Among the viruses, HPV has been reported to be associated with OSCC by many researchers (55–59). The HPV-16 integrates its genome in specific sites of HNSCC tumor. The integrated genome consists of seven early genes namely (E1-E7) and two late genes (L1-L2). E6 protein is involved in degradation of p53 while E7 protein induces cell cycle to go beyond restriction point (G1-S checkpoint) into S phase. Increased survival as well as accumulations of DNA damage and mutations over cell replications results in malignant transformation and development of carcinomas (59). The Epstein-Barr virus (EBV) was found to be involved in nasopharyngeal cancer (60, 61), OSCC and potentially malignant oral diseases including gingivitis, and periodontitis. This virus can establish latent infection in the B-lymphocytes and salivary gland cells (61). Using FaDu oral cancer cell line, researcher reported that Coinfection of HPV and EBV can induce the tumorigenicity of oral cancer (62). Many research group has been studying the association of herpes simplex virus (HSV), an adenovirus with oral cancer in clinical subjects and animal model (63, 64). A recent study has shown that HSV-1 is predominant in oral cavity and tumor tissue of OSCC subjects, however has no significance in OSCC cell survival and invasion (65). A population-based study showed hepatitis C virus (HCV), not hepatitis B virus (HBV) as risk factor for oral cancer (66). In addition to virus, fungal microorganism such as Candida has been identified in OLK lesions (50, 67). Study reported the high abundance of Candida (78.8% of all OSCC cases) and Saccharomyces (76.8% of all OSCC) in OSCC patients versus healthy control (68). Importantly, dysbiosis of oral bacteria due to the use of broad spectrum antibiotics may pave the way for the growth of hyperplastic Candidiasis that may contribute to OLK genesis (69, 70) as well as OSCC (71). Candida infection leads to induction of cytokines such as IL8 and TNF-α to activate TLRs which can interact with NF-kB inflammatory pathway in metastasis of oral cancer (72). Thus, dysbiosis creates a chronic inflammatory state, where bacteria, fungal metabolites, as well as the pro-inflammatory cytokines may lead to the genesis of OLK, and then subsequent malignant transformation.

Studies have shown that cancer staging could be achieved by the presence of specific microbiota, indicating the association of specific bacterial population in each stage of malignant progression (37, 73). Indeed a study by Yang et al. reported that the relative percentage of bacteria in saliva increased from stage 1 to stage 4 of OSCC patients as compared to healthy controls. The F. periodonticum percentage was found to be increased in oral saliva of OSCC patients from 1.66% in stage 1 to 2.41% in stages 2 and 3, further to 3.31% in stage 4. Some other bacteria including Parvimonas micra, Streptococcus constellatus, Haemophillus influenza, Filifactor alocis were also found to be increased in the same patients from stage 1 to stage 4 (73). Depending on the precancerous or OSCC stage, bacteria including Fusobacterium spp., Prevotella spp., Haemophilus spp., and Campylobacter spp. found in the samples express specific associated functional pattern (37).

Importantly, many commensal microbiota in oral cavity are reported to be healthy in context of carcinogenesis or OSCC progression (73–75). Researchers found P. Pasteri, a commensal bacterium present in oral saliva of healthy individual decreases in oral saliva of OSCC patients (73). This research group further found Streptococcus mitis and Haemophilus parainfluenza as prominent bacterial species associated with healthy individuals (73). Using OSCC cell lines, another group of researchers reported that S mitis bacterial lysate can inhibit OSCC cell proliferation (76). Another study observed a decrease of the presence of Actinobacteria, in tumor samples (74, 77). Actinomyces spp., known for its ability to safeguard the mucosa by releasing protease inhibitors that hinder tumorigenesis, becomes outnumbered due to the acidic TME and hypoxia (74). Furthermore, high relative abundance of the fungi Malasezzia, were shown to favor survival of OSCC patients (68). These studies indicate that many commensal oral microbiotas of oral cavity have the ability to reduce cancer progression.

