A protocol for creating endometriosis in rhesus macaques (Macaca mulatta)

Summary of the study outcome is provided Table 1 and Table 2 , and details for all 18 individual animals is provided in Table 3 ( Supp info ). As indicated above, one animal displayed advanced pre-existing endometriotic disease during laparoscopy in treatment cycle 1. This animal had multiple lesions >1 cm in diameter adjacent to the uterus, creating a large hemorrhagic cyst, and the endometriotic tissues were collected to provide samples for immunohistochemical analysis (see below). The remaining 17 animals were endometriosis-free at the first laparoscopic protocol. Figure 1a shows a representative laparoscopic photograph of the abdominal cavity of an animal without endometriosis at the time of first-seeding. Figure 1b shows a laparoscopic uterine puncture followed by transfer of the menstrual endometrium ( Figure 1c ) to the abdominal cavity. Menstrual samples were collected from the control animals. These consisted of blood and endometrial fragments. The endometrial fragments contained glands ( Figure 1d ) and stromal cells that stained positive for ERα ( Figure 1e ) and PGR ( Figure 1d ). As expected the PGR staining was minimal in the glandular epithelium of the menstrual fragments but strongly retained in the stromal cells, while ERa was observed in both glands and stroma. The 3 control animals that underwent uterine puncture without peritoneal seeding displayed filmy abdominal adhesions but failed to develop endometriotic lesions. In contrast, seeding menstrual endometrium resulted in the formation of red endometriotic lesions and adhesions in all 14 monkeys. The abundance of lesions at the end of the treatment cycles is summarized in Table 1 and the intra-abdominal distribution of lesions is shown in Table 2 . There was no significant difference in the abundance, or appearance, of lesions between animals with induced versus natural cycles. The induced disease presented as adhesions, small red, black or dark brown ("power burn,") or white lesions on serosal surfaces of the abdominal organs ( Figure 2 ). Red lesions were characterized as either small (5 mm), but none of the induced lesions were larger than 1 cm across. Ovarian adhesions are shown in Figure 2a . Small red lesions were significantly more abundant on the surface of the uterus, the uterine cul-de-sac ( Figure 2b ), bladder ( Figure 2c ) and omentum ( Figure 2d ) and peritoneum ( Figure 2e ) than near the oviducts ( Figure 2a ). Large red lesions were more evident on the uterus ( Figure 2 f ) and in the uterine cul-de-sac. In contrast brown lesions ( Figure 2g ) were most abundant on omentum and white lesions were evident on the bladder and uterus. Figure 2h and i show representative adhesions on the bladder and omentum. Histology and immunohistochemistry of the one spontaneous case of endometriosis revealed that the ectopic endometriotic lesions contained endometrium-like glands and stroma (not shown). Induced red lesions appeared histologically similar to spontaneous endometriosis with abundant glands and stroma that stained strongly positive for ERα and PGR and had abundant CD68 positive macrophages cells. As expected, the endometriotic gland and stroma of red lesions displayed nuclear staining for ERα and PGR ( Figure 3 ). Brown lesions also stained for ERα ( Figure 3e - f ) but PGR was greatly reduced. White lesions had minimal staining for ERα, and PGR was almost completely absent ( Figure 3i ). CD68-positive cells consistent with the morphology of macrophages were abundant in the subepithelial stroma of the lesions. Repeated seeding significantly increased the number and abundance of red lesions ( Table 1 ; P<0.05). In treatment cycle 2 (2 nd seeding) 10/14 animals displayed small (<5 mm) red lesions. By cycle 3 (3 rd seeding) 13/14 animals displayed lesions and the number of lesions increased significantly (P<0.05) by cycle 4. As cycle 5 and 6 progressed, the number of small red lesions stabilized while the abundance of brown and white lesions increased. Upon inspection in cycle 4, one cycle after the 3 rd seeding, all 14 of the treated animals displayed lesions

