A total of 82 CC tissues and 40 UC tissues were investigated in this study, these tissues were obtained from patients at the time of surgical resection or endoscopy. Normal colon tissues were obtained from 40 Chinese patients who had suffered from non-tumor diseases such as tediously long Colon and vascular malformation. All of the cases occurred between January 2010 and August 2011 at the First Affiliated Hospital of Nanjing Medical University (Nanjing, China), and all the samplers were conducted according to the principles expressed in the Declaration of Helsinki. Written documented informed consent for gene expression analyses of all tissues was obtained from all patients prior to surgery or endoscopy examination; this study and consent procedure was approved by the local ethics committee of the First Affiliated Hospital of Nanjing Medical University. The diagnosis and staging of colon cancer was assessed according to the AJCC (TNM) Staging System. The detailed clinical characteristics of the 82 colon cancer patients were listed in Table 1.

The fresh tumor tissues samples were collected from 82 cases of histologically confirmed colon cancer obtained from hospitalized cancer patients who underwent surgery. The tumor pieces were washed twice in RPMI 1640 (Invitrogen, CA). The fatty, connective, and/or necrotic tissues were removed from the tumor mass in a 10 cm dish. Next, the tissue was cut into 1–2 mm pieces in RPMI 1640, and the minced tumor pieces were transferred into a 15-ml or a 50-ml conical tube and incubated with a triple enzyme digestion medium that contained DNase (30 U/ml), hyaluronidase (0.1 mg/ml), and collagenase (1 mg/ml) for 2 hours at room temperature with gentle shaking. The samples were then resuspended in 10 ml RPMI 1640, filtered through a 70-μm cell strainer (BD), placed into several wells containing 1 ml of T-cell growth medium (RPMI 1640 with 10% human AB serum, supplemented with 5 U/ml human rIL-2) in a 24-well plate, the TILs was processed immediately for IL-22 detection by flow cytometry. To obtain IL-22(+)TILs, lymphocytes were isolated by flow cytometry, and maintained in T-cell growth medium. A polarization stimulation to Th22 cells was performed according Th22 condition described as previous report [³³] which can be briefly described as 50 ng/ml TNF-α, 20 ng/ml IL-6, 5 μg/ml anti–IL-12, and 5 μg/ml anti–IL-4 for 5 days, further mixed with Macrophage and NK cells isolated from TILs. We nominated this cell mixture as IL-22 (+) TILs, and preparing for in vitro and in vivo assays.

All animal experiments were performed in accordance with the guidelines of the Animal Care Committee, Nanjing Medical University. Immunodeficient nude mice (5–6 weeks of age) were purchased from Charles River Laboratories, China. All operation concerning animal was performed according to ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. Three groups, with 5 mice each, were injected subcutaneously with three cell mixtures respectively: Hct-116 cells alone, and Hct-116 cells plus TILs obtained from each of two CC patients (ratio 1000:1), respectively, at cell density of 3 × 10⁶ cells in 300 μl saline solution. Animals were sacrificed 6 weeks after transplantation, and the animals were monitored regularly for tumor occurrence throughout the entire experiment period.

Reverse transcription reactions were performed using the SuperScript First-Strand Synthesis System (Invitrogen, CA), and the RNA templates were treated with DNase to avoid genomic DNA contamination. To determine the relative level of cDNA in the reverse transcribed samples, real-time PCR analyses were performed using an Applied Biosystems 7300 Detection System (Applied Biosystems, CA). The primers sequence were Forward Primer: GCTTGACAAGTCCAACTTCCA, Reverse Primer: GCTCACTCATACTGACTCCGTG, with 140 bp amplification length for human IL-22, and Forward Primer: AAGGTGAAGGTCGGAGTCAAC, Reverse Primer: GGGGTCATTGATGGCAACAATA, with 102 bp amplification length for human GAPDH. The primers were synthesized by Genscript Inc. (Nanjing, China). Real-time PCR reactions were performed in accordance with the instructions of the SYBR® Premix Ex Taq™ kit. (Takara, Japan). Data were normalized with the GAPDH levels in the samples.

Proteins were extracted from cell and mouse tissues and quantified using a protein assay (Bio-Rad Laboratories, CA). Protein samples (30 μg) were fractionated by SDS-PAGE and transferred to a nitrocellulose membrane. Immunoblotting was conducted using antibodies against IL-22, IL-22RA1, total STAT3, p-STAT3(S727), Bcl-xl, CyclinD1, and VEGF (all purchased from Abcam Inc, MA). The results were visualized via a chemiluminescent detection system (Pierce ECL Substrate Western blot detection system, Thermo Scientific, IL) and exposure to autoradiography film (Kodak XAR film).

