Keywords

1 Introduction

Endometriosis is commonly believed to occur via retrograde menstruation, also known as the Sampson hypothesis [1], in which viable endometrial tissue flows retrograde through the fallopian tube and into the peritoneal cavity where it can attach to and invade tissues and organs within the cavity. At least 90 % of women experience retrograde menstruation, but endometriosis occurs in only 10–14 % of reproductive age women, suggesting that additional elements impact its etiology [2]. Since endometriosis impairs the quality of life of severely affected women, improved research and new treatment paradigms are needed.

Current medical therapies for endometriosis aim to decrease ovarian estrogen production and/or counteract estrogen effects with the use of GnRH agonists, progestins (including oral contraceptives), and androgens, but undesirable side effects limit their long-term use [2]. Likewise, the scope of surgical treatment is also limited by a high recurrence rate, which may eventually lead to extreme measures, such as removal of the uterus and ovaries. There is therefore a need for focused mechanistic research that can be translated into expanded therapeutic capability for this widely prevalent disease.

Animal models, which are important for elucidating the mechanisms underlying endometriosis and are used in the early stages of drug testing, usually rely on non-menstruating rodents with induced endometriosis-like lesions. One of the major limitations in endometriosis research is the paucity of robust animal disease models. Ideally, a disease model should mimic human disease and allow scientific investigation into the effects of both intrinsic (e.g., genes) and extrinsic (e.g., environment) factors on disease progression.

Several animal models of endometriosis have been established, most consisting of transplantation of endometrium into the peritoneal cavity, which is by far the most common site of disease. Besides rodents, primates, such as monkeys that spontaneously develop endometriosis or that have been transplanted intraperitoneally with endometrium, can be used to study drug candidates. Indeed, because endometriosis occurs spontaneously in monkeys, the only nonhuman animal model that has cyclic menstrual periods, such an animal model is undoubtedly the most reliable one [3–7]. However, high cost and difficulties maintaining these animals have led most researchers to concentrate their studies on smaller mammalian models, such as rats [8–12] and rabbits [13, 14]. Nevertheless, the murine model was considered as a possible candidate for this purpose, and these studies were restricted to nude [15, 16] or SCID mice [17], which were implanted with human endometrium (xenotransplantation). Since mice are much smaller than rabbits or rats, microsurgical procedures of tissue transplantation become even more complicated.

The murine model may provide advantages in terms of new therapeutic approaches. In recent decades, many knockout or transgenic mice have been generated. The availability of an endometriosis model in mice is crucial because it can be used to investigate some aspects of endometriosis.

2 Murine Endometriosis Model

Mice are the most common animal models capable of investigating the pathophysiology of endometriosis; however, they do not spontaneously develop endometriosis. To induce endometriosis in mice, endometrial tissue must be transplanted into the peritoneal cavity using several methods [18–21], which can be classified into two basic types, homologous and heterologous. Both models produce comparable phenotypes, which are then morphometrically evaluated.

In homologous or autologous models, normal endometrial tissue is transplanted into the peritoneal cavity of immunocompetent recipients and starts to grow in an estrogen-dependent manner. In almost all models, uterine endometrial fragments from a donor mouse are directly introduced via injection into the peritoneal cavity of an immunocompetent syngeneic recipient without suturing the implants. In heterologous models, human endometriotic lesions are transplanted into the peritoneal cavity of immunodeficient mice [22, 23]. Human endometriotic implants were sutured to the peritoneum of immunocompromised mice [24, 25]. This xenotransplantation model using the nude mouse is also used, but is limited by the lack of a normal immune system. Despite the advantage of being based on human endometrial tissues, the number of endometriotic lesions in the heterologous model that will develop varies from one animal to another. In both models, with or without suturing the endometrial implant, the drug’s influence on growth of endometrial or endometriotic transplants is evaluated.

