Role of Porphyromonas gingivalis in oral and orodigestive squamous cell carcinoma

1 |. INTRODUCTION

The nature of human existence has been the subject of philosophic debate for centuries, and more recently it has also become a topic of biological interest. Advances in our understanding of the human microbiome have led to the concept of humans as holobionts or symbiomes, integrated multitaxon units of biological organization. Indeed, an intimate relationship exists between host and microbiome, resulting in codependence to maintain a homeostatic state considered to represent health. Conversely, a dysbiotic interaction can lead to one of an ever-increasing list of pathologic conditions. While causal relationships remain to be definitively established in a number of instances, it is nonetheless apparent that many aspects of human health and disease have a microbiome component.

Carcinogenesis, by definition, involves a profound disruption of cell and tissue homeostasis. Though the etiology of cancer is complex and multifactorial, the microbiome is important, and a viral component may be the proximate cause of as much as a fifth of human cancers worldwide.¹ Carcinogenic viruses usually have latent forms, and include Epstein-Barr virus, hepatitis B and C viruses, human immunodeficiency virus type 1, human herpes virus 8, human papilloma virus, Merkel cell polyomavirus and human T-cell leukemia virus type 1. Human endogenous retroviruses have also long been suspected as oncogenic, although they may require transactivation by other viruses, such as Epstein-Barr virus.² By comparison, the role of bacteria in the etiology of cancer was historically more narrowly appreciated and less thoroughly investigated. That was to change, however, in the 1980s when pioneering and heroic (self-infection) work by Marshall and Warren³ established Helicobacter pylori as a cause of gastric inflammatory disease. This provided the foundation for studies establishing a link with stomach cancer, and by 1994 H. pylori became the first bacterial species to be recognized by the World Health Organization as a definite cause of cancer in humans. The H. pylori story precipitated a burgeoning of interest in the relationship between bacteria and cancer. At the forefront of this field of bacterial oncopathogenicity was epidemiologic correlations between periodontal disease and cancers of the head and neck, as well as of the colon and pancreas.^4–6 Moreover, interactions between oral bacteria and host cells were increasingly recognized as producing proliferative, prosurvival phenotypes recalcitrant to apoptotic cell death.^7–9 Although not universally accepted at the time,¹⁰ the notion that cancers and periodontitis both represent lesions that fail to heal¹¹ has gained traction, and considerable evidence has accumulated supporting a causal relationship between oral bacteria and cancers. Another interesting parallel to emerge is that the role of oral bacteria in head and neck cancers, as with periodontitis, depends more on community composition and action rather than any one specific organism.

