Microbiota profiles in oral biofilms revealed that P. gingivalis abundance was associated with an increased risk of EsoSCC development. In total, 57% of patients with EsoSCC were found to be infected with P. gingivalis. The presence of P. gingivalis was found to be associated with advanced clinical stages and a poor prognosis. It was also found to be associated with an elevated esophageal cancer incidence in a 4-nitroquinoline 1-oxide-induced mouse model and with an increased xenograft tumor growth. P. gingivalis infection increased interleukin (IL)‐6 production and it promoted epithelial-mesenchymal transition and the recruitment of myeloid-derived suppressor cells. Furthermore, inhibited IL‐6 signaling attenuated the tumor-promoting effects of P. gingivalis in 4-nitroquinoline 1-oxide-treated mice and xenograft mouse models.

Conclusions

Our data indicate that P. gingivalis may promote esophageal cancer development and progression. Direct targeting of P. gingivalis or concomitant IL-6 signaling may be a promising strategy to prevent and/or treat EsoSCC associated with P. gingivalis infection.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-020-00573-x.

Introduction

Esophageal cancer is a common gastrointestinal tumor type that constitutes a substantial public health burden [1]. Despite advances that have been made in diagnostics, esophageal cancer is often diagnosed at advanced stages, resulting in an overall poor survival [2]. Therefore, elucidation of the molecular mechanisms underlying esophageal cancer development and progression, as well as the identification of molecular biomarkers, are issues of considerable clinical importance.

Increasing evidence indicates that microorganisms may directly be linked to a significant number of human cancers [3]. Microbes present at mucosal surfaces can affect cancer growth and spread [3, 4], and dysbiosis has been implicated in the progression of aerodigestive tract malignancies [5, 6]. The gastrointestinal microbiome has also been shown to affect anti-tumor immunity [6, 7]. Specifically, the presence of certain species within the gastrointestinal microbiome has been associated with clinical outcomes in patients with esophageal cancer [8, 9]. The oral microbiome has been found to shape the esophageal microbiome, thereby potentially contributing to esophageal carcinogenesis [10]. This notion adds another layer of complexity to the relationship between the microbiome and esophageal cancer development. Recent studies have pointed to a link between inflammatory periodontal disease and a potential contribution of microorganisms to the development of esophageal cancer [7, 11, 12]. Immune modulations and altered gut microbiomes induced by oral dysbiosis have led to the association of esophageal cancer with periodontal disease. Adenocarcinoma and squamous cell carcinoma (SCC) are the predominant histologic subtypes of esophageal cancer. Although both have a poor prognosis, their clinical and molecular features vary considerably [13]. Relationships between the microbiome and esophageal adenocarcinoma development have extensively been investigated [5, 14], whereas the microbiome in patients with esophageal SCC (EsoSCC) is less well characterized. A better understanding of the role of the oral microbiome in EsoSCC may allow for the development of novel screening strategies to diagnose EsoSCC or to predict treatment outcomes in patients with EsoSCC.

Porphyromonas gingivalis is one of the main etiological factors in the pathogenesis of periodontal disease [15, 16]. P. gingivalis can disrupt the oral microbial equilibrium, resulting in a dysbiotic host-microbiota interaction. This, in turn, may play a key role in the development of periodontal disease and cancers of the upper aerodigestive tract [17]. Although a central role of P. gingivalis in orodigestive cancer is still a matter of debate, we propose that it may be an important predisposing factor in the microenvironment that directs the development and/or poor prognosis of EsoSCC. Therefore, we set out to examine associations of the oral microbiome composition with EsoSCC development/progression, and to elucidate the effect of P. gingivalis on the establishment of tumor characteristics.

Materials and methods

Study cohort

The study cohort consisted of 156 patients with EsoSCC who received curative care in accordance with the recommendations of the oncology team of our hospital. Specimens were collected at the time of diagnosis from the patients for immunochemical analysis and 45 EsoSCC samples were collected for qPCR analysis. Patient demographic, diagnostic, clinical stage and treatment data are listed in Table 1. In addition, oral biofilms were obtained from 34 patients with EsoSCC and 18 healthy donors. The oral biofilm samples were collected from each participant before antitumor treatment, and obtained by swabbing the dental plaques at the gingival margin on the molars with sterilized toothpicks. The study was approved by the Institutional Review Board of our hospital.

