Oestrogen-induced angiogenesis and implantation contribute to the development of parasitic myomas after laparoscopic morcellation

Uterine myomas (UMs) are the most common gynaecologic tumours, occurring in 40–70 % of women over 30 years of age [1]. For years, the only curative treatment for myomas is considered to be surgery including hysterectomy and myomectomy [2]. Laparoscopic procedures can be performed with decreased morbidity compared with exploratory laparotomy [3]. Steiner et al. introduced the first ‘electrical cutting device’ for laparoscopic removal of tissue from the abdominal cavity in 1993 [4]. However, power morcellation is certainly associated with tissue spreading per se resulting in intra-corporeal ectopic implants [5, 6]. The first case of a parasitic myoma (PM) after use of the laparoscopic morcellation was reported in 1997 by Ostrzenski [7]. With increasing frequency, laparoscopic and robotic-assisted myomectomies/hysterectomies are being performed in managing women with symptomatic myomas [5], a number of case reports were published that reported growth of PMs after morcellation.

The reported incidence of PMs after laparoscopic myomectomy was 0.20–1.25 % [8, 9]. Patients with PMs presented with multiple lesions and varying sizes (range, 0.8–30 cm). PMs largely occurred in the dependent part of the abdominal cavity, including intestines, peritoneum, omentum and port sites, and received abundant blood supply [8, 10‐17]. Most women (78.3 %) presented with symptoms, such as abdominal or pelvic pain, dyspareunia, abdominal distension, abdominal pressure, urinary frequency and constipation [13, 14, 17‐23]. In 10.1 % patients, debulking procedure, such as omentectomy, appendectomy, or bowel resection, was necessary to eliminate all PMs [18, 24, 25]. Thus, it is important to elucidate the pathophysiology of laparoscopic morcellator-induced PMs.

UM growth is dependent on the sex steroid hormones, oestrogen and progesterone [26]. Data from in vitro and animal models over decades suggest that oestradiol (E2) plays a central role in myoma growth via its receptor, oestrogen receptor α (ERα) [27, 28]. Most medical treatments reduce menstrual bleeding in patients with myoma, including gonadotrophin-releasing hormone agonist (GnRHa), aromatase inhibitors (AI), selective oestrogen receptor modulators (SERMs), progestins and selective progesterone receptor modulators (SPRMs) [1, 29, 30]. However, only GnRHa, AI and SPRMs can reduce both myoma volume and menstrual bleeding. Thus, it is hypothesised that exposure to oestrogen and progesterone could be a risk factor for the development of PMs [8, 31]. Until now, these evidences in managing PMs were based on case reports and small case series, which most received surgical treatment. Thus, the role of medical treatments for laparoscopically-induced PM remains largely unelucidated.

Due to the implantation and growth behaviour of PMs, we reasoned that oestrogen and angiogenesis might be involved in the development of PMs. In the present study, we simulated the laparoscopically-induced PMs xenografted mouse model to demonstrate the role of oestrogen-induced angiogenesis in the development of laparoscopically-induced PMs. Furthermore, we investigate the therapeutic effects of hormonal manipulation for PMs.

Until now, the treatment and prevention of laparoscopically-induced PMs are still challenges for clinical physicians. In this study, we employed a new animal model to examine the role of oestrogen in laparoscopically-induced PMs. First, implanted myomas possessed more ERα, PR, angiogenesis and proliferative property compared with pre-xenografted or non-implanted myomas. Second, depletion of oestrogen significantly decreased laparoscopically-induced PM implantations. The implantations, angiogenesis, and proliferative property of PMs were associated with the serum E2 levels. Third, sex steroid hormone modulator, AI, decreased the implantations, angiogenesis, and proliferative property. To our knowledge, this is the first report to demonstrate the involvement of oestrogen-induced angiogenesis in the development of PMs. These data highlighted the crucial role of oestrogen-induced angiogenesis and implantations in the development of PMs, and hormonal manipulation with AI might be potential remedies in preventing laparoscopically-induced PMs.

