CD206+ tumor-associated macrophages promote proliferation and invasion in oral squamous cell carcinoma via EGF production

Tumor-associated macrophages (TAMs) promote tumor progression and inhibit anti-tumor immune response by producing various mediators and preferentially express CD163, CD204, and CD206. However, the role of these TAM subsets in oral squamous cell carcinoma (OSCC) remains unclear. Here we investigated the expression and function of TAM subsets in OSCC, especially in cancer cell proliferation. Biopsy sample from 44 patients with OSCC were examined for the expression of TAM markers and EGF by immunohistochemistry. EGF production of TAM subsets isolated from OSCC patients was assessed by flow cytometry. We also examined the effect of conditioned medium from TAM subsets on the proliferation of OSCC cells. CD163+ cells were detected diffusely all over the tumor and connective tissue area, while CD204+ and CD206+ cells were mainly detected in/around the tumors. Flow cytometric analysis found that CD206+ TAMs strongly produced EGF compared with CD163+ and CD204+ TAMs. Cell proliferation and invasion of OSCC cells cultured with conditioned medium of CD206+ TAMs were strongly enhanced and inhibited by anti-EGFR. The number of CD206+ TAMs positively correlated with worse clinical prognosis. Our results revealed differences in localization and EGF production among these TAM subsets. CD206+ TAMs might play a critical role in the proliferation of OSCC via EGF production.

After taking the biopsy sample, 4% buffered formalin solution was used to fix the specimens and then implanted in paraffin to make blocks. 5-μm thick sections were prepared from the paraffin-embedded specimens, hematoxylin and eosin (H&E) were used to stain the sections to confirm the diagnosis and histologic grade by a professional oral pathologist. Stage of the tumor was classified as stated in the TNM classification of the International Union Against Cancer. WHO classification was used to determine the tumor histologic grade. According to the Yamamoto-Kohama clinical classification of OSCC the pattern of tumor invasion was assessed from the H&E-stained specimens as: grade 1 = well-defined borderline; grade 2 = cords, less-marked borderline; grade 3 = groups of cells, no distinct borderline; and grade 4 = diffuse invasion (4C = cord-like type invasion; 4D = widespread type invasion). Patients type and the clinicopathological characteristics of tumor are shown in Table 1. immunohistochemical analysis. First, we deparaffinized the sections in xylene and then hydrated by graded series of ethanol, the detail of the staining procedure were mentioned in our previous study 19 . Following primary antibodies were used to incubate the sections for overnight at 4 °C. Primary antibodies included rabbit anti-CD3 (Clone SP7; ab16669, Abcam, Tokyo, Japan, 1:200 dilution), rabbit anti-CD20 (Clone EP459Y; ab78237, Abcam, Tokyo, Japan, 1:100 dilution), mouse anti-CD163 (Clone 10D6; Novocastra, Newcastle, UK, 1:400 dilution), mouse anti-CD204 (Clone SRA-C6; TransGenic Inc, Kumamoto, Japan, 1:200 dilution), mouse anti-CD206 (Clone 5C11; Abnova, Taipei, Taiwan, 1:300 dilution), and rabbit anti-EGF (Clone ab9695; Abcam, Tokyo, Japan, 1:30 dilution). Antibody was washed by TBST and as a chromogen 100-400 µl DAB (Peroxidase Stain DAB Kit, nacalai tesque, Kyoto, Japan) was applied to each section. Finally, Mayer's hemalum solution (Merck KGaA, Darmstadt, Germany, 1:4 dilution) was used for counterstain and then sections were washed two times for 5 min each in dH 2 O. After dehydration, sections were mounted with coverslips. The numbers of CD3-, CD20-, CD163-, CD204-, CD206-, and EGF-positive cells were