2 |. BACTERIA ASSOCIATED WITH ORAL AND ORODIGESTIVE CANCER

Cancers of the head and neck region are predominantly squamous cell carcinomas,¹² and include carcinomas of the oropharynx (including the base of the tongue), generally referred to as oropharyngeal squamous cell carcinomas, along with cancers of the oral squamous cells present most frequently on the anterior of the tongue, lips, floor of the mouth, and gingiva.¹³ Human papilloma virus is the predominant risk factor for oropharyngeal squamous cell carcinomas, and the traditional risk factors of tobacco use and alcohol consumption are also operational.^14,15 While tobacco and alcohol are also risk factors for many oral squamous cell carcinomas, only a small fraction of them are attributable to human papilloma virus, and about 15% have no known risk factors.¹⁶ Indeed, smoking and alcohol consumption do not appear to be associated with gingival squamous cell carcinoma,¹⁷ and the lesions mimic the appearance of periodontal disease.¹⁸ Several studies have shown periodontitis to be a major risk factor for oral squamous cell carcinoma,^19–21 and periodontal pathogens have similarly been associated with the lesions. The catalog of organisms associated with oral squamous cell carcinoma has been extensively reviewed in several recent publications.^16,22–27 Clarity in this area, however, is hindered by the different sampling methodologies employed. Oral rinses, saliva, and serum only provide an indication of the presence of particular organisms, with no regard to the spatial relationship with the tumor. Tumor scrapings, homogenates, and swabs indicate surface colonization and/or penetration of the tumor site, whereas biopsy sections stained by immunohistochemistry show intracellular localization of the bacteria. These caveats notwithstanding, certain trends have emerged. Traditional periodontal pathogens, such as Porphyromonas gingivalis, Fusobacterium nucleatum, and Treponema denticola, tend to be positively associated with oral squamous cell carcinoma. In contrast, organisms more usually considered oral commensals, such as the mitis group streptococci, are often negatively associated and display anticancer properties in vivo.^{15,22,28–35} Interestingly, other more pathogenic streptococci, such as Streptococcus anginosus and peptostreptococci, can show a positive correlation with oral squamous cell carcinoma.^17,36,37 Accumulating evidence also implicates microbial dysbiosis in the upper digestive tract as a risk factor in the etiology of esophageal squamous cell carcinoma.^38–40 Further, there is decreased diversity of the oral microbiota along with enrichment of Porphyromonas and Prevotella in esophageal squamous cell carcinoma, compared with dysplasia and healthy controls. Higher levels of gram-negative anaerobes/microaerophiles are associated with esophagitis and Barrett’s esophagus, whereas a streptococcal-predominant microbiota resides in the normal esophagus.⁴¹ Salivary levels of P. gingivalis are associated with the progression of esophageal squamous cell carcinoma,³⁹ and P. gingivalis is overabundant in esophageal cancerous tissue.⁴² Moreover, detection of P. gingivalis in oral squamous cell carcinoma or in esophageal squamous cell carcinoma lesions is associated with a poor prognosis.^42–46 While further studies will increase precision in the field, the current pattern is one of a dysbiotic microbial community enriched in potentially carcinogenic organisms such as P. gingivalis and with an underrepresentation of homeostatic commensals, contributing to the development of oral squamous cell carcinoma and esophageal squamous cell carcinoma.

A theoretical mechanistic framework for a bacterial contribution to carcinogenesis includes modulation of the balance of host cell proliferation and death, creation of a proinflammatory microenvironment, and generation of carcinogenic metabolites.⁴⁷ As the majority of studies of bacterial function are at the single species levels, we shall address the potential tumorigenic activity of P. gingivalis as a monoinfection (Figure 1), before turning our attention to community aspects of the disease.

FIGURE 1.

Potential tumorigenic activity of Porphyromonas gingivalis in oral squamous cell carcinoma. A, Effector molecules of P. gingivalis that have been associated with a tumorigenic function. B, Tumorigenic effects of P. gingivalis on the oral epithelium. Green font denotes components that are activated or increased in messenger RNA or protein amounts. Red font denotes components that are suppressed or decreased in messenger RNA or protein amounts. EMT, epithelial-mesenchymal transition; LPS, lipopolysaccharide; NDK, nucleoside diphosphate kinase

3 |. CELL LIFE AND DEATH

Tumor cells, by definition, have the capacity for unregulated and unlimited proliferation. Enhancement of proliferation by bacteria is therefore a protumorigenic feature, and one prominently displayed by P. gingivalis. Transcriptional profiling of gingival epithelial cells transiently or persistently infected with P. gingivalis shows an upregulation of genes involved in cell proliferation.^48–51 Consistent with this, P. gingivalis can induce accelerated progression of primary gingival epithelial cells through the S-phase of the cell cycle by modulation of cyclin/cyclin-dependent kinase activity and by reducing the level of the p53 tumor suppressor protein.⁵² More rapid proliferation is dependent on the presence of the FimA protein,⁴⁶ the structural component of the major fimbrial adhesin of P. gingivalis. In addition to fimbriae, the gingipain proteases of P. gingivalis may also contribute to cell proliferation through activation of Notch signaling,⁵³ and through proteolytic degradation and activation of beta-catenin along with disassociation of the beta-catenin destruction complex.⁵⁴ Nuclear translocation and accumulation of active beta-catenin fragments drives the activity of the beta-catenin-dependent, pro-proliferative T-cell factor/lymphoid enhancer factor promoter.⁵⁴ P. gingivalis lipopolysaccharide can also activate the Wnt-beta-catenin pathway to stimulate proliferation of gingival progenitor cells.⁵⁵ P. gingivalis can thus deploy multiple effector molecules to manipulate host cell pathways that control cell division, a theme that recurs in many aspects of the interface between the bacterium and host, as shall be discussed further in subsequent sections.