Research of UMs is still challenged by the need for in vivo models. Until now, the laparoscopically-induced PMs in vivo model is lacking. At present, the commonly used transplantation model for myomas is the ELT3 cell injection model. Owing to a germ-line mutation of the tuberous sclerosis gene 2 (Tsc2), Eker rats spontaneously develop various neoplasms, including leiomyosarcoma [38‐40]. Nonetheless, the Tsc2-driven tumour development might not reflect the sporadic pathogenesis of myomas. An alternative model described by Hassan et al. [41], who provoked growth of myoma-like xenograft implanted subcutaneously in SCID mice, by transplanting human tumour tissue pieces overexpressing COX2 (cyclooxygenase 2) and VEGF through adenoviral transduction. However, that model exhibits substantial limitations because it is driven by the ectopic overexpression of COX2 [42‐44]. In recently, Drosch et al. report an approach by generating a myoma xenograft model through injection of human myoma–derived primary cells without genetic manipulation into myometrium of mice [45]. According to Drosch’s report, primary myoma cells are suited to generate fibroid-like xenografts for studying pathogenesis without genetic modifications [45]. Furthermore, one single stem cell is thought to give birth to a specific myoma (which is why it is called a clonal disease) [46]. Thus, it is reasonable that human UM xenografts are suited to generate PMs-like xenografts model without genetic modifications. For the first time, we could demonstrate this laparoscopically-induced PMs and setup a pneumoperitoneum system in SCID mice. SCID mice possess a combined congenital deficiency in T- and B-lymphocyte function, and cannot reject transplanted non-autologous tissues [32, 47]. Those tumours closely mimicked laparoscopically-induced PMs by their highly similar histology and growth characteristics. The xenografted tissues can be used to simulate aspects of the myoma-host microenvironment during the pathogenesis of PMs. Moreover, using different donors would also enhance the interpatient variability. Aside from these restrictions, the described model provides a powerful tool for basic research to investigate the pathogenesis of PMs, because this model omits any genetic modifications and closely resembles the pathogenesis of laparoscopically-induced PMs. This model can be used as a powerful tool to study mechanisms involved in the pathogenesis of laparoscopically-induced PMs and will open new opportunities for their treatment.

Although quite a lot is known about the factors contributing to myoma growth, the pathophysiology of PMs remains largely unelucidated [1]. Traditionally, PMs were supposed to be pedunculated subserosal myomas that were accidentally separated from the uterus, and subsequently attached to other organs in the pelvis for the provision of their blood supply [48]. Since the published case reports of PM after laparoscopic myomectomy, it was hypothesised that iatrogenic PMs could develop by seeding of retained small tissue fragments with stem cell after morcellation in the peritoneal cavity [22, 49]. Our result showed that implanted xenografts similar to primary UM histologically. Xenografts showed features of UM-like fusiform cells: a whirl-like pattern of the smooth muscle bundles and cigar-shaped nuclei determined by histology (Fig. 2c). In addition, SMA showed a strong abundance in the implanted xenografts and in the primary tumours (Additional file 5: Figure S3), confirming the smooth muscle differentiation of the xenografts. According to our case report and animal model, angiogenesis and cell proliferation increased in implanted xenografted myoma compared with primary or non-implanted xenografted myoma. Of note, Ishikawa et al. showed that the growth of tissue xenografts was strongly dependent on the cell density of the xenografts [22]. Hassan et al. showed VEGF provoked growth of myoma-like xenografts implanted subcutaneously in SCID mice [41]. VEGF-enhanced angiogenesis is also associated with an increase in vascular permeability, which results in an increase in the amount of growth factors and nutrients delivered to tumour cells [44, 50]. Thus, implantation, angiogenesis and cell proliferation are the important mechanisms for laparoscopically-induced PM growth.

UM are oestrogen- and progesterone-dependent monoclonal tumours that arise from the uterine smooth muscle tissue [51]. Data from in vitro and nonhuman animal models over decades suggest that E2 plays a central role in myoma growth via its receptor, ERα [28, 46]. Furthermore, the impact of sex hormones on the PM growth was illustrated by cases of rapid growth of PMs during pregnancy. Cucinella et al. [8] report a case of a woman where an asymptomatic PM was discovered during a cesarean section, 24 months after the procedure of laparoscopic myomectomy. Takeda et al. [52] published a case report of a woman who was diagnosed with a PM 2 years after the procedure of laparoscopic myomectomy. After 2 years of conservative treatment, the size of the mass remained the same, yet during pregnancy, the rapid growth of this mass was observed, supporting the sex hormonal impact on the growth of PM. In our study, the non-implanted myoma fragments disclosed decreased expressions of both ERα and PR compared with implanted lesions, and this result suggested the disseminated fragments with strong positive ERα and PR might be essential in the development of PM implantations. Our study also demonstrated that depletion of oestrogen decreased the angiogenesis, proliferation and implantations, and the result was correlated with previous studies which demonstrated shrinkage of benign metastatic leiomyomatosis after bilateral salpingo-oophorectomy (BSO) surgery [53‐56]. The duration of sex steroid hormone exposure after laparoscopic morcellation might also be a risk factor for the development of PMs [26]. According to the systematic review, 69 cases PMs after laparoscopic myomectomy by Meulen et al., it is hypothesised that prolonged exposure to sex steroid hormones such as hormonal replacement therapy, could be a risk factor for the development of PMs [8, 16, 23, 31, 57]. Our present study highlighted the crucial role of E2 and ERα and PR expressions in the development of laparoscopically-induced PMs.