Table 1.

Clinical characteristics of patients with EsoSCC

	No. of patients
	IHC- P. gingivalis (-)	IHC- P. gingivalis (+)	p value
Patients	67	89
Age
Range	35.8 ~ 78.7	37.3 ~ 78.5	0.069
Median	57.0	52.8
Differentiation
WD ~ MD	41	34	0.004*
PD	26	55
Clinical stage			0.004*
I-II	22	12
III-IV	45	77
LN metastasis			0.035*
negative	19	13
positive	48	76
IL-6 staining			< 0.001*
negative	46	34
positive	21	55
Response to neoadjuvant Tx			0.024*
Response	54	57
Non- response	13	32
Surgery s/p neoadjuvant Tx			0.894
Yes	21	27
No	46	62
Local‐regional recurrence /persistent			0.022*
No	38	34
Yes	29	55
Distant metastasis			0.028*
negative	47	47
positive	20	42

Abbreviations: Neoadjuvant Tx = neoadjuvant chemoradiotherapy;

Immunohistochemical staining

Formalin-fixed, paraffin-embedded tissues from 156 patients with EsoSCC, obtained at the time of diagnosis, were cut into 4-µm sections and subjected to immunohistochemical (IHC) staining. The sections were incubated overnight at 4 °C with primary antibodies directed against interleukin (IL)-6 and P. gingivalis (1:100). Additionally, esophageal specimens from 21 patients with EsoSCC containing malignant tissue and adjacent non-malignant epithelium were used to prepare tissue microarray blocks. For histological evaluation of P. gingivalis staining, the sections were examined by two observers who calibrated the staining by comparisons with printed IHC images and who were blinded to the clinical outcomes of the patients. The IHC scoring for P. gingivalis staining was obtained by multiplying the intensity score (0 = no staining, 1 = weak staining, 2 = moderate staining and 3 = strong staining) by the percentage of positively stained cells (0 = less than 10% positive cells, 1 = 10–30% positive cells, 2 = 30–60% positive cells and 3 ≥ 60% positive cells). As criterion we used a score of ≥ 2 as positive staining for P. gingivalis, which was defined by a receiver operating characteristic (ROC) curve that maximized the concordance between IHC and qPCR percentages using 45 EsoSCC samples and 9 non-malignant esophageal samples.

Microbiome analysis

DNA was extracted from oral biofilm samples using a procedure described in the Supplementary Methods section. Sequencing of multiplexed pooled libraries was performed on a MiSeq system (Illumina, San Diego, CA, USA). The quality-filtered reads from 52 samples were clustered into operational taxonomic units using the ribosomal database project framework 11.5 and were assigned taxonomies using QIIME V1.8.0. Alpha- and beta-diversities were assessed at the operational taxonomic unit level using unweighted and weighted UniFrac distances. Operational taxonomic unit differential abundances between healthy individuals and patients with EsoSCC were determined using linear discriminant analysis effect size and DESeq2.

DNA extraction and quantitative polymerase chain reaction (qPCR)

Genomic DNA was extracted from fresh esophageal samples obtained from 45 patients with pathologically confirmed EsoSCC and from 9 non-malignant esophageal tissues. Subsequently, the amount of P. gingivalis DNA was assessed using qPCR as reported before [8].

Cell and bacterial cultures

The human EsoSCC cell lines CE81T and TE2 were used. P. gingivalis (strain ATCC 33,277) was grown in Gifu Anaerobic Medium (GAM broth; Nissui, Tokyo, Japan). The respective EsoSCC cells were infected with P. gingivalis at a multiplicity of infection of 1:10 for 24 h or 72 h at 37 °C. Infected cells were used for subsequent in vitro experiments. Uninfected cells were used as controls. To determine the in vitro effects of the anti-IL-6 antibody, the different infected/non-infected cells were incubated with 5 µg/ml IL-6 neutralizing antibody or an isotype antibody for 48 h.