The sequencing based analysis may provide more insight into the association of microbiota in different cancer including OSCC (49, 76, 78, 79). Indeed in a sequencing analysis in clinical subjects, Yost et al. showed a significant alteration in the metabolic properties of certain anaerobic bacteria in the development of periodontitis induced dysbiosis (49). Application of next generation sequencing (NGS) technologies such as 16S rRNA sequencing has provided data on oral microbiota dysbiosis in oral cancer (74, 80). The 16s rRNA gene amplicon sequencing techniques has provided greater insight on dysbiosis in various oral cavity diseases (78) including periodontal diseases, OLK, dental caries, gingivitis and OSCC (81). Hooper et al. found heterogeneous viable bacterial population in the cancer cells and TME of primary OSCC using 16S rRNA sequencing (78, 82). Another 16s rRNA sequencing study revealed the presence of saccharolytic and acid-tolerant bacterial populations such as Prevotella melaninogenica, Staphylococcus aureus, Veillonella parvula, Micrococcus inside the tissues derived from tumorous specimens. Concurrently, in the adjacent non-tumor tissues, they found the presence of Moraxella osloensis, Prevotella veroralis, and Actinomyces spp (78). Furthermore, an interesting case control study by Perera et al, evaluated the potential role of the presence of microorganism in deep tumor tissue samples of OSCC patients. The cohort consisted of 25 OSCC cases and 27 fibro epithelial polyp (FEP, FEP is common reactive hyperplasia of oral mucosa) cases as controls from Sri-lanka. The 16S rRNA gene sequencing of V1-V3 region from tissue samples reported the presence of microorganisms such as Fusobacterium nucleatum subsp polymorphum, Pseudomoas aureginosa, and Campylobacter concisus in OSCC cases. The presence of Streptococcus mitis, Rothia spp. and Lautropia mirabilis was observed in the FEP cases. These bacteria showed pathways such as isoleucine biosynthesis, glycolysis/glucogenesis as well as base excision repair pathways (83). The functional prediction analysis in the OSCC cases reported inflammatory molecule lipopolysaccharides (LPS; bacterial endotoxin secreted by anaerobic bacteria) biosynthesis pathway and energy metabolism pathways. A single cell-based RNA sequencing method named INVADE seq used to study the microbial populations in OSCC clinical subjects (n=7) reported intratumoral microbial heterogeneity (84). Bacterial species were found to reside in specific intratumoral niche as well as Fusobacterium and Treponema sp. were found to modulate OSCC progression (84). A machine learning based study reported the presence of bacteria genera Prevotella, Stomatobaculum, Bifidobacterium in OSCC clinical subjects (n=54) which positively correlated with lymph node mediated metastasis aiding in OSCC (85). Furthermore, in clinical subjects (n=41), using a bioinformatic approach Arthur et al. reported that the OSCC associated microbiome comprised of Fusobacterium spp., Prevotella spp., Haemophilus spp., and Campylobacter spp. were in different proportion at different stages of malignancy than in control subjects (37). These sequencing data provides us with new opportunities to study the host-pathogen interaction between microbiota and oral cancer cells.

The molecular mechanism of bacterial induced OSCC initiation and progression may be due to the inflammation medicated host defense response. Using in-vitro assay of specific bacteria and OSCC cell lines, specific oral bacteria populations such as PG and FN have been identified to induce inflammatory cytokines along with cellular invasion of OSCC (86). Ha et al. reported that invasive ability of PG bacteria during chronic infection of Ca9-22 OSCC cells was mediated by cytokine IL-8 (77). Chronic infection of the OSCC cells by the PG bacteria up-regulated the expression matrixmetalloproteins (MMPs) such as MMP 1 and MMP 10 via IL-8. Importantly when the cytokine IL-8 was inhibited by specific siRNA, the expression of the MMPs were down-regulated and the invasive ability of the OSCC cells was also reduced (77). In another in-vitro study, Harrandah et al. reported increased secretion of cytokine IL-8, MMP-1, and MMP 9 in OQ01 cells (primary head and neck cancer cell line) infected by live FN bacteria. The increased secretion of IL-8 promoted cancer cell invasion. Interestingly, the secretion of IL-8 was mediated by LPS as inhibition of LPS by polymyxin B treatment reduced the secretion of IL-8 in the bacterial culture supernatant ( Figure 1 ) (40). More in-vitro studies using bacterial populations isolated from OSCC lesions, as well as the use of primary OSCC cells may provide valuable information about the host-pathogen interaction between oral bacteria and OSCCs. Importantly, such studies need to investigate the potential internalization of bacteria into the cancer cells, as clinical studies done in breast and colorectal cancer showed the presence of intracellular bacteria in cancer cells obtained from primary tumors (87). The presence of intracellular bacteria in OSCC has not yet been studied well. Using an in vitro model of the SCC-25 cancer cell line, we showed that FN present in the saliva of oral cancer subjects internalize to the ABCG2+/EPCAM+ population of SCC-25 cells (88, 89). We also recovered FN from the primary ABCG2+/EPCAM+ population of relapsed OSCC (90). However, it is not clear how a few intracellular bacteria may induce stemness and OSCC progression.

