Could Obesity be a Triggering Factor for Endometrial Tubal Metaplasia to be a Precancerous Lesion?

Background & Aims Endometrial tubal metaplasia (ETM) is mostly described in conjunction with unopposed estrogen levels, and its association with endometrial hyperplasia and endometrial carcinoma (EC) is striking. Obesity is a risk factor for endometrial hyperplasia and EC development. The aim of this study is to investigate the impact of BMI and serum estradiol level on expression of PAX-2, H-TERT, P16, Ki-67, and P53 in studied ETM in reference to benign endometrium and EC. Methods The study was conducted on the following groups: group (1) consists of 57 cases that had endometrial biopsies with histologically demonstrable ETM (typical or atypical) and all were subjected to serum estradiol levelling and body mass index (BMI) evaluation; group (2) had adjacent benign endometrial tissue as control; group (3) consists of 52 cases of conventional endometrial carcinoma and 16 serous carcinoma paraffin blocks which were collected and reevaluated. All included groups were immunostained for PAX-2, H-TERT, p16, ki67, and p53. Results The relation between BMI and serum estradiol level in group 1 and PAX-2, H-TERT, P16, and p53 was statistically significant, while their relation with atypia and ki67 expression was insignificant. Twenty-three ETM cases (40.4%) out of group 1 were all (100%) obese, 87% had high serum estradiol level, and 73.9% were postmenopausal and had a similar immunohistochemical profile as EC cases (group 3). Conclusions The presence of ETM regardless of the histologic atypia in obese postmenopausal patients with high serum estradiol level is an alarming sign. This implies that ETM might not be as benign as generally accepted, as under certain clinical conditions, it may turn into a potential premalignant lesion.


Introduction
Tubal (or ciliated cell) metaplasia of the endometrium is characterized by ciliated columnar cells with bland round nuclei and eosinophilic cytoplasm, similar to the cells normally seen lining the fallopian tube [1]. It is generally thought that endometrial tubal metaplasia (ETM) is a benign disease. However, some studies propose ETM to be a potential premalignant endometrial lesion and has an association with endometrial hyperplasia, endometrial intraepithelial neoplasia (EIN), and well-differentiated endometrial carcinoma (EC) [2].
Even though ETM does not always progress to neoplasia, a determined effort should be performed on its presence in endometrial samples to detect any coexisting neoplastic condition [3].
Obesity is thought to cause an increase in EC risk due to abnormal levels of several hormones, most notably increased estrogen levels. Estrogen is a known endometrial growth factor, and, after menopause, the primary source of estrogen is peripheral tissues, including adipose tissue [4]. Bessonova et al. [5] reported a significant increase in risk of developing endometrial cancer for each 5 kg/m 2 increase in body mass index (BMI) in women. Accordingly, postmenopausal obese women had an increase of both circulating estrone (E1) and estradiol (E2) fortyfold as compared to postmenopausal women with a normal BMI range [6].
PAX2 expression was described as a marker of the Mullerian duct derivatives as fallopian tubes, uterus, cervix, and upper vagina [7]. According to many studies, normal endometrial glandular cells have nuclear expression of PAX2 as revealed by immunohistochemistry. On the other hand, PAX2 expression was absent in EC and EIN implicating PAX2 as a possible key regulator in endometrial carcinogenesis [8,9].
Human telomerase reverse transcriptase (H-TERT) is a catalytic subunit with reverse transcriptase activity. hTERC is constitutively present in normal and cancer cells, whereas expression of H-TERT is almost exclusively limited to cancer cells [10].
Scattered patterns or focal expression of p16 positive cells in serous EC and tubal tissue was described by several authors. ey describe a pattern which is not seen in normal tissue and is different from diffuse p16 positivity as seen in high-risk HPV positive cervical intraepithelial lesions [11]. Inoue [12] reported that endometrial metaplasia, in general, was a precursor of variant types of endometrial carcinomas, based on the p53, PCNA, and ki67 overexpression in EC and endometrial metaplasia and the fact that EC are often accompanied by adjacent metaplastic epithelium [13]. However, it is, for now, not a generally accepted view that endometrial metaplasia is a precursor of endometrial cancer. e current study was designed to investigate the impact of BMI and serum estradiol level on expression of PAX-2, H-TERT, P16, Ki-67, and P53 in studied ETM in reference to benign endometrium and EC.

