Intratumoural expression of IL-6/STAT3, IL-17 and FOXP3 immune cells in the immunosuppressive tumour microenvironment of colorectal cancer Immune cells-positive for IL-6, STAT3, IL-17 and FOXP3 and colorectal cancer development

Abstract Immune cells in the tumour microenvironment (TME) interact with each other and with tumour cells to promote tumour development. IL-6/STAT3 pathway induces and maintains mainly pro-tumour TME. Macrophages and lymphocytes are positive for IL-6, STAT3 and IL-17, while FoxP3 cells are mainly regulatory cells. IL-17+ and FoxP3+ immune cells have impact on gut inflammation and tumourigenesis. The aim of this study was to determine IL-6+, STAT3+, IL-17+ and FoxP3+ immune cells in colorectal cancer (CRC) TME and explore their association with clinicopathological parameters and mismatch repair (MMR) status. The investigated samples were collected from 104 CRC patients. Immunohistochemistry for the aforementioned markers and microsatellite instability (MSI) markers was performed. MSI testing was used. The investigated immune cells were significantly more in the invasive front (IF) as compared to tumour stroma (TS). IL-6+ and STAT3+ immune cells were more in the early tumour stages as compared to advanced stages. IL-17+ and IL-6+ immune cells were more in well and moderately differentiated cancers. IL-6+, STAT3+ and IL-17+ immune cells prevailed in the TME in microsatellite stable patients and only FoxP3+ cells were fewer there. Higher numbers of STAT3+ cells correlated with longer survival. These results support the suggestion that the transition of normal colonic mucosa to CRC is marked by a shift of Th programme, leading to accumulation of Th17 cells and Tregs that sustain tumour cell growth through the IL-6/STAT3 signalling pathway.


Introduction
The development of colorectal cancer (CRC) is a result of the interactions between transformed cells and the host immune system [1]. Cancer growth is promoted by cancer cell tumour microenvironment (TME).
The immune system in the gut during CRC formation involves some specialized cell types that determine the TME there. Resident macrophages secrete retinoic acid (RA), inerleukin 10 (IL-10) and transforming growth factor-beta (TGF-β), which maintain immunosuppression and T regulatory cells (Tregs) presence. On the one hand, the IL-6/STAT3 signalling pathway is induced there to support tumour promotion [2]. On the other hand, IL-6 secretion drives Th17 cell polarization [3], which is prognostic of cancer development [4][5][6].
Th17 cell development requires two major cytokines, IL-6 and TGF-β. STAT3 activation is necessary for IL-17 production by CD4 + T lymphocytes [15,18,25]. IL-17 activates STAT3 to enhance tumour cell ability to resist apoptosis and to promote angiogenesis and neutrophil recruitment [26,27]. Therefore, IL-17 is mainly a pro-tumour acting cytokine [27,28]. The activities of IL-17 are studied mainly in cancer models in mice [1,29]. The first description of IL-17 protein expression in immune cells of human breast cancer tissue is largely restricted to macrophages [30]. The authors reveal that > 98% of IL-17-positive immune cells in tumour stroma are macrophages as defined by cytological criteria and by CD68 staining [30]. In addition to Th17 cells, other cells also have the ability to secrete IL-17 including CD8 + T cytotoxic cells (Tc17) [31], γδ T cells [32], natural killer T (NKT) cells [33], and some immune cells such as neutrophils and monocytes [34].
The transcription factor (TF) forkhead box P3 (FoxP3) is a master control gene for the development and function of Treg cells in mice and humans [35]. Naïve mouse CD4 + T cells convert into FoxP3 + Tregs (iTregs) when stimulated with TGF-β or retinoic acid (RA) [35]. In humans, iTregs are determined also as Tr1/Th9 cell infiltrates and they could react with soluble cytokines such as IL-10 and TGF-β [36]. The immunosuppressive cytokine TGF-β also prevents maturation of DCs by decreasing the levels of co-stimulatory molecules. Stimulated DCs via dectin-1, convert Tregs into IL-17-producing T cells [37]. The increased numbers of Tregs in tumour tissue are associated with poor outcome in solid tumours such as pancreatic [38] and hepatocellular [39] cancers. In contrast, increased FoxP3 + cells in tumour tissue have been associated with better prognosis in CRC [40].
The majority of CRCs (70-80%) have allelic losses, chromosomal amplifications and translocations with no microsatellite instability (MSI) and are defined as microsatellite stable (MSS) [41]. About 15% of sporadic CRCs display deficiency in the DNA mismatch repair (MMR) system. MSI refers to the accumulation of mutations in repetitive DNA sequences (microsatellites) which leads to loss of tumour suppressor genes and to activation of oncogenes [36]. The MSI-low (MSI-L) and MSI-high (MSI-H) subgroups in sporadic CRCs have small insertion and deletion mutations in repetitive DNA [41]. In MSI-L CRC patients there is an inactivation of tumour suppressor genes such as adenomatous polyposis coli (APC) gene and p53 gene, and an activation of K-RAS oncogene, while in MSI-H cases there is a low level of mutation in tumour suppressor genes [42,43]. Because of the high level of DNA instability, MSI-H cases have frameshift mutations in the repeat regions of genes such as TGF-β, IGF etc. [41]. As a result of coding microsatellite mutations there appeared multiple frameshift-derived antigens that ensure high immunogenicity of MSI-H tumours [36].
In the present study, we have investigated immunohistochemically immune cells-positive for IL-6, STAT3, IL-17 and FoxP3 in the tumour stroma (TS) and the invasive front (IF) of 104 colorectal cancers investigated for MSI and look for correlations with clinical parameters, tumour stage, MSI status and survival. Therefore, the distribution of various types of immune cells within a tumour may provide useful information for patient's prognosis.