Animal models and experimental design

All experimental procedures involving animals were approved by the Experimental Animal Ethics Committee of our hospital. Six-week-old C57BL/6 mice were used to establish a 4-nitroquinoline 1-oxide (4NQO)-induced cancer model [18, 19]. Briefly, 4NQO (100 µg/ml; tumor group) or solvent alone (control group) were added to the drinking water for 16 weeks. Esophageal lesions developed after 12–14 weeks, as described before [19, 20]. The mice in the tumor group were randomly divided into two subgroups: 4NQO and 4NQO + PG. Animals in the 4NQO + PG group were infected with P. gingivalis (200 µl bacteria per 10¹⁰ cells/ml) [21] three times per week for 2 weeks prior to 4NQO administration. Next, the mice were subjected 8 weeks to 4NQO treatment, followed by 10 weeks of bacterial infection (twice per week). In addition, athymic nude mice were used to establish esophageal cancer xenograft models. To this end, human esophageal cancer cells (1 × 10⁶ cells per mouse) were intraoperatively injected into the esophageal wall [20], after which orthotopic tumor growth was assessed 3 weeks after cell implantation.

Small-animal imaging

In vivo optical imaging was performed in tumor-bearing mice using fluorescence molecular tomography to measure tumor growth at different time points. A 2-DeoxyGlucosone 750 fluorescent probe was used for in vivo tumor imaging, based on enhanced glucose uptake by tumor cells compared to surrounding non-tumor tissues. After imaging, the presence of mouse esophageal lesions was evaluated through gross examination of tissue samples.

Flow cytometry

Myeloid-derived suppressor cells (MDSCs) are phenotypically heterogeneous cells with immune-modulatory functions. To determine the numbers of MDSCs, single-cell suspensions were prepared from murine spleens and analyzed by flow cytometry gated for Gr1 and CD11b, as described before [19, 20].

Immunofluorescence and enzyme-linked immunosorbent assays

In vitro and in vivo levels of IL-6 were determined using immunofluorescence staining and enzyme-linked immunosorbent assays (ELISA) in accordance with protocols described in the Supplementary Methods.

Results

Oral microbiome composition in EsoSCC patients

For characterization of the oral microbiome composition in patients with EsoSCC, oral biofilms were obtained from patients with EsoSCC and healthy volunteers. We found that the alpha-diversity (species richness) and beta-diversity (microbial composition) significantly differed between patients with EsoSCC and healthy individuals (Supplementary Fig. 1a-b). Adonis analysis confirmed the significant difference in overall oral microbial composition between patients with EsoSCC and healthy individuals (p = 0.001). To identify differentially enriched species, the linear discriminant analysis effect size method was used. Some Streptococcus species, Veillonella parvula and P. gingivalis, were found to be more abundant in the oral biofilms from patients with EsoSCC than in those from healthy volunteers (Fig. 1a). Moreover, for P. gingivalis, a significant difference was observed between the cancer and normal groups using DESeq2 (Cancer versus Normal: Fold change 17.97; adjusted p = 0.0034). To subsequently confirm that P. gingivalis was relatively more abundant in esophageal cancer tissues, P. gingivalis 16S rDNA levels were measured in fresh esophageal tissue specimens by qPCR. We found that P. gingivalis 16S rDNA was more abundant in EsoSCC specimens than in non-malignant esophageal samples (Fig. 1b).

Fig. 1.

Porphyromonas gingivalis is enriched in esophageal squamous cell carcinoma. a Cladogram showing differentially enriched species in the oral microbiome of patients with esophageal squamous cell carcinoma (EsoSCC) and matched healthy individuals. Species associated with EsoSCC are shown in color. Red, EsoSCC > healthy; Green, EsoSCC < healthy. Other species in networks are indicated by small outlined circles. b P. gingivalis 16S rDNA levels in esophageal mucosa of patients with EsoSCC and matched healthy individuals, as determined by qPCR. c Representative images showing immunohistochemical staining for P. gingivalis in EsoSCC tissues. d Concordance between IHC with an anti-P. gingivalis antibody and qPCR of P. gingivalis 16S rDNA in cancerous tissues from patients with EsoCC. e. Overall survival of patients with EsoSCC, according to the presence of P. gingivalis