Case Selection.
e study was conducted on three groups. In group (1), from June 2016 to June 2019, fiftyseven cases of benign endometrial biopsies with histologically demonstrable tubal metaplasia (typical or atypical) were identified, reviewed, and included in this study. ese samples were obtained from uterine curettage of 57 women aged from 28 to 72 years who were admitted to Gynecology and Obstetrics Department of Tanta University Hospitals complaining from dysfunctional uterine bleeding. After histological diagnosis of tubal metaplasia, all 57 patients were subjected to serum estradiol levelling and body mass index (BMI) evaluation. In group (2), the adjacent benign endometrial tissue was used as control. In group (3), for further comparison, Tanta Pathology Department archives were searched from January 2016 through to March 2019 and sixty-eight previously diagnosed EC (52 cases with conventional endometrial carcinoma and 16 cases with serous carcinoma) and study paraffin blocks were collected and reevaluated. Any cases of EC with adjacent ETM were excluded from this study.
is study was approved by the ethics committee of Faculty of Medicine, Tanta University. All included patients provided a written informed consent for the use of biological specimens for research purposes.

Histopathological Evaluation.
All included 57 tissue biopsies (groups 1 and 2) were fixed in formalin, embedded as paraffin blocks and cut at 4 μm, and stained with hematoxylin and eosin (H & E). After histopathological evaluation by two separate pathologists, they were diagnosed as follows: tubal metaplasia (nonmetaplastic endometrium having some ciliated cells and resembling fallopian tube) with adjacent 22 cases with disordered proliferative endometrium, 5 cases with atrophic endometrium, 6 cases with senile cystic fibrosis, and 24 cases with benign endometrial hyperplasia (used as control). Atypical tubal metaplasia was found in 24 cases. e criteria for cytologic atypia of atypical tubal metaplasia included pleomorphic and hyperchromatic nuclei and increased nuclear-to-cytoplasmic ratio. Atypical endometrial hyperplasia or EC was not identified in any studied cases. e collected 68 paraffin blocks (group 3) were also recut and stained with H and E, reevaluated, and diagnosed as 52 cases with conventional endometrial carcinoma and 16 cases with serous carcinoma.

Immunohistochemical Analysis.
All included groups were immunostained for PAX-2, H-TERT, p16, ki67, and p53 (Table 1). Briefly, all tissues were deparaffinized followed by blockade of endogenous peroxidases and antigen retrieval using Antigen Unmasking Solution (Vector; USA), followed by the primary antibody and then the visualization reagent (secondary goat anti-mouse immunoglobulin and horseradish peroxidase linked to a dextran polymer backbone). After rinsing with distilled water, the slides were incubated with DAB (3, 3-diaminobenzidine) substrate-chromagen solution and counterstained with hematoxylin. Negative controls were prepared using primary antibody with PBS and normal mouse or rabbit serum.

Staining Scoring.
Immunochemical results were independently scored by two experienced pathologists who were blinded to the patients' clinicopathological data and outcomes. e pathologists reviewed immunostained tissue sections under a light microscope. e ranges used for PAX-2 expression analysis were as follows: 0%, 1-25%, 26-50%, 51-75%, or 76-100%. Positive PAX-2 expression was only considered with a nuclear staining pattern. PAX-2 loss was scored as follows: (1) complete loss (0% cells staining), (2) partial loss (1-75% cells staining), and (3) minimal to no loss (76-100% cells staining) [9]. e H-TERT immunostaining score was calculated with a semiquantitative scoring system according to Qin et al. [14] (only diffuse nuclear was considered as positive). e intensity range was as follows: 0, no staining; 1, weak staining; 2, moderate staining; 3, strong staining. e extent of staining was presented as percentage of tumor cells that were stained: 0, <10% of tumor cells stained; 1, 10-50% of positive cells; 2, >50 and <75% of positive cells; 3, >75% of positive cells. An overall score was obtained as the product of the intensity and the extent of positive staining. Cases with 0 points were considered to be negative (0), cases with a score of 1-3 weakly positive (1+), cases with a score of 4-7 moderately positive (2+), and cases with a score of >7 strongly positive (3+).

