Inducing malignant transformation of endometriosis in rats by long‐term sustaining hyperestrogenemia and type II diabetes

Of the 90 rats, 10 (five from the Es group and five from the control group) died and were excluded from the experiment. Of these 10 rats, it was difficult to identify the lesions and the causes of death in four due to severe pelvic adherence; three died of severe intestinal obstruction, and the remaining three died 3 days after the injection of STZ. The EMs lesions were survived in all cases of the Es group and the control group. In most cases, EMs of the Es group grew with time, whereas the ones in the control group showed worsening atrophy with time. The average weights of the EMs lesions in the Es group 2, 4, and 8 months after surgery (424.6 ± 1.9, 453.9 ± 21.6, 523.6 ± 23.2 mg, respectively) were significantly greater than those of the control group in the corresponding months (402.2 ± 2.3, 298.9 ± 39.7, 220.0 ± 27.5 mg, respectively, P  < 0.05; Table 1 ). Compared with the volumes of the EMs lesions in the control group 2, 4, and 8 months after surgery (71.5 ± 6.8, 36.2 ± 6.9, 24.4 ± 7.1 mm 3 , respectively), those in the Es group were significantly greater in the corresponding months (88.7 ± 16.9, 210.8 ± 21.8, 394.8 ± 131.9 mm 3 , respectively, P  < 0.01; Table 1 ). Mean weight, mean volume, and macroscopic features of lesions in two groups of rats with surgically induced endometriosis (EMs) treated with estradiol and streptozotocin and fed a high carbohydrate-and-fat diet (Es group) or treated with placebo saline and fed a standard diet (control group) †Compared with the net weight of the EMs lesions in the control group, P  < 0.05. Compared with the volume of the EMs lesions in the control group, P  < 0.01. Compared with the atrophy rate of the EMs lesions in the Es group, P  < 0.05. Two cases of EMs in the Es group (8 months after surgery) were extraordinarily enlarged (Fig. 1 a). Grossly, the two were cystic, consisted of thickened capsule wall with papillary hyperplasia and sanguinopurulent fluid (H&E stain revealed malignancy). Endometriosis of the control group (Fig. 1 b) had the appearance of unilocular cystic structures filled with clear fluid. The uteri of the Es group showed severe edema, with no papillary hyperplasia, whereas uteri of the control group showed no obvious abnormality. Rat model of malignant transformation in surgically induced endometriosis (EMs). (a) Distinctively enlarged EMs lesion. (b) Viable EMs lesions of the control group (placebo saline and standard feed). (c) EMs of the control group (100×; 400×). (d) EMs with simple hyperplasia showing that the epithelial cells were increased in number, enlarged in size, and distributed in multiple layers, without atypical nuclei (100×; 400×). (e) Atypical EMs showing that the number of endometrial glands was increased, multiple layers were arranged compactly, cytoplasm was reduced, and the cell nuclei were enlarged and moderately irregular (100×; 400×). (f) Malignant transformation of EMs (endometrioid adenocarcinoma): the number of endometrial glands was increased extraordinarily with a disordered arrangement of cells, infiltration of inflammatory cells, reduced cytoplasm, and condensed nuclei irregular in shape and size with the back-to-back phenomenon. There were also bleeding points (100×; 400×). (g) Atrophy EMs showed an apparent decrease of epithelial and stromal cells (100×; 400×). (h, i) Ultrastructure of EMs. (h) Malignant endometriotic cells showing vacuole changes of mitochondria (arrow, 5000×). (i) Simple hyperplastic endometriotic cells of the Es group, treated with estradiol (5 mg/kg three times/week after surgery), streptozotocin (25 mg/kg, 1 month after surgery), and high carbohydrate-and-fat feed, at month 8 showing clearly enlarged and well-defined nuclei with reduced cytoplasm (arrow, 6000×). Histologically, the two distinctively enlarged EMs of the Es group (4.0%) were diagnosed as low-grade endometrioid adenocarcinoma, which showed a widespread cribriform arrangement with no invasion of the adjacent peritoneal muscular tissues. The corresponding eutopic endometria showed atypical hyperplasia. Three cases of EMs (6.0%) showed atypical hyperplasia with an increased number of endometrial glands, compactly arranged multiple