CD163+ macrophages infiltration correlates with the immunosuppressive cytokine interleukin 10 expression in tongue leukoplakia

Abstract Objective Accumulating evidence suggests that macrophages are involved in the immunoediting of oral squamous cell carcinoma but the role of macrophages in oral carcinogenesis is unclear. We aimed to clarify the role of macrophages in oral leukoplakia, which is the most common oral potentially malignant disorder from immunotolerance viewpoint. Materials and methods The study included 24 patients who underwent surgical resection for tongue leukoplakia. The relationships between macrophage markers and clinicopathological factors were assessed. Conditioned medium was harvested from the CD163+ human monocytic leukaemia cell line, THP‐1. The phenotypic alteration of human oral keratinocytes by the conditioned medium treatment was assessed using quantitative reverse transcription‐polymerase chain reaction and enzyme‐linked immunosorbent assay. Moreover, the clinical samples were evaluated using immunohistochemistry. Results Tongue leukoplakia tissues with high CD163+ macrophage infiltration were associated with significantly higher degrees of epithelial dysplasia, abnormal Ki‐67 expression and cytokeratin13 loss when compared with the tissues with low CD163+ macrophage infiltration. In vitro, CD163+ THP‐1 conditioned medium induced immunosuppressive molecules, especially interleukin‐10 (IL‐10) in human oral keratinocytes. The IL‐10 expression levels showed significant positive correlations with not only the numbers of FOXP3+ regulatory T cells but also that of CD163+ macrophages. Conclusions In tongue leukoplakia, CD163+ macrophages infiltration correlates with immunosuppressive cytokine IL‐10 expression.


| INTRODUCTION
The oral cavity plays an essential role in eating, speaking, swallowing, and facial aesthetics (Matsuhira et al., 2015). Oral squamous cell carcinoma (OSCC) is a solid tumour of epithelial origin that has been reported to affect approximately 400,000 people annually worldwide (Petersen, 2009). The mortality rate of OSCC has remained largely unchanged for the last several decades, with a 5-year survival rate <50% (Petruzzi, Cherubini, Salum, & de Figueiredo, 2017). The early detection of OSCC is thus very important for high patient quality of life.
It was reported that most of OSCCs pass through a premalignant stage (Liu et al., 2017). Oral leukoplakia (OL) is the most common oral potentially malignant disorder. OL occurs most frequently in the tongue. OL is defined as a white plaque of questionable risk having excluded other diseases or disorders that carry no increased risk for factor (Woo, 2019). OL is only a clinical term, and its presentation including clinical appearance, colour, and surface type can vary (Irani, 2016;Shafer & Waldron, 1961). It has been reported that approximately 3.5% (0.13-34.0%) of OLs develop into OSCC (Warnakulasuriya & Ariyawardana, 2016). Previous reports have suggested the proliferation marker Ki-67 and the squamous celldifferentiation markers cytokeratin (CK) 13 and CK17 may be the factors for biological malignancy of oral atypical epithelium (Kitamura et al., 2012;Kövesi & Szende, 2003;Mikami et al., 2015). For example, immunohistochemical Ki-67 expression have been sporadically confirmed in the second basal layer of normal epithelium, but many Ki-67 + cells are distributed to the basal and/or more superficial layers of oral precancerous lesions. Additionally, CK13 is expressed in normal epithelium but not precancerous lesions, whereas CK17 is expressed in precancerous lesions but not in normal epithelium (Yagyuu et al., 2015). However, the mechanism of OL development has not been completely elucidated to date. Thus, a better understanding of the underlying molecular mechanisms of oral carcinogenesis is necessary.
On the basis of this background, to clarify the specific roles of MΦs in oral carcinogenesis, we conducted immunohistochemical analyses with surgically resected tongue leukoplakia (TL) samples and in vitro assays with THP-1 human monocytic leukaemia cells and human oral keratinocytes (HOKs) from the viewpoint of immunotolerance.

| Tissue samples
A total of 24 cases of surgically resected TL treated at the Department of Oral and Maxillofacial Surgery, Kobe University Hospital, Japan were included. The patients were nine men and 15 women with an age range of 31-88 years and mean age of 63.2 years. None of the patients received adjuvant chemotherapy or radiotherapy before surgery. All resected specimens were fixed in 10% formalin and embedded in paraffin. Informed consent for their materials and data to be used was obtained from all patients, and the study was approved by the Kobe University Institutional Review Board.