Enzyme-Linked Immunosorbent Assay (ELISA)
Evaluation of Serum Estradiol and BMI Evaluation. All group 1 patients were subjected to serum estradiol levels analysis. A peripheral blood sample (10 ml) was collected in sterile tubes. After clotting the serum for 30 minutes at room temperature in a serum separator tube, it was centrifuged for 15 min and then analyzed immediately and frozen at −20°C for storage. Serum estradiol levels were quantified by an estradiol enzyme-linked immunosorbent assay (ELISA) kit as recommended by the instructions of the manufacturer (R&D Systems, Minneapolis, MN, USA) with normal range 50-300 ng/ml. Serum estradiol level of more than 300 ng/ml was considered high [15].

Statistical Analysis.
Statistical presentation and analysis of the present study were conducted, using the mean, standard deviation, and chi-square test, Student's ttest, and Mann-Whitney test by SPSS software (version 15.0; SPSS Inc., Chicago, Illinois, USA). Significant differences were considered as p < 0.05.
ere was a significant relationship between studied markers in the three different groups as illustrated in Table 3.    (Figures 1-4). e relation between BMI, age, menopausal status, serum estradiol level, PAX-2, H-TERT, P16, and p53 expression was statistically significant. On the other hand, the relation between BMI, history of oral contraceptive pills, presence of histologic atypia, and ki67 expression was statistically insignificant as shown in Table 4. Out of the 23 included obese patients (≥30 kg/m 2 ), 17 patients (73.9%) were postmenopausal, 20 patients (87%) had serum estradiol level of more than 300 ng/ml, 21 obtained endometrial samples (91.3%) were negative for PAX-2 expression, 20 (87%) samples were positive for both H-TERT and p53 expression, and 18 samples (78.3%) were positive for p16 expression.
Regarding serum estradiol level, there was a statistically significant relationship with age, menopausal status, PAX-2, H-TERT, and P16 and p53 expression, while the relation between serum estradiol level, history of oral contraceptive pills, presence of histologic atypia, and ki67 expression was statistically insignificant as illustrated in Table 5. Out of the 20 included patients (>300 ng/ml), 14 patients (70%) were postmenopausal, 18 obtained endometrial samples (90%) were negative for PAX-2 expression, 17 (85%) samples were positive for H-TERT, 16 samples (80%) were positive for p16 expression, and 18 samples (90%) were positive for p53 expression.
Histologic atypia in studied cases of group 1 was statistically insignificant with all other clinicopathological Journal of Obesity 5 characters and studied markers expression as described in Table 6.