Bacterial secretory products may activate inflammatory pathways such as TLRs, which may contribute to chronic inflammatory state of TME. Lipopolysaccharide (LPS), an endotoxin, is a key bacterial product that is secreted from the outer membrane of gram-negative bacteria (91). The LPS of oral bacteria FN-activated inflammatory cytokines IL-6, and TNF-α may affect development of cancer by influencing apoptotic pathways. Moreover, the LPS of oral bacteria PG can elicit TLR 4 response that prevents apoptosis and promote tumor cell proliferation as well as invasiveness ( Table 1 ) (98). Activation of TLRs in turn can activate NF-κB signalling in the TME of oral cancer and sustain a chronic inflammatory state (93), which may promote cancer progression. However, TLRs are also known as double edged sword in either favoring cancer progression (99), or inhibit cancer progression (100, 101). TLRs present in oral mucosa (102) can recognize the pathogenic microbiota and can trigger an immune response through the direct recognition of ligands derived from the microbiota (103, 104). Therefore, TLRs may recognize the pathogenic microbiota that induces cancer progression and activate an anti-tumor effect via eliminating the microbiota by triggering immune response. In an in vitro setting, S mitis bacterial lysate was reported to inhibit proliferation of three OSCC cell lines; Cal 27, SCC 25 and SCC 9 by up regulating the expression of the pro inflammatory cytokines IL 6 and TNF-α (76). These studies indicated that the up regulation of the pro inflammatory cytokines also have anti-cancer effect in OSCC.

To further evaluate the role of microbiota in OSCC initiation and progression, mouse models are required. Among several mouse models of OSCC, 4 NQO (4-Nitroquinoline 1-oxide) induced carcinogenesis mouse model is well characterized (95, 105). In this 4NQO mouse model, Gallimidi et al. demonstrated that co-infection by PG/FN promoted tumourigenesis through inflammatory response of TLRs (86). The new murine model comprising of FN and PG induced chronic periodontitis were subjected to 4-NQO exposure. It was found that the presence of the FN and PG microorganisms significantly up regulated the IL6-STAT-3 signaling axis, aggravating the OSCC progression. To corroborate the in-vivo study results, they carried out an in-vitro study using an assay of FN and PG co-infection of OSCC cells and found the enhanced activation of TLRs leading to high expression of IL-6. Importantly, the co-infection significantly changes the proliferation rate and characteristics of the OSCC cells (86). In another in-vivo study of 4-NQO induced carcinogenesis, transfer of oral bacteria from tumor bearing mice into germ free recipients significantly increased the numbers and sizes of tumors (46). While these mouse models are important to understand the complexity of bacterial and host cell interaction leading to carcinogenesis and tumor progression, their clinical relevance for human cancer is not yet clear. Future studies need to be conducted using humanized mouse model of oral cancer.

