Oesophageal squamous cell carcinoma–associated IL‐33 rewires macrophage polarization towards M2 via activating ornithine decarboxylase

Abstract Background The tumour microenvironment primarily constitutes macrophages in the form of an immunosuppressive M2 phenotype, which promotes tumour growth. Thus, the development of methodologies to rewire M2‐like tumour‐associated macrophages (TAMs) into the M1 phenotype, which inhibits tumour growth, might be a critical advancement in cancer immunotherapy research. Methods The expressions of IL‐33 and indicators related to macrophage polarization in oesophageal squamous cell carcinoma (ESCC) tissues and peripheral blood mononuclear cell (PBMC)–derived macrophages were determined. Inhibition of ornithine decarboxylase (ODC) with small interfering RNA was used to analyse the phenotype of macrophage polarization and polyamine secretory signals. CCK‐8, wound‐healing and Transwell assays were used to detect the proliferation and migration of ECA109 cells in vitro. The tumour xenograft assay in nude mice was used to examine the role of IL‐33 in ESCC development in vivo. Results This study showed the substantially elevated IL‐33 expression in ESCC tissues compared with the normal tissues. Additionally, enhanced infiltration of M2‐like macrophages into the ESCC tumour tissue was also observed. We observed a strong correlation between the IL‐33 levels and the infiltration of M2‐like macrophages in ESCC tumours locally. Mechanistically, IL‐33 induces M2‐like macrophage polarization by activating ODC, a key enzyme that catalyses the synthesis of polyamines. Inhibition of ODC suppressed M2‐like macrophage polarization. Finally, in vivo, we confirmed that IL‐33 promotes tumour progression. Conclusions This study revealed an oncogenic role of IL‐33 by actively inducing M2‐like macrophage differentiation; thus, contributing to the formation of an immunosuppressive ESCC tumour microenvironment. Thus, IL‐33 could act as a novel target for cancer immunotherapies.


| INTRODUC TI ON
In our immune system, macrophages constitute a major heterogeneous cellular population, which protects our body against several types of pathogenic infections, such as bacterial, fungal and viral. The macrophages located in the tumour microenvironment (TME) are classified as tumour-associated macrophages (TAMs). 1 The macrophage colony-stimulating factor (M-CSF), the C-C motif chemokine ligand 2 (CCL2) and the vascular endothelial growth factor (VEGF) specifically recruit circulating monocytes, which are eventually transformed into TAMs. 2 The differentiation of macrophages into the M1 type (classically activated) occurs under the influence of LPS or IFN-γ, 3,4 and into the M2 type (alternatively activated) through IL-4 and IL-13. 3 The M1 type macrophages engulf and destroy pathogenic microbes that cause viral, bacterial or fungal diseases. They also produce pro-inflammatory cytokines, such as IL-12 and TNF-α. 5 On the contrary, M2like macrophages produce anti-inflammatory cytokines, such as transforming growth factor-β (TGF-β) and IL-10, which promote the inhibition of these adaptive Th1 immune responses and the resulting inflammation. 6,7 Additionally, they exhibit an IL-12 low IL-10 high phenotype and are involved in angiogenesis, tissue remodelling, parasitic infections and tumorigenesis. [8][9][10] However, there are limited reports on the mechanism of M2-like macrophage differentiation and its role in human oesophageal squamous cell carcinoma (ESCC) is unclear.
IL-33, a pro-inflammatory cytokine from the IL-1 family, regulates the host immune response and promotes tumour growth. 11,12 Several studies have revealed that elevated IL-33 levels are probable prognostic markers and correspond to poor prognosis in many types of cancers. [13][14][15] Currently, research data indicate an association between IL-33 and the pathogenesis of ESCC. 16 Elevated IL-33 levels in ESCC are related to the invasion of Treg cells in the tumour. Moreover, the Treg cells express ST2, suggesting that IL-33 probably regulates the physiological activity of Treg cells through ST2. 17,18 Previous studies have indicated that IL-13/IL-4Rα signalling pathway is involved in the amplification and polarization of the alveolar and bone marrow-derived macrophages, resulting in airway inflammation. 19 Additionally, IL-33 is also known to regulate macrophage polarization. In a murine model, IL-33 improved CVB3induced viral myocarditis by inducing macrophage polarization to the M2 type. 20 IL-33 promoted IL-10 expression in macrophages through activating ERK 1/2 and STAT3, which subsequently promoted IL-10 transcription and thus contributed to the beneficial effects of IL-33 on macrophages. 21 Similarly, in the colon cancer TME, IL-33 promoted tumorigenesis by inducing macrophage polarization. 22 However, there are no reports on the in-depth investigation of the role of IL-33 in macrophage polarization to M2 type in ESCC, whether it promotes further tumorigenesis and its mechanism of action.
Here, we found that in patients with ESCC, the tumour tissue samples had elevated levels of IL-33 compared with non-tumour tissue samples. Additionally, an elevated count of CD206 + M2 type macrophages, which infiltrated the ESCC tumours, was observed, and this increase had a positive correlation with the secretion of IL-33. In vitro studies revealed that IL-33 favoured macrophage polarization towards the M2 phenotype, consistent with the expression of CD206 + by activating ornithine decarboxylase (ODC). Inhibition of ODC using small interfering RNA (siRNA) suppressed M2-like macrophages polarization. The supernatants from IL-33-induced M2-like macrophages promoted ESCC cell progression. Thus, these results describe the relevance of the IL-33/macrophage axis in the clinical treatment of ESCC.