These established endometriosis models in mice are under discussion. Indeed, mice do lack a menstrual cycle and do not develop spontaneous endometriosis. It should be established whether transplanting normal endometrium into the peritoneal cavity of a non-menstruating species reflects all pathophysiological aspects of human endometriosis. Although the murine endometriosis model is not exactly the same as human endometriosis, their endometriotic lesions develop in some ways the same as human endometriotic lesions.

The murine model of endometriosis is a versatile one used to study how the immune system, hormones, and environmental factors affect endometriosis and how endometriosis affects fertility and pain. A novel study design also enables the evaluation of molecular mechanisms that are critical for disease initiation [26, 27]. In Table 14.1, the recent representatives of mouse endometriosis models are shown. Additional and ideal models of endometriosis are needed. In the next section, we demonstrate the results in a murine endometriosis model that expands the capability of conducting both mechanistic and translational research.

Table 14.1 Representatives of mouse endometriosis models
Full size table

3 Our Animal Model of Endometriosis Using Homologous Mice

3.1 Care and Treatment

According to the implantation theory, we established the readily available murine model to evaluate the development of endometriosis-like lesions [27, 28]. Female mice (6 weeks of age, BALB/c) were purchased. Before initiating the experiments, animals were allowed to acclimate to the following conditions for 7 days. Mice were in a controlled temperature range (72–74 F) on a 12-h light, 12-h dark cycle. Mice were given food and water ad libitum. Recipient mice were ovariectomized through two 0.5-cm dorsolateral skin incisions and were then divided into two treatment groups, estradiol valerate (0.5 μg/mouse · week) in corn oil or only corn oil vehicle. Mice were dosed subcutaneously once per week for 2 weeks before inducing experimental endometriosis. Donor mice were primed 41 h before removing the uterus with pregnant mare serum gonadotropin (10 IU intraperitoneally). The donor uterus was removed en bloc after euthanasia, cleaned of excess tissue, and washed thrice in sterile PBS. The uterus was slit with a linear incision longitudinally and minced (≤1.5 mm). Recipient mice were anesthetized using isoflurane/oxygen and given buprenorphine (0.1 mg/kg) for pain management. A 0.5-cm right dorsolateral incision was made; minced donor tissue (1:2 donor uterus to host ratio) in 500 μL PBS was injected into the peritoneal cavity of the recipient and gently massaged to disperse the tissue. An equivalent amount (~100 mg) of minced tissue was transferred into all recipients. Mice were treated for 4 additional weeks with estradiol valerate or the vehicle. After 4 weeks, mice were euthanized with CO2, the peritoneal cavity was opened, and endometriosis-like lesions were removed. To assess the effects of drug candidates on ectopic uterine tissue, ectopic lesions were photographed to document in situ endometriosis-like lesions. Endometriosis-like lesions were visualized, dissected, measured, and weighed. Endometriosis-like lesions were removed and either fixed in 10 % formalin or snap-frozen on dry ice and stored at −80 °C until use.

4 Effects of Parthenolide on the Endometriosis-Like Lesions in a Murine Model

We undertook a study of feverfew as a potential therapy for endometriosis that illustrated the effectiveness of the murine model. The medical herb feverfew has long been used as a folk medicine for treating fevers, migraine, rheumatoid arthritis, and dysmenorrhea. Parthenolide is considered the primary bioactive compound in feverfew having antitumor and anti-inflammatory properties [41]. Parthenolide has produced anti-tumorigenesis effects against human acute myeloid leukemia and solid tumors, such as breast and pancreatic cancer [42–44]. We therefore used an experimental mouse model to evaluate the effect of parthenolide as therapy for endometriosis [28].

A murine model was established by injecting tissue suspension intraperitoneally. Endometriosis-like lesions had grown in the abdominal cavity of all mice. The deposits appeared as cystic lesions bulging under the serosal coat. Most of the lesions were observed around the abdominal incision, the intestinal membrane, and the renal capsule (Fig. 14.1a). The size of lesions ranged from approximately 2 to 8 mm in diameter (Fig. 14.1b). Histologic sections stained with hematoxylin and eosin were endometriotic in character. The monolayer epithelial cell lining of the cyst was revealed by HE staining (Fig. 14.1c, d) [28]. We confirmed that cytokeratin (a marker of epithelial cells) and vimentin (a marker of stromal cells) in the mouse endometriosis-like lesions were positive, whereas calretinin (that of mesothelial cells) was negative, indicating that these cystic lesions originated from the injected endometrial tissues, not the peritoneal cells.

