Evaluation of cytokine immunostaining in ovarian neoplasms and endometriomas

Overview

Objective: The objective of our study was to quantify and compare the immunostaining of IL-2, IL-5, IL-6, IL-8, and TNF-α in endometriomal tissue, non-neoplastic tumors, benign neoplasms, and malignant ovarian neoplasms. Materials and methods: The study involved 90 patients: 15 non-neoplastic ovarian lesions, 28 ovarian benign neoplasms, 28 ovarian malignant neoplasms, and 19 ovarian endometriomas were diagnosed. Immunohistochemistry was performed for cytokines IL-2, IL-5, IL-6, IL-8, and TNF-α and their concentrations were compared in these groups. Fisher‘s exact test was used, requiring a P-value of < 0.05 for significance. Results: IL-5 and IL-8 epithelial immunostaining is stronger in endometriomas than in ovarian cancer (P-values of 0.0046 and 0.0149, resp.). The stromal immunostaining of TNF-α, IL-5, IL-6, and IL8 is stronger in endometriomas than in ovarian cancer (P-values of 0.0008, < 0.0001, 0.0003, and 0.0006, resp.). Conclusions: Stronger immunostaining of some cytokines in endometriomas compared to ovarian cancer reflects an inflammatory and immune response that could be future targets for new discoveries about the infiltrative behavior of endometriosis.

Keywords:

Endometriosis – immunology – ovarian neoplasms – Cytokines

Introduction

Endometriosis is a benign dis ease of the female genital system that is characterized by the presence of ectopic endometrial tissue and may present in a diffuse manner with pelvic implants, and in a localized manner as in ovarian endometrioma. It is a chronic dis ease that is difficult to dia gnose. Literature data indicate a prevalence of 5% to 10% in women of childbearing age [1]. Ovarian endometriomas are found in 17–44% of patients dia gnosed with endometriosis [2].

Cytokines are the main mediators of the immune system. They have an immunological-regulatory role, which can be pro- or anti-inflammatory. To remain effective, the immune response requires a delicate balance between these actions. Thus, any dysregulation of cytokines becomes an important aspect in the progression of numerous pathological conditions, including endometriosis. It is known that in women with endometriosis, the peritoneal fluid contains a higher concentration of pro-inflammatory cytokines. However, further studies are needed to elucidate the role of cytokines in the mechanisms of endometriotic implants and their growth [3].

Studies report a link between endometriosis and ovarian cancer. Kok et al. (2015) showed that patients dia gnosed with endometriosis are 4-times more likely to have ovarian cancer [4]. In another study, it was shown that 19% of ovarian epithelial tumors were related to endometriosis [5].

In the search for new therapeutic possibilities, studies have been carried out on the role of immunology in the pathophysiology of endometriosis. These have led to the hypothesis of dysregulation in the immune system of women with endometriosis and resulting in failure to contain the growth of endometrial tissue in the pelvic cavity and even stimulation of angiogenesis in this tissue [3].

The objective of this study was to quantify and compare epithelial and stromal immunostaining of IL-2, IL-5, IL-6, IL-8, and TNF-a in endometriomal tissue, non-neoplastic tumors, benign neoplasms, and malignant ovarian neoplasms.

Materials and methods

Patients

Patients were evaluated and treated at the Pelvic Mass Outpatient Clinic of our service with an indication for surgical treatment according to pre-established criteria [6]. The histopathological results of paraffin-embedded tissues and the anatomopathological results were reviewed, and patients with a confirmed dia gnosis of non-neoplastic ovarian lesions (N = 15), benign neoplasias, (N = 28), malignant neoplasias (N = 28), and endometriomas (N = 19) were included in the study.

Exclusion criteria were secondary malignant ovarian neoplasms (metastasis), previous treatment, use of immunosuppressive medications, and comorbidity with dis eases that lead to immunosuppression and relapse.

The study was approved by the Research Ethics Committee of the institution (protocol 1,408) and informed consent was obtained from all participants.

Anatomopathological study

The anatomopathological evaluation and staging of the cases were performed according to the criteria of the International Federation of Gynecology and Obstetrics (FIGO) [7]. The histopathological study was performed by the Surgical Pathology Service on paraffin-embedded sections.

