{"paper_id":"c72389c3-e624-4214-b4c0-019d2ada1de5","body_text":"Endometriosis is a common chronic gynaecological disorder in women, with a\nprevalence of at least 10 % (Giudice and Kao, 2004; Brosens et al., 2016),\nwhich also occurs in several Old World primate species that have a menstrual\ncycle (Assaf and Miller, 2012; Barrier et al., 2007; Dick et al., 2003;\nWilkinson et al., 2008; Zondervan et al., 2004). It is defined by the\npresence of ectopic endometrial glandular and stromal tissue outside the\nuterine cavity, accompanied by nonspecific clinical symptoms, like\ndysmenorrhea, pelvic pain, or infertility, and has a highly variable\nmanifestation concerning the anatomical distribution as well as macroscopic\nand histological appearance (Abrao et al., 2003; Adamson, 2011; Mattison et\nal., 2007; Mounsey et al., 2006). Although extensive literature has been\npublished on possible pathogenetic mechanisms and promoting factors (Brosens\net al., 2016; Brosens and Brosens, 2000b; Defrere et al., 2008a; Gotte et\nal., 2011; Kobayashi, 2000; Sasson and Taylor, 2008; Schindler, 2007; Toki\nand Nakayama, 2000; Young et al., 2013) and many animal models in primate as\nwell as non-primate species have been more or less successfully\nintroduced (Defrere et al., 2008b; Dinulescu et al., 2005; Einspanier et\nal., 2006; Fazleabas et al., 2002; Gashaw et al., 2006; Grummer, 2006; Rier\net al., 1993), the etiology of endometriosis remains obscure. The\npathogenetic mechanism most widely accepted as a prerequisite for the\ndevelopment of endometriotic lesions is retrograde menstruation that is the\ndistribution and implantation of refluxed endometrial material within the\nperitoneal cavity (Sampson, 1927; for review: Brosens and Brosens, 2000b).\nThis theory is supported by the frequent involvement of pelvic organs and\nthe fact that this disease almost exclusively develops in women or female\nnonhuman primates that are menstruating. It has also been demonstrated that\nforced induction of retrograde menstruation in a baboon model contributed to\nthe development of the disease phenotype (Braundmeier and Fazleabas, 2009).\nHowever, the phenomenon of retrograde menstruation occurs in the majority of\nwomen, of which only a small percentage develop endometriosis (D'Hooghe and\nDebrock, 2002). Therefore, further pathogenesis theories, such as those of\ncoelomic metaplasia or embryonic rest, have been suggested. Coelomic\nmetaplasia means that mesothelial cells generally have the potential to\ndifferentiate into functional endometriotic tissue (Gruenwald, 1942),\ntriggered by an appropriate stimulus, e.g., certain humoral factors in\nmenstrual fluid or steroid hormones (Young et al., 2013; Giudice and Kao,\n2004). The embryonic rest theory proposes that, at puberty, there is\nan activation of cells of Muellerian duct origin at various sites within the\npelvic cavity (Batt et al., 1990). Both theories are supported by reports on\nprepubescent and adolescent girls (Brosens et al., 2016) and even rarely men\nundergoing hormone treatment diagnosed with endometriosis (Pinkert et al.,\n1979; Schrodt et al., 1980). In addition, several supplementary conditions\nare assumed to contribute to the development of the disease, including\ngenetic (Borghese et al., 2017; Thomas and Campbell, 2000; Zondervan et al.,\n2004), environmental (e.g., iron and dioxin; Lousse et al., 2009; Rier et al.,\n1993), or immune-mediated factors (Giudice and Kao, 2004). Increasing\nnumbers of publications also point to an important role of stem or\nprogenitor cells, respectively, in the pathogenesis of endometriosis (Forte\net al., 2014; Forte et al., 2009; Gotte et al., 2011; Matthai et al., 2006;\nSasson and Taylor, 2008).\nAnother striking feature of endometriosis is its frequently reported\nassociation with malignancies. It is proposed that certain aspects of this\ndisease are similar to those of malignant neoplasia, e.g., uncontrolled\nproliferation, infiltrative growth, and lymphogenic or vascular spread, and,\nthus, it might possess precancerous potential (Kobayashi, 2000; Mandai et\nal., 2009; Nezhat et al., 2008; Van Patten et al., 2010; Yoshikawa et al.,\n2000). But the ability to induce malignant transformation of other\ntissues, e.g., by persistent oxidative stress from endometriosis-dependent\nrecurrent hemorrhage, is also considered (Higashiura et al., 2012; Nishida et\nal., 2000; Tanase et al., 2013). In summary, there are many different facets\nof this disorder, with some of them not fully understood yet, painting a\nheterogeneous and very complex picture.\nSpontaneous endometriosis in nonhuman primates often remains unnoticed\nuntil the animal either suddenly dies from the fatal course of the disease or\ndeceases for other reasons (Dick et al., 2003; Mattison et al., 2007;\nZondervan et al., 2004). Also in women, an early, noninvasive diagnosis of\nendometriosis is challenging due to nonspecific symptoms and heterogeneity of\nthe disease pattern (Mehedintu et al., 2014). Hence, clinical staging and\nprognosis is often referred to localization and histological\ndifferentiation, and various classification systems have been introduced to\npredict clinical outcome in patients (Adamson, 2011). But, since the\npreferred diagnostic method for endometriosis is still the surgical\nendoscopic inspection of the abdominal cavity (mainly pelvic organs) with\nhistological evaluation of tissue biopsies, early, very small, or\ndisseminated lesions, which might have a divergent gross or histological\nappearance, are easily overlooked (Mounsey et al., 2006). Accordingly,\ncurrent studies have proven that there is a poor to no correlation between the\nextent of the diagnosed disease and its clinical symptoms or therapeutic\nsuccess (Mehedintu et al., 2014).\nIn order to receive a comprehensive overview of the heterogeneous\nmacroscopic and associated histological features in nonhuman primates and\nto gain further insights into its pathogenesis, a thorough and systematic\nmorphological characterization of spontaneous endometriosis in a large group\nof rhesus macaques from the breeding colony of the German Primate Center has\nbeen conducted, with special focus on varying histological grades of\ndifferentiation and the corresponding immunohistochemical profile of several\nsomatic, hormonal, and proliferation markers.