Fig. 14.1
figure 1

Cystic lesion in the murine endometriosis model. (a) An endometriosis-like lesion developed in the murine abdominal cavity. (b) Representative of a large excised lesion. (c) and (d) HE staining of lesion

Full size image

In view of the above considerations, we proposed a hypothesis that parthenolide may have an inhibitory effect on the development of endometriosis. After parthenolide treatment for 4 weeks, the total number of lesions (5.8 vs. 3.9/mouse) was significantly reduced, and the average weight (65.6 vs. 29.6 mg/mouse) and the surface area (50.3 vs. 25.7 mm2/mouse) of lesions were decreased by approximately 50 % of controls. To evaluate the proliferative activity of lesions, the ratio of Ki67-stained cells was calculated. In endometrial glands epithelia, the percentage of Ki67-positive cells decreased after parthenolide treatment (17.8 vs. 8.5 %).

5 Discussion

Convenient and reliable endometriosis animal models are needed to accelerate emerging therapeutic alternatives. According to the Sampson “implantation theory,” we established the syngeneic immunocompetent mice model. In this model, we and the other investigators provided crucial evidence of both the development of endometriosis-like lesion growth and donor tissue responsiveness. Recently, Pelch et al. described a detailed method of a mouse surgically induced endometriosis model by autotransplantation of uterine tissue [45].

The rodent endometriosis models using mice or rats are widely used in research, but may have limitations and may not mimic all aspects of human pathophysiology. For example, in homologous rodent models, “healthy” uterus is cut into fragments and transplanted into the peritoneum, whereas it has been suggested that the eutopic endometrium of women suffering from endometriosis may already be abnormal [2]. Nude mice lacking an intact immune system are employed in the heterologous model, which cannot mimic the inflammatory response normally seen in human endometriotic lesions [23]. Although heterologous rodent endometriosis models are responsive to drugs and manipulations that induce a hypoestrogenic state, such as ovariectomy, GnRH agonists, aromatase inhibitors, danazol, and selective estrogen receptor modulators, it may be difficult to analyze novel target families in which, for example, the murine ligand does not bind to the receptor of the human transplant.

Whereas cell-based in vitro experiments provide a framework for testing molecular mechanisms, eventually confirming their role in disease causality in vivo can only be accomplished by a suitable animal model. For a disease as diverse as endometriosis, a single animal model would unlikely be sufficient to represent the entire diversity in etiology, pathogenesis, and pathology. Each model has design-related strengths and limitations. For example, a recently described disease model consists of introducing endometrial tissue via direct injection into the peritoneal cavity of immunocompetent mice without suturing [21, 27, 28]. In mice, the injected tissue forms cyst-like endometriosis lesions; however, the injection method does not seem to work in rats because the tissue fails to attach to and invade the peritoneal cavity [9]. Since all endometriosis lesions in each model are attached to the peritoneum and/or mesenterium with or without suturing, evaluating attachment or invasion of lesions would be difficult. Furthermore, to quantify the injected endometrial fragments, this model may be insufficient to evaluate precisely the endometriotic lesions.

The rodent model is used extensively to study the etiology, pathology, and risk factors of endometriosis [21, 22, 27, 30, 33, 35, 46, 47] as well as to explore novel therapeutics [28, 29, 31, 32, 34, 36–40, 48–51]. In conclusion, this murine model of endometriosis provides an important tool to evaluate a therapeutic approach to the disease. This model will help to better understand disease evolution in the living animal and permit faster and more accurate characterization of a drug’s effect on experimental endometriosis.