Role of Inflammatory Markers in Prognosis of Oral Squamous Cell Carcinoma.

Background: This estudie evaluated the immunostaining of cytokines in oral carcinoma, in tissue of margin of surgical resecate (MSR) and metastatic lymph nodes, as well as their role in patient prognosis. Methods: A retrospective study was carried out in patients with oral squamous cell carcinomas, and sociodemographic and clinical-pathological data were evaluated. In addition, surgical site analysis of the patients was conducted by immunohistochemistry, using a tissue microarray for inflammatory (Tumor Necrosis Factor-alpha, Interleukin-1beta, Interleukin-6, interleukin-10), transcription NF-kappa B and CD68 markers. Immunoexpression was assessed qualitatively and quantitatively using ImageJ software, and data were correlated with the prognostic factors and patient survival rates. Results: There was a greater immunoexpression of inflammatory and CD68 cytokines in primary tumour and lymph node metastasis than in MSR. In a multinomial logistic regression model, patients with low education (p = 0.041) and a high histoscore for TNF-α (p = 0.021) showed a survival rate of 15.64 (95% CI = 1.13-217.24) and 6.81 (95% CI = 1.02-105.96). Conclusion: Therefore, despite there is an increased immunoexpression of cytokines in the primary tumour, only TNF-α was the inflammatory cytokine that influenced the survival of patients with oral cancer.


Introduction
Oral squamous cell carcinoma (OSCC) is most prevalent in the mouth and represents approximately 95% of mouth malignancies and is the 12 more prevalent in the world (Brasil, 2018).
OSCC is the result of chronic exposure to carcinogens and evolves from a series of genetic and epigenetic alterations (Durazzo et al., 2005). In this context, inflammation has been shown to play an important role in the fields of malignization, tumour growth, local invasion, angiogenesis and metastasis. Although most of these mediators are produced to activate the immune system against the tumour, they can stimulate tumour growth (Balkwill and Mantovani, 2001).
Tumour necrosis factor alpha (TNF-α) and numerous cytokines, such as interleukin 1 beta (IL-1β), 6 (IL-6) and 10 (IL-10), are associated with the activation of the cell cycle, leading to tumour progression and metastasis development (Eirón et al., 2012). The interaction between the stroma and the tumour parenchyma is essential for Sample Size Calculation Piva et al., (2013) showed that cases of OSCC that were the most poorly differentiated have a high expression of TNF-α in parenchyma cells (45.8%), compared with cases of OSCC that were well differentiated (18.2%). Based on the results of that study, we deemed it necessary to evaluate 66 patients in order to be able to reject the null hypothesis with 95% confidence and 90% power (Fleiss method with continuity correction). To account for the possibility of sample loss, the sample size was increased by 15%, bringing the total to 76 patients.

TMA and immunohistochemistry
The histologic slides were reviewed to select "host pots" with high cellularity, and we selected an area of the surgical site with a representative tumour microenvironment. For patients with lymph node metastasis, a representative tumour microenvironment in the sentinel lymph node was selected and a margin of surgical resecate (MSR) was too selected (Liu et al., 2016).
We used these areas (diameter: 2 mm) to make a TMA block using a Tissue Microarrayer (Quick-Ray UNITMA ® ). Then, microscopic slides (4 μm) were prepared conventional haematoxylin-eosin method and for immunohistochemistry.
After deparaffinization and rehydration, the tissue sections were submitted for immunohistochemical tecnique whit antigenic recovery was performed by heating in a citrate solution at pH 6.0, to peroxidase blocking with a 3% H 2 O 2 solution diluted in Phosphate-Buffered Saline.

Immunohistochemistry evaluation
Five microscopic fields (400x) were photographed (Leica DM 2000 ® ). The images were exported to ImageJ ® (cell counter command) to count the CD68 immunostained mononuclear cells and tumour cells exhibiting cytoplasmic or nuclear (for NK-kB) positivity for inflammatory markers.

Statistical approach
Data were analysed using statistical software SPSS (Statistical Package for Social Sciences) for Windows version 22.0, with a confidence level of 95% (P <0.05). The categorical data were analysed using a Chi-square test. The 5-year survival rate was also shown and analysed, and the mean and standard deviation were calculated and analysed using the Log-Rank Mantel-Cox test. The histoscores were shown as the standard deviation and compared using a Mann-Whitney or Kruskal-Wallis/Dunn test. A forward stepwise model was used to select variables that showed p<0.200 with the death rates for multinomial logistic regression. These results are shown as adjusted to hazard risk and a 95% confidence interval.