P. gingivalis infection correlates with poor prognosis in EsoSCC patients

To assess the relationship between P. gingivalis and EsoSCC prognosis, EsoSCC tissues obtained at diagnosis were stained for P. gingivalis using IHC (Fig. 1c). Positive cytoplasmic P. gingivalis staining was prominent in the epithelial cells. Subsequent IHC analysis of tissue microarray (TMA) specimens from 21 patients with EsoSCC revealed that P. gingivalis was more frequently present in the cancerous tissues than in the adjacent non-malignant esophageal mucosa (57% vs. 24%), suggesting an association between P. gingivalis infection and EsoSCC development. Moreover, we compared the results of IHC and qPCR assays for the presence of P. gingivalis in 45 esophageal cancer samples to determine the agreement between these two different methods. Significant concordance was observed between the IHC and qPCR assays for the detection of P. gingivalis infection in the cancerous tissues of the EsoSCC patients (Fig. 1d). We also assessed the relationship between P. gingivalis infection and the survival of patients with EsoSCC. The clinical characteristics of patients with EsoSCC are listed in Table 1. We found that in total 89 of the 156 specimens (57%) from the EsoSCC patients were P. gingivalis-positive. Importantly, we found that the presence of P. gingivalis was significantly associated with a poor differentiation and advanced disease (p < 0.05). In addition, we found that the presence of P. gingivalis significantly correlated with higher locoregional recurrences, distant metastases and reduced overall survival rates (Fig. 1e and Supplementary Fig. 2). In the multivariate Cox model adjusted for clinical features, P. gingivalis positivity was significantly associated with shorter survival (Table 2). These findings suggest that P. gingivalis infection promotes tumor aggressiveness and contributes to a poor prognosis in patients with EsoSCC.

Table 2.

Adjusted hazard ratio of variables associated with overall survival for patients with EsoSCC

Variable	HR	95% CI	P value
Clinical stage
I-II	Ref
III-IV	2.48	1.09–5.62	0.03*
P. gingivalis staining
negative	Ref
positive	2.14	1.17–3.92	0.014*
Response to
Neoadjuvant Tx
Response	Ref
Non-response	6.50	3.48–12.14	< 0.001*
Treatment
Definite CCRT	Ref
Surgery+/-   Neo/adjuvant Tx	0.59	0.35–1.02	0.063

P. gingivalis infection is associated with esophageal cancer progression in 4NQO-induced mouse tumors and tumor xenograft models

To further elucidate the relationship between P. gingivalis infection and esophageal tumorigenesis, a 4NQO-induced esophageal cancer mouse model was established. Esophageal lesions were observed after 12–14 weeks in 4NQO-treated mice. Most lesions constituted hyperplasia/papilloma, carcinoma in situ and invasive carcinoma (Fig. 2a). The presence of lesions after 12 weeks of 4NQO treatment was confirmed by fluorescence molecular tomography (Fig. 2b). In addition, we found that esophageal carcinomas, including carcinoma in situ and invasive carcinoma lesions, exhibited a significantly enhanced glucose uptake, compared to benign tumors (hyperplasia or papilloma). We also examined the link between tumor progression and the presence of MDSCs in the lesions and found that significantly higher numbers of MDSCs were present in tumors from mice with carcinomas, compared to tumors from mice with benign lesions (Fig. 2c). Figure 3a shows the presence of P. gingivalis infection in esophageal tissues from mice in the 4NQO + PG group. Importantly, we found that P. gingivalis infections were associated with enhanced glucose uptake, increased incidences of developing invasive carcinoma and MDSC recruitment in the 4NQO-treated mice (Fig. 3b-d).

Fig. 2.

Evaluation of esophageal tumor formation using histological examination and small animal imaging and its relationship with MDSC recruitment. a Gross examination of lesions from mice treated with 4NQO or vehicle for 16 weeks. 4NQO treatment promoted the development of hyperplasia, carcinoma in situ and invasive carcinoma. b Representative fluorescence molecular tomography images from 4NQO-treated mice that developed hyperplasia, carcinoma in situ or invasive carcinoma. Glucose uptake is shown (y-axis indicates relative ratio normalized to control signal). Data are presented as means ± standard errors of the mean. *p < 0.05. c CD11b⁺Gr1⁺ myeloid-derived suppressor cell (MDSC) infiltration in mouse esophageal hyperplasias or carcinomas, as determined by flow cytometry. Representative images are also shown. Data are presented as means ± standard errors of the mean. *p < 0.05

Fig. 3.

