American Journal of Primatology 69:901–916 (2007) RESEARCH ARTICLE Assessment of Menstruation in the Vervet (Cercopithecus aethiops) REBECCA L. CARROLL1, KUNIE MAH1, JOHN W. FANTON2, GWENDALYN N. MAGINNIS2, ROBERT M BRENNER1, AND OV D. SLAYDEN1 1 Division of Reproductive Sciences, Oregon National Primate Research Center, Oregon Health and Science University, Beaverton, Oregon 2 Division of Animal Resources, Oregon National Primate Research Center, Oregon Health and Science University, Beaverton, Oregon Vervet monkeys (Chlorocebus aethiops) are Old World nonhumans that display attenuated menstruation that requires detection by vaginal swab. The physiology underlying attenuated menstruation in this species has not been previously studied. To fill this gap, we evaluated endometrial cell proliferation, steroid receptor localization and expression of menstruationassociated matrix metalloproteinase (MMP) enzymes in vervets during natural and artificial menstrual cycles. The artificial cycles were induced by sequentially treating ovariectomized animals with estradiol (E2) and progesterone (P). Because menstrual flow is exceptionally light in this species, menses was detected by vaginal swab. We found that both natural and artificially cycled animals menstruated 3–5 days after the decline of P at the end of the cycle. As in other primates, P withdrawal at the end of artificial cycles triggered endometrial expression of MMPs, including MMP-1, 2, 3, 7, 10, 11, 13 and 26 transcripts. In both the natural and artificial menstrual cycle, menstrual sloughing was restricted to the upper one-fourth of the endometrium, and MMP-1 and 2 were strongly expressed by the stroma of the sloughing zone. MMP-7 was localized in the endometrial glands during late menses. As in macaques, epithelial cell proliferation was localized to the functionalis zone during the estrogen-dominated proliferative phase and to the basalis zone glands during the P-dominated secretory phase. Regulation of estrogen and progestin (or estradiol and progesterone) receptors was similar to that reported for macaques. Because strong similarities exist between the endometrium of vervets, macaques and women, we conclude that vervets can provide a useful animal model for studies on hormone regulation of menstruation. Am. J. Primatol. 69:901–916, 2007. c 2007 Wiley-Liss, Inc. Key words: menstruation; vervet; hormone action; matrix metalloproteinase enzymes; cell proliferation Contract grant sponsor: NIH grants; Contract grant numbers: HD07675, HD19182, HD18185 and RR00163. Correspondence to: Dr Ov D. Slayden, Division of Reproductive Sciences, Oregon National Primate Research Center, 505 NW 185th Avenue, Beaverton, OR, 97006, USA. E-mail: firstname.lastname@example.org Received 3 May 2006; revision accepted 28 September 2006 DOI 10.1002/ajp.20396 Published online 9 February 2007 in Wiley InterScience (www.interscience.wiley.com). r 2007 Wiley-Liss, Inc. 902 / Carroll et al. INTRODUCTION Old World nonhuman primates, especially macaques have long been the experimental animal model of choice for studies of hormone action in the endometrium. However, macaques are becoming increasingly difficult to obtain and it is important to assess other primate animal models for studies of the reproductive tract. The African green monkey or vervet (Chlorocebus aethiops) is a menstruating nonhuman primate with menstrual cycles of similar duration to those of women [Hess et al., 1979; Kudolo et al., 1986]. The potential use of vervets as an alternative to rhesus macaques for studies in reproduction have been explored previously [Eley et al., 1989; Hess et al., 1979; Kudolo et al., 1986]. Those studies revealed that menstruation is very light in vervets and that daily testing by vaginal swab is required to identify menstrual periods. The physiology underlying the attenuated menstruation in this species has not been previously studied. In several reviews, we have described morphological and physiological changes within the endometrium of rhesus macaques during both natural and artificial menstrual cycles [Brenner & Slayden, 1994; Brenner et al., 2002; Slayden & Brenner, 2004]. Like women, macaques naturally display 28-day menstrual cycles, and each cycle is marked by 3–4 days of frank menstrual bleeding. The macaque uterus is anatomically similar to the human uterus and the endometrium has been morphologically characterized as having four well-defined