Ethics statement
An informed consent was given by all patients before operation and was approved by the Ethical Committee of the Hospital.

Study population
Biopsies were collected from patients surgically resected for CRC having all clinical data and followed up, between 1996 and 2009 at the department of Surgery, University Hospital 'Prof St Kirkovich' Stara Zagora. A total of 104 patients, 53 males and 49 females, with a mean age of 66.5 ± 9.57 years (min. 35; max. 82 years) were investigated. The median follow-up time for the patients was 90.63 months. Clinical data of patients, i.e. gender, age, localization, stage and histological data of differentiation grade, histological type [44], inflammatory infiltrate in the IF, invasion and MMR status, are presented in Table 1. The Clinical data were obtained by a surgeon (D.C.) and survival data were collected by an oncologist (M. H.). Formalin-fixed paraffin-embedded tissue samples containing tumour IF stained routinely and immunohistochemically were observed by one pathologist (M.G.).
The main clinical and histological parameters of the patients with colorectal cancers are given in Table 1

Immunohistochemistry
Formalin-fixed paraffin-embedded tumour tissue samples cut to 4 μm thickness were processed as follows: sections were dewaxed in two xylenes at 56ᵒ C for 1 h, and were rehydrated. The avidin-biotin-peroxidase complex technique was used. The primary antibodies used are given in Table 2. The detection system was EnVision®TM FLEX + System, horseradish peroxidase (HRP) K8002. The reaction was visualized by a mixture of 3,3′-diaminobenzidine (DAB) (Sigma, St. Louis MO, USA). The sections were counterstained by Mayer's hematoxylin. The method has been described previously [45]. We chose tumour tissue sections that did not contain necrosis. Negative controls were processed using phosphate buffered saline (PBS) instead of primary antibodies.
The consecutive sections obtained from the same tissue were stained with anti-IL-17 or anti-CD68 antibody.

Cell counting
STAT3 + , IL-6 + , IL-17 + and FoxP3 + immune cells were counted in the tumour stroma and in the invasive margin, on 5 fields of vision each in the areas with most intense cellular infiltrates (hot spots) at a magnification of (x 320, 0.74 mm 2 area) with a Leica DM2500 microscope (Germany).