\n\nAll rhesus macaques ( Macaca mulatta ) included in this study came from the breeding colony\nof the German Primate Center in Göttingen, Germany, and were kept\naccording to the regulations of the European Parliament and the Council\nDirective on the protection of animals used for scientific purposes\n(2010/63/EU), the National Institutes of Health Guide for the Care and Use\nof Laboratory Animals (2010), and the applicable German Animal Protection\nLaw (“Tierschutzgesetz/Tierschutzversuchstierverordnung”). None of the\ninvestigated animals had ever been subject to any experimental intervention.\nThey were housed within metal and concrete indoor–outdoor facilities in\nlarge matrilineal groups of 25 up to 100 animals with one breeding male per\ngroup and fed commercial monkey diets supplemented with fresh fruits and\nvegetables as well as curd mash, enriched with vitamins and minerals. Water\nwas available ad libitum. All breeding macaques of the facility are annually screened\nfor  Macacine herpesvirus  1 (B virus), tuberculosis, and echinococcosis. Out of 154 adult female\nrhesus macaques from the breeding colony that were necropsied at the\nPathology Unit of the German Primate Center over a period of 10 years (from\n2001 to 2011), 9 cases had recorded spontaneous endometriosis, mainly\nconfirmed by histopathology of post mortem tissue. All of them were\nretrospectively evaluated for pedigree, medical, surgical and reproductive\nhistory, and cause of death as well as macroscopic appearance and distribution\nof endometriotic lesions at necropsy, while eight of them were additionally\nreinvestigated by histopathological and immunohistochemical analysis.\nDuring necropsy, representative tissue specimens of all major organs as well\nas additional tissue with grossly visible changes had been collected from\neach animal and were fixed in 10 % neutral buffered formalin, embedded in\nparaffin, and sectioned at approximately 4  µ m before being routinely stained\nwith hematoxylin–eosin (HE) stain. Occasionally, special stains, including\nthe periodic acid–Schiff (PAS)-reaction and Masson's trichrome stain, were used to\nfurther characterize particular features of epithelial and stromal cells as\nwell as the intercellular matrix. Immunohistochemistry (IHC) was performed on\nall paraffin-embedded sections that contained endometriotic lesions and/or\nnormal endometrium, using the avidin–biotin peroxidase method with an\nautomated immunohistochemical staining system (Ventana Discovery XT).\nAntibodies were directed against cytokeratin (MNF 116, broad spectrum CK,\n 1 : 100  monoclonal mouse anti-human, Dako), vimentin (V9,  1 : 100 , monoclonal\nmouse anti-human, Dako), smooth muscle actin (SMA; 1A4,  1 : 400 , monoclonal\nmouse anti-human, Dako), desmin (D 33,  1 : 100 , monoclonal mouse anti-human,\nDako), Von Willebrand factor (vWF; Factor VII-related antigen,  1 : 25 , Dako),\nestrogen receptor (ER; alpha 6F11,  1 : 10 , monoclonal mouse, Lab Vision),\nprogesterone receptor (PR; PgR636,  1 : 50 , monoclonal mouse anti-human, Dako),\nnuclear protein Ki67 (MIB-1,  1 : 50 , monoclonal mouse anti-human, Dako), and\np53 protein (DO 7,  1 : 50 , monoclonal mouse anti-human, Dako). In all\nimmunohistochemical staining procedures, diaminobenzidine (DAB) was used as\nthe chromogen and slides were counterstained with hematoxylin.\nFor morphologic evaluation, a histological classification was applied, which\nis also used in women, comprising four different forms of endometriosis\nbased on the degree of differentiation (Abrao et al., 2003; Kamergorodsky et\nal., 2009; Porto et al., 2015) described as follows. The first is a well-differentiated, glandular form with\npresence of surface epithelium or epithelium with glandular to cystic\nformations.\nThe cells are indistinguishable from those of normal endometrium\nduring different phases of the menstrual cycle (proliferative/follicular,\nsecretory/luteal, menstrual, and regenerative; Van Esch et al., 2008). The\nsecond is a\npure stromal form of endometriosis without any surface or glandular\nepithelium.\nThe stroma also mainly resembles that of normal endometrium\nduring different phases of the menstrual cycle. The third is a glandular pattern\nof mixed differentiation, with the epithelium being composed of both\ncylindric to columnar endometrial-like cells, low cuboidal to flattened\ncells, and\nundifferentiated cells, and also sometimes cells with other histological\nMuellerian patterns (serous or mucinous cells).\nAnd finally, the fourth is a poorly\ndifferentiated form that is characterized by a solely undifferentiated\nglandular pattern in which the surface epithelium or glandular/cystic\nformations are exclusively lined by low cuboidal to flattened,\nmesothelial-like cells or appear as small epithelial nests or islands.\nCase details of nine female rhesus macaques of the DPZ (German Primate\nCenter)\nbreeding colony with spontaneous endometriosis.\na  Animals are directly related.  b  Twins; n.a.: not announced; ret. plac.:\nretained placenta.\nEight rhesus macaques (no. 1–8) were systematically scanned for the presence\nand distribution of the described histological types of endometriosis within\nall collected representative histological HE sections. Every tissue/organ\ninvolved together with the corresponding grade and growth pattern (expansile\nvs. infiltrative) was recorded. Immunohistochemical expression patterns of\nnormal endometrium as well as epithelial and stromal components of different\nhistological types of endometriosis were evaluated semi-quantitatively\naccording to Ferreira et al. (2012). Staining for cytokeratin, vimentin,\nSMA, desmin, and vWF, was, respectively, scored in five categories: negative\n( - ) with < 5 % of cells stained, sparse (( + ) )  with 5 to\n25 % of cells stained, mild ( + )  with 25 to 50 % of cells stained,\nmoderate or heterogeneous ( + + )  with 50 to 75 % of cells stained,\nand strong ( + + + )  with > 75 % of cells stained. For\nestrogen and progesterone receptors, an “Allred score” was\nassigned (Allred et al., 1998; Harvey et al., 1999), consisting of a\nproportion score, representing the estimated proportion of positive-staining\ncells (0, none; 1 <  1 / 100 ; 2,  2 / 100 – 1 / 10 ; 3,  1 / 10 – 1 / 3 ; 4,  1 / 3 – 2 / 3 ;\nand 5, >  2 / 3 ), and an intensity score, representing the average\nstaining intensity (0, none; 1, weak; 2, intermediate; and 3, strong).