| Patients and tissue samples
Patients who were admitted to the Southern Hospital of Southern Medical University and who had not received chemotherapy/radiotherapy underwent surgical resection to provide tumorous and non-tumorous (minimum 5 cm from the tumour) oesophageal tissue samples. Histopathological studies confirmed the absence of cancer cell infiltration in the non-tumour tissue samples. All patients provided written informed consent before study initiation. The TNM staging system (the International Union Against Cancer, 8th edition) was used to identify the clinicopathological stages of tumours. The

Ethics Committee of the Southern Hospital of Southern Medical
University sanctioned this study.

| Immunohistochemistry
The ESCC tissue samples were embedded in paraffin, and 4 µm sections were obtained. Next, these sections were treated with normal goat serum for 30 minutes at 37°C, followed by overnight incubation at 4°C with the following primary antibodies: rabbit anti-CD206, rabbit anti-IL-33 or rabbit anti-CD68 antibodies (Affinity, USA). After washing, these sections were incubated for 30 minutes at 37°C with horseradish peroxidase (HRP)-conjugated anti-rabbit secondary antibodies (Affinity, USA). Two independent and experienced histopathologists who were blinded to clinical patient data reviewed each section. We identified five fields in each section and counted brown granule-containing cells at 400× magnification for each field using average values.

| Flow cytometry analysis
We collected blood samples from both healthy subjects and patients with ESCC, followed by isolation of human peripheral blood mononuclear cells (PBMCs) using Ficoll-Hypaque density gradient centrifugation (GE Healthcare, USA). We isolated CD14 + monocytes from PBMCs using the human CD14 Positive Selection Kit

| Cell culture and M2 macrophage induction
Fresh blood samples from healthy donors were collected, and PBMCs were isolated by density gradient centrifugation using Ficoll-Hypaque (GE Healthcare, NJ, USA). Isolation of monocytes and induction into M0 macrophages were done using the above procedures. Next, human recombinant IL-33 (PeproTech, New Jersey, USA) was used to induce M0 macrophages to differentiate into M2 macrophages for 24 hours. All cells were cultured at 37°C in a humidified incubator with 5% CO 2 .