Stemness is the key attribute of stem cells characterized by self-renewal and undifferentiated state (32, 107). It is an integrated molecular program comprising of signalling pathways, epigenetic and gene expression networks that sustains the self-renewal and undifferentiated state of stem cells (32, 107). Epigenetic mechanisms such as DNA methylation, histone modifications and non-coding RNA play a key role in the stemness of both normal and cancer stem cells (108, 109). These epigenetic mechanisms regulate transcription factors involved in stemness. The OCT 4, NANOG, SOX 2 are the transcription factors or genes that maintain stemness in embryonic stem cell (110) and may play a key role in cancer stemness too (22). Stemness in differentiated cells can also be induced by these transcription factors in combination with MYC, and KLF4 (110). Interestingly, stemness may be induced in differentiated cells and also in non CSCs by up-regulation of the stemness transcription factors (111). In hematopoietic stem cells (HSCs), stemness is regulated by development pathways such as Wnt and NOTCH pathways, as well as VEGF/HIF-1 autocrine pathways (112). These pathways regulate the integrated gene expression networks that maintain the stemness attractor state in the epigenetic landscape (112). Epithelial cells may enhance stemness to undergo epithelial mesenchymal transition (EMT), an important mechanism in developmental biology. In the EMT process, transcription factors play key roles, and these transcription factors may also contribute to the balance between stemness and differentiation (32).

Cancer stemness has been a key feature of CSC self-renewal to sustain tumor growth and progression (113). We have shown that upregulation of stemness genes in CSCs are maintained by a MYC-dependent HIF 2α stemness pathway ( Figure 1 ) that transiently suppresses the expression of p53 tumor suppressor gene (31). Inflammatory mediators are also reported to up regulate stemness genes in several cancer types (114–116). Toll-like receptors (TLRs) and nuclear factor-κB (NF-κB) ( Figure 1 ) are two key inflammatory mediators that contribute in cancer initiation and progression. Importantly, TLR3 and TLR4 are highly expressed in OSCC patients (21). Activation of these TLRs stimulates the expression of HIF1-α via NF-κB and thus, maintain the cancer stemness in OSCCs ( Figure 1 ) (21). In breast cancer cells, TLR3 trigger the co-activation of β catenin and NF-κB, which further enhances the expression of OCT3/4, NANOG, and SOX2 stemness genes (115). In hepatocellular carcinoma derived CSCs, LPS based inflammatory mediators were shown to enhance the expression of stemness genes OCT4 and NANOG via IGF-IR signalling pathway (114). CSCs enhance proliferation of the non CSCs in the tumor niche by inducing the expression of the multipotent stemness genes (117). Recently, Shin et al. reported that enhanced expression of OCT 4 in CSCs increases the tumorigenic potential of OSCC (25).EMT, the process of conversion of epithelial cells into mesenchymal phenotype may also induce cancer stemness (118). Epigenetic mechanisms are also involved in stemness and cancer stemness. A study reported that methylation of DNA in the NANOG promoter may induce a switch of non-CSCs to CSCs (119). Inflammatory pathways may promote abnormal DNA methylation leading to carcinogenesis and tumor progression (108, 109). Several histone modification enzymes such as KDM5B and KDM1A are key regulators of stemness and cancer stemness (120). Non-coding RNAs are also involved in phenotypic plasticity and cancer stemness (121). The role of these epigenetic modifiers in the head and cancer has been studied (122). Particularly, p16INK4a is the most hypermethylated gene in oral cancer (123), and this gene maybe a target of inflammation-induced epigenetic modification in cancer, as shown recently (108).

Many studies have shown the role of microbiota in the stemness of various cancer types as outlined in the Table 1 . However, only a few studies are reported on OSCC. A positive correlation was reported about the presence of Bacteroides fragilis and FN level with the expression of OCT-4, NANOG and SOX2 in colorectal cancer patients (94, 96) ( Table 1 ). Microbiota may also induce stemness by regulating the mechanisms of EMT. A study has shown that prolonged and repetitive infection with PG promotes EMT-like changes and stemness in OSCC cells (77). Another study, demonstrated that infection by FN induced high expression of oncogenes STAT 3, MYC, JAK 1 in OQ01 cells (primary head and neck cancer cell line) (40). HPV+ve oropharyngeal cancers (OPSCCs) is enriched in CSCs (124–127). Expression of CSC markers CD44 (128) and ALDH1 (126) were found to be higher in the HPV positive oral cancer patients compared to HPV negative oral cancer patients (126, 128). Hepatitis B virus (HBV) encoded X antigen (HBVx) was reported to activate stemness associated factors OCT-4, NANOG, β-catenin, KLF 4, and EpCAM as well as induce cell migration and sphere formation in hepatocellular carcinomas (HCC) (92).