\nThe proportion and intensity scores were finally added to obtain a total\nhormonal receptor score ranging from 0 to 8. The proliferation index (PI)\nwas determined by calculating the percentage of Ki67-positive nuclei within\nat least 500 cells. This was automatically estimated from digital slides\n(slide scanner Aperio CS 2, Leica, Germany) within representative regions in\nhighly positive areas without inflammation and/or necrosis, using an adapted\nnuclear algorithm (Aperio ImageScope Version 12.3.2.5030, Leica Biosystems).\nThe p53 staining was considered positive if more than 10 % of cells\nexhibited positive nuclear staining in 10 high-power fields (HPF, 40x\nmagnification), independent of staining intensity. For every IHC antibody,\nsemi-quantitative scores or PIs, respectively, were recorded in an\nexcel spreadsheet (Microsoft Office, 2010), and mean values were calculated\nfrom the eight animals analyzed. For calculation of the five\nsemi-quantitative categories, the scores ( - , ( + ) ,  + ,  + + ,  + + + ) \nwere transferred into corresponding numerical data from 0 to 4.\n\nCase details of all female rhesus macaques with spontaneous endometriosis,\nincluding age, cause of death, main macroscopic findings, pedigree, breeding,\nand medical history, are summarized in Table 1. All affected animals were\nmiddle-aged to old, ranging from 10.5 to 25 years, and, except for one, none of\nthem had a history of abdominal surgery. They were either euthanized for\ndifferent reasons, like the presence of an inoperable intraabdominal mass or\nsevere weakening, or died spontaneously due to severe hemoperitoneum,\nsepticemia, or age-related weakness. More than half of the monkeys (five of nine) were\ndirectly related to one breeding male (1585), either as a daughter or\ngranddaughter, with two of them being twins (nos. 8 and 9). As far as is known,\nall had multiple offspring with two to eight infants, but the last births\nhad always been long ago (4–11 years before death). Four cases had a\nhistory of stillbirth or premature delivery.\nOrgan distribution and growth pattern of different\nhistological types of spontaneous endometriosis in eight rhesus macaques\nfrom the DPZ breeding colony.\nRepresentative macroscopic pictures of spontaneous endometriosis\nin rhesus macaques ( Macaca mulatta ):  (a)  ovarian cyst (star) attached to the uterus\n(middle); note the normal size of the right ovary and red serosal plaques on\nthe urinary bladder, indicative of endometriotic lesions (animal no. 3);\n (b)  large multicystic mass with bloody contents comprising the entire uterus and\nadnexa (animal no. 5);  (c)  characteristic yellow-white to red\nplaque formation on serosal surface of a large multicystic intraabdominal\nmass (animal no. 6).\nRepresentative macroscopic changes during necropsy are depicted in Fig. 1.\nThe most common finding, present in eight of nine animals, was fibrous peritoneal\nadhesion of varying extent but always comprising at least the uterus,\novaries with adnexa, urinary bladder, and terminal parts of the colon (Fig. 1a). In five cases, additional cystic to multicystic masses were found,\neither on the ovary (Fig. 1a) or incorporating the entire uterus with adnexa\n(Fig. 1b). The cysts contained serohemorrhagic to bloody fluid and were\noften lined by polypoid to nodular projections. The serosal surface of\naffected organs generally showed multifocal to coalescing whitish-yellow to\nred, firm to soft plaques, strands or nodules of varying size, ranging from\nfew millimeters to several centimeters (Fig. 1a, c). In all animals, those\nlesions were present on the uterus, ovaries and ovarian tubes, or associated\nmasses, respectively. Further distribution on the mesometrium, urinary\nbladder, and colon, with respective mesenterial tissue, was present in most\ncases. Two cases (nos. 1 and 6) also showed lesions on the peritoneal surface\nof the diaphragm (Fig. 2a), with additional involvement of the liver,\nspleen, and stomach in animal no. 1. Another case (no. 4) had accordant\nchanges on the small intestine. Moreover, three monkeys (nos. 5, 6, and 9)\nrevealed remarkable thickening of the uterine wall with multiple transmural\nhemorrhagic cysts of varying size consistent with adenomyosis. Case no. 4 had\na large, poorly circumscribed, soft, intrauterine mass emanating from the\nendometrium, with a variegated beige to red cut surface and multifocal small\nhemorrhages indicative of endometrial hyperplasia. Additional macroscopic\nfindings related to endometriosis and representing the cause of death in\nthree animals were severe hemoperitoneum (no. 1) or suppurative peritonitis\nwith multiple abscesses and septicemia (no. 2 and 8), accompanied by rectal\nperforation in case 8. Furthermore, hydronephrosis due to ureter compression\nwas detected in two animals (no. 3 and 5).\nGrowth pattern and histological grades of differentiation\nin endometriotic lesions of rhesus macaques ( Macaca mulatta ):  (a)  diaphragm muscle with\nwidespread expansile and infiltrative endometriotic lesions on the\nperitoneal surface of animal no. 1, scale bar 1 mm, HE stain;  (b)  well-differentiated endometriosis with pseudostratified columnar glandular\nepithelium and cell-rich stroma in animal no. 6, scale bar\n100  µ m, HE stain;  (c)  pure stromal endometriosis without glandular formations on\nserosal surface of the colon (star) in animal no. 5, scale bar\n200  µ m, HE stain;  (d)  mixed-type endometriosis with poor and well-differentiated\nglandular epithelium in animal no. 1, scale bar 200  µ m, HE stain;\n (e)  poorly differentiated endometriosis with low cuboidal to flattened glandular\nepithelium in animal no. 7, scale bar 100  µ m, HE stain.