Interleukin 11 is upregulated in uterine lavage and endometrial cancer cells in women with endometrial carcinoma

Endometrial cancer tissue biopsies (N = 16) were collected from postmenopausal women undergoing total abdominal hysterectomy for endometrial carcinoma at the Monash Medical Centre Melbourne, Australia. The Human Ethics Committee approved the research project and informed consent was obtained from each patient participating in this study. Details of individual patients are provided in Table 1. All tissues were examined and tumors were graded histologically by a specialist gynecological pathologist according to the guidelines of the International Federation of Gynecology and Obstetrics (FIGO, 1998). In this system the presence of vascular/lymphatic invasion was noted and the depth of myometrial invasion was classified as either: no invasion, < 50% myometrial invasion or > 50% myometrial invasion. Biopsies of endometrium were also obtained from postmenopausal women (N = 4) undergoing minor gynaecological procedures unrelated to endometrial pathology. Histological examination by a specialist gynecological pathologist confirmed whether the endometrium from post-menopausal women was atrophic or active. A large majority of endometrium collected from postmenopausal women are atrophic as they are no longer under hormonal control so that with very little endometrium is present, thus we were very limited in the number of tissue samples of 'active' endometrium we could include in this study. We therefore also collected endometrium from proliferative phase endometrium (N = 10) as a second group of non-tumour normal controls. The control proliferative endometrial biopsies were collected at curettage from women between day 7 and 13 of their menstrual cycle that were scheduled for tubal ligation or were undergoing testing for tubal patency. Tissues were assessed by a pathologist and had no obvious endometrial pathology. The women had no steroid treatment or other medication for at least 2-3 months before the collection of tissue. Written informed consent was obtained from each patient and the study was approved by the Southern Health Human Research and Ethics committee. All endometrial biopsies were fixed overnight in 4% neutral buffered formalin, prior to routine paraffin embedding.

Uterine lavages (uterine washings) were collected from postmenopausal women (N = 4) and women with endometrial cancer above except for women with Grade 1 carcinoma where washings were collected from 4 women instead of 5 women (Grade 1-3 (N = 4; 5; 6 respectively) as previously described [24]. Uterine fluid from all women was concentrated 3-4 fold using Nanosep microconcentration devices with a 3 K cut-off (Pall Life Sciences, East Hills, NY). IL11 was then measured in the samples by ELISA as previously described [25].

Immunohistochemistry for IL11 and IL11Rα was performed as described previously [25] using a monoclonal anti-huIL-11 (5E3) and antihuIL-11Rα (4D12) antibodies (generous gifts from Dr. Lorraine Robb). Briefly, paraffin sections (5 μm) were dewaxed in histosol and rehydrated in a graded series of ethanol. Endogenous peroxidase activity was quenched by immersion in 3% H₂O₂ in methanol for 10 min. Non-specific staining was blocked using a blocking solution of 10% normal horse serum (Sigma-Aldrich Inc., Missouri, USA) and 2% normal human serum, diluted in 1× Tris-buffered saline (TBS) for 30 min. Primary antibodies were diluted to 4 μg/ml in blocking solution and applied for 18 h at 4°C. A non-immune isotype IgG negative control (R&D Systems Inc., Minneapolis, MN, USA) diluted to a matching concentration as the primary antibody, was also included for each tissue. Antibody localisation was detected by sequential application of biotinylated horse anti-mouse IgG (Vector Laboratories, Burlingame, CA, USA) diluted 1:200 in blocking solution for 30 min and an avidin-biotin complex conjugated to HRP (Vectastain ABC Elite kit; Vector Laboratories, Burlingame, CA, USA). The substrate used was diaminobenzidine (DAB) (Zymed, San Francisco, USA) forming an insoluble brown precipitate. Sections were then counterstained in Harris hematoxylin (Sigma Diagnostics, St. Louis, USA). Sections from normal endometrium were used as positive controls and included in each immunostaining run to provide quality control.

Immunohistochemistry for pSTAT3 and SOCS3 was conducted using polyclonal rabbit anti-mouse (Cell Signalling Technology Inc., MA, USA) and monoclonal rabbit anti-human (Clone C204) (Immuno-Biological Laboratories Inc., MN, USA) antibodies respectively as previously shown [9], at final concentration of 0.09 μg/ml and 1 μg/ml respectively.

Formalin fixed sections were deparaffinized in histosol and rehydrated in a graded series of ethanol. Endogenous activity was blocked by incubation in 3% H₂O₂ in methanol for 10 min. Non-specific staining was blocked using blocking solutions consisting of 10% normal swine serum (in-house) and 2% normal human serum for pSTAT3 and 10% normal goat serum (Vector Laboratories) and 2% normal human serum for SOCS3, each diluted in 1×TBS for 30 min. Primary antibodies were diluted in the appropriate blocking solution and applied for 18 h at 4°C. A non-immune isotype IgG negative control (R&D Systems) diluted to a matching concentration as the primary antibody, was also included for each tissue. Antibody localisation was detected by sequential application of biotinylated swine anti rabbit IgG (DAKO, Glostrup, Denmark) or biotinylated goat anti-rabbit IgG (Vector Laboratories) diluted 1:200 in blocking solution correspondingly for 30 min and an avidin-biotin complex conjugated to HRP (Vectastain ABC Elite kit, Vector Laboratories). The substrate used was diaminobenzidine (DAB) (Zymed), which forms an insoluble brown precipitate. Sections were then counterstained in Harris hematoxylin (Sigma Diagnostics). Sections from normal pre-menopausal endometrium were used as positive controls and included in each immunostaining run to provide quality control.

