Background
Since the discovery that
Helicobacter pylori plays a causative role in gastric adenocarcinoma, multiple other associations between specific bacteria and cancer have been reported [
1,
2], including
Salmonella typhi with gall bladder cancer [
3],
Streptococcus bovis with colon cancer [
4],
Chlamydophila penumoniae with lung cancer [
5],
Bartonella species with vascular tumor formation [
6],
Propionibacterium acnes with prostate cancer [
7], and
Escherichia coli with colon cancer [
8]. Esophageal cancer is the eighth most frequent tumor and sixth leading cause of cancer death worldwide. Whereas the majority of cases occur in Asia, particularly in central China, recent data suggest that the frequency of new cases is rising in Western Europe and the USA [
9,
10]. Two major histological subtypes of esophageal cancer have been identified including squamous cell carcinoma (ESCC), which is more common in developing countries, and adenocarcinoma, which is more common in developed nations [
11]. Esophageal cancer is characterized by difficulty of early diagnosis, rapid development and high mortality. Therefore, there is a considerable need to better understand causative agents in order to reduce the incidence and mortality of this disease. Like most cancers, a plethora of risk factors including age, gender, heredity, gene mutation, chemical exposure, and diet have been reported for esophageal cancer [
12,
13].
A potential contribution of microbes to the development of esophageal cancer is beginning to emerge. Pei et al. reported that
Streptococcus, Prevotella and
Veillonella are the most prevalent genera detected in esophageal biopsies [
14,
15]. Yang et al. have classified the esophageal microbiota into two subtypes: the
Streptococcus-dominated type I microbiome, which is mainly associated with a normal esophagus, and the type II microbiome, in which Gram-negative anaerobes predominate, which is associated with Barrett’s esophagus (BE) and esophagitis [
16]. A significant association between the inhabitants of the upper digestive tract microbiota and esophageal squamous dysplasia, a precursor lesion of esophageal squamous cell carcinoma, has also been reported [
17]. While there are several phylum-wide studies on the esophageal microbiota and the possible associations with reflux esophagitis, Barrett’s esophagus, and esophageal squamous dysplasia, there has been no research on the esophageal microbiota in patients suffering from ESCC, especially at species level, let alone the possible association of these bacteria with the development of ESCC.
The microbiome in chronic and severe manifestations of periodontal disease is enriched for Gram-negative anaerobic bacteria. Among these,
Porphyromonas gingivalis is a keystone oral pathogen which can invade epithelial cells, and interfere with host immune responses and the cell cycle machinery [
18‐
20]. Epidemiological studies have demonstrated that periodontal diseases and tooth loss are significantly associated several cancers such as oral cancer, gastric cancer, and pancreatic cancer and may even relate to survival [
20‐
24].
P. gingivalis-mediated immune evasion, apoptosis inhibition, carcinogen conversion, induction of MMP-9 and dysbiosis of the oral microbiota have all been posited as pro-tumorigenic mechanisms in the context of oral squamous cell carcinoma [
20,
25]. Since esophageal squamous cells are histologically similar to oral squamous cells and esophageal infection arising from the oral niche is highly plausible, we hypothesized that
P. gingivalis may be associated with ESCC. We set out to test this hypothesis using 100 ESCC subjects and 30 normal matched controls.
Discussion
To the best of our knowledge, the composition and potential role of the esophageal microbiota in the patients suffering from ESCC have not been investigated. Using three complementary approaches, we have established that antigens and DNA from P. gingivalis, a periodontal pathogen, can be detected in the epithelium of the esophagus of ESCC patients. The intensity of expression of whole P. gingivalis antigen, its unique protease Lys-gingipain, and detection of P. gingivalis-specific16S rDNA were all significantly higher in the cancerous tissue of ESCC patients than in the surrounding tissue or normal control sites. Moreover, our analysis indicates that the presence of P. gingivalis correlates with multiple clinicopathologic factors, including cancer cell differentiation, metastasis, and overall survival ESCC rate. These findings provide the first direct evidence that P. gingivalis infection could be a novel risk factor for ESCC, and may also serve as a prognostic biomarker for this prevalent cancer.