Studies of noncancer-derived epithelial cells thus reveal the potential for P. gingivalis to initiate uncontrolled proliferation. Investigation of the responses of cells derived from tumors shows that P. gingivalis can reinforce these phenotypes in transformed cells. P. gingivalis was found to increase proliferation of oral squamous cell carcinoma cell lines by regulating cyclin D1 expression through the micro–ribonucleic acid (microRNA) 21/PDCD4/AP-1 negative feedback signaling pathway,⁵⁶ and by FimA interactions with CXC chemokine receptor 4 and activation of phospho-AKT1-phospho-Forkhead Box O1 signaling.⁵⁷ Acceleration through the G1 phase of the cell cycle, along with upregulation of cyclins D1 and E, has also been reported in immortalized gingival epithelial cells.⁵⁸ Additionally, in oral tumor cells derived from the alveolus, P. gingivalis can enhance expression of alpha-defensins, which have been found to elevate proliferation through intersecting with epidermal growth factor receptor signaling.⁵⁹ Similarly, infection of esophageal squamous cell carcinoma cells with P. gingivalis stimulates cell proliferation and promotes tumor growth in vivo in a mouse xenograft model.⁴² Mechanisms that have been documented include upregulation of microRNA 194, which in turn targets the GRHL3 transcription factor and modulates GRHL3/PTEN/Akt signaling,⁶⁰ along with activation of nuclear factor kappa B and subsequent upregulation of cyclin D1, c-Myc, and matrix metalloproteinases.⁶¹

A host homeostatic mechanism to protect against relentless cell proliferation is programmed cell death, or apoptosis. Successful tumors, therefore, are able to avoid these apoptotic mechanisms, and antiapoptotic proteins such as Bcl-2 are often overexpressed in cancer cells while proapoptotic proteins, such as Bad, are inactivated.⁶² P. gingivalis has adopted a multitiered approach to the suppression of apoptotic cell death in primary epithelial cells. A secreted enzyme, nucleoside diphosphate kinase, functions as an adenosine triphosphate synthase and prevents adenosine triphosphate–dependent apoptosis mediated through the purinergic receptor P2X7.⁶³ Additionally, nucleoside diphosphate kinase phosphorylation of heat shock protein 27 curtails cytochrome C release and caspase-9 activation, thus stalling apoptosis.⁶⁴ Indeed, many of the antiapoptotic activities of P. gingivalis target the intrinsic apoptotic pathway at the mitochondrial membrane, in part through stimulation of signaling through the Janus kinase 1/Alpha serine/threonine kinase (Akt)/Signal Transducer and Activator of Transcription 3 pathway.^65,66 Increased expression of Bcl2, along with a decrease in proapoptotic factors such as Bax and Bad, tips the ratio of these interacting proteins toward stabilization of the mitochondrial membrane and resistance to apoptosis, and thus activity of the downstream caspases including caspase-9 and the executioner caspase-3 is suppressed.^46,65,67,68 In another tier of antiapoptotic activity, P. gingivalis upregulates the levels of microRNA 203, which leads to inhibition of the proapoptotic signaling molecule Suppressor of Cytokine Signaling 3.⁶⁹ The multipurpose transcriptional regulator Forkhead Box O1 is also a target of P. gingivalis, which through dephosphorylation of Forkhead Box O1 serine residues induces antiapoptotic programs in epithelial cells.⁷⁰ In immortalized oral epithelial cells, similar responses involving upregulation of Bcl2, along with activation of phosphatidylinositol-4,5-bisphosphate 3 kinase/Akt and beta-catenin-dependent signaling induces anoikis resistance in P. gingivalis–infected cells.⁷¹

In addition to contributing to the development of a tumor, resistance to programmed cell death has clinical implications, as many chemotherapeutic agents function through induction of apoptosis in transformed cells. Both in vitro and in vivo studies support this concept. For example, repeated infection of oral squamous cell carcinoma cells with P. gingivalis diminishes susceptibility to taxol⁷²; and in murine models, tumor xenografts composed of P. gingivalis–infected oral squamous cell carcinoma cells showed resistance to taxol through activation of Notch1 signaling.⁷³