layers, reduced cytoplasm, and enlarged and moderately irregular cell nuclei. The corresponding eutopic endometria showed simple hyperplasia. The malignant progression of the eutopic endometria was slower than that of the corresponding EMs, which was not synchronized. Endometriosis of the control group showed a similar appearance of the endometrium, consisting of cystically dilated glands with epithelium surrounded by stromal cells. Two cases of eutopic endometria showed diffuse squamous epithelium without columnar glandular epithelium (see Fig. S1). The results are shown in Table 2 and Figure 1 c–g. Histological features of lesions and eutopic endometria in rats with surgically induced endometriosis (EMs) treated with estradiol and streptozotocin and fed a high carbohydrate-and-fat diet (Es group) or treated with placebo saline and fed a standard diet (control group), n (%) In the Es group, another two cases of eutopic endometria (8%) showed squamous epithelium at 8 months. Compared with the atrophy rate of EMs of the control group, P  < 0.05. Transmission electron microscopy indicated that the malignant EMs showed remarkable abnormalities in cytomorphology and organelles including swollen mitochondria and vacuole changes (Fig. 1 h). In addition, compared to the simple hyperplastic EMs of month 4, the ectopic endometrial cells of month 8 with the same pathological change showed more active nuclei, and obviously enlarged and well-defined nuclei with reduced cytoplasm (Fig. 1 i). The ultrastructure of different pathological changes is shown in Figure S2. The atrophy and simple hyperplastic EMs were excluded from the proliferation and apoptosis analysis. In the Es group, the immunostaining of PCNA protein increased ( P  < 0.01), and the level of TUNEL positivity decreased ( P  < 0.01) progressively in the order of EMs, atypical EMs, and malignant EMs. The immunostaining of PCNA in EMs of the Es group was stronger than that of the control group (P  < 0.05), and the staining intensity of apoptosis in the Es group was lower than that of the control group (P  < 0.05) (Fig. 2 ). H -score value of apoptosis and proliferation of endometriosis (EMs) lesions in a rat model of malignant transformation, expressed as mean ± SEM. *Compared with EMs and atypical EMs of the Es group (treated with estradiol [5 mg/kg three times/week after surgery], streptozotocin [25 mg/kg, 1 month after surgery], and high carbohydrate-and-fat feed), P  < 0.01. #Compared to EMs of the control group (placebo saline and standard feed), P  < 0.05. (a, e) EMs in the control group (400×); (b, f) EMs in the Es group (400×); (c, g) atypical EMs in the Es group (400×); (d, h) malignant EMs in the Es group (400×). The samples with simple hyperplasia and atrophy were excluded. The staining of PTEN was weakened, and stainings of p-AKT and p-mTOR were intensified progressively in the order of EMs, atypical EMs, and malignant EMs ( P  < 0.05). The samples with the same pathological changes showed no significant difference in the abovementioned stainings, regardless of the treatment group or whether the samples were eutopic or ectopic ( P  > 0.05). The differences in the H -score for PTEN, p-AKT, and p-mTOR among groups are shown in Figure 3 and Table 3 . Statistical comparison of the expressions of phosphatase and tensin homolog (PTEN), phosphorylated protein kinase B (p-AKT), and phosphorylated mammalian target of rapamycin (p-mTOR) protein by an immunohistochemical method, P -values P  < 0.05, statistical difference; P  < 0.01, significant statistical difference. Control group, rats with surgically induced endometriosis (EMs) treated with placebo saline and fed a standard diet; Es group, rats with surgically induced EMs treated with estradiol and streptozotocin and fed a high carbohydrate-and-fat diet. Immunostaining of phosphatase and tensin homolog ( PTEN ) in benign (a), atypical (b), and malignant endometriosis (EMs) (c) (400×) in a rat model. Immunostaining of phosphorylated protein kinase B (p-AKT) in benign (e), atypical (f) and malignant EMs (g) (400×). Immunostaining of phosphorylated mammalian target of rapamycin (p-mTOR) in benign (i), atypical (j), and malignant EMs (k) (400×). H -score values of PTEN (d), p-AKT (h), and p-mTOR (l) immunostaining in EMs, atypical EMs, and malignant EMs.