functional zones [Bartelmez, 1951, 1957]. During the follicular or proliferative phase of the menstrual cycle, ovarian estrogen drives cell proliferation in the upper zones (Zone I–III; functionalis zone). During the luteal or secretory phase of the cycle, progesterone (P) inhibits mitosis and induces a secretory state in the functionalis. In women, the basalis zone (Zone IV) adjacent to the myometrium is relatively unresponsive to cyclic hormonal changes, but in the macaque the basalis zone’s glandular epithelium proliferates under the influence of P during the secretory phase of the menstrual cycle [Brenner et al., 2003; Okulicz et al., 1993; Padykula et al., 1989]. The endometrial functionalis is the zone that bleeds and sloughs during menses in nonhuman primates and women. In macaques and women, the decline of P at the end of the cycle is followed by increased expression of matrix metalloproteinase (MMP) enzymes [Matrisian et al., 1994; Rudolph-Owen et al., 1998]. The MMPs are capable of degrading the extracellular matrix and play a role in the dissociation of tissues in the upper endometrial zones (Zones I–III) during menstruation [Matrisian et al., 1994; Osteen et al., 2003; Zhang & Salamonsen, 2002]. The cyclic changes in the level of ovarian estradiol (E2) and P also tightly regulate the expression of steroid receptors in the endometrium [Brenner & Slayden, 1994]. For instance, in macaques, E2 increases both estrogen receptor (ER) a and progesterone receptor (PR) in the endometrial glands and stroma during the proliferative phase of the cycle. During the luteal phase, P greatly reduces ER in the glandular epithelium and stroma of the functionalis zone. P also acts to reduce PR in the glands of the functionalis zone, but low levels of PR protein are retained in the stroma. Interestingly, P does not reduce levels of ER or PR in the glands of the endometrial basalis zone. A second ER (receptor beta; ERb) is also expressed in the macaque endometrium. However, levels of ERb do not appear to change significantly during the menstrual cycle [Brenner & Slayden, 2004]. Our goal in this study was to assess the cyclic effects of E2 and P on the regulation of menstruation in the vervet and to compare the menstrual Am. J. Primatol. DOI 10.1002/ajp Endometrial Cycles in the Vervet / 903 endometrium in this species with the menstrual endometrium of the well-studied endometrium of rhesus macaques. For these studies we first evaluated endometria from a small number of naturally cycling adult vervets. To further explore the effect of P on endometrial cell proliferation, and P withdrawal on menstruation, we continued the studies on ovariectomized artificially cycled vervets, where the timing of P treatment was tightly regulated. Hormone data from the naturally cycling animals were used to design appropriate Silastic capsules for delivery of physiologically relevant levels of exogenous E2 and P, and artificial 28-day cycles were stimulated in the animals by treating them with implants of E2 and P. We report that the vervet endometrium differentiates into hormone-responsive functional zones that are similar anatomically to endometrium of macaques and women. Moreover, P withdrawal at the end of the cycle triggers expression of endometrial MMPs and tissue fragmentation during menses that is essentially similar to other primates. MATERIALS AND METHODS Animal Treatment Twenty adult female vervet monkeys (mean7SE body weight: 4.670.3 kg) were purchased from the Caribbean Primate Laboratory (St. Kitts, Eastern Caribbean). Animal care was provided by the veterinary staff at the Oregon National Primate Research Center (ONPRC) and all procedures were preapproved by the ONPRC Institutional Animal Care and Use Committee. Except where indicated all reagents used in this study were purchased from the SigmaAldrich Co (St. Louis, Missouri). Naturally cycling animals Blood samples were collected once in every 3 days from six vervets during natural menstrual cycles. Both blood sampling and assessment of vaginal swabs (see below) were collected on non-anesthetized animals that were trained to enter a box with a squeeze mechanism for light restraint. Serum samples from two natural cycles/animal were assayed for E2 and P to identify the follicular and luteal phases of the natural menstrual cycle. Because frank (external) menses was difficult to detect [Hess et al., 1979], the animals were inspected daily