Endometrial Flushing Tumor Necrosis Factor Alpha and Interleukin 2 Levels in Women with Polycystic Ovary Syndrome, Leiomyoma and Endometrioma: Comparison with Healthy Controls

There were 79 subjects and all of them tolerated endometrial fluid sampling well. No subject had early (uterine rupture, etc.) or late (pelvic infection, etc.) complications, and only one subject had pelvic pain that required two hours of observation. Table 1 summarizes the demographic data and baseline serum progesterone levels. The mean age of patients with PCOS, uterine myoma, endometrioma and the control group were 28.90 ± 5.45, 37.25 ± 2.73, 32.84 ± 6.62 and 32.15 ± 5.18, respectively. Compared to the control group, the mean age was lower in the PCOS (p = 0.045) and higher in the myoma patients in a statistically significant way (p = 0.001). The mean age difference was not statistically significant between the endometrioma patients and the control group (p = 0.717). As for mean body mass index (kg/m 2 ), relative to the control group, it was lower in the endometrioma group (p = 0.005) and similar for the uterine myoma (p = 1.0) and the PCOS groups (p = 0.205). Table 1  Demographic and baseline data including comparison of gynecologic diseases with control group. Mid-luteal phase serum progesterone level (ng/ml) was similar in the leiomyoma (9.61, p = 0.745) and the PCOS (9.47, p = 0.978) groups relative to the control group (8.75). It was, however, lower in the endometrioma group (6.26) than in the control group (p = 0.030) in a statistically significant way. Gravida and parity values were significantly lower in all three groups than the control group ( Table 1 ). Table 2 shows comparison of endometrial flushing fluid TNF-α and IL-2 levels of the three gynecologic disease groups with the control group. Comparisons of the TNF-α levels (ng/ml) in the PCOS and leiomyoma groups with the control group do not reveal significant differences (p values were 0.220 and 0.502, respectively). The endometrioma group had significantly higher TNF-α levels (338.2 ng/ml) than the control group (261.1 ng/ml, p = 0.004). However, mean IL-2 levels (ng/ml) were significantly lower in the PCOS (p = 0.0005), leiomyoma (p = 0.0002) and endometrioma patients (p = 0,0009) than in the control group (416.0 ± 123.4). Table 2  Comparison of gynecologic diseasesʼ IL-2 and TNF-α levels with control group. As the mean age values differ significantly between the groups, the association of age with TNF-α and IL-2 levels were investigated. The correlation between age and TNF-α (r = 0.126, p = 0.270) and between age and IL-2 (r = − 0.038, p = 0.741) were found to be insignificant.

The subjects in the study were patients who volunteered in Izmir Katip Celebi University Ataturk Education and Research Hospital outpatient gynecology clinic. The study was conducted between January and June 2013. Izmir Katip Celebi University Ethics Commission approved of the study which was funded by Izmir Katip Celebi University Scientific Research Projects Coordination Unit (2013-T1-TSBP-11). All subjects were informed in detail about the research project verbally and their written informed consent was obtained. The study was conducted with patients that were also part of another study 12 . This other study focused on the levels of structural proteins and progesterone regulated molecules (αVβ3 integrin, glycodelin and PGF2α). In contrast, the current study focuses on the comparison of 2 important immunomodulatory cytokines, TNF-α and IL-2 levels among healthy patients and patients with benign gynecological diseases. The subjects included in the study were between the ages of 20 to  40, and either had polycystic ovary syndrome, endometrioma and uterine myoma or were healthy subjects without any gynecological disorders. They were also required to have regular periods (between 28 – 34 days, variation ≤ 3 days between cycles). The subjects did not have any intrauterine devices, did not use hormonal contraception or medications related to their gynecologic disease or that may affect the endometrium, and had the intellectual capacity to comprehend the information about the study and provide written consent. The exclusion criteria for volunteers: smokers, subjects being pregnant, having pelvic infection symptoms, having luteal phase serum progesterone levels < 3 ng/dL seven days before their expected menstruation date and subjects opting out of the study. In addition, patients who had findings suggestive of pathologies associated with changes in ER markers, such as hydrosalpinx, adenomyosis, endometrial pathologies (submucous fibroids, endometrial polyps, etc.) were excluded from the study. The patients with concomitant gynecologic pathologies according to diagnosis criteria of stated below (i.e. patients with both myoma and endometrioma) were also not included. The evaluation of these pathologies were made based on transvaginal ultrasound (TV-USG) (Medison SonoAce, AM Company Co. Ltd., South Korea) and saline infusion