DNA isolation
Genomic DNA was obtained after dewaxing and rehydratation of 4 − 6 μm thick paraffin sections from tumour tissue (without necrosis) and from normal colon mucosa adjacent to the tumour. Two methods were used: (i) the standard method with proteinase K digestion followed by phenol-chloroform-isoamyl alcohol extraction and absolute ethanol precipitation of DNA; and (ii) a DNA extraction kit (Macheray-Nagel, Germany), based on digestion by proteinase K followed by silica column filtration and collection of DNA.

MSI analysis
A set of five polymorphic markers, BAT26, D2S123, D5S346, D18S35 and FGA, were analysed for MSI as described before [46]. The repeated markers were amplified from both normal and tumour tissues and were analysed via an automated fluorescence sequencer (ABI3130XI, Applied Biosystems) defined by the appearance of different alleles in the tumour DNA when compared to the corresponding normal DNA.
When two or more markers demonstrated to be unstable, the specimen was considered to be MSI-H; when only one marker was unstable the probe was considered to be MSI-L; and when no markers were unstable, the probe was considered to be MSS defined in the literature [47,48]. Imunihistochemistry for MMR proteins. MSI was evaluated as previously [49][50][51]. Probes that lack nuclear staining of lymphocytes, some stromal cells, or the nuclei of adjacent normal epithelial cells were considered uninformative [49,50]. A positive MSI screening status is evaluated when tumour cell nuclei lacked one or more of the proteins MLH1, MSH2, MSH6 and PMS2. A negative MSS status is defined when tumour cell nuclei are positive for positive for all four markers.

The immunohistochemistry interpretation
When there was nuclear staining, even patchy one, the probe was interpreted as positive. When there was absolute loss of nuclear staining, the probe was considered negative. In the present study, the expression of proteins was grouped as follows: (i) no loss of nuclear staining in tumour cells -MSS; (ii) loss of nuclear staining for at least one protein in tumour cells -MSI-L; (iii) loss of nuclear staining of two or more proteins in tumour cells -MSI-H [48,50]. We compared the data from the MSI analysis and immunohistochemistry for MMR proteins and established similarity in almost all probes (data not shown).

Double immunofluorescence for detection of IL-17/STAT3 and CD3/IL-6
The double staining was conducted as follows: the primary antibody cocktail against IL-17/STAT3 and CD3/ IL-6 were incubated for 1 h at room temperature. After that the sections were rinsed in PBS and incubated in a mixture of secondary antibodies, anti-rabbit IgG TRITC (T5268, Sigma Aldrich Inc.) and anti-mouse IgG FITC (F9006, Sigma Aldrich Inc.) for 1 h in a dark chamber at room temperature. At the end, the slides were washed in PBS, mounted in PBS/glycerol and observed under a microscope Leica DM2500 (Germany), as described earlier [52].

Statistical analysis
Statistical analysis was performed using SPSS, 16

IL-6-positive immune cells
Having in mind previous reports, IL-6 is secreted by immune cells in TME mainly by macrophages, lymphocytes, γδ T cells and NK cells [5,14]. IL-6 positivity was observed in the cytoplasm of immune cells having characteristics mainly of macrophages and lymphocytes as proved by co-localisation of IL-6/CD68 (not shown) and IL-6/CD3 ( Figure 3).

Th17-positive (Th17) immune cells
IL-17 immunoreactivity is found in cancer cell cytoplasm, in the cytoplasm of some normal epithelial cells and in the cytoplasm of immune cells with morphological characteristics of macrophages and lymphocytes [6,53,54]. In our study, Th17 + immune cells were co-localised on serial sections with CD68 ( Figure 4) The number of Th17 + immune cells was higher in the IF as compared to TS (14.67 ± 1.15 cells/mm 2 vs. 9.14 ± 1.39 cells/mm 2 , p < 0.0001), (Figure 1).