The veterinary staff of the Oregon National Primate Research Center (ONPRC) Division of Animal Resources and Research Support (ARRS) provided animal care and husbandry. The ARRS Surgical Services Unit, supervised by ONPRC surgical veterinarians, performed all procedures in dedicated facilities using aseptic techniques and comprehensive physiological monitoring. The ARRS clinical veterinary staff monitored animals with endometriosis daily for discomfort and provided analgesia, as indicated throughout the study. Female rhesus macaques ( Macaca mulatta ; 8-16 years of age; n=18) were inspected laparoscopically to identify animals with pre-existing endometriosis. The screening revealed one animal with advanced endometriotic disease. This animal was euthanized, and reproductive tract and endometriotic tissues were collected at necropsy to provide samples of naturally occurring disease. Three animals were used as sham surgery controls, and the remaining 14 macaques underwent menstrual seeding experiments as described below. Seven of the animals listed above were allowed to cycle naturally. The animals were checked daily by vaginal swab for menstrual blood to identify day one of menstruation. The weekly blood collection for assay of serum estradiol (E2) and progesterone (P) was used to verify ovarian cyclicity. The remaining 10 animals received capsules releasing estradiol (E 2 ) and then E 2 plus progesterone (P 4 ) to override natural ovarian cycles and synchronize the onset of menstruation with surgical procedures. The capsules were prepared from Silastic tubing (0.34 cm i.d.; 0.64 cm o.d.; Dow Corning; Midland MI). The ends of the capsules were sealed with medical-grade Silastic adhesive (Factor II, Inc., Lakeside AZ). The E 2 capsules were 5 cm long and were filled with 0.5g E 2 (Sigma Cat#E8875). The P 4 capsules were 6 cm long and filled with 0.65g P 4 (Sigma Cat#P0120). The capsules were washed in 1% gelatin (Sigma Cat # G1890) for 1 hour and then disinfected with 2% chlorhexidine (Nolvasan; Zoetis Inc, Kalamazoo MI) and rinsed with sterile water. The implants were soaked again in 2% chlorhexidine for 10 minutes immediately before placement in the animals. To insert the implants the animals were sedated with ketamine (10-18 mg/kg) and placed in ventral recumbency. Following sedation, the implant sites were shaved and cleaned with ChloraPrep (Becton, Dickinson and Company, Franklin Lake NJ; chlorhexidine/alcohol), 0.5 ml bupivicaine (0.5%) combined with 0.1 ml lidocaine (1%) was injected intradermally at the incision site, and a subcutaneous pocket was created by blunt dissection through a linear 1 cm skin incision in the parascapular area. The capsule was inserted and the skin closed with simple interrupted or intradermal sutures. The treatment provided E 2 (~80-100 pg E 2 /mL) and P 4 (5-7 ng P 4 /mL). Blood samples were collected weekly to confirm hormone release. E 2 implants were replaced if serum levels dropped below 80 pg/ml. P 4 releasing capsules were removed and replaced every 14 days to create controlled 28-day menstrual cycles. Removal of the P implant 14 days later (cycle day 28) stimulated menstruation 24-36 hours after P withdrawal. On day 2 of menstruation, 14 of the animals (7 artificially cycled and 7 intact) underwent laparoscopic intra-abdominal endometrial transplantation (seeding). For this procedure, monkeys were sedated with ketamine (10-20 mg/kg) followed by inhalant anesthesia (1-2% isoflurane). The animals were insufflated with CO 2 and placed in Trendelenburg position. At the onset of laparoscopy, the pelvic cavity was inspected ( Figure 1a ) for pre-existing endometriosis, adhesions and/or other abnormalities. During laparoscopy the uterus was stabilized at its base with grasping forceps, and a long 15 gauge needle was inserted into the abdomen, through the abdominal wall and uterine fundus, and into the uterine cavity ( Figure 1b ). Approximately 0.1-0.3 ml of menstrual endometrial tissue and blood were aspirated from the endometrium via a luer lock syringe. The menstrual mixture was immediately transferred onto the peritoneum of the intra-abdominal space along the dorsal and ventral aspect of the cervico-uterine junction and along the left and right uterine-ovarian ligaments and fat pedicles ( Figure 1c ). The syringe was flushed with Hanks Buffer Salt Solution (HBSS; 37° C), which was also deposited to the abdominal cavity with residual endometrial material from the syringe. Post-operative analgesia was provided for 48-72 hrs. following the surgical procedure using hydromorphone HCl (0.05-0.4 mg/kg, IM), and buprenorphine (0.01-0.1 mg/kg, IM). The laparoscopic "seeding" procedure was repeated for three menstrual cycles. By the last seeding, endometriotic foci were observed on the uterine fundus. Three animals treated with artificial cycles served as surgical controls. These animals underwent laparoscopic menstrual debris collection, but the aspirate was removed from the endometrium without transfer to the abdominal cavity. The sample was mixed with TissueTek OCT (Sakura Finetek, Torrance, CA) and frozen for immunohistochemical analysis of estrogen receptor alpha (ERα) and progesterone receptor (PGR). After the last 'seeding' cycle, all of the animals were laparoscopically accessed at menses for an additional three menstrual cycles, during which the abdominal cavity was inspected as described above without menstrual collection. Video recording during laparoscopy