\nAs shown in Table 2, all different histological types of endometriosis could\nbe detected in almost every animal, besides animal no. 2 and 4, which did not\nreveal poorly differentiated lesions. Within each of the four types, the\nendometriotic epithelium and stromal tissue either grew expansively on\nserosal surfaces or infiltrated the underlying tissue (Fig. 2a), apparently\nindependent of the organ involved, though infiltration was especially\npronounced in poorly differentiated mesenterial lesions. In\nwell-differentiated endometriosis, the surface and glandular formations were\nlined by simple to pseudostratified columnar epithelial cells (Fig. 2b)\noften showing cilia or apical blebs with PAS-positive secretion,\ncorresponding to the epithelial morphology of normal endometrium and\ndepending on the menstrual cycle of the animal. In less differentiated\nareas, surface and glandular epithelium was simple cuboidal to flattened\n(Fig. 2d, e), sometimes resembling mesothelial or even endothelial cells. The\nstromal component appeared very heterogeneous (Fig. 3a–e). Although\nwell-differentiated forms were generally characterized by a typical, highly\ncellular, and matrix-poor endometrial stroma sometimes accompanied by\ndistinct spiral arterioles (Fig. 3a), many other forms with fibrotic,\nmyxoid, fibroangioblastic, or myogenic phenotypes could be identified in all\nfour histological types. In two animals (no. 3 and 6), decidualization of\nendometriotic stroma was present to varying degrees, characterized by\nintermingled large, polygonal to epithelioid, and sometimes binucleated cells\nwith abundant finely granular, eosinophilic, and PAS-positive cytoplasm\nconsistent with decidual cells (Fig. 3f). Cystic structures within\nendometriotic lesions were regularly encountered, either resulting from\ndilated glandular formations with serous to mucinous contents due to\nexcessive secretion or, as blood-filled cavities, often accompanied by\ndesquamation of glandular epithelium, stromal inflammatory cell infiltrates,\nsuch as characteristic granular leucocytes, and necrotic debris in terms of\nmenstrual activity (Fig. 3g). Also interstitial hemorrhage of different\ndegrees was frequently observed, often with advanced chronicity reflected by\nextracellular accumulations of hemosiderin pigment (Fig. 3g). Furthermore, two monkeys (no. 6 and 3) revealed focally extensive,\npoorly circumscribed\nareas of extensive spindle cell proliferation within the uterine wall or\nadjacent mesentery, respectively.\nWithin the myometrium of case no. 6, large,\nplump spindle cells were arranged in fascicles or interwoven bundles,\npredominantly separated by a fine fibrovascular stroma with moderate amounts\nof fibrillar eosinophilic and sometimes vacuolated cytoplasm and central\noval to cigar shaped, vesiculated nuclei and prominent nucleoli. Cells and\nnuclei were fairly uniform with absent mitosis suggesting benign neoplasia\nof smooth muscle tissue (leiomyoma). In animal no. 3, spindle cell\nproliferations located in the mesometrium were rather pleomorphic, often\nembedded in moderate amounts of light eosinophilic, fibrillary to vacuolated\n(collagenous) matrix material with multifocal hemosiderin deposits and\nrevealed an\nincreased mitotic activity (1-2/HPF) indicating mesenchymal neoplasia of low\ngrade malignancy (Fig. 3g).\nThe intrauterine mass found in animal no. 4 was\nhistologically characterized by an expansile, non-encapsulated proliferation\nof well-differentiated glandular and stromal endometrial tissue with\nmultifocal cystic glandular ectasia and mild to moderate interstitial\nhemorrhage of varying chronicity, consistent with nodular\nendometrial hyperplasia.\nHistological variation of endometriotic lesions in rhesus\nmacaques ( Macaca mulatta ):  (a)  well-differentiated cell-rich endometriotic stroma with\nformation of spiral arterioles within mesometrium of animal no. 3, scale bar\n200  µ m, HE stain;  (b)  collagenous (fibrotic) stroma with\nwell-differentiated endometriotic glands on the urinary bladder of animal no. 2, scale bar 100  µ m, HE stain;\n (c)  loosely arranged (myxoid) endometriotic stroma cells with mild interstitial\nhemorrhage and glandular formations of mixed differentiation on the uterine serosal surface of animal no. 1, scale bar 100  µ m, HE stain;\n (d)  endometriotic\nstroma with distinct sprouting of fibroangioblasts, capillary formation and\nmild interstitial hemorrhage on the colonic serosal surface of animal no. 1,\nscale bar 100  µ m, HE stain (inset showing IHC vWF);  (e)  smooth muscle cells\nwithin well-differentiated endometriotic stroma in the mesentery of animal\nno. 1, scale bar 50  µ m, HE stain;  (f)  numerous decidual cells\n(“deciduosis”) within endometriotic stroma on the serosal surface of the\nurinary bladder in animal no. 6, scale bar 100  µ m, PAS reaction;\n (g)  endometriotic lesion with epithelial desquamation, fibrino-hemorrhagic\ndebris, and mild stromal inflammatory cell infiltrates indicative of\nmenstrual activity in mesometrium of animal no. 3, scale bar 100  µ m, HE\nstain (inset showing characteristic granular leucocytes);  (h)  pleomorphic\nspindle cell proliferation within collagenous matrix and hemosiderin\ndeposits adjacent to endometriotic lesions within the mesometrium of animal\nno. 3, scale bar 100  µ m, HE stain (inset showing two mitotic figures,\nscale bar 20  µ m).\nImmunohistochemical staining pattern of somatic markers in\nnormal endometrium and different histological types of endometriotic lesions\nfrom eight rhesus macaques with spontaneous endometriosis.\n-  negative (< 5 % positive cells). ( + )  sparse (5–25 % positive\ncells).  +  mild (25–50 % positive cells).  + +  moderate or\nheterogeneous (50–75 % positive cells).  + + +  strong (> 75 % positive\ncells).\nImmunohistochemical staining pattern of estrogen (ER) and\nprogesterone receptors (PR) as well as nuclear proteins Ki67 and p53 in\nnormal endometrium and different histological types of endometriotic lesions\nfrom eight rhesus macaques with spontaneous endometriosis.\nHormonal (Allred) receptor score: 0–8, sum of quotient (0–5) and\nintensity (0–3); Ki67 proliferation index (PI): positive nuclei/at least 500\ncell counts in highly positive areas; P53:  +  (positive) if > 10 % of all nuclei stain positive in 10HPF (regardless of staining\nintensity).