The endometrial carcinoma cells ECC-1, HEC-1A and Ishikawa cells were cultured in DMEM/F12 (1:1), McCoy's 5A and DMEM (Invitrogen, Victoria, Australia) respectively supplemented with 10% fetal calf serum (SAFC Biosciences, Victoria, Australia), 1% L-glutamine (Sigma-Aldrich Pty. Ltd) and 1% antibiotic-antimycotic (Invitrogen, Victoria, Australia). The non-cancer human endometrial epithelial (HES) cell line [26] was obtained from Dr. Douglas Kniss (Ohio State University, Columbus, OH). Cells were maintained in RPMI 1640 (Sigma-Aldrich Pty. Ltd) supplemented with 10% FCS, 1% L-glutamine and 1% antibiotic- antimycotic. Confluent cells were transferred into serum free medium for 24 hours prior to treatment

The endometrial cancer cell lines ECC-1, HEC-1A and Ishikawa and or HES cells were treated with diluents control, IL11 (1, 10, 100, 500 ng/ml) for 15 minutes or 4 hours. Phosphorylated STAT3 and total STAT3 abundance (15 min) and SOCS3 protein abundance (4 hours; treatment time was determined from previous studies in endometrial cells) were analysed by Western blot as previously described [9] and briefly as follows. Cells were grown to confluence, the medium aspirated and cells washed with ice-cold sterile PBS, twice on ice. Cells were lysed and scraped in ice-cold lysis buffer containing 50 mM Tris Base, 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 25 mM NaF, 25 mM β-glycerolphosphate, pH 7.5 and 2 μl/well protease inhibitors cocktail set III (AEBSF, HCl 100 mM, aprotinin 80 μM, bestatin 5 μM, E-64 1.5 mM, leupeptin hemisulfate 2 mM, pepstatin A 1 mM (Calbiochem, San Diego, CA, USA). Cell extracts were then centrifuged at 12000 rpm for 30 min at 4°C, and supernatant protein quantified using the BCA protein assay kit (Pierce, Rockford, IL, USA). Equal amounts of total protein were then resolved on SDS-PAGE gels and transferred to nitrocellulose membranes. All membranes were incubated with Ponceau S (Sigma) to ensure equal protein loading in all lanes. The membranes were blocked with 5% nonfat dry milk in Tris-buffered saline with 0.1% Tween (TBST) and probed separately with antibodies specific for phosphorylated STAT3 (Tyr705, Cell Signaling Technology Inc.) (1:1000), total STAT3 (Cell Signaling Technology), (1:1000) or SOCS3 (IBL Co. LTD, Gunma, Japan). The membranes were washed in TBST then incubated for 1 h with horseradish peroxidase (HRP)-conjugated rabbit secondary antibody (Dako Cytomation, Glostrup, Denmark) (1:1500). Finally, the HRP activity was detected using enhanced chemilluminescence reagent (Pierce, Rockford, IL, USA). To determine the specificity of IL-11 in the cells a specific IL11 antagonist was used (provided by Commonwealth serum Laboratories, Melbourne, Australia) [27].

All in vitro cell culture experiments were performed in two independent experiments in duplicate.

Positive staining was scored semiquantitatively by two independent observers, blind to the identity of the tissue, with an intensity score assigned as 0 (negative) to 3 (maximal staining intensity). All statistical analyses were performed using GraphPad Prism. Data was analysed by the non-parametric Kruskal-Wallis test followed by the Dunn's post-hoc test. Differences were considered significant at P < 0.05.

IL11 was detectable in 3 of 4 postmenopausal controls and in all the flushings from the Grade 1-3 tumours (Fig 1). IL11 levels in uterine flushings from women with Grade 1 cancers were higher than that of postmenopausal control women (p < 0.05) (Fig 1). Uterine flushings from proliferative phase control women ranged from very low to undetectable IL11 and were therefore not included in the results and data analysis. Similarly, an additional three uterine washings from postmenopausal women were assayed for IL11 but had undetectable IL11 and were not included in the data analysis. Overall, IL11 levels in uterine flushings in the cancer patients were higher than the postmenopausal controls although this did not reach significance (Fig 1). There was a sub-group of women with Grade 3 cancers that had very high levels of IL11 in the uterine flushings (Fig 1).