Immunohistochemistry study

Immunohistochemistry was performed for cytokines IL-2, IL-5, IL-6, IL-8, and TNF-a using standard techniques as described briefly below.

Slices (4 μm) of the paraffin-embedded tissues from the selected cases were adhered to silanized slides (ATPS – Silane, Sigma® A3648), and then processed using the Novolink™ Polymer Detection System (Leica Biosystems Inc.). The slides were kept in an oven at a temperature of 56 °C for a period of 24 hod, and then they were dewaxed in two xylol baths, remaining 5 min in each bath and dehydrated in three baths of absolute alcohol and one bath of 80% alcohol. The slides were then placed in a PBS bath (pH 7.2) for 5 min for hydration.

Heat-mediated antigen retrieval was then performed as fol lows. The slides were placed in cytology tubes containing a 10 mM citrate buffer solution (pH 6.0) or Tris-EDTA, according to the manufacturer‘s instructions, and placed in a distilled water bath for 30 minutes at a temperature of 100 °C. The tubes were then removed from the bath and placed on the bench for cooling to an ambient temperature of 22 °C.

The slides were placed on a rack and washed with distilled water. The slides were then wiped dry adjacent to the tissue sections while not allowing the tissue sections to dry out. The slides were placed on an incubation board and endogenous peroxidase was neutralized with a Peroxidase Blocker (3% hydrogen peroxide) applied to each tissue section for 5 min. The slides were then washed 3-times in a PBS buffer for 5 min per wash. The slides were then incubated with a Protein Blocker for 5 min and then washed 3-times in PBS buffer for 5 min per wash. Specific anti-TNF-a, anti-IL2, anti-IL5, anti-IL6, and anti-IL8 antibodies were diluted in Bovine Serum Albumin (BSA, Sigma®) and incubated with the prepared tissue sections in a humid chamber at 4 °C fol lowing the manufacturer’s specifications. These preparations of bovine albumin (2% BSA) and primary antibody serve to block nonspecific binding and reduce background color.

Positive controls were used for each antibody according to the manufacturer‘s directions. After incubation with the primary antibodies, the slides were washed twice with PBS buffer for 5 min per wash. The slides were then incubated with a post-primary antibody for 30 min and then washed twice with PBS buffer for 5 min. The slides were then incubated with a Novolink™ Polymer for 30 min and were washed twice with PBS buffer for 5 min per wash. The slides were then developed by adding chromogenic solution (DAB–diaminobenzidine), and incubating for 5 min fol lowed by a wash in running water and counterstaining with Harris’ hematoxylin. Finally, the slides were immersed in three baths of absolute alcohol for 10 s each to remove excess water, one bath of phenylated xylol, and three baths of xylol for 5 min each. Coverslips were added over the slides with Entellan® (Merck Millipore) mounting medium for further analysis.

Two observers evaluated the slides, observing immunostaining in the epithelium and stroma. The immunostaining intensity was assessed subjectively using a scale of 0–3:

• 0 (no staining);
• 1 (weak staining);
• 2 (moderate staining);
• 3 (strong staining) (Fig. 1).

Statistical analysis

The data were analyzed using the GraphPad Prism 6 software. Fisher‘s exact test was used to assess immunohistochemistry staining intensity between the tissue types. Statistical significance was considered when the P-value was ≤ 0.05. The Kappa test was performed to verify agreement among the observers. Discordant cases were reviewed together and results were obtained by consensus. Agreement between two observers was performed through kappa: k < 0.4: weak agreement; 0.4 ≤ k < 0.8: moderate agreement; 0.8 ≤ k < 1.0: strong agreement; k = 1.0: perfect agreement. The cut-off value used was immunostaining 1. Above immunostaining 1 (immunostaining 2 and 3), we consider strong immunostaining. Below or equal to 1 (immunostaining 0 and 1), we consider weak immunostaining. Sensitivity, specificity, and accuracy in endometriomas and ovarian cancer dia gnosis were calculated.

Results

The evaluated groups were: non-neoplastic ovarian lesions (N = 15), benign neoplasias, (N = 28), malignant neoplasias (N = 28), and endometriomas/endometrial cysts (N = 19).