\nComparison of representative immunohistochemical staining\npatterns between well- and poorly differentiated endometriosis: IHC for\ncytokeratin exclusively stains epithelial cells in well-differentiated forms\n (a)  compared to additional staining of several stromal cells in poorly\ndifferentiated endometriosis  (b) ; IHC for vimentin reveals strong staining\nof stromal cells with only sparse positive epithelial cells within\nwell-differentiated endometriosis  (c)  compared to distinct epithelial\nstaining in poorly differentiated forms  (d) ; IHC for PR\nintensely stains stromal and epithelial cells with slightly pronounced\nepithelial signals in well-differentiated endometriosis  (e)  and only\nsparsely disseminated positive stromal cells in poorly differentiated\nendometriosis  (f) ; IHC for Ki67 shows numerous positive cells predominantly\nwithin the epithelium of well-differentiated endometriosis  (g)  compared to\nfew positive stromal cells and virtually no epithelial staining in poorly\ndifferentiated forms  (h) ; hematoxylin counterstain, scale bars 100  µ m.\nEpithelial and stromal immunohistochemical staining patterns of the four\ndifferent histological types of endometriosis and normal macaque endometrium\nare summarized in Tables 3 and 4. While in normal endometrial tissue, as well\nas well-differentiated endometriosis, glandular and surface epithelium\nexclusively stained positive for cytokeratin with negative stromal cells,\nthe other types showed a gradual increase of cytokeratin-positive stromal\ncells with decline in differentiation (Fig. a, b). The IHC staining profile\nfor vimentin exhibited contrary results, with virtually negative epithelial\ncells and intensely positive stromal cells in normal endometrium and\nwell-differentiated endometriosis and progressive epithelial staining in\nmixed and poorly differentiated endometriosis (Fig. 4c, d). A similar trend,\nbut to a lesser degree, could be observed with myogenic markers (SMA and\ndesmin), though desmin staining was completely absent in normal endometrium\nor well-differentiated endometriosis, respectively. No obvious differences\nin staining pattern or intensity could be found with the vascular marker\nvWF, both between different types of endometriosis and also compared to\nnormal endometrium. The hormonal receptor scores for ER and PR were\ncomparatively high in normal endometrium and well-differentiated\nendometriosis (Fig. 4e), with slight emphasis on epithelial cells and\nconspicuous variation between individual animals. In less differentiated\nendometriosis, decreasing expression of hormonal receptors, especially\nwithin epithelial cells, could be observed (Fig. 4f). Additionally, PIs, demonstrated\nby the percentage of Ki67-positive cells, were almost twice as high in the\nepithelium of normal endometrium (40 %) and in well-differentiated\nendometriosis (34 %) compared to that of mixed (23 %) or poor (17 %)\ndifferentiation (Fig. 4g, h). However, this tendency was not discernible in\nstromal cells, with overall less Ki67 staining ranging from 9 to 23 % and\nhighest PIs in pure stromal endometriosis. Nuclear staining of p53 was\nconstantly positive in epithelial cells of endometriosis with mixed and poor\ndifferentiation, while the corresponding stroma and pure stromal\nendometriosis showed variable results. Well-differentiated endometriosis as\nwell as normal endometrium were always negative for p53. The described\nspindle cell proliferation of animal no. 3 revealed positive nuclear staining\nfor p53, together with a comparatively high PI (38 %) and intense IHC\nstaining for CK, vimentin, SMA, and desmin, but only little to no expression\nof hormonal receptors. Within the same animal, also sparsely dispersed\nsingle cells positive for CK, ER and PR, respectively, could be detected\nwithin medullary sinuses of lymphonodular tissue.\n\nThe results of this study provide a comprehensive overview of the clinical,\nmacroscopic, and microscopic manifestations of spontaneous endometriosis in\nrhesus macaques ( Macaca mulatta ), emphasizing important aspects of diagnostic relevance\nand possible pathogenetic implications. From the breeding colony of the\nGerman Primate Center, nine rhesus macaques presented with endometriosis\nover a period of 10 years. Related to the overall 154 adult female breeding\nrhesus macaques that were necropsied at the Pathology Unit during that time\nspan, this corresponds to an incidence of about 6 %. This is rather low,\ncompared to the reported, highly variable incidences of spontaneous\nendometriosis in other captive rhesus macaque colonies ranging from 0 % to\n45 % depending on the age and background of the animals investigated (Coe\net al., 1998; Mattison et al., 2007). Generally, it is assumed that\nincidences of endometriosis in nonhuman primates used for research purposes\nare likely underestimated, since many mild or incidental cases that occur\nwithin experimental studies remain undetected or unreported (Mattison et\nal., 2007). In addition, dysmenorrhoea, defined as pronounced pain during\nmenstruation, is one of the first clinical symptoms of endometriosis and\noften difficult to recognize in monkeys. This impedes an early intra vitam diagnosis of\nthe disease, so that in most studies only rather severe or fatal cases are\ntaken into account (MacKenzie and Casey, 1975). Accordingly, all rhesus\nmacaques from this study were of older age and exclusively presented with\nadvanced forms of endometriosis at necropsy, even though this was often not\nthe primarily suspected diagnosis. This suggests that younger females from\nthe population with milder forms of the disease might have been undetected.