by vaginal swab to detect menstrual bleeding. For this assessment, the animals were held in light restraint and a cotton swab was inserted fully (2 cm; no speculum was needed) into the vagina. Presence of blood on the swab was recorded as mense positive. Animals underwent hysterectomy during the proliferative phase (cycle day 10; n 5 1), secretory phase (cycle day 21; n 5 1) and the menstrual phase (cycle days 3–5; n 5 4) of the second natural cycle. Artificially cycling animals Fourteen adult animals were ovariectomized during the luteal phase of the cycle and treated sequentially with Silastic capsules (0.34 cm i.d.; 0.46 cm o.d.; Dow Corning, Midland, Michigan) filled with crystalline E2 or P to induce artificial cycles [Slayden et al., 1993]. For these cycles, a 2-cm E2 capsule was implanted subcutaneously (s.c.) for 14 days to stimulate an artificial proliferative phase. After 14 days of E2 priming a 3-cm P capsule was implanted for 14 days (leaving the E2 implant in place) to simulate the secretory phase of the cycle. Withdrawal of the P implant, leaving the E2 implant in place, completed the cycle. All the animals were treated with one 28-day cycle and then uteri were collected Am. J. Primatol. DOI 10.1002/ajp 904 / Carroll et al. by hysterectomy on days 0 (n 5 2), 1, 2, 3 (n 5 1 each), 4 (n 5 2), 5, 7 (n 5 1 each), 14 (n 5 2) and 21 (n 5 3) of the second cycle (Fig. 1). Tissue Handling The uterus from each animal was separated from the oviducts and cervix and cut into equal quarters along the longitudinal axes. Endometrium from the miduterine corpus from two quarters was further cut into 2 cm thick, full-thickness cross-sections, lumen to myometrium. These sections were either fixed for morphological studies and/or frozen for immunocytochemistry (ICC) (details are given below). Samples for histological analysis were fixed in 2% glutaraldehyde and 3% paraformaldehyde, embedded in glycol methacrylate sectioned (2 um) and stained with Gill’s hematoxylin [Bennett et al., 1976]. Endometrium from the mid-uterine corpus of the other half of each uterus was dissected from the myometrium under a dissecting scope with fine iris scissors, weighed and frozen for RNA isolation and analysis with focused cDNA arrays. Immunocytochemistry Detailed methods for ICC of nonhuman primate tissues have been previously published [Brenner et al., 2003; Slayden & Brenner, 2004]. Samples for ICC were microwave-stabilized, embedded in Tissue Tek OCT (Sakura Finetek USA, Inc., Torrance, CA) and frozen in liquid propane as previously described [Slayden et al., 1995]. Briefly, samples of fresh tissue were microwave irradiated for 7 s in an Amana Radarrange Touchmatic microwave oven (Amana, Iowa) mounted in Tissue Tek OCT and dropped into liquid propane, chilled in liquid nitrogen. Cryosections (5 um) were cut with a Leica CM30505 cryostat (Leica; Wetzlar, Germany) and thaw-mounted on Superfrost Plus (Fisher Scientific, Pittsburgh, PA) slides and held at 80 until all sections were cut. The slides were then placed on wet ice at 51C, and microwave irradiated again for 2 s. Localization for ER, and PR was accomplished with anti-human ER1D5 (NeoMarkers, Fremont, CA) and anti-human PR JZB 39 (courtesy of Geoffrey Greene; CA). Proliferating cells were identified with an antibody for Ki 67 (Dako Corp., Carpinteria, CA). MMPs 1, 2, and 7 were localized using mouse monoclonal antibodies (MMP1; R&D Systems, Minneapolis, MN; MMP 2 and 7; EMD Biosciences, San Diego, CA). The working concentration for each antibody (e.g., ER1D5, 1/50; JZB 39, 1 mg/ml; Ki 67, 1/200; MMP-1, 4 mg/ml; MMP-2, 4 mg/ml; and MMP-7, 1 mg/ml) was determined by antibody-specific serial dilution. Positive staining was detected with an avidin–biotin peroxidase kit (Vector Laboratories, Fig. 1. Timeline depicting artificial menstrual cycles. During the pretreatment cycle the animals were treated with E2 alone and then E21P to prime the endometrium for menstruation. Uteri were collected as shown during the treatment cycle. Am. J. Primatol. DOI 10.1002/ajp Endometrial Cycles in the Vervet / 905 Burlingame, CA), after which the slides were treated with 0.05% osmium tetroxide for 1 min, lightly counterstained with Mayers hematoxylin, and mounted with Permount (Fisher Scientific). Validation of ICC involved comparison of staining with each antibody on endometrium from the macaque and human samples. During the validation of the ICC for this