All the tissues were removed and fixed in 4% paraformaldehyde overnight at 4°C, processed, and sectioned at 5 μm thickness. The sectioned slides were stained immunohistochemically for IL-22, IL-22RA1, IL-23, CEA, and p-Stat3 (S727) (all purchased from Abcam Inc., MA) using techniques described previously [³⁴].

The peripheral lymphocytes and TILs were stimulated at 37°C for 5 hours with Leukocyte Activation Cocktail (BD Pharmingen). Thereafter, cells were stained with surface markers, fixed and permeabilized with IntraPre Reagent (Beckman Coulter), and finally stained with intracellular markers. Data were acquired on FACSVantage SE and analyzed with CellQuest software. The fluorochrome-conjugated mAbs against CD3, CD4, CD8, CD14, CD16, and CD56 are purchased from BD Pharmingen (CA).

Hct-116, Hct-116 + TILs, Hct-116 + IL-22(+)TILs, Hct-116 + IL-22(+)TILs + IL-22mAb, Hct-116 + IL-22(+)TILs + IL-6mAb (antibodies was all purchased from Abcam Inc., MA. 10 μg/ml anti-IL-22 or anti-IL-6 3 hours prior to the assay) and Hct-116 + IL-22(+)TILs + WP1066 (EMD Chemical USA, MA, 5uM for the concentration of WP1066) were cultured on coverslips such that they were rapidly dividing. Cells were incubated with 100 μM BrdU for 2 hours, the medium aspirated, and immediately fixed and permeabilized in cold methanol:acetone (1:1), then blocked with PBS/3% BSA (Sigma-Aldrich, MO), and incubated with primary BrdU monoclonal antibody (Sigma-Aldrich) diluted in 3% BSA/PBST (0.2% Triton X-100). Following incubation with rabbit anti-mouse FITC-conjugated secondary antibody (Sigma-Aldrich), slides were mounted in Mounting Medium (with 1.5 μg/ml DAPI) (Santa Cruz), and visualized with a fluorescense microscope (Axiovert 200; Zeiss, Stuttgart, Germany).

Hct-116, Hct-116 + TILs, Hct-116 + IL-22(+)TILs, Hct-116 + IL-22(+)TILs + IL-22mAb, Hct-116 + IL-22(+)TILs + IL-6mAb and Hct-116 + IL-22(+)TILs + WP1066 (same treatment to BrdU cooperation assay) were cultured with complete medium with 0.5 mM peroxide overnight. Next, the cells were harvested and fixed with 70% cold EtOH at −20°c overnight, then further analyzed by flow cytometry (FACSCaliburTM; BD Biosciences, NJ) using a PI/Annexin staining kit (Invitrogen, CA).

The results are expressed as mean ± SD. Comparisons between two groups were performed using the student’s t-test or the Mann–Whitney U test, as appropriate. All statistical analyses were performed using SPSS statistical software (version 13.0), and two-tailed t-tests were applied to all data unless otherwise specified, with P < 0.05 considered to represent a statistically significant result.

In order to investigate the expression of IL-22 in human colon cancer, tumor infiltrated leukocytes (TILs), which are the principal source of IL-22, were isolated from excised fresh tumor tissues. The total RNA was extracted from the TILs followed by the cDNA synthesis. Significant high expression of IL-22 were detected by real-time PCR in the TILs samples compared to peripheral blood mono-nuclear cells (PBMCs) (P = 0.00618, P < 0.01, by unpaired t-test) (Figure 1A). However, no significant difference was found between IL-22 expression and varied groups of clinical characteristics for colon cancer patients (Table 1). The distribution and localization of IL-22 proteins was verified by IHC in 82 cases of CC tissues, 40 cases of normal colon tissues, and 40 cases of UC tissues in which there was potential chronic inflammation for CC. IL-22 existed in the peri-tumor or non-parenchymal tissues around the intestinal gland area in most CC tissues (71 positive in a total of 82 cases, 71/82) and UC tissues (31/40), whereas normal colon tissue samples were almost all negative (5/40) (Figure 1B), which indicated that IL-22 acted as a driver of inflammation based on previous reports in IBD patients [³⁵]. Similar to the findings of real-time PCR analysis, all slides evaluated by software “Image-plus pro” (Ver. 5.0) also exhibited a significantly higher positive rate in tissues of both CC and UC compared to normal colon tissues (CC vs. Normal P = 0.00021, P < 0.01, UC vs. Normal P = 0.0062, P < 0.01, and CC vs. UC P = 0.0075, P < 0.01, by unpaired t-test) (Figure 1C). Furthermore, high expression of IL-22 in CC patients was also confirmed by flow cytometry (TILs vs. PBMC P = 0.00068, by paired t-test) (Figure 1D1, D2).