4 |. METASTASIS

Another hallmark of tumor malignancy is metastasis, a process by which cancer cells spread throughout the body. Early events in metastasis involve the epithelial-mesenchymal transition, whereby epithelial cells lose tight junctions and polarization, and acquire mesenchymal properties including motility and a cancer stem cell–like phenotype that is capable of seeding new tumors.⁷⁴ Epithelial-mesenchymal transition is a component of normal processes such as embryogenesis and wound healing and is controlled by a complex regulatory network that converges on a series of transcription factors, such as Zinc Finger E-box Binding Homeobox 1 and 2, Snail Family Transcriptional Repressors 1 and 2, and Twist. Originally considered a binary condition, epithelial-mesenchymal transition is now thought to represent a spectrum of states through which cells can transition in either direction.^75,76 P. gingivalis shows a remarkable ability to impact the activity of the pathways that control epithelial-mesenchymal transition and induce at least a partial shift toward the mesenchymal state^{29,72,77–79} (Figure 2). Zinc Finger E-box Binding Homeobox 1 is elevated through FimA-dependent signaling, whereas Zinc Finger E-box Binding Homeobox 2 is regulated by gingipain processing and activation of beta-catenin, along with dephosphorylation and activation of Forkhead Box O1.^29,79 P. gingivalis infection of epithelial cells also increases cancer stem cell markers, such as CD44 and CD133, and enhances migration.^{29,72,77–79} Migration and metastasis of epithelial cells can be facilitated by host matrix metalloproteinase enzymes, which degrade extracellular matrix and basement components. P. gingivalis can increase production of several matrix metalloproteinases, including matrix metalloproteinase-1, 2, 7, 9, and 10, from primary and transformed oral epithelial cells.^{29,72,78,80,81} Moreover, in invasive oral squamous cell carcinoma lines, P. gingivalis gingipains stimulate proteinase-activated receptors 2 and 4, which increases signaling through Extracellular signal-regulated kinase 1/2-Ets1, p38/Heat Shock Protein 27, and nuclear factor kappa B pathways, consequently elevating matrix metalloproteinase-9 proenzyme expression.^80,82 Activation of matrix metalloproteinase-9 is then enhanced by gingipain processing. Infection of oral squamous cell carcinoma cells by P. gingivalis also elevates cell migration, as well as tumorsphere formation, through integrin alpha V/Focal Adhesion Kinase signaling.⁸³ Epithelial-mesenchymal transition in esophageal squamous cell carcinoma cells is promoted by intracellular P. gingivalis, which induces the formation of transforming growth factor beta–dependent Smads–Yes1-associated transcriptional regulator–Tafazzin–TEA domain transcription factor 1 complexes. These protein aggregations orchestrate the activity of several transcription factors that promote epithelial-mesenchymal transition and stem-like traits.⁴⁴ Furthermore, P. gingivalis can augment the secretion and bioactivity of transforming growth factor beta through upregulation of its surface docking receptor, GARP.⁴⁴

FIGURE 2.

Impact of Porphyromonas gingivalis on pathways associated with epithelial-mesenchymal transition and the mitigating effect of Streptococcus gordonii. Transcription factors upregulated by P. gingivalis are shown, although the pathways have not been defined in all cases. MMPs, metalloproteinases. P indicates phosphorylation

