IL-33 facilitates proliferation of colorectal cancer dependent on COX2/PGE₂

The global gene expression data of 394 colorectal cancer samples were acquired from the Cancer Genome Atlas (TCGA) database (https://​gdc.​cancer.​gov/​). The clinical information of CRC patients was listed in Additional file 1: Table S1. Gene expression levels of IL-33, ST2 and COX2 included in the data were subjected to Kolmogorov-Smirnov (K-S) test of normality. The gene set enrichment analysis (GSEA) was performed using the program GSEA v2.2.0. Gene sets for GSEA were obtained from the Molecular Signature Database (MSigDb) (http://​www.​broadinstitute.​org/​gsea/​msigdb/​index.​jsp.). The Log2-rank test was used to make gene enrichment statistical comparisons. P-value (P < 0.05) was regarded statistically significant. Pearson’s correlation test was performed using SPSS software with COX2 and ST2 expression levels extracted from the downloaded data. P-value (P < 0.05) was regarded statistically significant. For experimental data, statistical analysis was performed using the GraphPad Prism 5 software. Student's t-test was used for comparison of paired groups. Multiple group comparisons were performed using analysis of variance, ANOVA.

PGE₂ and human recombinant IL-33 were purchased from ProteinTech. The mouse recombinant IL-33 was purchased from Pepro Tech. The following antibodies were used: ST2 antibody (R&D systems), PGE₂ antibody (Cayman), COX2 antibody (Abclonal), and control IgG (Santa Cruz). The following chemical reagents were used: SB203580 (Cayman, 10 μg/mL), PD98059 (Cayman, 20 μg/mL), SP600125 (Cayman, 10 μg/mL), BIX01294 (MCE, 2 μM), 5-Aza (Sigma, 10 μM), SC-560 (Cayman, 0.1 μg/mL), Celecoxib (Sigma, 20 μg/mL) and BAY11–7082 (Cayman, 10 μM).

Primary CRC cell lines were isolated as described previously [4, 29] from fresh human CRC tissues of three patients. Human CRC cell line HT29 was bought from American Type Culture Collection (ATCC). Murine CRC cell line MC38 was provided by Dr. Weiping Zou (Michigan, USA) and was tested in 2011 [30]. These cells were all cultured in RPMI1640 medium with 10% fetal bovine serum, 100 IU/mL penicillin and 100 μg/mL streptomycin at 37 °C in a cell incubator with a 5% (v/v) CO2 humidified atmosphere. C57/BL6 wild type mice were bought from Beijing HFK Bioscience Co. Ltd. C57/BL6 IL-33 transgenic mice were from Dr. Zhanguo Li (Beijing, China) and Dr. Lianfeng Zhang (Beijing, China) [4]. All mice were housed in specific pathogen free (SPF) animal room of Tongji Medical College.

Six-week-old IL-33 transgenic mice and wild-type male C57/BL6 mice were used in tumor growth experiments. Each mouse was inoculated with 1 × 10⁶ MC38 cells subcutaneously on the back. Once visible tumors generated, tumor sizes were measured every 2 days. Tumor volumes were calculated by the formula V = 1/2 × length×width² (mm³). For comparing the tumor growth rates, seven mice were set in each group. Another same set of experiment was performed for harvesting tumor tissues.

Immunohistochemistry staining was performed as described previously [31]. The tumors removed from the wild-type or IL-33 transgenic mice at Day 22 post tumor inoculation were fixed with 4% formaldehyde and embedded with paraffin. The sections were labeled with anti-Ki67 antibody (Arigo, 1:200) and anti-PCNA antibody (Boster, 1:200). Quantification of Ki67 and PCNA expression was independently performed by two pathologists. The positively staining cells were quantified by ImageJ software. Twenty CRC tissues and adjacent normal tissues were obtained from surgery in Union hospital (Wuhan) with the permission of each patient. The ST2 staining was performed with anti-ST2 antibody (R&D, 1:200).

Primary CRC cells were seeded in 96-well plates (6000 cells per well) and incubated with RPMI1640 medium with IL-33 (0, 50, 100, 200 ng/mL) or PGE₂ (50 ng/mL). Cell viability was measured with the Cell Counting Kit-8 (Biosharp) at 24^th, 48^th or 72^nd h. The curves of cell viability were plotted by the absorbance of each time point.