Positive immunostaining for IL11 was detected in all cancer tissues examined; meanwhile IL11Rα staining was present in all cancer tissues except two from grade 3 (Fig 2A and 2B). In all tissues, IL11 and IL11Rα immunoreactivity was mainly localised to epithelial cells of tumour origin (Fig 3A-C and 4A-C respectively). Very little immunoreactive IL11 and IL11Rα was seen in the stromal compartment of the tumours (Fig 3A-C and 4A-C respectively). By contrast, only low levels of IL11 and IL11Rα staining was evident in epithelial cells while stromal cells were negative in endometrium from postmenopausal women and proliferative phase endometrium (Fig 2A and 2B, Figure 3F and 4E). IL11 immunostaining was significantly higher in epithelial tumour cells from Grades 1 and 3 but not Grade 2 tissue compared to endometrial epithelial cells from proliferative phase women (p < 0.05 in G1 and p < 0.001 in G3, Fig 1A) but was higher only in Grade 1 compared to postmenopausal women. There was no significant difference in IL11 staining in epithelial cancer cells between the tumour grades (Fig 2A). IL11Rα staining was higher in epithelial tumour cells in Grade 1 and 2 but not Grade 3 compared to endometrial epithelial cells from control postmenopausal women (p < 0.05, Fig 2B). Similar to IL11, there was no significant difference in IL11Rα staining in epithelial cancer cells between the tumour grades (Fig 2B).

IL11 staining was also present in vascular endothelial and smooth muscle cells in Grade 3 tumours (Fig 3D-E). Similarly, positive staining for IL11Rα was seen in vascular smooth muscle and endothelial cells in Grade 3 tumours but not in Grades 1 and 2 (Fig 4D). No staining for IL11 and IL11Rα was seen in vascular endothelial and smooth muscle cells in postmenopausal endometrium or proliferative phase endometrium (data not shown). Intense staining for IL11 was seen in subpopulations of leukocytes infiltrating the cancer glands in four of the six Grade 3 tumours (Fig 3E). Secretory phase endometrium served as positive controls for IL11 and IL11Rα and demonstrated positive IL11 and IL11Rα staining in glandular epithelium as previously reported (Fig 3G and Fig 4F) [12]. No immunostaining was detected in the IL11 and IL11Rα negative controls (Fig 3H and 4G).

Staining for pSTAT3 was detected in epithelial and stromal compartments in the endometrial carcinomas and postmenopausal endometrium (Fig 5A-C and 5G). pSTAT3 immunostaining was low in the postmenopausal epithelium and proliferative phase. pSTAT3 was significantly increased in the cancer epithelium in Grades 1 and 2 compared to proliferative phase epithelium (Fig 2C). pSTAT3 staining in the tumour stroma was low to moderate but was minimal in the endometrium from postmenopaual women (Fig 5A-C and 5G). While there was an increase in pSTAT3 immunostaining intensity in the Grades 1 and 2 compared to postmenopausal epithelium, it did not reach significance (Fig 5C). There were no statistical differences in tumour stroma between cancer grades and also between each cancer grade and postmenopausal endometrium (data not shown). SOCS3 localised primarily to the endometrial cancer epithelium in all grades of carcinomas (Fig 5D-F). There was minimal staining for SOCS3 in endometrial epithelial cells from the postmenopausal women (Fig 5H). SOCS3 in proliferative phase epithelium was significantly higher compared to epithelium in post-menopausal controls and all Tumour Grades (Fig 2D). However, there were no significant difference in SOCS3 staining in the epithelial tumour cells between tumour grades (Fig 2D).

Overall, all the human endometrial cancer cell lines (ECC-1: 5 pg/10⁶ cells, HEC-1A: 3 pg/10⁶ cells, Ishikawa cells: 6 pg/10⁶ cells) and the endometrial epithelial cell line HES, secreted very low levels of IL11 under serum free conditions. The cells were subsequently cultured in serum free conditions to examine the effect of IL11 on pSTAT3/STAT3 and SOCS3 protein abundance.

The effect of IL-11 on pSTAT3 and STAT3 in human endometrial epithelial cancer cell lines was examined by Western blot (Fig 6A top). Addition of IL-11 to ECC-1 cells weakly stimulated pSTAT3 at 100 pg/ml while there was no activation with all other concentrations. By contrast, IL11 stimulated pSTAT3 from 1-10 ng/ml in HEC-1A and 1 ng/ml in Ishikawa endometrial carcinoma cells respectively compared to diluent control treated cells (Fig 6B and 6C). STAT3 protein abundance was not affected at any IL11 concentration tested in all carcinoma cell lines (Fig. 6A).

To determine the effect of IL-11 on SOCS3 protein abundance, endometrial carcinoma and non-carcinoma (HES) cells were treated with IL-11 for 4 hours and SOCS3 abundance examined at 0 (before treatment) and 4 hrs as previously described [9]. SOCS3 protein abundance in ECC-1 cells did not change with addition of IL11 (Fig. 7A). In HEC-1A (Fig. 7B) and Ishikawa carcinoma cells (Fig. 7C), there was an upregulation of SOCS3 protein following the addition of 100 ng/ml IL11 compared to respective controls. In non-carcinoma HES cells, SOCS3 protein increased after addition of IL11 from 1-500 ng/ml (Fig. 7D). Addition of IL11 antagonist with 100 ng/ml IL11 reduced SOCS3 protein compared to controls (Fig. 7D).