Ethical Correlations
This study conformed to ethical principles and was accepted by the ethics committee of the Hospital Haroldo Juaçaba under protocol 1.552.674

Clinical-pathological and socio-demographic data and 5-year survival of OSCC
Of the 76 patients evaluated in this study, the majority were male (n=50, 65.8%), were more than 60 years old (n=43, 56.6%) and had tumours in the floor of mouth(n=35, 46.1%). The 5-year survival rate was 86.6% (n=66) and the survival mean was 49.11±25.13 months; these rates were not influenced by sex, age or localization of the tumour (Table 1).
The stage that was most prevalent was III/IV (n=60, 78.9%), the histologic gradation most prevalent was OSCC moderately differentiated (n=71, 93.4%), and the therapeutic approach most used was surgery and radiotherapy or surgery, radiotherapy and chemotherapy (n=32, 43.2%). The stage and histologic gradation did not influence survival, but the cases treated only with surgery showed superior survival compared to those subjected to combination treatments (p=0.010) ( Table 1).
Of the 76 patients evaluated in this study, the majority of the patients were illiterate or had incomplete primary schooling (n=39, 51.3%), were mixed race (n=61, 80.3%), were from rural areas of the state (n=51, 67.1%), were married (n=55, 72.4%) and were smokers (n=30, 39.5%). Only education level significantly influenced survival; patients who were illiterate or who had incomplete primary schooling showed lower survival rates (p=0.014) ( Table 1).

Immunohistochemistry profile for inflammatory markers in OSCC, lymph node metastasis and MSR squamous epithelium and 5-years survival rates
The CD68 mononuclear positive cells were significantly higher in tumour than in MSR squamous epithelium connective tissue and lymph node metastasis. The lymph Inflammatory Markers in Prognosis of Oral Carcinoma  node metastasis samples had significantly more CD68 mononuclear positive cells than did the connective tissue of MSR squamous epithelium (p<0.001) (Figure 2). The NF-kB immunostaining was significantly higher in tumour and lymph node metastasis than in MSR squamous epithelium (p<0.001), and the TNF-α immunostaining was significantly higher in tumour than in lymph node metastasis and MSR squamous epithelium (p<0.001). IL-1β, IL-6 and IL-10 immunostaining was significantly higher in tumour and lymph node metastasis than in MSR squamous epithelium (p<0.001) (Figure 2). NF-kB, IL-1β, IL-6 or IL-10 immunostaining and CD68 mononuclear positive cells in living patients did not differ in MSR squamous epithelium, tumour or lymph node metastasis when compared to that in deceased patients. However, deceased patients showed a higher TNF-α immunostaining in tumour (p=0.032) and lymph node metastasis (p=0.009) compared to living patients (Table 2).
Lower TNF-α immunostaining in patients was not associated with stage and education level (p=1.000). However, in the higher TNF-α immunostaining patients, the III/IV stages were most prevalent in those with low education level (p=0.033, HR = 5.61,   IC95% = 1.21 -25.76). Additionally, the patients with higher TNF-α immunostaining and lower education levels were 3.82 (IC95% = 1.24 -11.8) ( Table 3) and and most were treated with surgery, radiotherapy and chemotherapy, rather than with surgery alone (p=0.026) ( Table 4).