sonography (SIS) that were routinely performed on patients after endometrial fluid sampling. The study was conducted in case control design. For each group, the first 20 patients that attended Izmir Katip Celebi University Ataturk Education and Research Hospital between January and June 2013, satisfied the inclusion criteria and did not violate the exclusion criteria were included in the study. During that period, we only had 19 endometrioma patients. The first group in the study included 20 subjects with polycystic ovary syndrome who all had a specific phenotype (hyperandrogenism plus ultrasound criteria according to Rotterdam criteria 13 ) and did not have ovulatory disorders according to their serum progesterone levels on the 21st day of their cycle measured on their first evaluation. The second group included 20 patients with at least one intramural leiomyoma with 3 to 10 cm diameter as measured on TV-USG. The third group included 19 endometrioma patients who satisfied all three of the following criteria: pelvic pain, dysmenorrhea, at least one unilocular cyst larger than 3 cm on TV-USG with diffuse homogeneous ground-glass echoes due to probable hemorrhagic debris. Lastly, the control group included 20 otherwise healthy women who were chosen among the patients who attended to outpatient gynecology clinic with tubal ligation demand. They were required to be at reproductive age, have no gynecologic symptoms and no pathologic findings on TV-USG (inclusive of leiomyoma, endometriotic nodules, polycystic ovaries, abnormally thickened endometrium or free fluid in the Douglas pouch). Fertility state of the patients were not used as criteria in defining the groups. The main outcome measure of the study was TNF-α and IL-2 levels in endometrial flushing fluid. The levels of these markers in PCOS, leiomyoma and endometrioma patients were compared with the control group. We measured levels of markers by endometrial flushing which is a commonly used method in the clinical literature to assess intra-uterine microenvironment and molecular mechanisms. In particular, the patients were cautioned against having sexual intercourse until the sampling of endometrial fluid in the current menstruation cycle and they were called to the clinic for the sampling seven days before their expected menstruation date on the time corresponding to their midluteal phase. After confirmation of ovulation based on blood progesterone levels, endometrial fluid sampling was performed following a procedure similar to saline infusion sonography technique and adopted from an earlier study 14 . The patients were placed in dorsal lithotomy position. A bivalve disposable speculum was inserted into the vagina and to eliminate potential contaminants, the cervix was cleaned with sterile saline. Then, a menstrual regulating cannula (4 mm diameter) was introduced into the uterine cavity through the cervical canal. A syringe was then used to instill 5 ml of sterile saline into the uterine cavity. The uterine contents were aspirated quickly. The procedure was repeated two times, after which the samples were collected in separate Eppendorfs, centrifuged and then stored at − 80 °C. The subjects were kept under observation for about half an hour after the endometrial fluid sampling. After endometrial fluid samples were collected from all patients, TNF-α and IL-2 levels were measured by using Eastbiopharm (Hangzhou Eastbiopharm Co. Ltd./China) ELISA kits (TNF-α lot: 20130924, IL-2 lot: 20130924) in BioTek ELISA devices. For all 79 patients, the same kits were used for the same markers until the end of data collection. For statistical analysis, SPSS 23.0 (SPSS Inc., Chicago, IL) software was used. The normality of the data was tested with Shapiro-Wilks test and homogeneity of variances was tested with Leveneʼs test. Accordingly, Mann-Whitney U Test or Independent Sample T Test were used to compare the three groupsʼ TNF-α and IL-2 levels with the control group. As the mean age values differ significantly among the groups, to check whether the differences of TNF-α and IL-2 levels between the groups depend on age or not, correlation analysis with Spearman correlation coefficient was performed. Two-tailed p-values below 0.05 were assumed to be statistically significant. Values are presented in median (min–max) [mean ± SD] format.

In PCOS and leiomyoma, TNF-α levels were not found to be significantly different from the control patients. However, decreased IL-2 levels in PCOS, leiomyoma and endometrioma patients and increased TNF-α levels in endometrioma patients might be playing a role in infertility and impaired endometrial implantation associated with these diseases. These findings provide hints for molecular mechanisms underlying these diseases and contributes to the literature that investigates ways to diagnose these diseases with auxiliary methods for clinical purposes. More studies on the relationship between these cytokines and infertility are needed to shed light on this question.