| Immunofluorescence
We induced the M0/M2 macrophages, at a concentration of 5 × 10 5 cells/well, using the method described in the previous section. After washing with PBS, the cells were treated with 5% goat serum in PBS for 30 minutes, followed by incubation with rabbit anti-human ODC and arginase 1 (ARG1) primary antibody (Affinity, USA) diluted in the antibody dilution. This antigen-antibody complex was coupled with a tetramethylrhodamine-conjugated goat anti-rabbit secondary antibody (Zhongshan Biotechnology).
Finally, the PBS-washed cells were visualized using a microscope (Olympus, Japan).

| siRNA transfection of macrophages
We induced the M0 macrophages, at a concentration of 5 × 10 5 cells/ well, using the method described in the previous section. Next, these macrophages were transfected for 6 hours with either nonsilencing control or ODC-targeting siRNA (40 pmol/mL), following the manufacturer's protocol (GenePharma, Shanghai, China). Posttransfection, the cells were grown for another 24 hours in fresh RPMI-1640 medium containing FBS (10%), followed by a 24 hours treatment with IL-33 (20 ng/mL) to induce their differentiation into M2 macrophages.

| Quantitative RT-PCR
The TRIzol reagent (Vazyme, Nanjing, China) was used for total RNA extraction from the cultured cells, following the manufacturer's protocol. The cDNA was obtained by reverse-transcribing the isolated RNA (500 ng/10 µL) using a reverse transcription kit (Vazyme, Nanjing, China) and was diluted in 10 µL of nuclease-free water. A Roche LC480 System was used to perform Real-time PCR by mixing

| Western blot analysis
We used the RIPA Lysis Buffer (Fudebio, Hangzhou, China) containing phosphatase/protease inhibitors (Fudebio, Hangzhou, China) to extract total protein from the cultured cells following the manufacturer's method. SDS-polyacrylamide gel electrophoresis (10%; SDS-PAGE) was performed to separate the protein samples (20 µg), which were then transferred to Millipore polyvinylidene difluoride (PVDF) membranes. These PVDF membranes were kept in overnight incubation at 4°C with anti-CD206, anti-ODC, anti-IL-33, anti-iNOS, anti-CD68, anti-ARG1 and anti-GADPH primary antibodies (Affinity, USA), followed by incubation at room temperature for 60 minutes with horseradish peroxidase (HRP)-conjugated secondary antibodies (Affinity, USA) that were diluted to 1:5000 in 5% skim milk. After washing with TBST, the membranes were treated for 1 minutes with SuperSignal™ West Dura Extended Duration Substrate (Fudebio, Hangzhou, China) and were visualized through chemiluminescence.

| Collection of supernatants
We induced the M0 macrophages using the method described in the previous section. The resultant M0 macrophages were treated with IL-33 (20 ng/mL) for a period of 24 hours to induce their differentiation into M2-like macrophages. Next, fresh RPMI-1640 medium (400 µL/well) containing FBS (10%) was added to both M0 and M2 macrophages, and we harvested the cell-free supernatants from M0-or M2-like macrophages post another 24 hours culture.

| Determination of polyamines by highperformance liquid chromatography (HPLC)
1.0 mL of macrophage supernatant was collected, followed by the addition of 1 mL of 2 mol/L NaOH solution and 5 µL of benzoyl chloride, followed by vortex mixing, and then allowed to stand for 20 minutes. 2 mL of saturated sodium chloride solution and 2 mL of chloroform were added, followed by vortex mixing for 1 minutes, then centrifugation at 2200 g for 10 minutes, and finally, the lower organic phase was collected and blown dry with nitrogen. The resulting white solid was dissolved in 0.5 mL chromatographic methanol, then placed in a 50°C water bath for 8 hours, then 0.2 mL 2 mol/L NaOH solution and 2 mL chloroform were added, followed by vortex mixing, followed by centrifugation at 2200 g for 10 minutes, and finally the lower organic phase was collected and blown dry with nitrogen. The resulting white solid was dissolved with 0.5 mL chromatographic methanol, and filtered with 0.22 µm organic microporous membrane. Chromatographic conditions are as follows: detection wavelength: 234 nm; column temperature: 25°C; mobile phase: 55% methanol-45% water; flow rate: 1.0 mL/min; and injection volume: 20 µL.