Discussion
This study evaluate the immunostaining of cytokines in OSCC, in tissue of margin of surgical resection (MSR) and metastatic lymph nodes and the role of these cytokines in patient prognosis and survival. The clinical data showed a high prevalence of men who were less than 60 years old, an unusual age in epidemiological studies (Heo et al., 2012;Odveig et al., 2014;Liu et al., 2016). As previously described, the most prevalent tumour localization was in the tongue and floor of mouth (Odveig et al., 2014;Satochi et al., 2014), the most frequent stages were III/IV (Durazzo et al., 2005) and the main risk factor was smoking (Iamaroon et al., 2004).
The inflammatory process is vital for mammals, acting as a protector against infectious agents, traumatic injuries and cancer development. The recognition of lesions made by microorganisms, trauma or disorderly tissue growth leads to the production of many inflammatory mediators, such as cytokines (Landskron et al., 2014   However, in chronic processes, a pathologically favourable microenvironment is created, promoting the development of a series of diseases and contributing to tumour progression (Iamaroon et al., 2004). In addition to clinical-epidemiological factors, we have shown the profile of the immunoexpression of inflammatory cytokines in the tumour, perilesion and metastasis in the lymph node and have observed that, in general, the immunoexpression of these cytokines is greater in the primary tumour or in lymph node metastasis than in MSR. A similar result has been shown in the literature, although the role of these cytokines has not yet been fully elucidated (Zhang et al., 2014).
In our study, we did not observe a change in the immunoexpression profile of inflammatory cytokines from the primary tumour to lymph node metastasis, showing that after malignization, the immunoexpression of inflammatory mediators increases in relation to MSR and remains at a similar level when there is metastasis to the lymph node. Studies have shown that proinflammatory cytokines such as IL-1 beta, TNF-alpha, and IL-6, produced by the malignant cells themselves, are associated with increased keratinocyte motility, and consequently nodal metastasis (Voiculescu, 2016).
IL-10, an anti-inflammatory cytokine, has been correlated with antitumour activity in some studies6, not only in squamous cell carcinoma but also in other tumours, such as breast and ovarian carcinomas (Ranelletti et al., 2001). In our study, IL-10 was not associated with a better prognosis, although its immunoexpression is prevalent in tumour tissue and metastasis in the lymph node, as previously described (Arantes et al., 2016).
In addition to the characterization of cytokines, infiltration of immunological cells has been used to evaluate the prognosis of malignancies . Our study shows CD68-positive cells, a marker that has been used as a pan-macrophage marker (Arun et al. 2009), to be more present in primary tumour and metastasis in lymph nodes than in MSR, as previously described (Madrek et al., 2012). Although this is not associated with better or worse prognosis, a similar result was found in the study by Fang et al., (2017). Perhaps phenotypic macrophage differentiation, M1 or M2, is necessary to better evaluate the role of cytokines in the progression of squamous cell carcinoma (Alvez, 2018).
These clinical data are complemented by a new approach. We have shown that the immune profile of some pro-inflammatory cytokines of innate response is increased in the tumour or lymph node metastasis, compared to MSR squamous epithelium, which was also demonstrated (Balkwill and Mantovani, 2001). Despite the role of these mediators remaining unclear, the augmentation of expression of inflammatory markers has been shown in the literature (Zang, 2014). Balkwill and Mantovani (2001) demonstrated that inflammatory cells and cytokines in peritumour stroma contribute more to the progression and development of the tumour than to the host immune response and that macrophage activation can lead to more DNA damage, anti-neoplastic gene suppression, angiogenesis and local destruction (Landskron et al., 2014).
In support of these data, in our study, the patients with high TNF-α demonstrated lower survival. TNF-α is involved in both carcinogenesis and tumour progression, assisting in angiogenesis and tumour invasion (Eirón et al., 2012). In cell culture of OSCC, TNF-α reduces cell viability (Ni et al., 2015), and many chemotherapies lead to cell death by activation of TNF-α receptors mediated by Bcl-2 and caspase-independent pathways (Chan et al., 2012). TNF-α receptor activation can activate cell death leading to apoptosis, NF-kB induction, and overproduction of IL-1β, IL-6 and other chemokines (Popa et al., 2007).
However, this inflammatory microenvironment leads to macrophage migration and production of oxygen-reactive species, leading to metastasis in lymph nodes (León et al., 2015). TNF-α autocrine activation is associated with tumour growth (Landskron et al., 2014) through mechanisms such as increased cell proliferation that are dependent on NF-κB activation (Tang et al., 2017); reduction of apoptosis and inactivation tumour suppressor genes (Zhang et al., 2014); increased synthesis of matrix metalloproteinases, which is closely related to the local invasion and the distance of malignant cells (Tang et al., 2017); induction of metastasis and angiogenesis (Zhou et al., 2015); and increased expression in endothelial cells of proteins that mediate the adhesion of VCAM-1 (Song et al., 2012).
This tumour growth results in a greater tumour staging that is linked to worse prognosis and higher mortality (Iamaroon et al., 2004). A previous study conducted by this group of researchers (Dantas et al., 2016) also observed that schooling directly influences patient survival. Oral cancer is related to low education, which may be partly due to poor access to disease information in general, including the diagnosis and treatment of oral cancer (Atununes et al., 2001).
It was also observed, similar to the present study, that patients with low education and high immunoexpression for TNF-α showed a worse survival rate, in addition to greater staging and use of surgery associated with radiotherapy and chemotherapy as a treatment of choice (Krishnan et al., 2014). In previous studies, patients with stage III and IV tumours expressed more TNF-α; in the researchers' country, patients with advanced disease stage are patients with a low level of education (Ferreira et al., 2012;Dantas et al., 2016).
In conclusion, this work shows that there is an increased immunoexpression of inflammatory cytokines and macrophages in the primary tumour in relation to the MSR, but without any major changes in metastasis to the lymph nodes. We have also shown that TNF-α, as the only inflammatory cytokine that directly influences the survival of patients with oral cancer, should be considered for further studies, possibly proving to be a predictor of prognosis of squamous cell carcinoma. This study, along with further studies, may also verify the social profile behind the worst prognosis of this neoplasia. Therefore, taking into consideration that patients with a low level of schooling show lower survival, this study shows that there is a fundamental need for public health policy intervention in populations with this profile. In terms of the limitations of this work, we cannot demonstrate by which mechanisms TNF-α acts on the progression of the tumour and the consequent increase in the mortality of these patients.