P. gingivalis infection promotes tumor progression in 4NQO-induced mouse tumors. a Presence of P. gingivalis infection in esophageal tissues from mice in the 4NQO + PG group. b-d Effects of P. gingivalis infection on esophageal tumor formation in 4-NQO-treated mice determined by fluorescence molecular tomography analysis of glucose uptake (b), increased incidence of developing invasive carcinoma (c) and flow cytometry analysis of MDSCs (d). Data are presented as means ± standard errors of the mean. *p < 0.05

P. gingivalis enhances esophageal cancer cell invasion and stemness

To examine the relationship between P. gingivalis infections and cancer cell invasiveness, in vitro invasion assays were performed using P. gingivalis-infected human EsoSCC cells. We found that P. gingivalis-infected CE81T and TE2 cells exhibited a profound increased invasive potential, compared with control cells (Fig. 4a). Epithelial-mesenchymal transition (EMT) plays a critical role in cancer cell invasion and metastasis [22]. Concordantly, we found that P. gingivalis-infected EsoSCC cells exhibited enhanced EMT-associated characteristics, including elevated β-catenin and matrix metalloproteinase-9 expression levels, accompanied by reduced E-cadherin expression levels (Fig. 4b). Cancer stem cells (CSCs) are critical for cancer progression [23], including that of upper aerodigestive tract cancers. Here, we found that P. gingivalis-infected human esophageal cancer cells exhibited elevated expression levels of the CSC markers CD44 and ALDH1 (Fig. 4c-d and Supplementary Fig. 3). In addition, we established an orthotopic xenograft model to examine the effect of P. gingivalis infection on human esophageal tumor growth. We found that P. gingivalis infection significantly increased glucose uptake and augmented EsoSCC tumor growth and EMT (Fig. 4e and Supplementary Fig. 4).

Fig. 4.

P. gingivalis enhances esophageal cancer cell invasion and stemness. a Effects of P. gingivalis infection on esophageal cancer cell invasion. Representative images and quantitative plots are shown; y-axis indicates relative ratio normalized to number of invading control cancer cells. b Effects of P. gingivalis infection on EMT-associated protein levels. Changes in E-cadherin, matrix metalloproteinase-9 and β-catenin protein levels were determined by immunofluorescence (4′,6-diamidino-2-phenylindole [DAPI], blue; target proteins, green). c Effects of P. gingivalis infection on expression of CSC-related proteins CD44 and PD-L1 in TE2 cells, as determined by flow cytometry. d Effects of P. gingivalis infection on expression of CD44, ALDH1 and PD-L1 evaluated by immunofluorescence (DAPI, blue; target protein, green). e Effects of P. gingivalis infection on tumor growth evaluated in a xenograft esophageal cancer model. *p < 0.005

P. gingivalis infection is associated with enhanced IL-6 expression in esophageal cancer cells

Previously, we reported that IL-6 overexpression is associated with a poor prognosis of EsoSCC patients and that the IL-6/signal transducer and activator of transcription 3 (STAT3) axis plays a critical role in CSC formation and EMT [24, 25]. Furthermore, IL-6/STAT3 pathway activation has previously been implicated in microbiome-induced tumor progression [3]. Hence, we set out to investigate the relevance of IL-6 signaling in the pro-EsoSCC effects of P. gingivalis. We found that P. gingivalis infection induced IL-6 expression in cancer cells and led to significantly elevated IL-6 levels in the cell culture supernatants (Fig. 5a-b). In addition, we found that the presence of P. gingivalis in primary cancer specimens was significantly associated with increased IL-6 levels (Fig. 5c and Table 1). Bacterial infection-induced autophagy is presumed to promote cancer development and resistance to anti-cancer therapies [3, 26]. Additionally, P. gingivalis has been shown to induce autophagy and pro-inflammatory cytokine production [27]. Here, we found that P. gingivalis-infected cells exhibited elevated autophagy levels, indicated by conversion of LC3-I to LC3-II, and elevated IL-6 production levels (Fig. 5d-e and Supplementary Fig. 5). Furthermore, pre-treatment with the autophagy inhibitor 3-methyladenine (10 mM for 1 h) attenuated the expression of IL-6 in P. gingivalis-infected cells. These results suggest that autophagy activation may contribute to the elevated IL-6 production observed in P. gingivalis-infected EsoSCC cells.