The median age of the study participants was 46 years (35–82 years) in the non-neoplastic group, 43 years (20–61 years) in the endometrioma group, 48 years (18–69 years) in the benign neoplasm group, and 51 years (25–73 years) in the group of malignant neoplasms.

The staging of malignant neoplasms according to FIGO were: IA (nine patients; 32.14%), IB (two patients; 7.14%), IC2 (two patients; 7.14%), IIA (one patient; 3.57%), IIB (one patient; 3.57%), IIIB (one patient; 3.57%), IIIC (ten patients; 35.71%), and IVB (two patients; 7.14%). The histological grading of the malignant neoplasms was: grade 1 (11 patients; 39.28%), grade 2 (11 patients; 39.28%), and grade 3 (six patients; 21.42%).

The histological types of ovarian malignancies were: borderline mucinous tumors (nine patients; 32.14%), serous papillary cystadenocarcinomas (eight patients; 28.57%), mucinous cystadenocarcinomas (two patients; 7.14%), serous cystadenocarcinomas (two patients; 7.14%), and one patient each (3.57%) for adenocarcinoma, papillary anaplastic adenocarcinoma, endometrioid adenocarcinoma, grade III poorly differentiated adenocarcinoma, clear cell carcinoma, borderline serous tumor, and atypical borderline proliferative endometrioid tumor.

The histological types of benign ovarian neoplasms were: serous cystadenomas (16 patients; 57.14%), mucinous cystadenomas (nine patients; 32.14%), mucinous cystadenomas with Brenner‘s tumor (two patients; 7.14%), and serous papillary cystadenoma with Brenner‘s tumor (one patient; 3.57%).

The histological types of non-neoplastic ovarian lesions were: simple cysts (five patients; 33.33%), theca-lutein cysts (four patients; 26.66%), follicular cysts (three patients; 20%), inclusion cyst (one patient; 6.66%), hemorrhagic cyst (one patient; 6.66%), and corpus luteum hematoma (one patient; 6.66%).

Immunohistochemical study

Agreement between observers from the Kappa test ranged from strong to perfect. Tab. 1 compares the cytokine epithelial immunostaining in malignant and benign ovarian neoplasms, endometriomas, and non-neoplastic lesions. There was stronger TNF-a immunostaining in benign neoplasms compared to malignant neoplasms (P = 0.01), and in endometriomas compared to benign neoplasms (P < 0.0001). With IL-2, there was stronger immunostaining in malignant neoplasms compared to benign neoplasms (P = 0.0001), malignant neoplasms compared to endometriomas (P = 0.0002), and in benign neoplasms in relation to endometriomas (P = 0.0009). With IL-5, there was stronger immunostaining in benign neoplasms compared to malignant neoplasms (P = 0.02), in endometriomas compared to malignant neoplasms (P = 0.0046), and in endometriomas compared to non-neoplastic lesions (P = 0.0035). With IL-8, there was stronger immunostaining in endometriomas than in malignant neoplasms (P = 0.0149).

Tab. 2 compares cytokine stromal staining in malignant and benign ovarian neoplasms, endometriomas, and non-neoplastic lesions. For TNF-a, there was stronger immunostaining in benign neoplasms in relation to malignant ovarian neoplasms (P = 0.001), stronger immunostaining in endometriomas compared to malignant neoplasms (P = 0.0008), and stronger immunostaining in endometriomas compared to benign neoplasms (P < 0.0001). With IL-2, there was stronger immunostaining in malignant neoplasms than in benign ovarian neoplasms (P = 0.004), and stronger immunostaining in benign neoplasms compared to endometriomas (P = 0.0434). For IL-5, there was stronger immunostaining in benign neoplasms compared to malignant ovarian neoplasms (P = 0.0003), stronger immunostaining in endometriomas than in malignant neoplasms (P < 0.0001), stronger immunostaining in endometriomas than in benign ovarian neoplasms (P = 0.009), and stronger immunostaining in endometriomas than in non-neoplastic lesions (P = 0.0135). With IL-6, there was stronger immunostaining in endometriomas than in malignant ovarian neoplasms (P = 0.0003), and stronger immunostaining in endometriomas than in benign neoplasms (P < 0.0001). For IL-8, there was stronger immunostaining in malignant than in benign neoplasms (P = 0.004), and stronger immunostaining in endometriomas than in malignant ovarian neoplasms (P = 0.0006).