counted in 4 mm 2 sections from five independent high-power microscopic fields (400×, 0.0625 μm 2 ). Double and triple immunofluorescence analysis. The slides were microwaved in AR6 buffer (Opal-4 Color Manual IHC kit; PerkinElmer, Waltham, MA, USA) and cooled for 30 min. Sections were then incubated in Antibody Diluent/Blocking Buffer (Opal-4 Color Manual IHC kit; PerkinElmer) for 10 min at room temperature and then incubated with primary antibodies (as listed before). Double staining was performed with CD3/ EGF, CD20/EGF, CD163/EGF, CD204/EGF, and CD206/EGF, and triple staining was performed with CD163/ CD204/CD206 for 2.5 h at room temperature. Samples were washed three times in TBST for 2 min each and then incubated in Polymer HRP (Ms + Rb), (Opal-4 Color Manual IHC kit; PerkinElmer) Samples were rinsed three times in TBST for 2 min each and then incubated in Opal Fluorosphore working solution (Opal-4 Color Manual IHC kit; PerkinElmer) for 10 min at room temperature. Samples were rinsed in TBST, microwaved in AR6 buffer and then mounted with VECTASHIELD MOUNTING MEDIUM FOR FLUORESCENCE WITH DAPI (Vector Laboratories, Burlingame, CA, USA). In double staining, CD3, CD20, CD163, CD204, and CD206 were labeled with Opal 520 Fluorosphore (Green) and EGF was labeled with Opal 570 Fluorosphore (Red). In triple staining CD163, CD204, and CD206 were labeled with Opal 520 Fluorosphore (Green), Opal 570 Fluorosphore (Red), Opal 690 Fluorosphore (White), respectively. For taking photomicrographs we used a light microscope with a digital camera (BZ-9000 series; Keyence, Tokyo, Japan). The ratio of EGF-positive cells in immune cells was defined as the ratio of the number of each EGF + immune cells (yellow) to the whole number of each immune cells (yellow and green) in a 4-mm 2 field of view, from five separate areas. flow cytometric analysis. For detection of intracellular EGF, 5 OSCC patients peripheral blood mononuclear cells (PBMCs) (5 × 10 5 cells/ml) were collected and cultured in PBS and then stimulated with 40 ng/ml PMA (phorbol 12-myristate 13-acetate; Wako, Tokyo, Japan), 4 μg/ml ionomycin (Ionomycin Calcium; Wako) and Brefeldin A Solution (BioLegend, San Diego, CA, USA, 1:1000 dilution) for 6 h in a moistened chamber at 37 °C with 5% CO 2 supply.
Next, the harvested cells were washed with eBioscience Flow Cytometry Staining Buffer (Thermo Fisher Scientific, Waltham, MA, USA). After rinsing, the cells were incubated at normal temperature for 20 min in dark with PE anti-human CD163 antibodies (Clone GHI/61, IgG1, κ ; BioLegend), FITC anti-human CD204 antibodies (Clone REA460, IgG1; Miltenyi Biotec, Bergisch, Gladbach, Germany), and PerCP/Cy5.5 anti-human Separation and culture of tAM subsets. PBMCs from a healthy donor were collected and by using EasySep TM Human Monocyte Isolation Kit (STEMCELL Technologies Inc, Vancouver, Canada) CD14 + monocytes were isolated. Then, the isolated CD14 + cells (5 × 10 6 cells/ml) were differentiated into M2 macrophages by CellXVivo Human M2 Macrophage Differentiation Kit (R&D System) in a humidified chamber at 37 °C in a 5% CO 2 atmosphere for 6 days. After differentiation M2 macrophages were collected and washed.  