\nIn this respect, breeding populations with a high incidence of endometriosis\nshould be monitored by careful records of menstruation history and routine\ndigital rectal examination in order to identify some early changes (Fanton\net al., 1986). A well-known risk factor increasing the incidence of\nspontaneous endometriosis in captive nonhuman primates is abdominal\nsurgery, including caesarean sections, ovarian follicle aspiration, and\nembryo transfers (Coe et al., 1998; Dick et al., 2003). Since all examined\nanimals, except one, did not have a history of surgical intervention, this\nhad not been a crucial factor for the development of the disease in the\npresent study. In contrast, a familial link for endometriosis with increased\nrisk for first-degree relatives, as reported for both nonhuman primates\n(Zondervan et al., 2004) and women (Bischoff and Simpson, 2000), could be\nconfirmed since more than half of the affected rhesus macaques (five of nine) were\ndirectly related to one breeding male. An underlying polygenetic inheritance\nis thereby assumed, which means that the disease is caused by the cumulative\neffect of several different genes acting in concert with first degree\nrelatives holding more of the susceptibility genes than the average\npopulation (Bischoff and Simpson, 2000). Although the incidence of\nendometriosis is generally higher in nulliparous females (Schindler, 2007),\nwhich is also reported for the rhesus macaque, e.g., with 78 % in a cohort\nstudy of 72 animals (Fanton et al., 1986), all animals with spontaneous\nendometriosis from the investigated breeding colony had multiple offspring\nin the past. However, the time span between last birth and death was long\n(4–11 years) in all of them, and four animals had a history of abortion or\nstillbirth, indicating a disease-related impairment of fertility. Evidence\nfor infertility associated with endometriosis is also described in several\nanimal studies as well as human patients (reviewed in Giudice and Kao, 2004). As shown, for example,\nby data from IVF (in vitro fertilization) treatment, this is mainly\ncaused by poor ovarian reserve in advanced disease, low oocyte quality,\nimpaired embryo survival, and poor implantation capacity, most probably due\nto adverse effects of peritoneal fluid containing high concentrations of\ncytokines, growth factors, and activated macrophages in patients with\nendometriosis (Barnhart et al., 2002; Olivennes, 2003; Taketani et al.,\n1992).\nThe course of the disease and clinical symptoms with endometriosis are\nhighly variable and closely associated with the nature and distribution of\nlesions (Fanton et al., 1986; Mounsey et al., 2006). According to the\nliterature, three different types of endometriosis are recognized in humans:\nperitoneal endometriosis, deep infiltrating endometriosis, and ovarian cysts\n(also named endometriomas; Adamson, 2011; Young et al., 2013), and it is\nassumed that at least the pathogenesis for ovarian endometriosis might be\ndifferent from that of peritoneal endometriosis (Brosens and Brosens, 2000a;\nD'Hooghe and Debrock, 2002). All three forms were also detected in the\nrhesus macaques of the present study; however, a clear separation between\nanimals could not be confirmed. The major common macroscopic feature, found\nin eight of nine cases, corresponds to peritoneal endometriosis, characterized by\nfibrous adhesions and variable peritoneal plaque formation, at least\ncomprising the pelvic organs. Ovarian cysts and large cystic masses were\nadditionally present in two or three animals, respectively. And, as shown by\nhistology, there was also more or less deep infiltration of endometriotic\ntissue present in all monkeys, regardless of the organ or tissue involved.\nThis overlap of the three different endometriosis forms suggests at least\nsome commonalities in pathogenesis. Together with the fact that the pelvic\nperitoneum is also the most common site for endometriotic lesions in over\n80 % of human patients (Mahmood and Templeton, 1991), a crucial role of\nthe peritoneum in the pathogenesis of endometriosis is obvious. In a recent\nreview (Young et al., 2013), different peritoneal factors were highlighted\nthat significantly influence the development and progression of\nendometriosis, emphasizing the implantation theory as the main pathogenetic\nmechanism. The authors suggest that the establishment of endometriosis in\nthe peritoneal cavity requires refluxed endometrial tissue or cells, which\nare only able to adhere, infiltrate, and proliferate in cooperation with\ncertain altered properties of the peritoneum. These alterations refer to\nmesothelial expression of adhesion molecules (e.g., cadherins and integrins) and\nchanges in mesothelial cell morphology facilitating ectopic cell attachment\nand invasion together with tissue remodeling via matrix metalloproteinases\nand epithelial–mesenchymal transition (EMT). Furthermore, enhanced secretion\nof pro-inflammatory cytokines by activated peritoneal macrophages and\nmesothelial cells as well as immune evasion by impaired clearance mechanisms\nare assumed to support the establishment of endometriotic lesions, together\nwith concurrent peritoneal cell proliferation and differentiation (for\nfurther detail, see Young et al., 2013).\nIn the present study, various histological features of endometriotic lesions\nwere identified, both in terms of epithelial differentiation and the stromal\nphenotype ranging from cell-rich endometrial to collagen-rich fibrotic or\nloosely arranged myxoid forms to spindle cells with pronounced myogenic\n(SMA + )  or angiogenic (vWF + )  differentiation. This could be explained by\ninduced EMT (as described above) and proliferation of either peritoneal\nmultipotent mesenchymal stem cells or by endometrial mesenchymal stem-like\ncells that have been demonstrated in menstrual fluid, extrauterine\nendometrial implants (Figueira et al., 2011; Gotte et al., 2011; Matsuzaki\nand Darcha, 2012), and normal endometrium, respectively (Gargett et al.,\n2016). However, both ways most likely depend on the microenvironment and\nchronicity of the lesion. Additionally, in normal endometrium, it is recognized that\nthe microenvironment determines the phenotype of endometrial tissue: active\ncyclic metaplasia of endometrial stromal cells (ESCs) into myofibroblasts and vice\nversa is observed in basal layers, whereas in superficial layers\nhormone-dependent secretion, vascularization, or bleeding is pronounced\n(Brosens and Brosens, 2000a). Accordingly, atypical or less differentiated\nforms of pure stromal endometriosis in rhesus macaques could easily be\nmistaken for spindle cell proliferation of other origin, such as\nretroperitoneal fibromatosis (RF), leiomyoma, inflammatory myofibroblastic\ntumors,\ninflammatory pseudotumors, or peripheral nerve sheath tumor\n(Bielefeldt-Ohmann et al., 2005). Thus, a careful evaluation of histological\nspecimens in several sections should be performed for identification of\npossible glandular or cystic structures. And, if applicable,\nimmunohistochemistry for epithelial markers (e.g., CK) should be carried out\nin cases of equivocal, more or less blood-filled cavities for discrimination\nbetween endothelial lining and flattened epithelium of an endometriotic\ncyst.