Sensitivity, specificity, and accuracy in endometriomas and ovarian cancer dia g- nosis were calculated. The epithelial immunostaining of IL-5 and IL-8 is stronger in endometriomas than in ovarian cancer. For epithelial IL-5, sensitivity, specificity, and accuracy were 89.5%, 53.6%, and 68.1%, resp. For epithelial IL-8, sensitivity, specificity, and accuracy were 94.7%, 39.3%, and 61.7%, resp.

The stromal immunostaining of TNF-a, IL-5, IL-6, and IL-8 is stronger in endometriomas than in ovarian cancer. For stromal TNF-a, sensitivity was 84.2%, specificity was 67.9%, and accuracy was 74.5%. For stromal IL-5, sensitivity, specificity, and accuracy were 78.9%, 82.1%, and 80.8%, resp. For stromal IL-6, sensitivity was 84.2%, specificity was 71.4%, and accuracy was 76.6%. For stromal IL-8, sensitivity, specificity, and accuracy were 94.7%, 53.6%, and 70.2%, resp.

Discussion

Association between endometriosis and ovarian malignant tumors have been reported [8–10]. The pathophysiology of endometriosis is not well established. There are multiple and complex endocrine-immunological interactions involved in this process, and the immune response participates in this context. The study of the role of the immune response in this dis ease may lead to the discovery of new therapeutic targets [11,12].

Endometriosis is a dis ease with a prolonged inflammatory response. The presence of ectopic tissues in the peritoneal cavity, without a clear mechanism, reaches a state of equilibrium with a high level of pro-inflammatory cytokines and immune cells [13]. It is well established that women with endometriosis have immune dysfunction in the form of intensified local and systemic inflammation [3,14]. Studies have shown that inflammation is a risk factor for ovarian cancer, and the inflammatory response is involved in almost all stages of tumor development [15,16]. In tumors, histological analysis has shown varying degrees of infiltration of immune cells, suggesting the recruitment of these cells in response to neoplastic proliferation [17], leading to chronic inflammation and tumor development and progression [18].Thus, local inflammation and immune dysregulation characterize both endometriosis and cancer. Like cancer, endometriosis resembles a chronic wound that does not heal [19]. The lymphatic vessels invaded in endometriosis can serve as a channel for the infiltration of immune cells, which further increases the inflammatory state in the endometriotic microenvironment [13].

Inflammatory behavior of the peritoneal fluid in patients with endometriosis is complex, and some studies have demonstrated the role of cytokines in this environment [12,20,21]. One study used mass cytometry to assess immune cell compartments in endometriosis, demonstrating the dynamic spectrum of cell signatures across dis ease stages [20]. Modifications to this immune microenvironment can account for the mechanisms of recruitment and functional regulation of Tregs in endometriotic lesions [21]. In our study, we observed and measured epithelial and stromal immunostaining of IL-2, IL-5, IL-6, IL-8, and TNF-a in endometriomal tissue, non--neoplastic tumors, benign neoplasms, and malignant neoplasms of the ovary. Our results show endometriotic tissue, both epithelial and stromal, is rich in immunological cytokines.

Immune disorders have been suggested to contribute to the development and progression of endometriosis, creating a microenvironment that supports the survival and implantation of endometriotic cells [22–24]. In our study, there was stronger epithelial and stromal TNF-a immunostaining in benign neoplasms compared to malignant neoplasms, and in endometriomas compared to benign neoplasms. For stromal immunostaining, we observed stronger immunostaining in endometriomas compared to ovarian cancer. It is suggested that the inflammatory response in endometriosis increases due to cytokines such as TNF-a [25,26]. The presence of TNF-a polymorphisms increases the risk of endometriosis and its development [27,28], as suggested by elevated levels of TNF-a in peritoneal fluid and the positive regulation of TNF-a in peritoneal macrophages and peripheral blood monocytes [29,30].