5 |. ANGIOGENESIS

Developing tumors require a blood supply to provide nutrients. Additionally, to metastasize, tumors need a route to enter the circulation, and this is provided by tumor blood vessels. Thus, angiogenesis is necessary for both tumor development and metastasis. Interleukin-1beta (IL-1β) and tumor necrosis factor, which are secreted by epithelial cells in response to P. gingivalis,^70,84–86 are proangiogenic. IL-1β activates endothelial cells to produce vascular endothelial growth factor, and tumor necrosis factor contributes to an angiogenic microenvironment.⁸⁷ Lipopolysaccharide from P. gingivalis also increases production of Angiopoietin-like Protein 2 from transformed gingival epithelial cells.⁸⁸ Angiopoietin-like Protein 2 can induce angiogenesis and also increase IL-1β and tumor necrosis factor production through an autocrine loop. P. gingivalis lipopolysaccharide evokes angiogenic responses directly in endothelial cells by activating the mitogen-activated protein kinase Extracellular signal-regulated kinase 1/2.⁸⁹ Proangiogenic outcomes will be reinforced by lower levels of the angiostatic cytokines CXC motif ligands 9, 10, and 11,⁹⁰ which P. gingivalis suppresses transcriptionally through downregulation of interferon regulatory factor 1 and reducing the amounts of Stat1.⁹¹ Angiogenic properties are also exhibited by the GroEL protein of P. gingivalis, which stimulates endothelial nitric oxide synthase production and p38 mitogen-activated protein kinase signaling. Consequently, GroEL enhances migration of endothelial progenitor cells and promotes angiogenesis and tumor growth in animal models.⁹²

6 |. INFLAMMATION

In the periodontal space, inflammation is an important ecologic determinant that releases nutrients through the destruction of host tissue. P. gingivalis is an inflammophilic organism that utilizes these nutrients for growth.⁹³ Interestingly, to avoid adverse effects of inflammation, P. gingivalis can selectively suppress bactericidal aspects, thus creating a dysbiotic inflammatory microenvironment.^8,94 Uncontrolled inflammation is also considered a major driver of tumorigenesis,⁹⁰ as inflammation can disrupt stromal integrity, and cytokine-promoted proliferation will facilitate tumor growth.⁹⁵ Hence, an inability to resolve inflammation provides a mechanistic link between periodontitis and carcinogenesis. Various cytokines and CXC family chemokines have been reported to increase oral squamous cell carcinoma growth and increase cancer cell migration.^96,97 Serum levels of interleukin-6 (IL-6) are higher in oral squamous cell carcinoma patients and associated with a worse prognosis,⁹⁸ and CXC motif ligand 8 (interleukin-8) is increased in the saliva of patients with oral squamous cell carcinoma.⁹⁹ Of note, epithelial cells possess pathways to constrain cytokine production and maintain homeostasis. For example, A20 (TNFAIP3), a ubiquitin-editing enzyme, dampens IL-6 and CXC motif ligand 8 production through modulation of nuclear factor kappa B activity.¹⁰⁰ Moreover, the serine phosphatase SerB of P. gingivalis antagonizes production of CXC motif ligand 8 from primary gingival epithelial cells through dephosphorylation of the p65 subunit of nuclear factor kappa B.¹⁰¹ Similar restraint mechanisms are also observed in innate immune cells. P. gingivalis activation of Janus kinase 3 curtails production of IL-6 and tumor necrosis factor through ubiquitination dependent Wnt3 degradation.¹⁰² Nonetheless, despite the presence of both host and bacterial mechanisms to dampen cytokine/chemokine production, once cells have become transformed, P. gingivalis stimulates secretion of IL-6 and CXC motif ligand 8.¹⁰³ In addition to canonical activation of pattern-recognition receptors, cytokines/chemokines can be stimulated by P. gingivalis through modulation of microRNA expression. For example, P. gingivalis can suppress mi-205–5p production, which promotes cytokine synthesis through the Janus kinase/Signal Transducer and Activator of Transcription pathway.¹⁰⁴ The implications of increased cytokine levels are many. IL-6 and CXC motif ligand 8 can increase matrix metalloproteinase levels and cell invasiveness, as well as modulate expression of genes involved in regulation of the cell cycle and apoptosis.^90,105 CXC motif ligand 8 can stimulate proliferation through transactivation of the epidermal growth factor receptor.¹⁰⁶ Epidermal growth factor itself can also be induced by P. gingivalis, which will contribute to the induction of epithelial-mesenchymal transition, as will transforming growth factor beta 1 and tumor necrosis factor, which are also upregulated by P. gingivalis.^77,107

The interleukin-23/interleukin-17 axis, which regulates homeostasis in the periodontium,¹⁰⁸ is strongly protumorigenic, at least in colorectal cancer.¹⁰⁹ P. gingivalis can incite production of these cytokines,^110,111 although the carcinogenic potential has yet to be investigated.