\nTo the authors' knowledge, this is the first study that systematically\ncompares the presence, distribution, and immunohistochemical characteristics\nof variable histological grades of differentiation in spontaneous\nendometriosis, both in human and nonhuman primates. The fact that all\nhistological types were found more or less next to each other, or even\nmerged, within the same animal strongly suggests parallel events of\nsequential differentiation processes. Furthermore, less differentiated\nhistological grades, especially purely stromal forms but also the flattened\nepithelial components of endometriosis with mixed or poor differentiation,\nwere sometimes difficult to distinguish from serosal tissue. Together with\nthe demonstrated immunohistochemical co-expression of epithelial and\nmesenchymal markers (CK, vimentin, sometimes together with SMA and desmin),\nmost obvious in poorly differentiated endometriosis and resembling distinct\nmesothelial cell properties, an induced differentiation of peritoneal cells\ninto endometrial tissue, is conceivable, supporting the theory of coelomic\nmetaplasia. The consistently reported presence of pluripotent mesenchymal\nstem cells (MSCs) in peritoneal tissue (Carmona et al., 2011; Mutsaers et\nal., 2015) could therefore serve as an essential prerequisite. Moreover,\nperitoneal mesothelial cells provide such a high degree of plasticity that,\nif placed in the appropriate microenvironment, they have the potential to\ngenerate various other mesenchymal-derived cell types (Gotloib et al.,\n2007). Hence, it is assumed that the different histological grades represent\na subsequent graduation of differentiation in the course of time, with poorly\ndifferentiated types representing newly formed, immature lesions and\nwell-differentiated types being older, fully differentiated forms, rather\nthan being the outcome of dedifferentiation processes. The possible\nunderlying mechanism of induced mesothelial differentiation into endometrial\ntissue (epithelial and stromal) might be exerted by direct cell-to-cell\ncommunication, probably via exosomes containing epigenetic regulatory\nmolecules (e.g., miRNAs). Recent insights on the communication of tumor cells\nwith adjacent mesenchymal tissue support this hypothesis (Webber et al.,\n2015). Moreover, exosomes isolated from endometrial stromal cells\nwere capable of inducing angiogenic effects in human umbilical vein\ncells,\npointing out that exosomes derived from ESCs play paracrine roles in the\ndevelopment of endometriosis (Harp et al., 2016). Thus, exosomes might work\nas effective intercellular communication modulators in endometriosis,\nexerting direct epigenetic effects on surrounding tissue, so that the\nendometriotic cells themselves do not need to “invade” the environment and\nfurther proliferate but rather stimulate the adjacent cells (e.g., MSCs) to\nbecome one of their kind. This assumption would also be supported by the\npresent finding that proliferative activity in less differentiated grades of\nendometriosis was comparatively low.\nIn one of the two animals (no. 3) with conspicuous spindle cell proliferation\nin addition to endometriotic lesions, the exact cell origin is ambiguous.\nThe histological appearance of streams and interwoven bundles of plump\npleomorphic spindle cells within up to moderate amounts of collagenous\nmatrix most likely resembles low grade fibrosarcoma, although\nretroperitoneal fibromatosis or other intestinal stromal tumors can also\nhave a similar morphology (Bielefeldt-Ohmann et al., 2005). RF has been\ndescribed in macaques with simian acquired immunodeficiency syndrome (SAIDS)\nassociated with simian retrovirus type D (SRV-2) infection (Marx and\nLowenstine, 1987), and is presumably caused by a gammaherpesvirus\n(retroperitoneal fibromatosis herpesvirus, RFHV; Bielefeldt-Ohmann et al.,\n2005; Rose et al., 1997). Lesions are typically restricted to the ileocaecal\njunction, mesenteric root, and mesenteric lymph nodes, and tumor cells are\nusually immunohistochemically positive for vimentin, vWF, and SMA (Fikes and\nO'Sullivan, 1995). However, the viral status for SRV-2 and RFHV of this\nrhesus monkey was not known and the pleomorphic spindle cells were positive\nfor CK, vimentin, SMA, and desmin, suggesting an additional epithelial\nphenotype as also observed in the undifferentiated forms of concurrent\nendometriosis. Therefore, MSCs from endometrial or peritoneal tissue, being\ncapable of pluripotent differentiation characterized by immunohistochemical\nexpression of more than two histogenetically unrelated antigens, could have\nbeen the cell of origin in this case. A similar immunohistochemical\nexpression pattern is also described for certain forms of mesothelioma\n(Fassina et al., 2012), but calretinin, a diagnostic marker for\nmesothelioma, has not been tested in the present case, so the definite\ndiagnosis remains speculative.\nThe other spindle cell tumor (animal no. 6) that was found in the present\nstudy was consistent with leiomyoma and occurred together with multifocal\nadenomyosis, which is defined as the presence of ectopic endometrium within\nthe myometrium. Besides in women, adenomyosis has been reported in several\nnonhuman primates and rarely other non-primate species (Barrier et al.,\n2007; Baskin et al., 2002; Greaves and White, 2006; Wilkinson et al., 2008).\nIt is frequently accompanied by surrounding myometrial hyperplasia or\nconcurrent leiomyoma (Barrier et al., 2007; Wilkinson et al., 2008). Some\nauthors consider adenomyosis as a different form of endometriosis, and its\npathogenesis is likewise obscure (Brosens and Brosens, 2000a).