was used to record the presence of lesions, during seeding and inspections. When the study was completed, all of the animals were euthanized and the abdomen inspected for residual lesions. Tissues from these animals were archived for future study. The presence of endometriotic lesions was observed during laparoscopy and recorded. The videos were reviewed and the presence of lesions assessed by two viewers to reduce observational bias. Reproductive tracts and all endometriotic lesions were collected from the animals at the end of the protocol. Endometriotic lesions were dissected free from surrounding tissue, microwave irradiated for 7 seconds, imbedded in TissueTek OCT (Sakura Finetek, Torrance, CA) and frozen in liquid propane. Menstrual tissue fragments were mixed with TissueTek OCT and frozen liquid propane. The OCT embedded blocks were cryosectioned at 7 microns and thaw mounted onto Superfrost Plus slides, fixed in 2% paraformaldehyde in PBS (pH 7.30) and immunohistochemistry was performed as previously described. 39 The sections were stained with specific antibodies against ERα (ThermoFisher, Grand Island NY Cat #MA5-13191; 1:400 dilution), PGR (ThermoFisher Cat # MS-192-P0; 1:200) and CD68 (Agilent, Santa Clara CA; M081401; 1:100). Immunostaining for these epitopes has been validated previously by our laboratory. 40 Staining controls included the following: 1, omitting the primary antibody, 2, omitting both the primary and secondary antibodies, and 3, including an irrelevant primary antibody (anti-Br(d)U). Photomicrographs were captured using a Zeiss AxioImager A.1 microscope (Carl Zeiss Inc., Oberkochen, Germany) with planapochromatic lenses fitted with a Leica DFC 480 camera (Leica, Wetzlar, Germany). The frequency of endometriosis in the control (sham) and menstrual-seeded animals at completion of the protocol was analyzed by Fisher's exact test. The abundance of red, brown and white lesions in artificially cycled and naturally cycling animals at each seeding and inspection procedure was analyzed by analysis of variance for repeated measures. 41

Studies on the pathogenesis of the endometriosis requires the use of experimental animals. Herein we validated a protocol for protocol for creating endometriosis in macaques that recapitulates the outcome of retrograde menstruation. This method is a valuable tool for researchers investigating the onset of endometriotic disease and for preclinical studies into novel therapeutics.

Studies on the pathogenesis of endometriosis requires the use of experimental animals. This work validated a surgical protocol for creating endometriosis in rhesus macaques that recapitulates in part the outcome of retrograde menstruation. Earlier studies to develop a macaque model for induced endometriosis reported variable success. 37 , 38 Creating the disease in macaques via transcervical collection of menstrual endometrium is not feasible because the complex anatomy of the macaque cervix. 42 Our uterine puncture technique provides a “surgically sterile” approach. Moreover, it is possible that some deep functionalis zone or basalis cells are also transferred with menstrual samples. Endometrial stem cells are preported for the deeper zones of the endometrium. 43 Transfer of endometrial stem cells residing in the deeper zones could improve the success of this technique. 44 Endometriosis studies in rodent models fall into two broad categories. One category involves the auto-transplantation of uterine tissues to ectopic sites or the syngeneic transplantation of similar tissues in C57BL/6 mice. A refinement of this approach is to create a menstruation-like event in the mice and to transfer the endometrial debris into hormone-primed C57BL/6 mice. To create the menstruation-like condition, the mice treated with progesterone and decidualization is stimulated with a foreign substance to create a pseudo-pregnant condition. Endometrial breakdown is then initiated by progesterone withdrawal 45 . It is important to note that spontaneous decidualization in human and nonhuman primate endometrium does not require trauma by embryo implantation. The other broad category involving rodents includes studies where human endometrium is heterotransplanted into immunocompromised rodents (e.g. athymic nude mice, severe combined immunodeficient mice; and, Rag2 targeted knockout mice). Clearly, the immunodeficient nature of the animals that make transplantation possible may not accurately reflect immune environment in women with endometriotic disease. Retrograde menstruation, which is the reflux of menstrual debris containing endometrial tissues through the fallopian tubes into the pelvic cavity, is widely accepted as the origin of endometriotic tissues. 