Programmed death-ligand 1 (B7-H1, CD274) plays a crucial role in the control of T-cell function and survival. Programmed death-ligand 1 expression is upregulated in invasive oral squamous cell carcinoma cells,¹¹² and tissue samples of oral squamous cell carcinoma express both programmed death-ligand 1 and another ligand of the programmed death receptor programmed death-1, namely B7-DC.¹¹³ In both oral squamous cell carcinoma cell lines and primary gingival epithelial cells, P. gingivalis can upregulate programmed death-1 and B7-DC.¹¹⁴ Mechanistically, peptidoglycan from P. gingivalis, packaged in outer membrane vesicles, is internalized by host cells and induces ligand expression by activating serine/threonine kinase RIP2–dependent signaling.¹¹⁵ Enhanced programmed death-ligand 1 and B7-DC expression may lead to anergy and apoptosis of activated T-cells and contribute to the resistance of tumor cells to host immune responses.¹¹⁴

Immunoevasion can also be promoted by P. gingivalis through modulation of macrophage activity. P. gingivalis inhibits phagocytosis of oral squamous cell carcinoma cells by macrophages, and in an animal model increases the relative amount of protumorigenic M2 macrophages in the tumor-associated macrophage population.¹¹⁶ Additionally, secretion of the tumor-enhancing molecules interleukin-1alpha, CC chemokine ligand 3, and chemokine ligand 5 is elevated in macrophages challenged with P. gingivalis. The ability of P. gingivalis to modulate macrophage function may depend on the amount of sphingolipid in the bacterial membrane.¹¹⁷ P. gingivalis also disrupts immune surveillance by generating myeloid-derived dendritic suppressor cells, which functionally resemble myeloid-derived suppressor cells associated with oncogenesis. Through interaction with Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin, the Mfa1-component fimbriae of P. gingivalis mediate invasion into monocytes, which promotes differentiation to apoptosis-resistant indoleamine 2,3-dioxygenase–competent myeloid-derived dendritic suppressor cells. These myeloid-derived dendritic suppressor cells induce immune tolerance through increased FOXP3⁺ T_reg responses.⁵⁷

7 |. TOXIC METABOLITES

P. gingivalis secretes a variety of metabolic end products as a result of its asaccharolytic metabolism; however, study of the carcinogenic potential of these is scant. Volatile sulfur compounds, such as hydrogen sulfide, methyl mercaptan, dimethyl sulfide, and dimethyl disulfide, are cytotoxic, and hydrogen sulfide in particular may also be genotoxic and stimulate cell proliferation.¹¹⁸ Short-chain fatty acids, such as butyrate and propionate, are produced in abundance by P. gingivalis and influence the physiology of epithelial and immune cells through serving as energy sources.^119,120 Hence, an imbalance in the levels of short-chain fatty acids in the tumor microenvironment has the potential to impact cell proliferation and differentiation; however, the matter requires experimental investigation. Butyric acid produced by P. gingivalis can contribute to activation of the Epstein-Barr virus lytic cycle. Butyric acid inhibits histone deacetylases, thus increasing histone acetylation and the transcriptional activity of the Epstein-Barr virus BZLF1 gene, which encodes ZEBRA, a master regulator of the transition from latency to the lytic replication cycle.¹²¹

8 |. IN VIVO STUDIES

In vivo studies support the carcinogenic potential of P. gingivalis. In the 4-nitroquinoline-1-oxide tongue squamous cell carcinoma model, P. gingivalis increased both the size and the number of tumors.¹²² The development of carcinomas was associated with enhanced free fatty acid production, both in the tongue and in the serum of 4-nitroquinoline-1-oxide–treated mice, a shift that can also be observed in oral squamous cell carcinoma.¹²² Promotion of oral squamous cell carcinoma progression by P. gingivalis in the 4-nitroquinoline-1-oxide model has been corroborated in an independent study, which further showed P. gingivalis invasion increased infiltration of oral lesions with immunosuppressive CD11b⁺ myeloid cells and myeloid-derived suppressor cells.¹²³ In a subcutaneous transplantation model of oral squamous cell carcinoma in mice, repeated infection with penicillin/streptomycin-treated P. gingivalis resulted in an increase in tumor growth and volume.¹¹⁶ Similarly, in a murine floor-of-mouth model, mice injected with P. gingivalis–challenged oral squamous cell carcinoma cells exhibited a greater tumor burden.⁸³ Oral infection of conventional mice has been found to enhance Zinc Finger E-box Binding Homeobox 1 levels in gingival tissues, indicating the potential for P. gingivalis to initiate epithelial-mesenchymal transition in vivo.²⁹