The fungi has both cancer development and anti-cancer properties (129). Candida has been extensively reported for cancer progression (130), whereas, Ganoderma lucidum polysaccharides reduce cancer stemness by inhibiting EMT (131). These studies indicate that bacteria, virus as well as fungi may induce cancer stemness in OSCC.

The molecular mechanisms of microbiota-induced cancer stemness are not yet clear. Cancer stemness may be induced by bacterial metabolites such as LPS and Lipooligosaccharide (LOS) (91, 97). Oral bacterium, Trepenoma denticola has been reported to induce stemness, by enhancing cellular migration and tumorosphere formation. This contributes to an aggressive OSCC phenotype via LOS mediated TLR4/MyD88 and integrin/FAK signalling pathways ( Table 1 ) (97). LPS is known to induce immune response, and many studies have shown that some bacteria, such as P. multocida are dependent on LPS for infection (132). Several inflammatory cytokines IL1β and TNFα were reported to be stimulated by LPS of oral microbiota (133). Interestingly, different inflammatory cytokines were reported to upregulate the MMPs production which triggers cancer cell invasion (134). LPS meditated TLR 4 expression induces EMT (135) and stemness (90) thereby aiding in tumour progression and metastasis (135).

These findings indicate that oral bacteria secreting LPS may induce stemness in oral cancer cells mainly by activating inflammatory pathways. Importantly Ha et al. showed that PG invaded OSCC cells led to increased secretion of IL-8 and VEGF as well as increased stemness and expression of CD44 (77) ( Table 1 ). Microbe induced TLR expression in oral cancer cells may increase the transcription of NF-κB that in turn promote cancer stemness ( Figure 1 ). HPV was reported to promote tumor growth by enhancing cancer stemness via miR-181a/d regulation in OPSCC cells (136). Thus, oral microbiota may employ various mechanisms to modulate stemness in the oral cancer cells.

Microbial interaction with stem cell niche defense may promote carcinogenesis. The ASC phenotype exhibits transient suppression of the p53 tumor suppression gene, but exhibits very high transcriptional activities of NANOG, SOX2 and OCT4. The high expression of these factors may provide a self-sufficiency state by yet unknown mechanisms (110). Therefore, the ASC phenotype could be the target for malignant transformation. Interestingly, in a 4-NQO model of oral mucosa stem cell reprogramming to ASCs ( Table 1 ) (141), the addition of FN or HPV-16 markedly increased the malignant transformation of oral mucosa stem cell-derived ASCs to CSCs (141). Notably, we identified FN together with HPV 16 has the ability to reprogram human primary CD271+ oral keratinocyte progenitor into ASC state (89) ( Table 1 ). Thus, the microbial interaction with the ASC mediated stem cell niche defense may play important role in the progression of pre-cancerous lesions to malignant growth.