| Morphological evaluation
Three pathologists (M. S., Y. K. and H. Y.), blinded to the patients' clinical data of the patients, performed the grading of epithelial dysplasia based on the modified squamous intraepithelial neoplasia system (Yagyuu et al., 2015;Yagyuu et al., 2017). Briefly, the degree of dysplasia was divided into high (moderate/severe dysplasia or carcinoma in situ) and low (no/mild dysplasia) grades. In this study, not only the basaloid-type but also the differentiated-type carcinoma in situ were classified as high grade.

| Immunohistochemical evaluation
We used a modified version of the immunoglobulin enzyme bridge technique with the Linked Streptavidin-Biotin Kit (DakoCytomation, Glostrup, Denmark) as described elsewhere (Shigeoka et al., 2013).  (Yagyuu et al., 2015). Briefly, the Ki-67 expression was evaluated as a second basal layer or an unclear or basal layer and/or more superficial layer. The CK13 and CK17 expressions were evaluated as follows: positive, loss, or unclear. MΦs were each counted in subepithelial areas up to 100 μm from the basement membrane. CD163 + , CD204 + , and CD206 + round cells were counted as MΦs. Three high-power fields (×400) were randomly selected, and the mean number was calculated (Sato et al., 2005;Yagyuu et al., 2017;Zhang et al., 2003). The median MΦ number in subepithelial areas was used to divide the patients into high and low groups. IL-10 immunoreactivity of 24 TL tissue samples was divided into high and low immunoreactivity in comparison with that of corresponding normal oral epithelium. Three pathologists (M. S., M. N., and H. Y.) who were blinded to the clinical data performed these evaluations.

| Cell cultures
HOKs were purchased from ScienCell Research Laboratories Before the medium was changed, to remove TPA sufficiently from the well, we aspirated well and washed using PBS, three times in the present study. After 2 days, the supernatant was harvested, centrifuged, and stored in aliquots at −80 C.

| Enzyme-linked immunosorbent assay
Human IL-10 concentrations were measured by the Quantkine ELISA Human IL-10 Immunoassay (R&D, MN, USA) according to the manufacturer's instructions. The optical density of each well was read at 450 and 540 nm. The concentration of IL-10 was calculated using a standard curve and the measured absorbance. .

| Statistical analysis
We used the χ 2 test to analyse the relationships between the patient's clinicopathological features and the immunohistochemistry results.
Statistical comparisons were performed using the paired t test. All in vitro assay was performed three times independently. A p value < .05 was considered statistically significant. All statistical analyses were carried out using SPSS Statistics Ver. 21 software (IBM, Chicago, IL, USA).

| Macrophage infiltration was observed in TL
MΦs expressing CD163, CD204, or CD206 immunoreactivity were detected in all TL tissues examined. CD163 + MΦs and CD206 + MΦs were distributed in the subepithelial stroma, especially beneath the basement membrane, whereas no CD204 + MΦs were observed.

| The infiltrating number of CD163 + MΦs was closely associated with clinicopathological factors of the patients with TL
We next determined whether CD163 + MΦs and CD206 + MΦs had any statistical associations with the clinicopathological factors of the patients with TL (Table 1)

| CD163 + THP-1 CM induced IL-10 expression in HOKs
On the basis of the immunohistochemical findings, we hypothesised that infiltrating CD163 + MΦs in the subepithelial areas of TL tissues    Table 2).