9 |. COMMUNITY ACTION

Similar to periodontal diseases, there is a diverse and complex microbial community associated with oral squamous cell carcinoma and esophageal squamous cell carcinoma.^16,22,24,27 Bacteria in communities communicate with one another and exchange nutritional substrates to optimize physiology and metabolism. Functional specialization among community participants leads to codependence, and the community, rather than individual species, emerges as the pathogenic unit. Host responses to bacterial communities, including epithelial derived, are clearly distinct from responses to individual constituent organisms.^85,124,125 Community pathogenicity, or nososymbiocity, is well studied in periodontal disease, and similar concepts appear applicable to the microbial involvement in tumorigenesis.^14,35,126 As documented herein, P. gingivalis clearly has the potential to impinge upon the host in a manner consistent with a carcinogenic outcome. What, then, of the ability of other community participants to modulate the action of P. gingivalis? Synergistic interactions have been demonstrated in the 4-nitroquinoline-1-oxide model, in which coinfection with F. nucleatum and P. gingivalis promotes cancer progression.¹²⁷ The ability of organisms, such as many oral streptococcal and Neisseria species, to generate the carcinogen acetaldehyde from alcohol could also enhance the tumorigenic potential of heterotypic communities.^128,129 On the other hand, interactions with oral streptococci may mitigate the tumorigenic potential of P. gingivalis. S. gordonii can prevent P. gingivalis–induced gingival epithelial cell proliferation,⁹⁴ and through suppression of P. gingivalis–induced Zinc Finger E-box Binding Homeobox 2 upregulation⁷⁹ (Figure 2). Mechanistically, S. gordonii activates the TAK1-NLK pathway, which negatively regulates Forkhead Box O1–dependent regulation of Zinc Finger E-box Binding Homeobox 2. Indeed, coinfection of gingival epithelial cells with S. gordonii overrides much of the transcriptional program differentially regulated by P. gingivalis and returns the cells closer to the homeostatic state.^130,131 One of the major pathways transcriptionally induced by P. gingivalis, but suppressed in a dual-species context with S. gordonii, is the Notch signaling pathway.¹³¹ Activation of Notch signaling proceeds through increased expression of the Notch1 receptor and the Jagged1 agonist. Following Jagged1-Notch1 engagement, the Notch1 extracellular domain is cleaved by P. gingivalis gingipain proteases to activate signaling. The Notch downstream effector olfactomedin 4, a prosurvival glycoprotein,¹³² participates in epithelial cell migratory, proliferative, and inflammatory responses to P. gingivalis. Olfactomedin 4 has relevance in the oral cavity, as it accumulates in the secretome of head and neck squamous cell carcinomas and is a potential biomarker for the disease.¹³³ Antagonism by S. gordonii involves inhibition of gingipain activity by secreted hydrogen peroxide, which prevents cleavage of the Notch1 extracellular domain (Figure 3). Such abilities position S. gordonii, and potentially other peroxide-producing streptococci, as a homeostatic commensal in tumorigenesis, an organism that functions to maintain eubiosis.⁹⁴ It is interesting to note that in periodontitis S. gordonii has properties of an accessory pathogen, an organism that enhances the virulence of more overtly pathogenic species,^134,135 reinforcing the importance of ecologic context in the pathogenic potential of an organism or community. Antagonism between oral streptococci and P. gingivalis is consistent with many of the in vivo studies showing a trend of reduced streptococci and elevated P. gingivalis associated with tumors. However, as with periodontal disease, it is difficult to discern whether disease results from the relative abundance of the organism or if the relative abundance of the organism reflects its competitiveness in the disease microenvironment.

FIGURE 3.