A number of aspects of the interaction of
P. gingivalis with host epithelial cells provide a plausible molecular basis for potential
P. gingivalis-mediated carcinogenesis [
20,
25,
26]. First, chronic inflammation
per se is associated with the development of cancer [
27], and, for example, prolonged IL-6 signaling and STAT3 activity is known to be pro-tumorogenic [
28,
29]. In this regard, both our group and others have demonstrated that
P. gingivalis activates JAK2 and GSK3β pathways, thus increasing the production of IL-6 in epithelial cells [
28,
30]. Secondly,
P. gingivalis can promote tumorigenesis by secreting a nucleoside diphosphate kinase (NDK). NDK from
P. gingivalis antagonizes ATP activation of P2X
7 receptors, and thus reduces IL-1β production from epithelial cells [
31]. Since IL-1β is critical for priming IFNγ-producing, tumor–antigen-specific CD8
+ T cells, NDK from
P. gingivalis could promote the immune evasion of tumor cells [
20]. Moreover, NDK-mediated degradation of ATP also suppresses apoptosis dependent on ATP activation of P
2X7 receptors [
32]. Thirdly,
P. gingivalis inhibits epithelial cell apoptosis by a number of mechanisms, including activation of Jak1/Akt/Stat3 [
33,
34], enhancing the Bcl2 (antiapoptotic): Bax (proapoptotic) ratio, blocking the release of the apoptosis effector cytochrome c, and the activation of downstream caspases [
35]. Moreover,
P. gingivalis can upregulate microRNAs, such as miR-203, which suppress apoptosis in primary gingival epithelial cells [
36]. In concert with suppression of apoptosis,
P. gingivalis can accelerate progression through the cell cycle by manipulation of cyclin/CDK (cyclin-dependent kinase) activity and reducing the level of the p53 tumor suppressor [
37]. Lastly, in oral squamous cell carcinoma (OSCC) cells,
P. gingivalis promotes cellular migration through activation of the ERK1/2-Ets1, p38/HSP27, and PAR2/NF-κB pathways to induce pro-matrix metalloproteinase (MMP)-9 expression [
25]. Apart from all the above, another possible mechanism for
P. gingivalis induced carcinogenesis is the metabolism of potentially carcinogenic substances. For example,
P. gingivalis converts ethanol into its carcinogenic derivative, acetaldehyde, to levels capable of inducing DNA damage, mutagenesis and hyperproliferation of the epithelium [
38,
39], which could help explain the epidemiological evidence associating heavy drinking and development of some cancers [
20].
While it is possible that
P. gingivalis infection initiates or is a co-factor in the transformation of esophageal epithelial cells, the possibility that cancer tissues represent a preferred microenvironment for
P. gingivalis cannot be excluded. Thus, while our results reveal a positive association between infection with
P. gingivalis and the progression of ESCC,
P. gingivalis is not yet established as a novel etiological agent or co-factor of ESCC. Should
P. gingivalis prove to cause ESCC, the implications are enormous. It would suggest (i) that improved oral hygiene might reduce ESCC risk, (ii) that screening for
P. gingivalis in dental plaque may identify susceptible subjects, and that (iii) antibiotic use, or other anti-bacterial strategies, may prevent ESCC progression. Should the clear association between
P. gingivalis infection and ESCC turn out to be better explained by physiological conditions inside ESCC cells being more amenable to
P. gingivalis survival and growth, this would imply that attenuated
P. gingivalis or non-pathogenic bacteroidetes strains that contain eukaryotic lysins may represent a novel and effective therapeutic approach for ESCC. In this regard, several studies have attempted to take advantage of the oxygen-limited conditions present in malignant cells to develop anaerobic, non-pathogenic bacteria for the delivery of cancer cell cytolysins [
40]. These include
Clostridium novyi for the treatment of melanoma and modified
Bifidobacterium longum carrying 5-fluorocytosine for the treatment of breast cancer [
41‐
43]. Hence, further studies to determine if
P. gingivalis infection promotes the initiation and progression of ESCC are required.
Finally
, colonization by
P. gingivalis promotes the conversion of a symbiotic to a dysbiotic of oral microbiota, a process considered critical for the progression of periodontal disease [
44]. Dysbiosis of the microbiota in the esophagus could potentially cause or exacerbate the severity of esophageal disorders [
18]. Thus, a further possibility to be tested is that esophageal infection with
P. gingivalis leads to shift in the microbiome involved in the development of esophageal cancer.
Competing interests
The authors declare they have no competing interests.
Authors’ contributions
SG, SL, and HW contributed to the experimental studies; SG, XF and HW contributed to the study design, supervision of experiments, and manuscript review; ZM, SL, TS, MZ, XZ, PZ, GL, FZ, XY, and RJ contributed to the collection of samples, the acquisition of clinical data, and the supervision of the experiments; FZ, HW, DAS, and RJL conceived of the study and prepared the manuscript; ZM, SL, TS, MZ, XZ, PZ, GL, FZ, XY, and RJ performed surgical treatments, and patient follow-ups; JP, DAS and RJL contributed to the study design and manuscript review. All authors read and approved the final manuscript.