HPV-related and unrelated oropharyngeal cancer: Contribution of oral pathobionts Treponema denticola and Porphyromonas gingivalis to carcinogenesis

The incidence of oropharyngeal squamous cell carcinoma (OPSCC) associated with human papillomavirus (HPV) has increased in western countries in the last decades, whereas the number of HPV unrelated OPSCC cases has remained stable or even declined. The prognosis of patients carrying an HPV-positive tumor is considerably more favourable than that of patients with HPV-negative disease. Differences between these two OPSCC subgroups with distinct etiology, prognosis, biology, and genetic landscape have been studied widely. Numerous trials examine the potential for treatment de-escalation of HPV-positive OPSCC to improve quality of life while maintaining acceptable survival outcomes. The diagnostic biomarkers for the treatment de-escalation decision remain, however, unavailable. Further, as OPSCC is often detected at an advanced stage owing to a lack of symptoms in the early stages, there is a similar need to identify and validate diagnostic biomarkers for earlier detection. The microenvironment of the OPSCC in oropharyngeal cavity is exposed to the influence of oral microbiota. The bacteria of the oral microbiota may influence the cancer development by inducing and maintaining local and systemic chronic inflammation cascades, by supressing host immune responses, and by producing carcinogenic metabolites. Besides the acknowledged etiological role of HPV in OPSCC, several oral pathogens including Porphyromonas gingivalis (Pg), Treponema denticola (Td), Fusobacterium nucleatum (Fn) Streptococcus mitis, and Streptococcus anginosus have been associated with oral and gastrointestinal tract cancers. The role of these bacteria in carcinogenesis is, however, unclear. This study aimed to investigate the prevalence and role of oral pathobionts Td and Pg in the development and progression of HPV-related and HPV-unrelated OPSCC in two patient cohorts available in Helsinki University Hospital. The possible underlying mechanisms influencing carcinogenesis were explored by studying biomarkers related to host immune responses, toll-like receptors (TLRs) and matrix metalloproteinases (MMPs), in tumors and in the tumor microenvironment (TME). In addition, activation of transcription factor NF-κB cascade, involved in control of transcription of proinflammatory cytokines, was investigated in OPSCC cell lines for validation of the previously reported prognostic role of TLR5. The goal was to find biomarkers useful in prevention, screening, and individual treatment planning of OPSCC, as well as in evaluation of treatment response and prognosis. Clinicopathological data for patient cohorts I and II were derived from university hospital records. The immunoexpression of species-specific enzymes Td chymotrypsin-like protease (Td-CTLP) and Pg arginine-specific cysteine proteinase RgpA (Rgp) were used as surrogate biomarkers for Td and Pg, respectively. Td-CTLP, Rgp, TLR2, 4, 5, 7, and 9, and MMP-8 and 9 immunoexpression was evaluated by immunohistochemistry. The Cox proportional hazards model served in univariable and multivariable survival analysis. The proportional hazard assumptions were tested with KM curves (studies I, II, IV). In study III, death by other cause was included in the Cox proportional hazards model as a competing event with death by OPSCC, and subdistribution hazard ratios (SHR) were calculated. NF-κB cascade activation via TLR5 was studied in vitro in two human OPSCC cell lines and immortalized human keratinocytes (HaCat). Flagellin served to stimulate TLR5 in cultured cells, followed by analysis of the activity of the NF-kB signaling cascade with In-Cell Western for IκB and p-IκB. Td-specific enzyme Td-CTLP and Pg-specific enzyme Rgp were immunoexpressed in OPSCC in both HPV-positive and HPV-negative disease. Td-CTLP and Rgp were detectable in tumor cells and in the cells of surrounding stroma. Correlation profiles of the whole patient cohort and HPV-positive and HPV-negative subgroups differed when correlations between Td-CTLP and Rgp, and TLR2, 4, 5, 7, and 9, and MMP-8 and 9 were evaluated. In multivariable survival analysis TLR2, tumoral MMP-9 and Rgp in inflammatory cells were independent prognostic factors for DSS: Strong TLR2 and tumoral MMP-9 immunoexpression predicted poor prognosis among HPV-negative patients, whereas strong immunoexpression of Rgp in inflammatory cells was associated with favourable outcome in the same group. Immunoexpressions of tumoral MMP-9 and Rgp in inflammatory cells, but not TLR2, were similarly prognostic among all OPSCC. Td-CTLP and Rgp were immunoexpressed in both HPV-positive and HPV-negative OPSCC, reflecting current or former presence of Td and Pg in tumors and surrounding TME. Both pathobionts contributed to the immune TME, acting in concert with known etiological factors in progression of both OPSCCs. Based on biomarker correlation profiles and survival analysis results, the role of Td and Pg differ in HPV-related and unrelated OPSCC. Possible events influencing the immune TME, and thus, OPSCC development and progression may be: 1) Rgp-related increase in tumoral TLR4 immunoexpression, especially in HPV-positive OPSCC; 2) Rgp activation of host responses that lead to increased MMP-9 expression and impaired prognosis among all OPSCC, especially in HPV-negative OPSCC; 3) Rgp-related host response involving tumor infiltrating inflammatory cells and promoting tumor suppression and improved prognosis among all OPSCC, especially in HPV-negative disease. These findings strengthen the view of oral pathobionts as factors in oral carcinogenesis and emphasize the urgency of oral health measures in prevention of oral cancer. Further studies on molecular mechanisms and underlying causalities are needed, however, for the development of effective prevention and treatment strategies for both OPSCCs.