species, we found that several antibody preparations directed against human epitopes, and that work well in macaques and human samples, failed in our trials to produce reliable positive staining in the vervet. Those antibodies included PR-8 (Neomarkers) and MMP-3 (Neomarkers). Photomicrography Low-power micrographs were photographed with an Olympus OM-system 38 mm macro lens (Olympus Optical, Tokyo, Japan) on Ektachrome 64-T film (Eastman Kodak, Rochester, NY) and digitized with a Polaroid SprintScan 35 scanner. High-power micrographs were captured through Zeiss planapochromatic lenses with the Optronics DEI-750TD CCD camera (Optronics Engineering, Goleta, CA). Digital images were adjusted for sharpness and contrast with Adobe Photoshop (Adobe Systems, Seattle, WA) and photomicrographs were printed with an Epson Stylus Photo 1200 printer (Epson, Tokyo, Japan). Total RNA Isolation and Focused Gene Arrays Samples of endometrium were thawed in ten volumes of TRIzol reagent (InVitrogen, Carlsbad, CA), immediately homogenized with a Polytron tissue homogenizer (Brinkman, Instruments, Westbury, NY) and processed following the standard TRIzol protocol. The TRIzol extracted samples were treated with RNase-free DNase (Qiagen, Valencia, CA) on RNeasy Mini columns (Qiagen). Concentration of the eluted RNA was determined by UV absorbance on a 640B spectrophotometer (Beckman Instruments Inc., Fullerton, CA) and RNA integrity was checked by fractionation on denaturing agarose gels and visualized with ethidium bromide. Total RNA was analyzed on low-density cDNA arrays (GEArray Q Series Human Extracellular matrix and adhesion molecules kit; SuperArray, Bethesda, MD) following the manufacturers instructions. The array membranes were prehybridized in 2 ml hybridization buffer for 2 h at 601C. During pre-hybridization, 2 mg total RNA was reverse-transcribed and amplified using the GEArray AmpoLabeling-LPR kit in the presence of [32P] dCTP (Perkin Elmer, Shelton, CT). Similar incorporation of [32P]dCTP among the samples was verified by electrophoresing 1 ml of labeled preparation on a 1% agarose gel, exposing the gel to X-ray film. The remaining labeled sample was denatured at 941C and hybridized to the arrays in a roller hybridization incubator (Robbins Scientific, Sunnyvale, CA) overnight at 601C. After high stringency washing the arrays were visualized by phosphor imaging with a Bio-Rad Molecular Imager FX (Bio-Rad Laboratories, Hercules, CA) and quantified with ScanAylze (Stanford University, Stanford, CA). For each array the mean intensity for the negative controls (PUC 18 and no cDNA) was calculated, and this signal level plus two SD of the mean intensity was considered background level on the blot. The background was subtracted from signal for each target cDNA and the relative signal for each gene on the membrane was calculated as the ratio of intensity of target gene to the intensity of RPL-13a, a ribosomal protein transcript. Am. J. Primatol. DOI 10.1002/ajp 906 / Carroll et al. Sample Size Limitations As indicated above, we were limited in the number of animals available for these studies, and in some cases only one animal was available at each time point. Traditional statistical analysis has therefore not been undertaken. On the basis of our studies in macaques, three clearly defined phases (menstrual, proliferative and secretory) of the cycle can be identified. We have combined samples when possible that fit into these phases to provide a database for our descriptive results below. RESULTS Natural Menstrual Cycles Figure 2 shows the E2 and P concentrations in the naturally cycling vervets. The mean (7SE) concentration of E2 was during 82736 pg/in the follicular phase, and 34712 pg/ml in the luteal phase. Serum P levels were o1 ng/ml during the follicular phase and 5.471.3 during the mid-luteal phase. These values for E2 and P during these cycles were similar to other menstruating nonhuman primates including macaques [Brenner & Slayden, 1994]. The average vervet follicular phase was 16.574.6 days and the average luteal phase was 1573 days. Frank menstrual bleeding (e.g., menstrual blood on the vulva or cage floor) was never noted for any of the animals in this study. Bleeding was detectable by vaginal swab between 2 and 4 days after serum P levels dropped below 1 ng/ml at the end of the cycle. Menstrual periods averaged 2.470.3 days. On the basis of these data, the animals displayed a