The present study explored a feasible way of inducing malignant transformation of rat EMs in combination with hyperestrogenemia and type II diabetes. The results indicated that three cases of ectopic endometria had atypical hyperplasia (two cases in 4 months and one case in 8 months), and two cases had endometrioid adenocarcinoma (both in 8 months), which were similar to the pathological appearances of human disease. The process of pathological changes, from benign to atypical and then to malignant change, is consistent with that of the human disease, especially the occurrence of atypical EMs (3/50). Ogawa et al . 24 found atypical EMs in 78% EAOC, and it can act as a premalignant lesion. Atypical EMs was also reported to have spatial and chronological association with MTOE. 25 , 26 It may represent the transition from benign EMs to malignant carcinoma. In the Es group, most EMs lesions grew along with time, the ectopic endometrial cells had rigorous hyperplasia, and the organelles were active. In the control group without intervention, atrophy worsened along with time, indicating estrogen dependency and the significant role of hyperestrogenism in this study. In our previous study, we injected once only with estradiol, without STZ and HCF-feed. After 8 months, only 9 of 14 cases of EMs presented with simple hyperplasia, with no malignant or atypical cases (unpublished results, 2014). This illustrated the indispensible role of type II diabetes in this study, and the combination of both treatments was more effective and feasible. Although STZ and estradiol were both give i.p., special attention was paid during the surgery and post-surgery treatment to assure that both the eutopic and ectopic endometrial side was not directly exposed to the peritoneal cavity and carcinogenetic stimulations. So the route of administration was not supposed to take an influential part in the hyperplastic progression of disease. Rare reports have been published regarding the correlation between STZ with MTOE, but indeed the possibility cannot be ruled out that STZ might be an unknown interference, which may be worthy of further investigation. The inactivation of the PTEN gene and the following activation of the PI3K–AKT–mTOR pathway were reported to be associated with MTOE. 27 Silencing PTEN could lead to the abnormal activation of PI3K and the phosphorylation of AKT, a serine/threonine kinase involved in cell growth and survival. Activated AKT (or p-AKT) could further activate mTOR by phosphorylation, which could regulate cell stress response and cell cycle. 12 As PTEN mutations are seen early in endometrioid ovarian carcinoma and endometrioid ovarian cancer is thought to arise from EMs, it has been proposed that somatic genetic alterations in the PTEN gene may be an early event in MTOE. 6 The ovarian endometrioid adenocarcinoma with an activated PI3K–AKT–mTOR signaling pathway is more likely to be of a low grade instead of a high grade. 28 In our present research, PTEN expression was progressively reduced in the order of EMs, atypical EMs, and malignant EMs. The most remarkable abnormality was seen in malignant cases, with conspicuously decreased PTEN and increased p-AKT and p-mTOR. These results implied a possible role of the loss of PTEN and the following activation of the PI3K–AKT–mTOR pathway in the development of the present rat model of MTOE and a shared molecular alternation of the present rat model of MTOE with human low-grade endometrioid adenocarcinoma. Recently, there is growing evidence pointing to hereditary predisposition and biological characteristics of the eutopic endometrium as the determinative factors in the pathogenesis of MTOE. 29 Vigano et al . 30 proposed that the malignant transformation of eutopic and ectopic endometria shared similar incidence and pathological type, and that abnormal gene expressions in the eutopic endometrium could be related to MTOE. Clinically, it was reported that patients with ovarian tumor arising from EMs suffered higher risks of endometrial pathology (endometrial adenocarcinoma, hyperplasia, and polyps). 14 Studies have also shown that women with EAOC will also have a simultaneous endometrial adenocarcinoma, and almost 90% of synchronous tumors identified in the ovary and in the endometrium were of the endometrioid cell type. 31 , 32 In the present study, we found that the pathological grade of the eutopic endometria corresponding with malignant EMs was higher than that with atypical EMs. The former was atypical hyperplasia, whereas the latter was simple hyperplasia. It is assumed that rats with malignant EMs might have abnormal gene expressions in their eutopic endometria. Such endometrium and its EMs cysts may be more sensitive to carcinogenic stimulations and thus more susceptible to malignant transformation. Moreover, the EMs lesions and the corresponding eutopic endometria in the Es group showed the same progressively reduced tendency with regard to PTEN , p-AKT, and p-mTOR expressions (no malignant case in the eutopic ones). But in a certain pathological grade of either eutopic or ectopic endometria, the expressions of the above proteins showed no statistical differences. These results contributed to imply the similarity and correlation between malignant transformation of EMs and endometrium. However, as two independent disorders, MTOE and endometrial carcinoma must have many distinctions. Malignant transformation of EMs commonly occurs at reproductive age and seems to be associated with oxidative stress induced by free iron and heme accumulated during periodic and repeated bleeding, and inflammation. 6 In contrast, type I endometrial carcinoma often occurs in premenopausal and younger postmenopausal women, is strongly associated with the estrogen-related pathway, and arises in association with unopposed estrogen stimulation. 33 In the present study, we found that the progression of hyperplasia of EMs was more advanced than the corresponding eutopic endometria, indicating a higher risk of malignization of EMs. This phenomenon may be attributed to the specific local microenvironment and enhanced sensitivity to carcinogenic risk factors of EMs. Studies on humans and rats both proposed that there were more activated inflammatory cells and cytokines in the site of EMs than the corresponding eutopic endometria, 34 , 35 indicating an inflammatory microenvironment of EMs. Inflammatory cells and their secreted cytokines can stimulate angiogenesis and cell proliferation, inhibit cell apoptosis, facilitate invasion and metastasis, and produce active oxygen capable of inducing DNA damage and mutation. 36 So inflammation can prompt growth and invasion of EMs, which constitutes a link between EMs and ovarian cancer. 2 In addition, the electron microscopy of the malignant endometriotic cells in our study revealed the presence of mitochondrial swelling and vacuole changes, which reflected that the EMs lesions might grow in a hypoxic microenvironment and implied a possible role of the hypoxic damage to mitochondria in the relatively higher malignant transformation risk of the endometriotic lesions. Apart from these, ectopic survival and growth might induce a higher sensitivity to various simulations. 37 , 38 Wu et al . 38 found that ectopic endometrium was at least 100 times more sensitive to interleukin-1β, compared with eutopic endometrium. Therefore, the hypersensitive EMs lesions might exaggerate the carcinogenic simulations and trigger the cascade chain reaction of oncogenesis. In summary, the present study found that the combination of hyperestrogenemia with type II diabetes in a rat EMs model can induce MTOE, which showed similar morphological changes and molecular abnormalities to human disease. The method of induction is easy and feasible, and provides a good foundation for establishing an animal model of MTOE. Although eutopic endometrium and EMs showed similar responses and pathological changes to the carcinogenic simulations, EMs had higher susceptibility and risk for oncogenesis. Therefore, malignant transformation of EMs and eutopic endometrium may share hereditary correlations, which are worthy of further investigation. Future studies might also focus on raising the malignant transformation rate by improving the induction method and exploring the genetic molecular mechanism of MTOE in more depth.