| Cell proliferation and migration analysis
We co-cultured the ECA-109 cell line (5 × 10 3 cells/well) with the supernatants of M2-like macrophages in 96-well plates (NEST Biotechnology, Wuxi, China) to study cellular proliferation. We used F I G U R E 1 M2 macrophage infiltration and IL-33 production are enhanced with close correlation in oesophageal squamous cell carcinoma (ESCC). A, Representative images of IL-33 + cell, CD206 + cell and CD68 + cell in non-tumour and tumour tissue (scale bar = 50 μm). B, Western blot analysis of IL-33 and CD206 expression in non-tumour and tumour tissues, and GAPDH was used as a reference control. C, IL-33, CD206 and CD68 were measured by RT-PCR in non-tumour and tumour tissues; the correlation of CD206 and IL-33 in ESCC tissues; N = 20, R 2 = .54, P < .01. D, M1 (CD68 + CCR2 + ) and M2 (CD206 + CX3CR1 + ) populations in peripheral blood from those with ESCC (n = 48) and healthy controls (n = 33), as measured by flow cytometry. The difference in M1/M2 ratio between two groups was analysed. E, The population of M2-like macrophages subset in peripheral blood between ESCC patients and healthy controls showed significant difference instead of M1-like macrophages. F, ROC curve was used to analyse to assess the diagnostic value in ESCC. **P < .01; ***P < .

| Animal experiments
We procured female BALB/c nude mice

| Statistical analysis
All statistical analyses were done using spss v20.0. The data were represented as mean ± SEM (standard error of the mean). Student's t test was performed to determine the statistical significance of the differences between the two groups. The multi-group data analysis was done using ANOVA. The Pearson correlation analysis or the linear regression analysis was used to evaluate the correlations between parameters. Data analysis was done using two-tailed tests; P < .05 was regarded as statistically significant unless stated otherwise.

M2-like macrophage infiltration and IL-33 secretion
We performed immunohistochemical studies to assess the IL-33, CD68 and CD206 levels in the tumour and non-tumour tissue samples from patients with ESCC. IL-33 is mainly located in the cytoplasm of ESCC cells, while CD68 and CD206 are mainly found in the tumour stroma. Substantially elevated levels of IL-33 and CD206 but not CD68 were observed in the tumour tissue samples than the non-tumour tissue samples ( Figure 1A). We performed WB to determine IL-33 levels to verify the previous results and found considerably elevated secretion of IL-33 in tumour tissues compared with the non-tumour tissues ( Figure 1B). Moreover, we found a positive correlation between IL-33 expression and the number of CD206 + macrophages ( Figure 1C). However, CD68 + macrophages were not correlated with IL-33 production. Table 1 presents the baseline clinical and pathological characteristics.  Figure 1F).

| IL-33-induced M2-like macrophage differentiation occurs via ODC activation
Here, we treated cells with ODC-targeting siRNA, followed by the repetition of the IL-33-induced M2-type macrophage polarization test to understand the regulatory role of ODC activation in the IL-33 induced M2-type macrophage polarization. We found substantially downregulated levels of IL-10 and TGF-β while upregulated levels of IL-12p35 in the ODC siRNA group than the control siRNA group ( Figure 4A). Thus, these results indicated that IL-33 induced M2 macrophage differentiation with the IL-10 high TGF-β high IL-12 p35 low phenotype through ODC activation. The inferences from immunofluorescence ( Figure 4B), flow cytometry ( Figure 4C) and WB ( Figure 4E) also verified these results. We further used HPLC to detect the content of polyamines (putrescine, spermidine and spermine), downstream of ODC, in the supernatant of induced macrophages, and found increased levels of putrescine, spermidine and spermine in the supernatants of IL-33-induced M2 macrophages and decreased levels in the ODC siRNA group ( Figure 4D).