We observed stronger stromal immunostaining of IL-6 in endometriomas when compared to malignant and benign ovarian neoplasms. Similarly, Tsudo et al. (2000) detected a significant increase in IL-6 expression in endometriotic tissue derived from cell stroma [31]. A study demonstrated a significant production of IL-1b, IL-6, and TNF-a levels in endometriotic tissue and endometrium, with significant differences between tissue types, indicating a deviating cytokine pattern in both endometriotic tissue and endometrium from women with endometriosis compared with that from healthy controls [32].

In the present study, IL-8 immunostaining was strong in endometriomas, both in epithelial and stromal components, and in agreement with two studies that analyzed the immunostaining of this cytokine in ectopic endometrial tissues. These researchers suggested that IL-8 and its receptor system would be involved in the pathogenesis of endometriosis [33,34]. A stronger stromal immunostaining of IL-8 was observed in malignant neoplasms compared to benign neoplasms. Increased expression of IL-8 is associated with increased cell proliferation, angiogenesis, and metastases during cancer progression, with a worse prognosis noted in prostate and ovarian cancer [35].

The association of a panel of cytokines (IL-1a, IL-1b, and IL-6) in endometrial fluid aspiration was able to predict stage 3–4 endometriosis [36]. Peritoneal fluid is also an interesting environment for the study of endometriosis. One study evaluated some cytokines in the serum and peritoneal fluid of women with endometriosis. IL-8 and TGF-b levels were higher in patients with endometriosis compared to the control group [37]. In our study, there was stronger IL-8 immunostaining in malignant than in benign neoplasms, and stronger immunostaining in endometriomas than in ovarian cancer.

IL-2 modulates the expression of receptors for other cytokines and transcription factors, promoting or inhibiting cytokine cascades that correlate with each T helper cell differentiation state [38]. Velasco et al. (2007) demonstrated that clinical treatment with IL-2 in rats significantly reduced the size of endometriosis implants. Thus, this interleukin could become an effective alternative for the non-hormonal treatment of endometriosis [39]. In our study with IL-2, there was stronger epithelial and stromal immunostaining in malignant neoplasms compared to benign neoplasms, and in benign neoplasms in relation to endometriomas. Only epithelial staining was stronger in malignant neoplasms compared to endometriomas.

There are few conclusions in the literature on the role of IL-5 in the pathophysiology of ovarian neoplasms. In an immunohistochemical evaluation of endometriosis tissues, an increase in eosinophil peroxidase (a marker of eosinophil degranulation) and IL-5 expression was observed during the evaluation of eutopic endometrial tissue [40]. In our study, there was stronger epithelial and stromal immunostaining in benign neoplasms compared to malignant neoplasms, in endometriomas compared to malignant neoplasms, and in endometriomas compared to non-neoplastic lesions. Only in tumor stroma, there was stronger immunostaining in endometriomas than in benign ovarian neoplasms.

It is well established that the study of bio markers can contribute to the management of ovarian cancer, distinguishing benign and malignant pelvic masses and monitoring the response to treatment [41]. Similarly, the histopathological associations between endometriosis and epithelial ovarian cancer may have prognostic value and help to clarify some pathogenic mechanisms of endometriosis that are not yet clearly understood. The presence and type of cellular infiltrate in endometriotic tissue and its location can be considered in a study of prognostic factors.

The main limitation of the study was heterogeneity of the histological types evaluated, the presence of a certain type of cellular infiltrate in endometriotic tissue, and its location can be considered in a study of prognostic factors. Another limitation was the inclusion of borderline tumors in the group of malignant neoplasms. On the other hand, this is the first study in the literature that evaluates an immunohistochemical panel of cytokines both in the stroma and in the epithelium of ovarian tumors and endometriomas. Our findings demonstrate significantly different levels of cytokine expression in the two tissue microenvironments, the epithelium and stroma. This work may also be the basis for further studies on the inflammatory response and immunology of endometriomas, and their relationship with carcinogenesis.

Conclusion

Epithelial immunostaining of IL-5 and IL-8 is stronger in endometriomas than in ovarian cancer. The stromal immunostaining of TNF-a, IL-5, IL-6, and IL8 is stronger in endometriomas than in ovarian cancer. The stronger immunostaining of some cytokines in endometriomas compared to ovarian cancer reflects inflammatory and immune responses that could be future targets for new discoveries about the infiltrative behavior of endometriosis.