cell proliferation assay. HSC-2 cells (1 × 10 4 /well) were cultured in a 96-well dish with CM from CD163 + , CD204 + , and CD206 + cells for 4 days at 37 °C with 5% CO 2 with or without Anti-EGFR antibody (Clone ab231; Abcam, Tokyo, Japan 1:100 dilution). DMEM (Life Technologies Corporation) with 2% FBS (Sigma-Aldrich) was used as a control. Cell proliferation was examined by the XTT Cell Proliferation Assay Kit (Cayman Chemical cell invasion assay. HSC-2 cell lines were cultured for 4 days in CM of CD163 + , CD204 + , and CD206 + cells and used for invasion assays by using Corning ® BioCoat ™ Matrigel ® Invasion Chambers (Corning Incorporated, Corning, NY, USA) As follows the product's guidelines, three Matrigel and three control inserts were prepared for a 24 well culture dish. The CM (750 μl) of CD163 + , CD204 + , and CD206 + TAM was used as a chemoattractant in each well subsequently and the insert was filled with 500 μl of DMEM (Life Technologies Corporation) without serum and HSC-2 cell line (2.5 × 10 4 cells/insert). The cells were cultured for 22 hours in a humid chamber with 5% CO 2 atmospheric state at 37 °C temp. Then the inserts were removed from the wells and Matrigel was cleaned off. After that cells were fixed with 100% methanol (Junsei Chemical Co.,Ltd., Tokyo, Japan) and then stained with Mayer's hemalum solution (Merck KGaA, Darmstadt, Germany, 1:4 dilution) and Tissue-Tek eosin (Sakura Finetek Japan Co., Ltd, Tokyo, Japan) (Supplementary Method 1).
The numbers of cells on the membrane were counted in 4 mm 2 sections from five different high-power microscopic fields (400×; 0.0625 μm 2 ). The average number of cells in the Matrigel insert membrane was divided with the average number of cells in the control insert membrane, and then multiplied with 100 to calculate the invasion percentage.
Statistical analysis. JMP software version 11 (SAS Institute, NC, USA) was used to carry out the statistical analyses. The significant differences between each group were assessed by Mann-Whitney U test, Kruskal-Wallis test, Wilcoxon test, and Spearman's rank correlation coefficient test. For progression-free survival (PFS) and disease-specific survival (DSS) evaluation Kaplan-Meier method was used and by using the Cox hazard test curve comparisons were calculated. Receiver operating characteristic (ROC) curve was used to evaluate the sensitivity, specificity and the areas under the curve (AUC). In all analyses, P values ≤ 0.05 were considered statistically significant.  Fig. 1. Expression of CD163 was diffusely detected in tumor stroma and around tumors, while expressions of CD204 and CD206 were mainly detected in and around tumors (Fig. 1A). Triple immunofluorescence staining confirmed that the distribution of these TAM markers in OSCC tissues was quite different (Fig. 1B).

co-localization of tAM markers and the eGf proangiogenic growth factor in oScc tissues.
Immunohistochemical staining showed that the expression of EGF, a proangiogenic growth factor, was strongly detected throughout the tumors and epithelial cells (Fig. 1A). To clarify whether TAM subsets express EGF, double immunofluorescence staining with TAM markers and EGF was performed. As shown in Fig. 1C, all three TAM markers were partially co-localized with EGF. Moreover, lymphocytes (CD3 or CD20-positive cells) also were partially co-localized with EGF, but EGF was mainly expressed on TAMs, especially CD206 + TAMs (Supple Fig. 1). expression levels of proangiogenic growth factor produced by tAMs. To determine whether CD163, CD204, and CD206-positive cells expressed EGF, we isolated the three TAM subsets in PBMCs from 7 patients with OSCC and examined the expression levels of EGF. As shown in Fig. 2A,B, CD206 + cells expressed higher levels and numbers of EGF intracellularly compared with CD163 + and CD204 + cells. These results suggest a difference in the expression of EGF among the TAM subsets. To determine the production of EGF by each TAM subset, we also evaluated the concentration of EGF in the CM of each TAM subset. The concentration of EGF in CM of CD206 + cells was higher than that of CD163 + and CD204 + cells (Fig. 2C).

Association of tAMs with proliferation and invasion of oScc cell lines.