Fig. 5.

P. gingivalis infection is associated with enhanced IL-6 expression in esophageal cancer cells. a IL-6 and p-STAT3 levels were evaluated by immunofluorescence. Representative images are shown (DAPI, blue; IL6, green; p‐STAT3, red). b IL-6 levels in the supernatants of P. gingivalis-infected cells, as determined by enzyme-linked immunosorbent assay (ELISA) analysis. Data are presented as means ± standard deviations from three independent experiments; *p < 0.05. c IL-6 levels in primary human esophageal cancer specimens are positively associated with the presence of P. gingivalis. Representative images of positive and negative IL-6 and P. gingivalis staining are shown. d Relationships between P. gingivalis infection, autophagy and IL-6 signaling. Levels of LC3 I-II, IL-6 and CD44 in TE2 cells were determined by immunofluorescence; representative images are shown (DAPI, blue; LC3/IL-6, green; β-catenin/CD44, red). e LC3 I-II and IL-6 protein levels after P. gingivalis infection determined by Western blotting

Role of IL-6 in tumor invasiveness of P. gingivalis-infected esophageal cancer cells in vitro

Based on the observed positive link between IL-6 expression and P. gingivalis infection in EsoSCC, we next set out to examine whether IL-6 inhibition can attenuate the tumor-promoting effects of P. gingivalis. We found that in vitro treatment with an anti-IL-6 antibody for 48 h caused an impairment of EsoSCC cell invasion. It also suppressed the expression of β-catenin and ALDH1 in P. gingivalis-infected EsoSCC cells (Fig. 6a-c). These data suggest that attenuating IL-6 signaling reverses EMT, CSC development and tumor invasiveness in P. gingivalis-infected esophageal cancer cells.

Fig. 6.

Blockade of IL-6 signaling attenuates tumor invasiveness and stemness in P. gingivalis-infected esophageal cancer cells. a Effects of IL-6 inhibition on the invasion in P. gingivalis-infected esophageal cancer cells; y-axis shows relative ratio normalized to number of invading control cancer cells. b-c Effects of calcitriol treatment on the expression of EMT-associated proteins. β-catenin levels were determined by immunofluorescence (b), while ALDH1 expression was evaluated by flow cytometry (c)