mean (7SE) menstrual cycle length of 28.873.4 days. Artificial Menstrual Cycles Because of bleeding variably between days 2 and 4, after serum P levels dropped below 1 ng/ml at the end of the cycle, we treated ovariectomized animals with E2 and P implants better control the effects of P in this species. Silastic capsules were designed to provide serum hormone concentrations within the normal physiological range for cycling vervets. Treatment with implants of E2 and P resulted in 81722 pg/ml E2 and 570.4 ng P/ml in serum, which were within the normal range for naturally cycling vervets (see Fig. 2). Similar to the natural cycles above, no frank menstrual bleeding was noted in the artificially cycled animals. Mense was detectable by vaginal swab on days 3 through 6 of the cycle for all the animals. For instance, 0/10 animals were swab-positive on day 2, 6/10 were swab-positive on day 3, 7/10 were swab positive on day 4, 7/10 were Fig. 2. Hormone profiles from an adult naturally cycling vervet during the menstrual cycle. Values represent the mean1SE hormone level from six animals. The values have been aligned relative to the peak E2 value. Am. J. Primatol. DOI 10.1002/ajp Endometrial Cycles in the Vervet / 907 swab positive on day 5 and 3/10 were swab positive on day 6. Because all of the animals began menstruating on either day 2 or early on day 3, the use of hormone-controlled artificial cycles obviates the need to conduct daily vaginal swabs to time the onset of menstruation in this species. Cyclic Changes in Endometrial Histology The vervet uterus was anatomically similar to the uterus of macaques [Slayden & Brenner, 2004]. The mean (7SE) uterine weight in the vervet 3.0870.26 g. Figure 3 shows representative photographs of vervet uteri from animals during the natural menstrual cycle that have been cut along the longitudinal axis, after the cervix was removed. The cervix (not shown) was elongate, approximately three times the length of the uterus, and frequently bisected during hysterectomy. The upper cervical canal collected during hysterectomy appeared less tortuous than that of macaques. Figure 4 shows photomicrographs of hematoxylin and eosin-stained sections of endometrium from vervet monkeys during the natural cycle. Histological analysis of the vervet endometrium revealed no obvious cycle-specific differences between the naturally cycling and artificially cycled animals. This observation is similar for macaques, and women where the only ovarian factors required for menstrual cyclicity are the ovarian steroids E2 and P [Brenner & Slayden, 1994; Hodgen, 1983]. Therefore the analysis described below applies to both naturally and artificially cycled vervets. Secretory phase endometria During both the luteal phase of the natural cycle and on days 7 and 28 of the artificial cycle, the glands of the functionalis zone were hypertrophied and deeply sacculated and the epithelial cells of the secretory phase glands in the functionalis zone were notably pseudostratified (Fig. 4a and e). Pseudostratification and glandular sacculation became more notably pronounced during the late secretory phase (not shown). This is strikingly different from the glands in macaques and women as they always maintain a simple columnar epithelium [Brenner & Slayden, 1994]. Mitotic cells were essentially absent from the functionalis zone during the secretory phase. To analyze better proliferating cells in the endometrium, we stained sections from the animals during the artificial cycle for Ki-67 antigen, a marker of Fig. 3. Macrophotographs of vervet uteri collected at different phases of the natural menstrual cycle. In each case the cervix and oviducts have been removed and the uterus is cut in half along the longitudinal axis. In each photo a double-headed arrow shows the endometrial thickness, and a line has been drawn to indicate the approximate endometrial–myometrial border. The endometrium was thickest during the secretory phase of the cycle (a). The decline in P levels during the menstrual phase resulted in menstrual breakdown, which is clearly evident in the inner third of the menstrual endometrium (b). No bleeding was evident in the proliferative phase endometrium (c). Am. J. Primatol. DOI 10.1002/ajp 908 / Carroll et al. Fig. 4. Photomicrographs showing endometrial histology from glycol methacrylate sections of vervet endometrium from natural menstrual cycles. Similar histology was observed