46 While this theory, commonly referred to as the Sampson Hypothesis, has not been unequivocally proven, it is a logical explanation that can be recreated in animal models including the baboon. 22 , 47 In the baboon procedure, menstrual endometrial currettings are laparoscopically inoculated to the abdominal cavity of the baboon. The resulting ectopic endometrium displays red, "power burnt," black or dark brown, and white serosal lesions with endometrium-like glands and stroma similar to the natural disease in both baboons and women. Artificially induced endometriosis in the baboon has been investigated over the last 20 years and continues to have promise for assessing pathogenesis of the disease. Long-term interventional studies in the baboon are limited by cost, animal availability, housing constraints, and the lack of well-developed assisted reproductive technologies (ART). Macaques have many advantages as an animal model. First, macaques naturally develop painful endometriotic disease, and therefore provide a model for studies targeting pain relief. 48 Second, assisted reproductive technology (ART) including in vitro fertilization has been comprehensively developed in macaques and the use of ART provides valuable methods to assess endometriosis–induced infertility. Third, macaques are the preferred primate species of the pharmaceutical industry; thus, anti-endometriotic agents evaluated in the macaque will be well positioned for application in clinical trials. Currently, there are no reliable serological tests for endometriosis, so it is not feasible to accurately diagnose the onset or early-stage disease in women or NHPs. 24 In rhesus macaques, naturally occurring endometriosis is not identified until it is very extensive, often obliterating the entire abdominal cavity. Because there is no available screening for minimal or mild disease, it is very important to develop a model for induced endometriosis for controlled study of the pathogenesis and treatment of endometriosis. Our ability to induce manageable disease in the rhesus macaque indicates that this animal is a suitable model for studying endometriosis. The lesions identified under this protocol closely resemble endometriosis in women, making this a good translational model. Therefore, it is not feasible to identify NHPs with spontaneous early-stage disease, and most cases are identified at an advanced stage, immediately preceding, or at necropsy. This represents a significant roadblock in studying the etiology of the disease. To address this roadblock, we have now an established protocol for creating subclinical endometriosis in macaques. This technique recapitulates the outcome of retrograde menstruation, and in theory, produces the most relevant model for early stage disease. The resulting lesions has endometriosis-like histology, stain positive for ovarian steroid receptors similar to lesions in women. The endometriotic sites distribute in the lower pelvic floor and cul-de-sac similar to non-ovarian endometriosis in women. While retrograde menstruation appears to be the source of endometriotic tissues, factors that mediate the establishment of lesions remains enigmatic. Retrograde menstruation is reported to occur in ~90% of women whereas endometriotic disease develops in only about 10%. Prior research implicates that aberrant inflammation contributes to the onset and progression of endometriosis diseases 49 - 52 . In theory, retrograde menstruation can cause inflammatory responses in the pelvic cavity, and this is supported by our observation here that multiple seeding events are required to induce the disease in macaques. Even if retrograde menstruation occurs in more than 90% of menstruating women, and the acute inflammation should be resolved in most women by the following menstrual cycle. Thus factors that increase aberrant inflammation occurs in endometriosis patients, which creates a chronic inflammatory environment in the pelvic cavity that promotes endometriotic disease. The observation that nearly all of the animals in this study developed endometriosis supports the premise that multiple surgical seeding events exacerbates pelvic inflammation. The role of pelvic inflammation in the development of endometriotic disease is evidenced in our recent reports that factors associated with polycystic ovarian syndrome including an obesogenic western style diet and exogenous androgen can lead to increased incidence of endometriosis in macaques 53 . Recently, neuropeptide S receptor has been identified in women and macaques and may provide a novel target for treatment of endometriosis. 54 Our induced model in the macaque will provide a working system that can be explored for addressing factors that increase the likelihood of developing the disease. This model also helps validate imaging techniques when laparoscopic manipulations reveal the localization of induced lesions, therefore advancing studies on diagnostic methods. Diagnostic delays are very common for endometriosis due to a lack of techniques and non-specific symptoms. 55 Endometriotic lesions, especially microscopic ones, are commonly missed during laparoscopic surgeries, resulting in incomplete removal. 56