Discussion
ETM is mostly described in conjunction with unopposed estrogen levels, and its association with simple and complex endometrial hyperplasia and well-differentiated EC is striking [17]. Some authors suggested that the presence of atypical endometrial hyperplasia is an alarming sign for suspecting of malignant transformation [11,18]. On the other hand, other authors thought that the presence of architecturally complex glands in ETM, even with no atypia, is often associated with endometrioid adenocarcinoma and consequently should be managed as a case of complex endometrial hyperplasia [19,20]. To our knowledge this the first study to assess the impact of BMI and serum estradiol level on ETM with or without presence of atypia and their role as precursor of EC. e current study investigates the impact of BMI and serum estradiol level on expression of PAX-2, H-TERT, P16, Ki-67, and P53 in studied ETM in reference to benign endometrium and EC.
In the present study, PAX2 expression was absent in 23 cases (40.4%) out of group 1, 5 cases (8.8%) of group 2, and 57 cases (83.8%) of group 3. is relation was statistically significant. Also, the relation between PAX-2 expression and BMI and serum estradiol level in group 1 was statistically significant, while its relation with atypia was insignificant (p � 0.962). Our   findings are similar to those of Cao et al. [8] and Allison et al. [9] in terms of PAX2 expression in benign endometrial lesions and EC. ey found that PAX2 protein expression had progressive loss along the spectrum from hyperplasia to EC, but unfortunately, they did not include ETM cases in their studies. Wu et al. [21] described PAX2 expression in endometrial cancer cell lines. ey stated that tamoxifen and estrogen could activate PAX2 mRNA expression in endometrial cancer cell lines but not in normal endometrial samples. is increased expression was associated with cancer-linked hypomethylation of the PAX2 promoter.
ese results appeared to offer a mechanism to explain the increased incidence of endometrial cancers in women with unopposed estrogen in obese patients with high serum estradiol level especially in the postmenopausal state.
In the current study, H-TERT expression was positive in 21 cases (36.8%) out of group 1, 4 cases (7%) of group 2, and 58 cases (85.3%) of group 3. is relation was statistically significant. Also, the relation between H-TERT expression and BMI and serum estradiol level in group 1 was statistically significant, while its relation with atypia was insignificant (p � 0.503). Some researches were in accordance with these results which stated that H-TERT expression was limited almost exclusively to cancer cells reporting that the immunohistochemical reactivity of TERT was present in numerous human cancers, including uterine endometrioid and serous carcinoma, but not in benign lesions such as disordered proliferative endometrium and benign endometrial hyperplasia [10,22]. On the other hand, Simon et al. [3] reported that H-TERT immunopositivity was present in uterine serous carcinoma, and in contrast, atypical tubal metaplasia was completely negative for TERT. is could be attributed to a smaller number of ETM cases (only 16) who were all atypical. ey  Journal of Obesity did not include other cases of ETM without atypia in their study.
In this study, P16 expression was positive in 23 cases (40.4%) out of group 1, seven cases (12.3%) of group 2, and 23 cases (33.8%) of group 3. is relation was statistically significant. Also, the relation between p16 expression with BMI and serum estradiol level in group 1 was statistically significant, while its relation with atypia was insignificant (p � 0.620). Horree et al. [11] proposed that TM of the endometrium could be a potentially premalignant lesion. ey explained their finding that expression of p16 is regarded as a carcinogenetic event in many tumors, including those in the gynecological tract [23,24]. Further, Umezaki [25] suggested that tubal metaplasia should be considered as a neoplastic lesion of uterine cervical glandular lesions that may have the potential to undergo malignant transformation. p16 is also aberrantly expressed in TM in some ovarian inclusion cysts. ese cysts have been proposed to be precursors of ovarian cancer [26].
For Ki67 expression, it was positive in 55 cases (96.5%) out of group 1, 47 cases (82.5%) out of group 2, and in all cases of group 3 (68 cases, 100%). is relation was statistically significant. However, the relation between ki67 expression with BMI, serum estradiol level, and histologic atypia in group 1 was statistically insignificant. It is well known that Ki-67 is expressed exclusively in proliferating cells. Studies have also shown that Ki-67 proliferation indices are increased in numerous human malignancies, including high-grade EC as well as atypical endometrial hyperplasia, as compared with benign endometrial lesions [18,27]. On contrary to our results, Simon et al. [3] reported that the majority of atypical tubal metaplasia either is negative for Ki-67 or shows a low proliferation index similar to that seen in typical tubal metaplasia, and this could be explained by their small sample size.
For P53 expression, it was positive in 26 cases (45.6%) out of group 1, 11 cases (19.3%) out of group 2, and 33 cases (48.5%) out of group 3. is relation was statistically significant. Besides, the relation between p53 expression with BMI and serum estradiol level in group 1 was statistically significant, while its relation with atypia was insignificant (p � 0.829). is was in agreement with Simon et al. [3] who   showed that atypical tubal metaplasia displayed negative or focal, weak immunoreactivity with p53. e presence of weak and heterogeneous p53 immunoreactivity in ETM could be a consequence of DNA damage. While intense, diffuse and homogeneous p53 staining favors carcinoma [1].
In the present study, it was found that 23 ETM cases (40.4%) out of group 1 were all (100%) obese (more than 30 kg/m 2 ), 87% had high serum estradiol level of more than 300 ng/ml, and 73.9% were postmenopausal and had a similar immunohistochemical profile as EC cases (group 3).
is suggests that certain clinical conditions could be triggering factors for ETM to be a precursor lesion for EC. Cases of ETM, especially those with certain previous conditions, need clinical followup in the form of regular uterine ultrasound examination of endometrial thickness and subendometrial vascularity every three months at least.

Conclusions
e presence of ETM regardless of the histologic atypia in obese postmenopausal patients with high serum estradiol level is an alarming sign for patients strict followup. is implies that ETM might not be as benign as generally accepted, as under certain clinical conditions, it may turn into a potential premalignant lesion. After all, studies on a larger scale should be performed to validate the precancerous potential of ETM under those certain conditions. Data Availability e data supporting the results are present in the discussion and presented in the references section.