In this case-control study, endometrial flushing fluid levels of TNF-α and IL-2 were compared for subjects with polycystic ovarian syndrome, endometrioma and uterine leiomyoma with the healthy controls. According to our findings, PCOS and leiomyoma patients had significantly lower IL-2 and similar TNF-α levels relative to healthy patients, while endometrioma patients had higher TNF-α levels and lower IL-2 levels. In our 2017 study, we had chosen to evaluate three cell signaling molecules that have impact on endometrial receptivity with mostly paracrine effect 12 . That study included structural proteins and progesterone-regulated molecules. This previous study found that αVβ3 integrin and glycodelin levels were higher in the endometrioma and PGF2α was higher in the PCOS patients than the control group in a statistically significant way 12 . Beside these molecules, cytokines also play an important role for progesterone-regulated ER and embryo development. The motivation for investigating cytokines comes from studies that indicate its role in fostering a healthy intrauterine environment through their autocrine, paracrine and endocrine (systemic and immunomodulatory) effects. For example, TNF-α, an immunomodulatory cytokine is conjectured to be an appropriate signal for the closing of the implantation window and preparation for menstrual shedding 15 . In vitro studies claimed that TNF-α protects the embryo from teratogenic effects 16 and regulates invasion characteristics of trophoblast by regulating MMP 2 and MMP 9 17 . On the other hand, an animal study of TNF-α has found it to be a pro-inflammatory cytokine that triggers immunological pregnancy loss 18 ,  19 . While there are many studies that systemically investigate the relationship of serum TNF-α levels with benign gynecologic diseases, TNF-α levels within uterine cavity of women with gynecologic diseases has not received sufficient attention. Endometrioma or endometriosis are important examples of such diseases which have received limited attention. Minici et al. 20 and Iwabe et al. found that the concentrations of TNF-α were significantly higher in peritoneal fluid from patients with endometriosis than that of patients without endometriosis 21 . Iwabe et al. also conjectured that TNF-α might inhibit the development of preimplantation embryo and impair sperm function 21 . In another in-vitro study, the adherence of endometrial stromal cells to mesothelial cells increased significantly when mesothelial cells were exposed to TNF-α nutrients. This finding highlights TNF-αʼs role in the development of endometriosis 22 . Furthermore, it suggests that, the higher level of TNF-α that the current study identifies might also be playing a role in the pathogenesis of infertility associated with endometriosis. Thus, patients with endometriosis might benefit from the administration of TNF-α receptor blockers as mentioned before 23 . There is no study that investigates the TNF-α levels in the endometrial cavity of PCOS patients. The current study finds that TNF-α level in uterine irrigation fluid of PCOS patients is not different from the control group. Nevertheless, the relationship between TNF-α levels and PCOS has been subject to a number of earlier studies. Younis et al. found that the blood TNF-α levels of PCOS and unexplained infertility patients are significantly higher than the healthy control group 24 . Other studies have also found that TNF-α is higher in normal-weight women with PCOS and is even higher in overweight/obese adolescents with PCOS 25 ,  26 . While Thathapudi et al. found that TNF-α system might contribute to the pathogenesis of hyperandrogenism, obesity, and insulin resistance in PCOS 27 , a meta-analysis found that TNF-α polymorphism may not be associated with PCOS risk 28 . Despite the studies that did not find differences in the expression of endometrial genes that play a role in implantation between intramural myoma patients and fertile groups 29 , problems related to ER and fertility were always suspected and studied in patients with myomas. TNF-α is an important cytokine for ER and it was also suggested that it might be playing an important role in regulating apoptosis in leiomyoma cells 30 . Inagaki found that the mean level of TNF-α in uterine cavity was higher in leiomyoma patients than adenomyosis and endometrial polyp patients and the control group 31 . Nevertheless Ben-Nagi et al. found that TNF-α levels in mid-luteal phase uterine flushing of submucosal myoma patients is lower than the control group, but the difference was not statistically significant 14 . The results of the current study are consistent with findings of the study by Ben-Nagi et al. 14 . However, more research is needed to clarify the relationship between TNF-α levels and feto-maternal immunologic regulation in myoma patients. T regulatory cells (T reg s) play an important role in implantation and determination of peripheral tolerance 32 . T reg s are critical for the success of implantation and their absence might end pregnancy in early-mid gestation 33 . Animal studies have shown that IL-2 modulates T regS and contributes to T regS survival, proliferation and function. Consequently, trials of low-dose IL-2 treatment in mice substantially improved pregnancy outcome 34 . Nevertheless, Guneet Makkar et al. compared natural cycle and stimulated cycle patients and concluded that increased expression of IL-2 in the peri-implantation endometrium may account for the lower implantation rate in excessive responders (subjects with E 2 concentration on the day of hCG was > 20 000 pmol/) 35 . In contrast to these results based on endometrial biopsy, the IL-2 levels in uterine flushing were not significantly different 35 . The findings in the literature on the IL-2 level in peritoneal fluid of endometriosis patients and its relationship with infertility are inconsistent. Some studies find IL-2 level is