Association of tumour stages (I + II and III + IV) with IL-6 + and STAT3 + immune cells
The number of STAT3 + immune cells in the IF was higher in the early stages (I + II) as compared to the same type of cells in the advanced stages (III + IV) (21.40 ± 2.96 cells/mm 2 vs. 9.06 ± 2.25 cells/mm 2 , p = 0.012). Similarly, although without statistical significance, IL-6 + immune cells had higher density in the IF of tumours of patients with early stage (I + II) as compared to those with advanced stage (III + IV) (35.28 ± 3.93 cells/mm 2 vs. 29.27 ± 5.61 cells/mm 2 , p = 0.550) ( Figure 5).    When STAT3 + and IL-6 + immune cells were analysed in the TS, we obtained analogous findings. The number of STAT3 + and IL-6 + immune cells was higher in early stages (I + II) as compared to the advanced stages (III + IV) ( Figure 5). These data show that STAT3 + and IL-6 + immune cells are more abundant in early stages as compared to advanced stages of CRC.

Associations with the grade of differentiation
most of the tumours were with moderate or high differentiation (80.8%), while only 19.2% of the tumours were with low differentiation or gelatinous (Table 1). When we compared the cell counts labelled with the studied markers, we found that the number of IL-17 + cells was significantly higher in the IF of cancers with moderate or high differentiation grade than in the IF of cancers with more unfavourable differentiation grade (low differentiation or gelatinous) (p = 0.003) ( Figure 6). Similar findings, but without reaching statistical significance, were obtained for IL-6 + cells in the IF (p = 0.096) and for IL-6 + and IL-17 + cells in tumour stroma (p = 0.119 and p = 0.133, respectively) ( Figure 6). These results are in agreement with the previous results in Figure 5, i.e. IL-17 + and IL-6 + cells were more abundant in well and moderate differentiated cancers.

Associations with patients' survival
The patients with higher STAT3 cell density (above the median of 12.69 cells/mm 2 ) in the IF tended to survive Figure 5. comparison of the density of Stat3 + and il-6 + immune cells between the tumours of patients with early stages (i + ii) and those with advanced stages (iii + iV) of cRc. the data are presented as means ± Sem (standard error of the mean), and the p-values were obtained by mann-Whitney u test.
longer as compared to the patients with lower STAT3 cell density in the IF (mean of 105 ± 8.425 months vs. 97.54 ± 11.415 months, χ 2 =3.38, p = 0.066, Log rank test) ( Figure 7A). By analogy, the higher density of IL-6 + cells in IF (above the median of 26.53 cells/mm 2 ) was associated, although not significantly, with longer survival after the surgical therapy of the patients with comparison to the lower number of such labelled cells (mean of 125.28 ± 10.44 months vs. 83.68 ± 8.76 months, p = 0.102, Log rank test) ( Figure 7B).

Correlations with the MSS/MSI status
There was a marginal significance for a higher number of IL-6 + immune cells in the IF of tumours of MSS patients as compared to MSI patients (40.43 ± 5.20 cells/mm 2 vs. 32.83 ± 4.19 cells/mm 2 , p = 0.053). In addition, the number of IL-17 + immune cells was significantly higher in the IF in MSS patients as compared to MSI patients (17.45 ± 2.00 cells/mm 2 vs. 13.38 ± 1.38 cells/mm 2 , p = 0.039). the STAT3 + immune cells in the   TS showed a similar trend but it was statistically non-significant: they were more abundant in MSS patients as compared to MSI patients (15.53 ± 3.56 cells/mm 2 vs. 11.99 ± 1.73 cells/mm 2 , p = 0.344). Only the FoxP3 + cells were fewer in the TS of MSS patients compared with those with MSI tumours (8.25 ± 1.63 cells/mm 2 vs. 10.61 ± 1.51 cells/mm 2 , p = 0.121) (Figure 8).