However, TILs are white blood cells originated from the blood stream, they are composed of various of inflammatory cells, including lymphocytes, macrophages, natural killer cells (NK) etc. Therefore, in order to identify the phenotypic features of cells expressing IL-22 in colon cancer, IL-22 was co-stained with other cellular surface markers such as CD3, CD4, CD8, CD68, and CD3 + CD16 + CD56 in 32 cases of TILs isolated from human colon cancer tissues. This finding indicated that both CD4 and CD8 positive lymphocytes expressed IL-22 (4.80 ± 1.81% for CD4 positive cells and 4.01 ± 1.72% for CD8 positive cells), and a small portion of NK cells were positive for IL-22 (3.01 ± 1.29%%). A small population of macrophages (CD68 positive cells, 3.92 ± 1.12%) unexpectedly secreted IL-22 (Figure 2). We sought to isolate IL-22 expressing TILs for further functional studies in accordance with the distribution of IL-22 in TILs. Lymphocytes, macrophages, and NK cells were isolated by flow cytometry respectively. Lymphocytes were maintained in vitro after polarizing stimulation to Th22 [³³], and then further mixed with macrophages and NK cells into a cell mixture, we nominated this cell mixture as IL-22(+) TILs.

Because IL-22 is one of the hallmark cytokines secreted by Th17 cells, we analyzed three essential cytokines in the microenvironment that recruit and induce Th17 proliferation. Numerous studies have verified the level of IL-1β and TNF-α in human colon cancer [^20-23]. Our findings indicated that expression of IL-23, which has been rarely reported previously, was detected by IHC analysis in both CC and UC. These results indicated that significant upregulation of IL-23 was detected in both CC and UC (CC vs. Normal control, P = 0.0027, P < 0.01, and UC vs. Normal control, P = 0.011, P < 0.05, CC vs. UC, P = 0.017, P < 0.05, by Mann–Whitney U test), whereas relatively weak expression was detected in only a small portion of normal colon tissue samples (4/40) (Figure 3A1, B1-B3). The positive region of IL-23 in CC and UC was confined mostly to tumor and intestinal epithelial cells. In addition to IL-23, we identified another key molecule, IL-22RA1, which is necessary for signal transmission; the localization was similar to IL-23 and it was identically overexpressed in tumor and intestinal epithelial cells of UC (CC vs. Normal control, P = 0.0044, P < 0.01, and UC vs. Normal control, P = 0.0084, P < 0.01, CC vs. UC, P = 0.0072, P < 0.01, by Mann–Whitney U test), whereas it was nearly negative in normal colon tissues (7/40)(Figure A2, B4-B6). As the downstream effects of IL-22, activation of STAT3 was accessed by staining with phosphorylated STAT3 at the residue of S727. STAT3 was activated in both UC (34/40) and CC (68/82), which is a significantly higher rate compared to its activation in normal colon tissue (3/40) (CC vs. Normal control, P = 0.0051, P < 0.01, and UC vs. Normal control, P = 0.0081, P < 0.01, CC vs. UC, P = 0.022, P < 0.05, by Mann–Whitney U test) (Figure 3A3, B7-B9).