This unique altruistic defense mechanism of stem cell niche may be hijacked by CSCs to defend their niche (138). CSCs may reprogram MSCs to ASC phenotype to modulate a cytoprotective and immunosuppressive TME (142). CSCs may make transition to a higher state of stemness to defend the niche against stress such as oxidative stress, chemotherapy, as well as pathogen invasion (138). We showed that CSCs might reprogram to a transient but highly inflammatory stemness state when exposed to extreme hypoxia/oxidative stress (22), and chemotherapy (111); the phenotype could protect CSCs and non-CSCs in the niche from oxidative stress (111) as well as pathogen invasion (23). Therefore, we termed the phenotype as tumor stemness defense (TSD) phenotype (23). Our ongoing work suggests that the TSD phenotype can be isolated from primary OSCCs obtained from stage III and stage IV locally advanced cases (23) suggesting potential clinical implications of the TSD phenotype. Moreover, we found that while FN can induce TSD phenotype in SCC-25 derived CSCs and transform non CSCs to CSCs ( Figure 2 ) (88, 95), Myocobacteria can induce apoptosis of the TSD phenotype (23). Thus, bacteria may either exert pro or anti-tumor effects based on their effect on the CSC niche defense. Interestingly, a retrospective analysis of OSCC subjects revealed less frequent relapse in the FN positive subjects possibly due to enhanced immunogenicity by FN (143). Future studies are required to investigate the microbes/CSC niche defense interaction to understand microbe’s contribution in cancer stemness or cancer reduction. Potential future studies include microbial interaction with stemness pathways that switch non-CSCs to CSCs (117), epigenetic pathways that promote stemness (144), MYC-HIF-2α stemness pathways that reprogram CSC to TSD phenotype (22, 88) and embryonic stemness related transcription factors that promote OSCC progression (25). In this context, we have developed an in vitro model of FN and OSCC derived CSC host/pathogen interaction to study the microbiota-induced TSD phenotype ( Figure 2 ).

Following the rapid approval of a few immune check points’ inhibitors (ICI) as immunotherapy agents for melanoma, lung cancer and renal cancer, there is a growing interest to study the mechanisms of immunosuppression in other tumor types, where ICI has failed to show satisfactory clinical response. ICI based immunotherapy effectiveness in head and neck squamous cell carcinoma is very poor due to immunosuppressive TME that hinders T cell infiltration (145). Here we shall discuss the possible ways that the stem cell niche defense and microbes contribute to immunosuppressive TME in OSCC. A growing body of research suggests the pathological communication between CSCs and immune cells that shape the immunosuppressive TME in diverse tumor types (146–149). Briefly CSCs are intrinsically immunosuppressive (149, 150) and reside in the immunopriviledge niche of hypoxia (32). Several stemness pathways active in CSCs such as TGFβ, Wnt/β-catenin and NOTCH pathways inform immune cells for immune evasion (146–149, 151). In this context, only a few studies have been done on the immunosuppressive properties of HNSCC and OSCC derived CSCs. One study found that CD44+ cells of OSCC cross-talk with macrophages to promote stemness activities (152) while another study reported STAT3 mediated expression of PDL-1 expression in CD44+ CSC leading to suppression of T cell mediated tumor immune response in OSCC. We found that OSCC derived TSD phenotype of CSCs shape immunosuppressive TME by secreting immunosuppressive cytokines and also reprogramming mesenchymal stem cells and innate immune cells (142). Although the physiological basis of CSC-immune cell crosstalk is not yet clear it is possible that CSCs hijack the reciprocal communication between adult stem cell and immune cell in the stem cell niche. Both hematopoietic and mesenchymal stem cells interact with the immune cells to maintain stem cell homeostasis (153). The crosstalk involves stemness pathways and therefore it is not surprising that the CSC niche defense hijacks the stemness pathways such as TGF β and Wnt- β-catenin to evade tumor immune response. However, it is not yet clear if and how microbes define CSC and immune cell crosstalk. CSC may adopt intrinsic mechanisms of innate immune cells to defend the niche against microbial invasion. We demonstrated that M tuberculosis and BCG infected CSC activate the innate immune defense of bystander apoptosis to eliminate intracellular microbes. This mechanism of CSC niche defense can be exploited to eliminate the TSD phenotype of CSC obtained from diverse tumors (23). CSC may also reprogram microbes to enhance tumor stemness and immunosuppressive TME ( Figure 2 ). These reprogrammed intracellular microbes may then modulate crosstalk between CSC and immune cells and aid in resistance against immunotherapy as well as chemotherapies (106). Indeed, cisplatin treated oral cancer CSCs (23) exhibit NK cell suppressive activity in the presence of Fusobacterium nucleatum (89). Then chemotherapy may also modify the crosstalk among the microbes, CSC, and immune cells in the niche. Understanding the molecular and cellular mechanism of the crosstalk between CSC niche defense and microbes may better inform the reciprocal communication between CSCs and immune cells and consequently develop new therapeutic avenues.