To prevent PMs, it is important not to disperse tissue fragments during intra-corporeal power morcellation and to inspect all the surgical fields during laparoscopy with caution [58, 59]. Until now, no effective method to prevent tissue dispersion from power morcellation has been established [60‐62]. Most patients with symptomatic PMs after laparoscopic surgeries at their reproductive age, and BSO might not be indicated in all patients for the purpose of prevention of PMs [63, 64]. Thus it is necessary to find medical prevention and treatment for those patients. Leiomyomatosis peritonealis disseminata (LPD) is a disease that is characterised by the presence of many (sub-) peritoneal smooth-muscle nodules disseminated through the omentum and peritoneum. The nodules are thought to originate from metaplasia of subperitoneal mesenchymal stem cells without a history of laparoscopic myomectomy. Although LPD is a different entity when compared with UMs, there might be a relation between morcellation of UMs and an iatrogenic development of this disease. Several studies which presented that LPD shrink under sex steroid hormone manipulation, including BSO [53‐55], menopause [65], megestrol [66], SERM [55], AI [55, 67], and GnRHa [55, 56, 68]. Laparoscopic myomectomy might result in the growth of multiple PM nodules mimicking LPD [31]. Thus it is reasonable to test the prevention and therapeutic effect of sex hormone manipulation on PMs after laparoscopic surgery.

Our xenografted mouse model demonstrated that AI significantly decreased the implantations rate and weight of implantations, while GnRHa and SERM did not change the implantation rate.

Aromatase is a cytochrome P450 enzyme (CYP-19) that allows the transformation of androgens into oestrogens. AIs are compounds that interact with the hormone-binding site of the molecule (exemestane) or with its catalytic subunit (anastrozole, letrozole). This is supported mostly by clinical data showing that AIs are as effective as GnRH analogues in reducing myoma volume in premenopausal women. Letrozole reduces fibroid volumes by 46 % (vs. 32 % in the GnRH analogue group) [69]. In peripheral tissues, including skin and adipose tissue, and the ovaries, aromatase catalyses the formation of oestrogen, which reaches UM tissue through the circulation. In addition, aromatase in myoma tissue converts androstenedione of adrenal or ovarian origin to oestrogen locally [70]. This is possible reason AI is more effect than GnRHa in our study.

After binding to the GnRH receptors, GnRHa induces subsequent stimulation of gonadotropin secretion followed by desensitisation, and thus they delay the gonadotropic axis blockade. In vitro, GnRHa inhibits cell proliferation and induces apoptosis [71]. Even though, GnRHa is the most effective agent to correct anaemia, and reduce fibroid volume by 50 % after 2–3 months treatment [1, 2]. However, in our study, GnRHa did not reach the comparable results with AI and SPRM in lowering uterine myoma implantations. The possible reason was that because GnRHa takes approximately 1–3 weeks to obtain a fully hypogonadotropic hypogonadal state to decrease serum E2 level [72, 73], while we sacrificed mice only after 3 weeks GnRHa treatment. Thus, further study of the suitable timing of GnRHa treatment for PMs is needed.

SERMs are nonsteroidal oestrogen receptor ligands with agonist or antagonist effects depending on the tissue. Raloxifene (SERM) has been approved for the treatment and prevention of post-menopausal osteoporosis [74]. Three randomised controlled trials have evaluated raloxifene in premenopausal women with confirmed UMs [75]. Two of these trials, including 215 women, showed the therapeutic efficacy of raloxifene, but the third did not. This may be due to the rise in E2 secretion observed in premenopausal women following SERM treatment. There was no therapeutic effect on PMs in our study.

Our study had some limitations, including lack of long-term follow-up of treatment of different ligands, and lack of serum hormone level changes of progesterone. In addition, we did not examine the expression of the somatic mutations in the mediator complex subunit 12 (MED12) gene which might be associated with 50 % of UM [76].

Until now, there is still no effective method to treat and prevent PMs after laparoscopic power morcellation. The identification of tumourigenic factors will give insights into the pathogenesis of UMs and might open new possibilities for drug testing. Based on the present study, PMs are oestrogen-dependent and AI might be potential remedies to prevent PMs after laparoscopic procedures.