To investigate the effect of IL-22 on tumor growth and metastasis, we designed an in vivo tumorigenesis assay using the subcutaneous cell transplantation model. Hct-116, a colon cancer cell line, was co-transplanted with TILs isolated from two CC patients with different IL-22 expression levels according to analysis by FCMs (Figure 4A). Cell suspensions consisting of a 1:1000 cell mixture ratio (TILs: Hct-116) were injected subcutaneously into BALB/c nude mice. TILs from two CC patients promoted tumor growth and metastasis. Significant increase in tumor volume was evident when Hct-116 cells were co-transplanted with TILs compared to Hct-116 cells alone; moreover, our findings indicated that elevated IL-22 expression was correlated to enhanced tumor volume (TILs1, 1.63 ± 0.23 cm³ vs. Hct-116, 0.34 ± 0.19 cm³, P = 0.0019, P < 0.01; TILs2, 2.01 ± 0.30 cm³ vs. Hct-116, P = 0.0048, P < 0.01, TILs1 vs. TILs2, P = 0.022, P < 0.05 by unpaired t-test; Figure 4B1 and B2). Tumor tissue was analyzed by IHC with IL-22 staining, it exhibited that IL-22 expressing TILs isolated from two patients have proliferated underneath the mice skin (Figure 4C1, C2). Increased phosphorylation of STAT3 (S727) in tumor tissues was detected; upregulation of cyclinD1 was also found to be the effect of the activation of STAT3 signaling, which potentially explains the tumor growth enhanced by TILs. BCL-XL, an anti-apoptosis gene, was similarly elevated in the tumor tissues with co-transplanted with TILs, which were also downstream transcripts of STAT3 activation and potentially explains the tumor growth enhancement (Figure 4D).

We investigated metastasis by analyzing the lymph nodes that emerged around the tumor tissue (2/6 in the TILs1 group and 4/6 in the TILs2). We confirmed that Hct-116 invaded the lymph nodes with IHC staining for carcino-embryonic antigen (CEA), which is negative in normal lymph nodes (Figure 4E). No visceral metastasis was found in all groups. Based on this result, we analyzed VEGF expression in the tumor tissues, and our findings indicated that VEGF expression was elevated in correlation with TILs (Figure 4D); increased VEGF expression is also a downstream target of STAT3 activation.

To further investigate the mechanism of IL-22 on tumor growth, we utilized in vitro studies (Figure 5A). Hct-116 cells were co-cultured with TILs and IL-22(+)TILs. In order to exclude the effect of IL-22 secreted by IL-22(+)TILs, IL-22 was blocked, moreover, the effect of TILs secreted IL-6 which can induce activation of STAT3 was blocked as well, in addition, the expression of both IL-22 and IL-6 from TILs and IL-22(+)TILs were verified by ELISA in the supernatant of their cell culture (Additional file 1: Figure S1). The effects of IL-22 on tumor cell proliferation and anti-apoptosis were then investigated. The proliferation capability of each group was assessed by the BrdU cooperation assay. Both TILs and IL-22(+)TILs significantly enhanced the proliferation ability of Hct-116 cells (Hct-116 + TILs vs. Hct-116, P = 0.026, P < 0.05; Hct-116+ IL-22(+)TILs vs. Hct-116, P = 0.0072, P < 0.01, by unpaired t-test) (Figure 5B1-B3, B8); however, when IL-22 and IL-6 were blocked, the percentage of proliferating cells (BrdU positive cells) decreased dramatically (Figure 5B4, B5 and B8). Additionally, CyclinD1 expression also decreased in correlation with fewer dividing cells (Figure 5D). IL-22(+)TILs and TILs also significantly enhanced the anti-apoptosis ability of Hct-116 cells, which decreased the percentage of apoptotic cells significantly induced by peroxide in comparison to Hct-116 cells (Hct-116 + TILs vs. Hct-116, P = 0.032, P < 0.05; Hct-116+ IL-22(+)TILs vs. Hct-116, P = 0.0086, P < 0.01, by unpaired t-test) . However, when the effect of IL-22 and IL-6 were blocked by their antibody, apoptosis increased (Figure 5C). To explain this, the anti-apoptosis gene BCL-XL increased with the existence of TILs and IL-22(+)TILs, and decreased when IL-22 was blocked (Figure 5D). To further confirm the effect of IL-22 is depend on STAT3 activation, WP1066 which is a STAT3 specific inhibitor was used, decreased proliferation and increased apoptosis were investigated even with the existence of IL-22(+)TILs (WP1106 vs. Hct-116 for BrdU cooperation assay P = 0.0031, P < 0.01, and for apoptosis assay, P = 0.0047, P < 0.01, by unpaired t-test) (Figure 5B6, B8 and C). Evidently, all of these changes were originated from the activation or blocking of STAT3 signaling. Therefore, in this study, we introduced a STAT3 activator, IL-22, that exhibits a tumor enhancement mechanism similar to IL-6.

ND: not determined.

* by the student's t-test or the Mann–Whitney U test.