Ninety adult female Sprague–Dawley rats (age, 8–12 weeks; weight, 250–300 g), were provided by the Experimental Animal Center of Shengjing Hospital of China Medical University (Shenyang, China). The animals were fed standard feed and housed in a controlled environment (22 ± 2°C) with 12:12 h light:dark cycles. This study was approved by the Ethics Committee of Medical Scientific Research and Technology, Shengjing Hospital of China Medical University (2013PS140K). The rats were randomized into four groups: (i) the Es group ( n  = 55), autologous endometrium transplantation; 19 (ii) the control group ( n  = 25), autologous endometrium transplantation; (iii) the negative control group ( n  = 5), autologous epiploon transplantation; and (iv) the blank control group ( n  = 5), “open–close” surgery. Details are given in Document S1. After surgery, the rats of the Es group were treated with 17β-estradiol (5 mg/kg, three times/week) i.p. and provided with HCF-feed (10% fat, 20% glucose, and 70% standard feed). One month after surgery, a single dose of STZ (25 mg/kg) was injected i.p. Three days later, fasting blood glucose measurement higher than 250 mg/dL was diagnosed as diabetes. The rats in the control, negative control, and blank control groups were treated with only vehicle saline and provided with standard feed. All the i.p. injections were carried out on the opposite side of the EMs foci to minimize direct simulation. The rats were killed randomly by cervical dislocation 2 months (Es group, n  = 5; control group, n  = 6; negative control group, n  = 5; blank control group, n  = 5), 4 months (Es group, n  = 20; control group , n  = 7), and 8 months (Es group, n  = 25; control group, n  = 7) after the surgeries. Details are given in Document S1. For H&E staining, paraffin-embedded samples were sectioned at 5-μm thickness, stained with H&E, and examined under a light microscope. The EMs lesion was determined by the histological confirmation of glandular epithelial and stromal cells. For electron microscopic analysis, Epon 812 epoxy resin-embedded samples were cut into slices of 50-nm thickness, and then the histological ultrastructure of endometrial cells was observed with a transmission electron microscope. Details are given in Document S1. Cell proliferation was assessed by detection of proliferating cell nuclear antigen (PCNA; Keygen Biotech Company, Nanjing, China). Apoptosis was evaluated with a TUNEL detection kit (Keygen Biotech Company, Nanjing, China). Both detections were carried out following the manufacturer's instructions. The level of apoptosis was established by counting the number of positively stained cells (brown in color). For details of the procedure, see Document S1. The reagents and dilution of the primary antibody were as follows: PTEN (1:100, Cell Signaling Technology, Danvers, MA, USA), p-AKT (1:50; Santa Cruz Biotechnology, St Cruz, CA, USA) and p-mTOR (1:50; Cell Signaling Technology). The secondary antibody and 3,3′-diaminobenzidine-tetrachloride kit were purchased from Zhongshan Goldenbridge Biotechnology (Beijing, China). The assessment criteria for results of the proliferation analysis, apoptosis analysis, and immunohistochemistry were as follows. Five view fields under a high power lens of the microscope (400×) were randomly chosen to observe all the cells in the view fields. The H -score scoring method was used (details provided in Doc. S1). 23 SPSS version 16.0 software (Armonk, NY, USA) was used. The weight, volume, and H -score value were described by mean ± SEM. The t -test for two independent samples was used to calculate the intergroup difference. The χ 2 -test and Fisher's exact probability test were adopted for comparative analysis of the count data. P  < 0.05 was considered significantly different.