Primary CRC cells, HT29 cells or MC38 cells were seeded in 12-well plates (2 × 10⁵ cells per well) and incubated with the following reagents: human/mouse recombinant IL-33 proteins (0, 50 or 100 ng/mL), celecoxib (20 μg/mL), ST2 antibody (2 μg/mL) or BAY11–7082 (10 μM) for 24 h. Three parallel wells were set for each treatment. Total RNA was isolated with TRIzol reagent (Invitrogen) and was reversely transcribed into complementary DNA with RNA reverse transcriptase (Vazyme). Real-time PCR was performed on ABI StepOne Plus Detector System (Applied Biosystem). Relative mRNA expression of human genes was normalized to GAPDH, and for mouse genes the mRNA levels were normalized to mouse Hprt gene. Each experiment was repeated three times and representative results were shown. The primers used are listed in Additional file 1: Table S2.

CRC cells were seeded in 12-well plates at a density of 500 viable cells per well. Then the cells were incubated in RPMI1640 medium with recombinant IL-33 protein (added at Day 1, 3 and 5), celecoxib, the ST2 antibody or the BAY11–7082. Colonies were photographed and counted at Day 10 or 15 to allow all wells undergoing different treatments to generate visible colonies. Three parallel wells were set for each treatment. Each experiment was repeated three times and representative results were shown.

Primary CRC cells, HT29 cells or MC38 cells were seeded in 6-well plates (5 × 10⁵ cells per well). The CRC cells receiving different treatments were scraped and collected by low speed centrifugation and lysed using Cell Lysis Buffer. The Western blotting was performed as previously described [32]. Blots were performed with a COX2 antibody (Abcolonal), ST2 antibody (R&D systems), NF-κB P65 antibody (CUSABIO) and β-actin antibody (Proteintech). Specific bands were detected using ECL detection reagents (Millipore, USA). Each experiment was repeated three times and representative results were shown.

Primary CRC cells seeded in 6-well plates (5 × 10⁵ cells per well) were incubated in RPMI1640 medium or RPMI1640 medium containing rhIL-33 protein (100 ng/mL) for 24 h. The culture supernatants were collected. The concentration of PGE₂ was measured by the Parameter^TM PGE₂ assay kit (R&D) according to the manufacturer’s instructions.

Primary CRC cells (5 × 10⁶) were collected from culture plates by low speed centrifugation and made into signal-cell suspensions. Primary CRC cells were stained with PE-conjugated specific antibody to ST2 (bs-2382R, Bioss, China) and PE-conjugated isotype IgG (bs-0295P-PE, Bioss, China). The sample was diluted into a concentration of 2 × 10⁶ cells/mL for sorting. The sorting was performed by a high speed flow sorter (FACSAria II, BD). The sorting system was fluxed using ethanol 70% for 10 min to reduce pollution. Sheath fluid was using autoclaved and filtered (0.22 μm) phosphate saline buffer (1 × PBS). Sorting rate typically was 3000 events/s, and cells were acquired at a rate of 300–500 cells/s. By comparing to the negative control, the PE positively stained subset was gated for sorting as the ST2-positive subset of primary CRC cells. The rest cells were collected and used as ST2-negative primary CRC cells. Cells were recovered into a 15 mL Eppendorf tube with washing buffer. Before being used for experimental assay, the sorted ST2-positive and ST2-negative cells were subjected to flow cytometry analysis with the same program and testing with the same gating condition used for sorting. Purity of sorted cell subsets was > 90% as verified by flow cytometry. Flow cytometry data was analyzed by Flow Jo 7.6.1 software.

To investigate the signaling of IL-33 in CRC, we analyzed the gene expression data from TCGA Data Portal that consisted of 394 CRC samples. Enrichment analysis revealed that the gene signature of cell proliferation was significantly correlated with IL-33 expression (Fig. 1a; Additional file 1: Table S3). This indicates that IL-33 might regulate the proliferation of CRC cells. Thus, we performed experiments to test this notion. By inoculating MC38 CRC cells in animals, we found the tumor growth in IL-33 transgenic mice was more rapidly than that in wild-type mice (Fig. 1b). The immunohistochemical staining showed that the cell proliferation marker Ki67 and the proliferating cell nuclear antigen (PCNA) were expressed significantly higher in the tumors generated in IL-33 transgenic mice than in the tumors from wild-type mice (Fig. 1c, d). The increased Ki67 and PCNA expression in the tumors from IL-33 transgenic mice were verified by Western blotting (Fig. 1e). To determine whether IL-33 facilitated the proliferation of CRC cells directly or via regulating other factors in vivo, we incubated primary CRC cells isolated from human cancer tissues with recombinant IL-33 protein. We found that IL-33 increased cell viability of primary CRC cells and upregulated the expression of Ki67 and PCNA in a dose dependent manner (Fig. 1f, g). To confirm the direct effects of IL-33 on the proliferation of CRC cells, we performed colony formation assay with human CRC cell line HT-29, mouse CRC cell line MC38 as well as the primary human CRC cells. IL-33 significantly facilitated the colony formation of all the three types of cells (Fig. 1h, i, j). Therefore, we concluded that IL-33 accelerated proliferation of CRC.