We next examined the effect of co-culturing CM of the TAM subsets (CD163 + , CD204 + , and CD206 + cells) with OSCC cells in vitro (Fig. 3A). We first evaluated cell division rates of three OSCC cell lines, including SQUU-A, SQUU-B, and HSC-2, and found that HSC-2 cells showed a high cell division rate compared with SQUU-A and SQUU-B cells (Supple Fig. 2). We therefore selected HSC-2 cells for subsequent co-culture experiments.
Cell proliferation assays revealed a highly increased viability of HSC-2 cells co-cultured with CM of CD206 + cells in comparison with cells co-cultured with CD163 + and CD204 + cell CM. Furthermore, we found that anti-EGFR antibody significantly reduced the viability of HSC-2 cells co-cultured with CM of TAM subsets, www.nature.com/scientificreports www.nature.com/scientificreports/ especially CD206 + cells (Fig. 3B). We also confirmed the effect of EGFR expression on HSC2 cells co-cultured with each TAM subset CM, and there were no significant differences among TAM subsets (Supple Fig. 3). Similar results were observed in invasion assays. The invasion activity of HSC-2 cells co-cultured with CM of CD206 + cells was more enhanced than the invasion of cells co-cultured with CD163 + and CD204 + cell CM (Fig. 3C).
Association of TAM subsets with clinical and pathological findings of OSCC patients. We next examined the association of TAM subsets with the clinicopathologic factors of OSCC patients. Each number of TAM subsets was counted by single staining. As shown in Table 1, the number of CD163 + and CD204 + cells did www.nature.com/scientificreports www.nature.com/scientificreports/ not show significant associations with any clinicopathologic finding. However, the OSCC patients with clinical stage, clinical T classification, and cervical nodal metastasis showed a significant increase only in the number of CD206 + cells.

Associations of tAM subsets with clinical outcomes and prognosis of oScc patients.
To evaluate the correlation between TAM subsets and the clinical prognosis of OSCC patients, survival rates were estimated by using Kaplan-Meier method. We divided the OSCC patients into low and high expression groups according to the mean number of TAM subsets. In the PFS and DDS, there were no significant differences in low and high CD163 + and CD204 + expression. However, patients with high CD206 + expression had a significantly more unfavorable outcome than those with low expression group (Fig. 4A). According to the above results, we speculate that CD206 is a potential indicator to describe the malignancy of OSCC, so we performed statistical analysis by the ROC curve of clinical prognosis. The AUC area of DFS-related ROC curves of CD163 + , CD204 + , and CD206 + expression were 0.542, 0.534, and 0.659, respectively, and that of DSS-related ROC curves of CD163 + , CD204 + , and CD206 + expression were 0.570, 0.576, and 0.615, respectively (Fig. 4B). These results indicate higher sensitivity and specificity of CD206 + expression for adverse prognosis. Moreover, univariate analysis revealed that PFS and DSS were associated with YK criteria and number of CD206 + cells (Table 2).
Together these results suggest that CD206 + cells play a key role in the invasion and metastasis in OSCC.

Discussion
In 1908, Metchnikov et al. first demonstrated that macrophages were efficient phagocytes and played major roles in inflammation and natural cellular immunity 23 . In the 1970's, macrophages were considered to be notable effector cells that functioned in the cytotoxic killing of tumor cells 24 . Macrophages are divided into two functionally distinct subtypes: classically activated (M1) macrophages, which are stimulated by T helper type 1 (Th1) responses, and alternatively activated (M2) macrophages, which stimulated by Th2 responses 7,8 . M1 macrophages produce pro-inflammatory cytokines and contribute to tumor suppression, whereas M2 macrophages express anti-inflammatory cytokines and have been shown to contribute to tumors by promoting angiogenesis, immunosuppression and activation of tumor cells 25 . Recent studies have described M2-polarized macrophages as TAMs, which highly express markers, including CD163, CD204, and CD206 [26][27][28] . CD163 is a member of the scavenger receptor cysteine-rich family class B and is mainly expressed on mature tissue macrophages 29 . The main function of CD163 is the binding of the hemoglobin-haptoglobin complex. In addition, CD163-positive macrophages infiltrate in inflammatory tissues and were involved in the resolution of inflammation 30 .