significantly higher in infertile women 36 ,  37 while others do not find a significant difference 38 . In patients with PCOS, the stimulated production of IL-2 was found to be lower in peripheral circulation, irrespective of BMI 39 . Krishna et al. found that T reg s which play an important role in efficient reproductive function due to their immunosuppressive effect are lower in PCOS patients and explained this finding with the inherent hyporesponsiveness to IL-2. It has been conjectured that IL-2 based therapies may help counter PCOS complications 40 . There are few studies on IL-2 that investigate ER in human endometrial tissue or endometrial flushing fluid. A 2013 study found a lower expression of IL-2 in curettage material of women with spontaneous miscarriage relative to women with voluntary termination of pregnancy and highlighted involvement of specific Th1 cytokines on the implantation process 41 . The current study contributes to this debate and is also the first study to investigate endometrial flushing IL-2 levels for ER in PCOS, leiomyoma and endometrioma patients. The results show that during the implantation period IL-2, which plays an important role in the regulation of Th, T reg s, Tc, and NK cells, is lower in endometrioma, PCOS and leiomyoma patients than the control group in a statistically significant way. More studies are needed to better understand the relationship between the reduced IL-2 levels and impaired receptivity in PCOS, leiomyoma and endometrioma patients. Limitations of the current study include the differences between groups regarding baseline demographic data and not using the infertility state as a criterion for including patients in different benign gynecologic disease groups. These limitations make it difficult to associate these markersʼ levels with infertility and reach a conclusive result. At the same time, age is not correlated with TNF-α and IL-2 levels, suggesting that the differences between groups are not driven by age differences. Another potential concern is that analyzing ER markers with immunohistochemistry using endometrial tissue as in the previous studies 41 ,  42 would have given more accurate results than the uterine flushing technique we employ. On the flip side, endometrial flushing technique is cheaper, less invasive and more acceptable to patients, and hence would arguably be more feasible for routine. Lastly, the timing for carrying out the endometrial biopsies might be a concern. The literature commonly describes implantation window as LH + 6 to LH + 11, collects tissue specimens between these days 43 , and test the urine for LH surge 14 . We did not have the chance to use these kits on patients. To address this concern we narrowed down the criteria for patientsʼ periods (regular periods between 28 – 34 days, variation ≤ 3 days between cycles). Endometrial fluid sampling was carried out 7 days before the following presumed menstruation start after confirmation of ovulation based on blood progesterone levels (≥ 3 ng/dl).

Endometrial receptivity (ER) indicates the period when luminal epithelium is suitable for blastocyst implantation. Development of receptive endometrium depends on the satisfactory change of estrogen-processed endometrium in response to progesterone secretion. It has been found that integrins and their ligands (e.g. osteopontin), mucins, growth factors (HB-EGF), homeobox transcription factors (HOXA gene products), lipids, cytokines (Leptin, LIF, IL-1, IL-11) and other molecules take part in this process 1 ,  2 ,  3 and the focus of this study, cytokines regulate potent intercellular signals that conduct feto-maternal communication in endometrial cells and have a critical importance in this process. There are ongoing studies that aim to develop alternative treatment strategies in order to balance cytokine levels by external cytokine support or inhibition 4 . One cytokine that has been investigated for this purpose is tumor necrosis factor alpha (TNF-α) 5 . Existing studies suggest that TNF-α inhibits blastocyst vitality, growth and metamorphosis before implantation 6 . Interleukin-2 (IL-2), another cytokine secreted by T cells during antigen presentation responsible for antigen specific cell proliferation, is also under study. It increases the differentiation and proliferation of other immune cells (growth and anticore synthesis stimulant for NK, LAK and B cells) and apoptotic death of T-cells activated by antigen 7 . There are few studies in the literature that investigate the relationship between IL-2 and ER, even though it has a close relationship with the cells that have roles in feto-maternal immune regulation 8 ,  9 . Benign gynecological disorders including endometriosis, hydrosalpinx, leiomyoma, and polycystic ovary syndrome (PCOS) are associated with decreased fertility cycle outcomes and impaired ER 10 . An important debate in the literature is whether ER marker levels are different in the endometrium of women with infertility-causing diseases compared with healthy women. One method for assessing endometrial micro-environment is measuring the marker levels in endometrial flushing fluid. While endometrial flushing has no certain diagnostic use in clinics and is not as accurate and sensitive as immunohistochemistry, it is non-invasive and easy to implement and it has been widely used in the clinical literature to assess intra-uterine microenvironment and to comment on molecular mechanisms. It is well tolerated by the patients and preovulatory use of it was suggested as a treatment model for women with unexplained fertility 11 . This study compares TNF-α and IL-2 levels in the endometrial flushing fluid of patients with polycystic ovary syndrome, endometrioma and leiomyoma with healthy women. It investigates whether endometrial receptivity markers differ across these groups and attempts to provide insights into the molecular mechanisms of these diseases.