\nOne rhesus macaque (no. 4) revealed an intrauterine mass histologically\nconfirmed as endometrial hyperplasia. Endometrial hyperplasia or polyps are\nreported in perimenopausal women or older nonhuman primate females, often\nassociated with endometriosis, and can be induced by unopposed estrogen\n(Baskin et al., 2002; Bennett et al., 2009). Another striking endometrial\nchange, caused by an increase in progesterone levels or induced by\nmechanical or chemical irritation (Marston et al., 1971), is decidualization\nof the endometrial stroma, an essential feature of the implantation stage of\npregnancy in rhesus macaques and other primates (Beck et al., 2014). It\ninvolves stromal proliferation and decidual differentiation, characterized\nby large polyhedral cells with abundant faintly eosinophilic cytoplasm,\ndistinct cell borders, and a centrally located, large, round to ovoid nucleus,\noften with a prominent nucleolus. Binucleation is regularly present, and the\ncytoplasm reveals a strong PAS-positive reaction due to high glycogen contents\n(Wadsworth et al., 1980). However, also ectopic decidualization or\n“deciduosis”, with groups of decidual cells located on serosal surfaces\noutside the uterus, occurs in the vast majority of pregnant women, usually\nas an incidental finding that regresses postpartum within 4 to 6 weeks, but sometimes\nleading to spontaneous fatal intraperitoneal hemorrhage (Kinra et al., 2006;\nO'Leary, 2006). Deciduosis can occur concurrently with endometriosis, as\nreported for macaques (Atkins et al., 2016; Beck et al., 2014) and women\n(O'Leary, 2006), and has also been detected in two animals of this study to\nvarying extent. While in animal no. 3 only few small foci of decidual cells\nwere found admixed with endometriotic tissue in the mesometrium, the other\ncase (no. 6) revealed widespread decidualization of expansively growing\nendometriotic stroma located on the uterus, mesometrium, and urinary\nbladder. The same animal, being the only one of the examined population,\nunderwent abdominal surgery 3 months prior to death and had a history of\npremature delivery with retained placenta. This suggests either\nsurgically derived irritation or hormonal imbalances as possible causes for\ndecidualization. Furthermore, decidual-like cells can also arise from\nmesothelial tissue independently of pregnancy or endometriosis, as known\nfrom rarely described deciduoid mesothelioma in human patients with\npredilection to the peritoneum and female preponderance in contrast to other\ntypes of mesothelioma (Shia et al., 2002).\nEndometriosis is a hormone-dependent disorder, which is often treated with\ncontraceptive steroids, progestagens, agonists of gonadotropin-releasing\nhormone (GnRH), and androgens in order to limit further growth and activity\nof ectopic endometriotic tissue (Giudice and Kao, 2004; Mounsey et al.,\n2006). Accordingly, the immunohistochemical staining of steroid receptors is\na typically reported feature of endometriotic tissue in different species\n(Assaf and Miller, 2012; Fazleabas et al., 2003; Slayden and Brenner, 2004)\nand was also evident in the present study. However, the overall hormonal\nreceptor scores in epithelial cells were considerably higher in normal\nendometrium and well-differentiated endometriotic lesions than in less\ndifferentiated forms. Furthermore, the level of hormonal receptor scores\napparently correlated with the percentage of epithelial cells positive for\nKi67, which has also been described by others (Toki and Nakayama, 2000),\nand, thus, a hormone-dependent proliferative activity of endometriotic\nglandular tissue, like in normal endometrium, is indicated. Nonetheless,\nprevious analysis of Ki67 staining patterns in human endometriotic tissue\ncompared to uterine endometrium revealed significantly reduced proliferative\nactivity in endometriotic epithelium (Scotti et al., 2000). Though data on\nconcurrent expression of steroid receptors and information about\nhistological differentiation of endometriotic specimens were not provided,\nrather less differentiated forms of endometriosis were examined in that\nstudy, as is recognizable from the published pictures.\nThe other cell-cycle-related antigen tested in the present study was p53,\nalso known as tumor protein TP53 or “guardian of the genome”, here showing\nconsistent positivity in epithelial cells of endometriotic lesions with\nmixed or poor differentiation. As an important regulator of the cell cycle,\ne.g., by inducing cell-cycle arrest, activating DNA repair mechanisms, or\ninitiating apoptosis, p53 becomes activated in response to numerous\nstressors, including but not limited to DNA damage, as well as oxidative\nstress or dysregulated oncogene expression (Surget et al., 2013).\nInactivation of p53 by mutation or other genetic alterations leads to loss\nof function and can cause uncontrolled cell proliferation (Burns et al.,\n1991). Correspondingly, p53 overexpression has been detected in a wide\nvariety of human malignancies, including ovarian endometrioid carcinoma\n(Harlozinska et al., 1996). Therefore, nuclear accumulations of mutated p53\nprotein are often supposed as a carcinogenic change. Accordant to our\nresults, an overexpression of p53 has also been recognized in epithelial\ncells of several human endometriotic lesions (Toki and Nakayama, 2000).\nHowever, in those lesions neither mutations nor microsatellite alterations\ncould be detected, so that the p53 accumulations are regarded as wild type,\nand, thus, a role other than oncogenic is considered. Additionally, the fact that the\nimmunohistochemical p53 staining in our study did not correspond to PIs of\nthe different histological grades indicates that the constantly detected\naccumulations of p53 in mixed or poorly differentiated epithelium are not\nrelated to uncontrolled cell proliferation.\nAltogether, the results of the present study point out the diagnostic\nchallenges of endometriosis in nonhuman primates related to clinical,\nmacroscopic, and histological findings and provide novel insights into\npossible pathogenetic mechanisms derived from morphological characteristics,\nemphasizing the hypothesis of coelomic metaplasia. However, further research\nis necessary in order to verify the underlying mechanism of induced\nperitoneal differentiation into endometriotic tissue, especially\nregarding cell-to-cell communication via exosomes, which might potentially\nreveal new therapeutic approaches.","source_license":"CC0","license_restricted":false}