| IL-33-induced M2-like macrophage supernatants support the migration of ESCC cells
We isolated the supernatants from IL-33-induced M2-like macrophages and studied their effect on cellular proliferation and migration of the ECA109 cells. We observed that co-cultivation of ECA109 cells with the IL-33-induced M2-like macrophages resulted in a

33-induced repolarization of M2-like macrophages
We subcutaneously injected IL-33 in the areas surrounding tumour mass every alternate day using PBS as the control. We found that IL-33 resulted in an enhanced tumour volume compared with PBS ( Figure 6A).
Similar results were observed for tumour weight ( Figure 6B). We did not observe a significant difference in the body weight of mice after the treatment, which indicated the absence of severe side effects due to this treatment ( Figure 6C). Additionally, the IL-33 group had a shorter survival rate compared with the control group ( Figure 6D). We observed less necrotic area in the tumour mass of the control group, as evidenced by the H&E staining analysis ( Figure 6E).

| D ISCUSS I ON
The TAMs are those macrophages in the tumour microenvironment (TME), which are mainly chemoattracted by monocytes in the blood system by chemokines in TME. The most common chemokines are CSF-I, CCL-2, CCL-3, CCL-5 and VEGF. 23 Here, we found elevated levels of IL-33 in the tumour tissue samples, which were also associated with the upregulated density of CD206 + M2-like macrophages (R 2 = .65). Thus, we concluded that in ESCC, elevated expression of IL-33 at the tumour site was correlated with the accumulation of macrophages and M2 macrophage differentiation ( Figure 1).
Next, we created an IL-33-based differentiation system for in vitro stimulation of the macrophages to understand the basic mechanism of IL-33-induced macrophage differentiation. IL-10 and TGF-β are known to be primarily produced by the M2-type macrophages.
ODC gene is the first rate-limiting enzyme in the process of polyamine synthesis, which has high activity in the tissues with vigorous cell growth and responds quickly to growth-stimulating factors. Recent can be decarboxylated by ODC to produce putrescine, which is further converted into spermidine and spermine. The knockdown of ODC considerably decreased the levels of IL-10, TGF-β and polyamine (putrescine, spermidine and spermine) levels; however, the IL-12p35 expression was restored on silencing ODC in the IL-33-induced M2-like macrophages, consistent with the previous reports ( Figure 4). Thus, the ODC pathway, which promotes M2-like macrophage differentiation, was involved in the negative regulation of IL-12p35 expression by IL-33. These results indicated that the high secretion of IL-33 induced the infiltration of M2-like macrophages in ESCC through the ODC pathway.
Additionally, we found a strong correlation between IL-33 secretion in the ESCC tissue samples and the tumour TNM stages.
A significant elevation in cellular migration and proliferation was observed in the ESCC carcinoma cells that were co-cultured with the supernatant of the M2-type macrophages in vitro compared with the control group. However, the knockdown of ODC attenuated the M2 supernatant-induced cellular proliferation and migration ( Figure 5).
On the contrary, we observed an IL-33-induced immunosuppressive response in tumour-bearing mice with no side effects ( Figure 6).
Thus, IL-33 secretion in the tumour sites induced M2-like macrophage differentiation and favoured tumour growth and tumour metastasis in ESCC.
Thus, the results of this study suggest that M2-like macrophage differentiation is promoted by the elevated secretion of IL-33 in the tumour sites via ODC activation during the establishment and progression of ESCC, and blocking IL-33 or ODC may constitute a novel therapeutic route for patients with ESCC.

CO N FLI C T O F I NTE R E S T S
The authors declare that there are no conflict of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this article.