Endometriosis is an incurable, gynecological disorder affecting more than 10% of reproductive-age women and greater than 35% of women with pelvic pain or infertility. 1 , 2 It also affects menstruating Old-World nonhuman primates (NHPs), including apes, macaques, and baboons. 3 - 6 It is defined as the presence of endometrium-like tissue, often referred to as lesions, outside the uterus. 7 - 10 The heterotopic tissues include endometrium-like glands and stroma that undergo cyclic steroid hormone-dependent cell proliferation, angiogenesis, and menstruation-associated bleeding. 11 , 12 Endometriotic lesions typically involve the peritoneal surface of the reproductive tract organs, including the ovaries, fallopian tubes, uterus, and cervix, as well as the pouch of Douglas (e.g., pelvic cul-de-sac; rectovaginal pouch), omental peritoneum, and the intestine. Lesions can also occur at extra-peritoneal sites, including the lung, urinary tract, and brain. 13 - 16 Endometriosis in women exists as distinct entities: superficial peritoneal endometriosis, ovarian endometrioma, and deep infiltrating endometriosis. 17 Peritoneal endometriosis includes lesions on serosal surfaces of the abdominal organs (e.g., red, "power burnt," black or dark brown, and white lesions). 18 Deep infiltrating endometriosis is an extensive intrusion of lesions that extend >5 mm beneath the peritoneal surface. 19 In contrast, ovarian endometriomas are cysts filled with thick brown fluid (e.g., chocolate cysts). Since the three endometriosis presentations have striking histological and physiological differences, it has been postulated that the different forms represent divergent disease etiologies. 20 However, the histo-pathological differences may result from the intra-lesion microenvironment during disease progression, and recent genetic studies support a common origin for ovarian and peritoneal endometriosis. 21 Retrograde menstruation, the reflux of menstrual debris containing endometrial tissues through the fallopian tubes and into the pelvic cavity, is widely accepted as the origin of endometriotic tissues. 22 This hypothesis has not been proven unequivocally, but it is logical and can be recapitulated in animal models. 23 Animal models are indispensable for studies on the pathogenesis of endometriosis because it is currently not feasible to diagnose the onset of the disease in women. 24 A clinical diagnosis of endometriotic disease is often delayed by 4 to 11 years between the first appearance of symptoms and their final definitive diagnosis in women. Endometriosis models employing rodents are often the first choice for study because of their relatively low cost. However, rodent models have limitations, the most important being that rodent species do not naturally menstruate or develop spontaneous endometriotic disease. 25 Thus, findings derived from rodent models may not translate well to clinical application. 26 Old World NHPs that display menstrual cycles similar to women and develop spontaneous endometriosis constitute the most physiologically relevant animal models for preclinical studies. 27 - 29 The prototypic method for inducing the disease in NHPs is a protocol developed in the baboon where menstrual endometrium is collected via a transcervical biopsy and deposited under laparoscopic guidance into the peritoneum. 27 , 30 - 32 While the baboon model has facilitated studies on disease pathogenesis and endometriosis-induced dysregulation of the eutopic endometrium, research in baboons is limited by animal availability, cost, and housing the constraints for large NHPs. 30 , 33 - 36 Macaques are the preferred NHP animal model for many research institutions and pharmaceutical industries. However, developing a rhesus macaque model for endometriosis has been ongoing for over 60 years with variable success. 37 , 38 Failure in creating the disease in naive monkeys via methods described for baboons appears to be methodological because the complex anatomy of the macaque cervix prevents reliable trans-cervical catheterization. Moreover, maintaining a sterile transcervical collection of menstrual fragments is challenging in these species. Considering the factors referenced above, this work aimed to develop a surgical method to induce endometriosis in rhesus macaques through the repeated transfer of menstrual endometrium to the abdominal cavity. Since endometriosis is considered an estrogen-dependent disorder clinically, we also sought to characterize the hormonal responsiveness of the induced lesions.