CD204 is a prototypic member of a family of structurally different transmembrane receptors conjointly termed as scavenger receptors and is primarily expressed on macrophages and dendritic cells 31 . CD204 recognizes modified lipid proteins, and exogenous pathogen-associated molecular patterns, and apoptotic cells. We previously reported that CD163 + CD204 + TAMs promote T-cell apoptosis and immunosuppression via IL-10 and programmed death-ligand 1 in OSCC patients 19 .
CD206 contributes to lipid metabolism, atherogenesis, and metabolic processes 32 . CD206 is a C-type lectin, alternatively termed as the macrophage mannose receptor, that is, generally expressed by tissue macrophages, dendritic cells, and specific lymphatic or endothelial cells. CD206 plays an important role in immune homeostasis, but its high expression has been increasingly detected in the tumor microenvironment 33 .
TAMs play important functions in the tumor microenvironment, with roles in invasion, proliferation, and metastasis 34 . Several studies indicated that TAMs secrete various kinds of cytokines including growth factor, for example EGF and IL-6, and EGF strongly induces tumor cell proliferation [15][16][17] .  Table 2. Univariate analysis of progression-free survival and disease-specific survival in advanced stage. Statistically significant differences between groups were determined by Cox proportional hazard model (*P < 0.05). HR, hazard ratio; CI, confidence interval.
In this present study, we examined the expression of TAM subsets in OSCC tissues and the relationship between the expression of TAM markers and EGF production and tumor progression. Immunohistochemical staining demonstrated marked differences in the localizations of CD163 + , CD204 + , and CD206 + cells in OSCC tissues. Interestingly, CD206 + cells, from PBMCs of OSCC patients, strongly produced EGF compared with the other TAM subsets. In addition, OSCC cells showed high viability and invasion activity after co-culture with CM of CD206 + cells and a significant decrease in their viability with anti-EGFR antibody. These results suggest that CD206 + TAMs might play an important role in the proliferation and invasion of OSCC via EGF secretion. Goswami et al. 35 found a paracrine loop between breast carcinoma cells and macrophages through EGF and colony stimulating factor 1 production, leading to increased carcinoma cell invasion. While this paracrine loop might be also involved in the invasion activity in OSCC cells, further examinations are needed to clarify the expression of other cytokines secreted by TAM subsets in OSCC.
Next, we examined the involvement of TAMs in the prognosis of OSCC patients. The numbers of CD206 + TAMs in OSCCs were positively correlated with several clinicopathologic factors including clinical stage, clinical T classification, mode of invasion, and cervical nodal metastasis. Notably, a high number of CD206 + TAMs was significantly correlated with poor prognosis. Recently, Dong et al. 36 reported that the high presence of CD206 + TAMs in hepatocellular carcinoma markedly correlated with aggressive tumor phenotypes and were associated with poor prognosis. These results were consistent with our results. In contrast, in adult T-cell leukemia/lymphoma, both CD204 + and CD206 + TAMs were not associated with clinical outcome 37 . These controversial results may be due to different tumor types.
In conclusion, we demonstrated that CD206 + TAMs promote proliferation and invasion of OSCC cells via EGF and high numbers of CD206 + TAMs predict unfavorable clinical prognosis in OSCC patients. However, it is still necessary to elucidate the involvement of other cytokines secreted by TAMs in the tumorigenesis. A more detailed knowledge of the role of each TAM subset could lead to the development of novel pharmacological approaches targeting TAMs or their products, and inhibiting the tumor progression and metastasis.