Discussion
To our knowledge, our study is the first to access the number of immune cells positive for STAT3, IL-6, IL-17 and FoxP3 in a series of 104 sporadic CRCs and discuss their influence on the tumour development. There is interest in identifying a molecular target of prognostic importance for CRC. The IL-6, IL-17 and FoxP3 markers of immunosuppression have been previously associated with CRC patients' prognosis and tumour progression [7,12,36]. The investigated immune markers were assessed quantitatively by an experienced pathologist, so we avoided mistakes in recognizing types of immune cells, irregular staining density and chose the appropriate areas for cell counting in the IF and TS without necrosis, mucin pools or irrelevant tumour mass and invasive margin.
STAT3 + immune cells have been less investigated [56]. TILs secrete multiple cytokines such as IL-11, IL-6, IL-22 etc. that trigger STAT3 activation in tumour cells [1]. The activation of STAT3 and NF-κB signalling pathways is detected in both tumour cells and in tumour infiltrating immune cells in early colonic lesions of APC min/+ mice and can regulate the behaviour of immune cells in the tumour [57]. The STAT3 + immune cells control the secretion of IL-17, IL-6 and TNF-α [56]. We found an increase of STAT3 + immune cells in early stages (I + II) in CRC and a positive correlation between STAT3 + immune cells and IL-17 + immune cells in the IF as reported by other authors [57]. Moreover, STAT3 mRNA and STAT3 protein expression are found to be markedly increased in invasive CRCs (with lymph node metastases) of 32 patients [58].
For the first time we show an association between more STAT3 + immune cells and longer patients' survival. In one study, a positive correlation between STAT1 and STAT3 expression in tumour cells in 414 CRCs by immunohistochemistry is correlated with better prognosis [59]. STAT3 activity has been associated with favourable clinical prognosis in some other cancers such as breast cancer [60] and prostate cancer [61]. In contrast, p-STAT3 expression is indicative for poor prognosis in gastric cancer [62]. It was supposed that genetic and environmental factors might contribute to pro-tumour or anti-tumour role of STAT3 and to its association with disease progression [59].
It is known that IL-6 is secreted mainly from cancer associated fibroblasts (CAFs) and tumour cells in CRC [63]. IL-6-positive immune cells are already demonstrated in CRC cancer tissue [64]. Our investigation shows that IL-6 + immune cells are increased in the IF as compared to TS, similarly to all other investigated immune cells.
The association between IL-6 expression and tumour staging shows that IL-6 + immune cells are significantly increased in early tumour stages of CRC as compared to advanced stages. In CRC patients flow cytometry has shown that TILs secrete increased levels of IL-6 and other tumour-promoting cytokines, such as IL-7, IL-21, IL-22, TNF-α, that trigger STAT3 and NF-κB in tumour cells [1]. Although, IL-6 or its receptor sIL-6R positivity in tumour cells is associated with CRC disease progression [5], to our knowledge no investigation has been made of IL-6 + immune cells in the tumour and its development. IL-6 + cytoplasmic expression in CD45 + TILs, in tumour cells and in Peyer's patches and mesenteric lymph nodes has been reported [2]. The authors, like us, concluded that IL-6 is significantly expressed in the immune cells. Tumour cells and stromal cells (TAMs, MDSCs, CD4 + T cells and CAFs) secrete IL-6. Through a trans-signalling mechanism -the binding of IL-6 to sIL-6R and to gp130 activates down-stream signalling molecules that drive pro-tumourigenic activities in the cells in thr TME [25].
In the presence of IL-6 secreted by TAMs, CAFs and some T cells, Tregs can convert into Th17 cells and produce IL-17 [63,65]. Th17 cell development is induced by a combination of IL-6 and TGF-β, and by the IL-12 family member IL-23 [16]. STAT3 is a key TF in the activation of Th17 [10,66]. Although IL-17producing cells have been described as increased in number in certain tumours, such as gastric cancer [6], ovarian cancer [67] and prostate cancer [34], it is controversial whether IL-17 promotes or inhibits cancer progression [68].