Interplay between Porphyromonas gingivalis and Streptococcus gordonii in the activation of Notch1 signaling and upregulation of olfactomedin 4 (OLFM4). Activation of Notch signaling is induced through increased expression of the Notch1 receptor and the Jagged1 (Jag1) agonist. In addition, Jagged1 is released in response to P. gingivalis, leading to paracrine activation. Following Jagged1-Notch1 engagement, the Notch1 extracellular domain is cleaved by P. gingivalis gingipain proteases. Antagonism by S. gordonii involves inhibition of gingipain activity by secreted hydrogen peroxide. Olfactomedin 4 is involved in epithelial cell migratory, proliferative and inflammatory responses to P. gingivalis. ICD: intracellular domain

One characteristic of the “community as pathogen” model is that the identities of constituent bacteria are less important than the functions provided by the metagenome. Study of the metatranscriptome of human oral squamous cell carcinoma tumors found microbial metabolic activities such as iron transport, tryptophanase activity, peptidase activities, and superoxide dismutase were better correlated with disease than was community composition.¹³⁶ In support of this concept, a comparison of microbiotas associated with oral squamous cell carcinoma in different countries revealed functional rather than compositional similarities.¹³⁷ Moreover, studies of the microbiota in the 4-nitroquinoline-1-oxide model show consistent patterns of metabolic signatures associated with disease.¹³⁸ Genes associated with bacterial chemotaxis, flagellar assembly, and lipopolysaccharide biosynthesis in particular have been correlated with oral squamous cell carcinoma,^{33,34,137,139} and lipopolysaccharide biosynthesis genes are also enriched in gingival squamous cell carcinoma patients.

10 |. CULPRITS OR INNOCENT BYSTANDERS

The etiology of cancer is multifactorial, involving multiple genetic and environmental predisposing factors, which may operate in a temporally defined manner. Defining the contribution of any individual factor is thus fraught with difficulty. Deciphering the role of bacteria becomes even more complicated when incorporating the notion that heterotypic communities operate as a functional unit.¹⁴⁰ Nonetheless, compelling models that accommodate the existing data are being developed. Al-Hebshi et al²⁷ have proposed the “Passenger-Turning-Driver” microbiome model for oral squamous cell carcinoma. This holds that the oral microbiome is not involved in the initiation of disease. Rather, the nature of the tumor microbiome is a “passenger” event resultant from selection within the tumor microenvironment. As this competitively fit microbiome develops, there is increased expression of proinflammatory components and a transition to a dysbiotic, or “driver,” state that enhances tumorigenesis by sustaining chronic inflammation. A related model, which has also been applied to colorectal cancer,¹²⁶ derives from the polymicrobial synergy and dysbiosis model of periodontal disease.¹⁴¹ Polymicrobial synergy and dysbiosis posits that driver mutations in host cells begin to establish a tumor microenvironment that selects for a microbial community that can vary among sites. In instances where there is a relative decrease in homeostatic commensals, such as S. gordonii, and a relative overabundance of gram-negative anaerobes, such as P. gingivalis, the community will have tumorigenic potential. Suppression of programmed cell death, stimulation of uncontrolled epithelial cell proliferation, along with a more mesenchymal, migratory phenotype all contribute to tumor development. Dysbiotic inflammation further contributes to oncogenesis while also sustaining colonization by inflammophilic organisms such as P. gingivalis through a reciprocating feed-forward loop. In terms of the passenger/driver metaphor, this may be characterized as the “hitchhiker-turned-car-jacker” model. These models have several features in common, and both may be operational in different contexts.

11 |. CONCLUSIONS

What was once considered at best fanciful and at worst “fake science” is now undeniable: bacteria such as P. gingivalis have a role to play in certain cancers, particularly those of the oral cavity and orodigestive region. The nature of that role requires further research, including large longitudinal and intervention studies. In one sense, however, it is immaterial whether certain bacteria have a causal role or simply have a competitive advantage in the tumor microenvironment. These cancers are often only detected at advanced stages; therefore, any consistent associations occurring between the microbiome/metatranscriptome and disease may provide a foundation for discovering novel targets for early detection and diagnosis.