Discussion

There is increasing evidence that the gut microbiome plays a crucial role in various diseases, including cancer [3, 4] and, specifically, the gut microbiome has been shown to affect the development and progression of gastrointestinal tract cancers. Although the esophageal microbiome has a composition similar to that of the oral microbiome, key taxonomic differences have been reported [10, 28]. Poor oral hygiene and periodontal disease have been reported to increase esophageal cancer risk [29]. In our country, EsoSCC constitutes more than 90% of all esophageal cancers. In the present study, we examined the relationship between the oral microbiome and EsoSCC. We found significant differences in microbiome composition (β-diversity) and species richness (α-diversity) between EsoSCC and healthy samples. We also identified several bacterial species significantly associated with EsoSCC risk. The abundance of the periodontal pathogen P. gingivalis in oral samples has previously been linked to elevated risks of oral SCC and EsoSCC [28, 30, 31]. We found that the presence of P. gingivalis in dental biofilm samples was more frequent in patients with EsoSCC than in healthy individuals. Furthermore, P. gingivalis 16S rRNA was more abundant in EsoSCC specimens than in non-malignant tissues. Others have suggested that ongoing inflammatory responses instigated by periodontal pathogens leads to an increased risk of chronic disease and cancer [30, 32]. Of the bacteria thought to be pathogenic in periodontal disease, P. gingivalis has recently emerged as a risk factor in cancer etiology. P. gingivalis has extensively been studied due to its unique ability to invade epithelial cells and to survive in blood and host tissues, where it modulates host immune responses and promotes tumor development [17, 33–35]. Broad surveillance studies have shown that elevated serum P. gingivalis IgG levels may be associated with an increased orodigestive cancer mortality [36, 37]. Multiple clinical and experimental studies have revealed various degrees of associations between P. gingivalis and cancers of the oral cavity and gastrointestinal tract, including esophageal cancer and pancreatic cancer [17]. A recent review has highlighted the relationship between P. gingivalis and oral SCC [31] and that changes in EMT and cell invasion associated with P. gingivalis infection may be implicated in oral carcinogenesis. Therefore, we investigated the relationship between P. gingivalis and esophageal cancer progression. We found that 57% of EsoSCC patients were positive for P. gingivalis and that P. gingivalis infection was associated with a poor differentiation, advanced stages, tumor recurrences and a poor overall survival. To further investigate the link between P. gingivalis infection and esophageal cancer progression, we examined the effects of P. gingivalis on tumor growth in immunocompetent mice, using a 4NQO-induced esophageal cancer model. In 4NQO-treated mice, P. gingivalis infection was associated with enhanced glucose uptake in esophageal lesions and an elevated incidence of esophageal carcinoma development. EMT and cancer stemness are essential for the progression of epithelial tumors, including esophageal cancer [22, 38]. Here, we found that P. gingivalis infection promoted EMT-associated alterations and induced the expression of CSC markers in esophageal cancer cells. These findings suggest that the presence of P. gingivalis plays a crucial role in esophageal cancer development and progression.

P. gingivalis and F. nucleatum have been implicated in the pathogenesis of various types of gastrointestinal malignancies and to be associated with their prognosis. Co-infection with P. gingivalis and F. nucleatum has been reported to affect oral cavity SCC progression, but the role of this interplay in EsoSCC needs further investigation. In the present study, we examined the role of P. gingivalis in EsoSCC development. Our model may offer the opportunity to the further explore this interplay in the future.

Several mechanisms involved in the pro-tumorigenic effects of the microbiome have been described, including chronic inflammation [4, 7]. It has recently become apparent that inflammatory responses represent a key link between the presence of a pathogen and cancer development. Notably, certain gut microorganisms have been found to induce tumor-promoting inflammation. Furthermore, IL-6/STAT3 pathway activation has been implicated in microbiome-induced tumor progression [3, 39]. In a previous study, we found that elevated IL-6 levels were associated with a poor prognosis in patients with EsoSCC [24, 25]. We also found that IL-6 induced the expression of CSC- and EMT-related markers, while establishing an immunosuppressive tumor microenvironment. In the present study, we examined the relationship between P. gingivalis infection and IL-6 production in EsoSCC. In the 4NQO-induced tumor model, we found that P. gingivalis infection was associated with a higher IL-6 expression in esophageal tissues and an enhanced recruitment of MDSCs. In addition, we observed a positive association between the presence of P. gingivalis and IL-6 levels in primary EsoSCC patient specimens.

P. gingivalis has been reported to induce autophagy and a link between IL-6 secretion and autophagy induction has been observed in various human cancers [40, 41]. In this study, we found that both IL-6 production and autophagy were induced in P. gingivalis-infected esophageal cancer cells. Furthermore, we found that P. gingivalis infection modulated the local EsoSCC microenvironment by promoting MDSC recruitment, thus allowing cancer cells to escape from host immune responses [42]. Conversely, inhibition of IL-6 abrogated the induction of EMT- and CSC-related alterations in P. gingivalis-infected cancer cells. Based on these findings, we presume that P. gingivalis promotes esophageal cancer development and progression in an IL-6-dependent manner.

Taken together, we conclude that P. gingivalis can promote esophageal cancer development and progression, at least partly by increasing IL-6 production in the esophageal mucosa. Therefore, next to good oral hygiene, IL-6 signaling targeting should be considered in current treatment approaches for P. gingivalis-positive EsoSCC.

Funding

This work was support by Chang Gung Memorial Hospital. Grant CMRPG6H0561-2 (to M.F. Chen).

Compliance with ethical standards

Conflict of interest

The authors declare that they have no competing interests.