| DISCUSSION
Several studies have shown that the infiltration of CD163 + MΦs positively correlates with epithelial dysplasia and the malignancy of oral precancerous lesion (Mori, Haraguchi, Hiori, Shimada, & Ohmori, 2015;Stasikowska-Kanicka et al., 2018b). Additionally, previous studies have shown that the number of Tregs and the levels of IL-10 expression increase in oral precancerous lesions (Goncalves et al., 2017;Sun et al., 2016). However, to the best of our knowledge, this is the first study to demonstrate the association of CD163 + MΦs infiltration with IL-10 expression in TL.
The results of our analyses demonstrated that the number of infiltrating CD163 + MΦs but not the numbers of CD206 + MΦs and CD204 + MΦs had significantly positive correlations with the degrees of epithelial dysplasia, abnormal Ki-67 expression, and CK13 loss in TL tissues. These results are in agreement with previous findings that CD163 + MΦs are the major TAMs in OSCC and that a high number of CD163 + MΦs correlates with poor prognosis (Fujii et al., 2012;Wang et al., 2014). It was reported that CD163 + MΦs are distributed not only in the cancer stroma but also within the cancer nest in OSCC (Usami et al., 2013). On the other hand, it has been shown that the majority of infiltrating CD163 + MΦs were distributed in the subepithelial stroma in OL (Mori et al., 2015). In accordance with these reports, we observed herein that CD163 + MΦs were distributed in the subepithelial stroma, especially beneath the basement membrane.
Many reports have shown the importance of a direct interaction between MΦs and cancer cells (Komohara et al., 2013;Komohara, Ohnishi, Kuratsu, & Takeya, 2008;Usami et al., 2013). We speculate that MΦs adhere to cancer cells due to the breakdown of the base- reported that CD163 + TAMs in OL co-express CD163 and STAT1, suggesting that the TAMs in oral premalignant lesions possess an M1 phenotype (Mori et al., 2015). Essa et al. (2016)  We observed that the expressions of immunosuppressive molecules, especially IL-10 in HOKs were induced by CD163 + MΦ-like cells CM. Moreover, in our immunohistochemical analysis, IL-10 was detected in epithelial cells of TL tissues.
It is reported that TAMs induce the infiltration and differentiation of Tregs and that Tregs induce MΦ polarisation into the M2 phenotype (Guan et al., 2007). Sun et al. (2016) reported that IL-10 expression levels gradually increased during the early stages of OL and in OSCC (Sun et al., 2016). Kubota et al. (2017) revealed that TAMs promote T cell regulation via IL-10 and PD-L1 production in OSCC (Kubota et al., 2017). Our present findings also showed that the expression level of IL-10 was significantly positively correlated with the infiltration of Tregs in TL. It is speculated that IL-10 induced by CD163 + MΦs may contribute to the infiltration of Tregs during the development of oral carcinogenesis.
There are three limitations in this study. First, the sample size (n = 24) was small. At our hospital, the first choice of treatment for leukoplakia occurring in parts other than the tongue (including the gingiva or the palate) is vaporisation with a CO 2 laser instead of surgical resection; it is thus difficult to perform several histopathological examinations. Second, we have not verified the effect of IL-10 on Tregs using in vitro assays. Mori et al. (2015) proposed that the infiltration of MΦs and T cells into epithelial lesions may be involved in early morphological changes in the development of dysplasia (Mori et al., 2015). To overcome these limitations, it may be necessary to undertake an accurate study with a large sample size or an in vitro study with not only HOKs and THP-1 cells but also with T cells. Third, we did not analyse the nature of the CD163 + MΦ-like cells CM that induced the IL-10 expression in HOK. We speculated that the humoral factor(s) from CD163 + MΦs is one of the important inducers of IL-10 in TL. It is reported that IL-10 is produced by monocytes or T cells (Del Prete et al., 1993;Konjevic, Vuletic, Mirjacic Martinovic, Larsen, & Jurisic, 2019). There are many reports on the inducer of IL-10. The induction of IL-10 from Tregs is reported to be dependent on TGF-β, but IL-10, IL-2, and IL-4 also help to promote its optimal production (Josefowicz, Lu, & Rudensky, 2012;Ouyang & O'Garra, 2019).
In mouse models of malaria, IL-27 is an important regulator of IL-10 producing Type 1 regulatory T (Tr1) cells (Kumar, Ng, & Engwerda, 2019). Kumar et al. (2019) also showed that Type I interferons (IFNs) are also critical regulators of IL-10 production by Tr1 cells (Kumar et al., 2019). It is showed that the increasing plasma IL-10 levels observed during exercise are mediated by IL-6 from contracting muscles (Steensberg, Fischer, Keller, Moller, & Pedersen, 2003). Comprehensive analysis of CD163 + THP-1 CM using cytokine array or cDNA microarray should be further conducted.
Overall, our results indicated that CD163 + MΦs might play important roles in the development of TL via immunosuppression. Immunological approaches targeting IL-10 could be effective for the establishment of novel therapies of OL.
F I G U R E 4 High expression of IL-10 show a significant positive correlation with the numbers of CD163 + MΦs and regulatory T cells in tongue leukoplakia (TL). (a,b) Immunohistochemical images of IL-10 expressions in normal epithelium of representative TL cases. (c,d) Immunohistochemical images of high and low IL-10 levels in representative TL cases. (e-h) The expression levels of CD163 + cells and FOXP3 + cells are stronger in the IL-10 high group compared with the IL-10 low group (with original magnification: ×400 and scale bars: 20 μm; inset magnification: ×100) T A B L E 2 Expression levels of IL-10 in tongue leukoplakia and the associations with the infiltration of MΦs and regulatory T cells .