We identified the distribution of IL-17-producing cells in various zones in radically resected CRC specimens. Th17 + lymphocytes are derived from CD4 + naïve T cells in lymph nodes [34,69]. In addition, IL-17 positivity has been observed mainly by macrophages [29,30]. In our study, we like others [30], co-localised IL-17producing cells with CD68 + cells and found that the majority of them are macrophages. IL-17 + cells were statistically significantly increased in number in the IF of the tumour. Using immunohistochemistry, some authors have shown increased numbers of IL-17 + immune cells in tumour tissues in CRC as compared to normal tissue [54]. Contrary to our study, flow cytometry of peripheral blood has shown that IL-17 + cells are increased in advanced stages (III + IV) but the investigation was done in a quite small number of patients (n = 40) [54]. Our study included a larger number of patients (n = 104) and associations with a lot of clinicopathological data, a long follow -up period, and many immune cell counts. No significance of Th17 + cells for tumour progression in 20 patients has been observed by others [53].
The present study provides information of increased FoxP3 immune cells in IF of the tumours as compared to TS. FoxP3 + cells were fewer in TS of MSS patients as compared to MSI ones. Some authors have shown better tumour characteristics when high FoxP3 numbers in CRC in a large cohort of MSS and MSI CRC patients are investigated [70]. Moreover, significantly higher FoxP3 + immune cells in MSI-H CRCs are demonstrated that parallel with enhanced number of cytotoxic CD8 + T cells that impact the effective immune response [71]. FoxP3 + Tregs are considered to be immunosuppressive and have been associated with poor outcome in several malignancies such as pancreatic cancer [38,72] and hepatocellular cancer [39] or with better survival in CRC [70,73]. We detected that the number of FoxP3 + Tregs was increased in TME of MSS CRCs by immunohistochemistry. Some authors [36], have observed higher expression levels of FoxP3 mRNA in MSS patients. Using RT-PCR gene expression analysis of FoxP3, IL-10 and TGF-β in 65 CRC patients, it has been shown that Foxp3 mRNA is increased in CRC tumour tissue as compared to normal one from early to advanced stages [74]. By immunohistochemistry, highest levels of FoxP3 in the cytoplasm of tumour cells have been found in early CRC stages, where FoxP3 + immune cells are decreased (an inverse correlation) [75].
We conducted this study in order to examine the immunosuppressive immune cells in CRC and their presence in association with MSI/MSS status. A deficient DNA mismatch repair system is observed in 15% of CRC. The persistence of mismatch mutations is mainly in microsatellites [76]. Colorectal cancers having MSI-H possess a distinct immunologic phenotype and histological appearance [77]. The MSI-H phenotype has been associated with Crohn's like reaction [78], with elevated number of CD8 + T lymphocytes [79], and with elevated number of FoxP3 + T helpers [72]. On the other hand, some authors have shown increased expression levels of mRNA for FoxP3, IL-17, IL-1β, IL-6 and TGF-β in MSS CRC patients as compared to MSI CRC patients [36]. Increased numbers of FoxP3 + Tregs using immunohistochemistry in CRC has been reported in MSS patients [70]. Moreover, it has been reported that the density of FoxP3 + Tregs is similar in MSI and MSS probes [80]. The decreased numbers of FoxP3 + Tregs in CRC tissue could be associated with blocking of tumour-promoting inflammation [81].
The transition of normal colonic mucosa to CRC is marked by a shift of Th programme, leading to accumulation of Th17 cells and Tregs that sustain tumour cell growth through the action of the IL-6/STAT3 signalling pathway. The complexity of the immune response in CRC is not well understood concerning the immune interactions between tumour antigens and the cells of the innate and adaptive immune system. Further investigations are necessary to elucidate the complexity of TME for tumour development.

Conclusions
The studied immune cells (STAT3, IL-6, IL-17 and FoxP3) were statistically significantly more abundant in the IF of CRC as compared to TS. Therefore, the immunosuppressive immune cells prevailed in tumour margins of CRC supporting tumour development. Concerning tumour stages, we found that the IL-6 + and STAT3 + immune cell counts increased in the early stages, implying that the immunosuppressive TME presents even from the beginning of tumourigenesis. The MSS CRCs contained fewer FoxP3 + Tregs. Only the STAT3 + and IL-6 + immune cell counts correlated with longer survival.

Availability of data
The data that support the findings of this study are available from the corresponding author (MG), upon reasonable request.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was financially supported by the National Science Fund, Bulgarian, Research grant number KP-06-H23/2 from 17.12.2018