We next sought to investigate the mechanism how IL-33 facilitated CRC proliferation. We screened tumor proliferation associated signals: DNA and histone methylation and prostaglandin E2 (PGE₂) synthesis using inhibitors. The IL-33-induced Ki67 and PCNA were detected when the primary CRC cells were treated with the P38 inhibitor SB203580, the MAPK/ERK kinase (MEK) inhibitor PD98059, the c-Jun N-terminal kinase (JNK) inhibitor SP600125, the histone methyltransferase inhibitor BIX01294, the DNA methyltransferase inhibitor 5-Aza, COX1 selective inhibitor SC-560, and the COX2 selective inhibitor celecoxib. We found that in celecoxib treated primary CRC cells IL-33 did not elevate Ki67 or PCNA (Fig. 2a, b). In CRC cell lines HT-29 and MC38, celecoxib also effectively abrogated the IL-33-induced elevation of Ki67 and PCNA (Fig. 2c, d). COX2 functions as a key enzyme in the synthesis of PGE₂ that potently accelerates tumor proliferation [33‐35]. These indicate that COX2/PGE₂ might mediate the proliferation promoting function of IL-33. In accordance with this notion, IL-33 incubation increased COX2 mRNA and protein levels in the primary CRC cells in a dose dependent manner (Fig. 2e, f). CRC cells incubated with IL-33 produced significantly higher level of PGE₂ (Fig. 2g). The artificially synthesized PGE₂ increased the cell viability of the primary CRC cells (Fig. 2h), verifying its function in promoting tumor proliferation characterized previously. To confirm the autocrine of PGE₂ mediated the IL-33-induced acceleration of proliferation, we performed colony formation with CRC cells incubated with a PGE₂ neutralizing antibody as well as the inhibitor celecoxib. Both PGE₂ neutralizing antibody and celecoxib blocked the increase of colony numbers induced by IL-33 (Fig. 2i). Together, IL-33 facilitated CRC proliferation via increasing PGE₂ production.

We previously reported that cultured CRC cells expressed functional IL-33-recceptor ST2 [4]. Here we further detected the receptor in human CRC. By immunohistochemical staining, we found positive ST2 expression in most of  CRC samples (19/20), whereas ST2-positive staining was rarely observed in the adjacent normal tissues (Fig. 3a). The ST2 expression in CRC tissues was also verified by Western blotting (Fig. 3b). Although ST2 has been identified as IL-33 receptor, IL-33 can function in an ST2-independent fashion [36]. We checked whether IL-33 promoted CRC proliferation through its receptor by using an ST2 blockade antibody. ST2 blockade abolished the increase of colony numbers caused by IL-33 incubation (Fig. 3c). The antibody treatment also suppressed the IL-33-elevated Ki67 and PCNA levels (Fig. 3d). Thus, we demonstrated that IL-33 facilitates CRC proliferation by signaling through its receptor ST2.

When ST2 was blocked, IL-33 did not upregulate the mRNA and protein levels of COX2 in primary CRC cells or HT-29 cells anymore (Fig. 4a, b). As shown by flow cytometry assay, 12.3% of primary CRC cells were obviously ST2 positive (Fig. 4c). To further verify COX2 induction was dependent on ST2, we sorted the ST2-positive subset and the ST2-negative subset from primary CRC cells (Fig. 4c). The folds of COX2 induction caused by IL-33 were markedly higher in ST2-positive CRC cells than in ST2-negative CRC cells (Fig. 4d). By analyzing TCGA data, we found that COX2 levels in human CRC were positively correlated with ST2 expression (Fig. 4e). These demonstrate that IL-33 induces COX2/PGE₂ via ST2-mediated signaling.

Previous studies have revealed the COX2 gene activation relies on assistance of several transcriptional factors including NF-κB, NF-AT, C/EBP and CREB [37‐39]. IL-33/ST2 axis could directly stimulate NF-κB nuclear factors in different cell types [40]. Thus, we checked whether the NF-κB signaling mediated the process where IL-33/ST2 axis induced COX2. As expected, the NF-κB specific inhibitor blocked the induction of COX2 mRNA and protein caused by IL-33 in primary CRC cells, HT29 cells and MC38 cells (Fig. 4f, g). In line with this, the NF-κB specific shRNA (shP65) repressed COX2 mRNA and protein levels induced by IL-33 (Fig. 4h, i). Therefore, we concluded that IL-33/ST2 axis induces COX2 through NF-κB signaling.