during the artificial cycle. In the full thickness micrographs (a–d), a black line has been drawn to delineate the endometrial–myometrial border. (M 5 mitotic cells; Ap 5 apoptotic cells). Secretory phase. The secretory endometrium was thick and the glands sacculated (a) and there was prominent peudostratification (PS) of the endometrial glands of the functionalis zone. Peudostratification was not pronounced in the basalis zone. No mitotic cells were evident in the functionalis zone, but abundant mitotic cells were observed in the basalis zone during the secretory phase. Menstrual phase. Menstrual sloughing was evident on days 3–5 of the artificial cycle. During early menstruation (b, f, j) there was clear evidence of tissue break down in the upper functionalis zone and apoptotic cells were clearly evident in the basalis zone. During late menstruation, menstrual breakdown was decreased and the healing endometrial stroma was densely packed, apoptotic cells were still abundant in the basalis zone. Proliferative phase. During the proliferative phase the functionalis glands were less sacculated and nonstratified. Abundant mitotic cells and some apoptotic cells were evident in the functionalis zone. In the basalis zone, there was still evidence of apoptosis, but little evidence of mitosis. Inset shows an enlargement of a spiral artery. proliferating cells [Gerdes et al., 1984; Okulicz et al., 1993; Slayden et al., 1993]. Figure 5 shows representative sections stained with Ki-67. During both the mid (d21) and late (d0) secretory phase, essentially no Ki-67-positive cells were detected in the glandular epithelium of the functionalis zone, strongly supporting the lack of mitotic activity in this zone during the secretory phase. In contrast, 3276% of the epithelial cells of the basalis zone were Ki-67 positive during the secretory phase (days 21 and 28 combined). The vervet endometrial basalis was not strikingly pseudostratified, and lack of epithelial stratification provided a marker of lower margin of the functionalis zone for counting Ki-67-positive cells. Unlike the functionalis zone, the basalis zone also contained abundant mitotic cells. Secretory phase cell proliferation was confirmed on the animal from the natural luteal phase (above), which had 31% of the cells in the basalis zone epithelial cells positive for Ki-67 (not shown). In this Am. J. Primatol. DOI 10.1002/ajp Endometrial Cycles in the Vervet / 909 Fig. 5. Photomicrographs showing Ki-67 imunostaining, during the artificial menstrual cycle in the vervet. Functionalis zone Ki-67 positive cells were minimal at the end of the secretory phase (a) and early menstrual phase (b, c) of the cycle. Functionalis zone cell proliferation increased during late mense (d) and was maintained at a high level throughout the proliferative phase on days 5, 7 and 14 (e, l and m). P suppressed functionalis zone Ki-67 by mid-secretory phase (n, d 21). In contrast, in the basalis zone, Ki-67 positive cells were abundant during both the mid- (r) and late secretory phase (a). respect, the vervet is very similar to macaques in which P stimulates basalis zone cell proliferation. Also like macaques, the vervets had well-developed endometrial spiral arteries. In macaques, these arteries grow during the secretory phase. In the vervet, Ki-67-positive cells were observed in the endometrial stroma surrounding these arteries (not shown) indicating that these vessels grow during the secretory phase as in macaques and other primates. Menstrual phase endometria During the natural menstrual cycle, menstrual breakdown was clearly evident during fresh tissue dissection of the uterine cavity during the menstrual phase of the natural (Fig. 3b). In general, menstruation was preceded by endometrial shrinkage and then stromal tissue fragmentation in the upper functionalis zones of the natural cycle (Fig. 3a–c). This was also strikingly evident by day 3 of the artificial cycle (Fig. 5c). In the Vervet menstrual flow was very light, detectable only by swab, and menstrual breakdown was more limited to the very uppermost endometrial zones than observed in macaques [Rudolph-Owen et al., 1998]. Menstruation was slightly more prolonged in vervets during the natural cycle and some breakdown was still evident during late menses (Fig. 4c and g) 5 days after P levels declined below 1 ng/ml. Am. J. Primatol. DOI 10.1002/ajp 910 / Carroll et al. Premenstruation and menstruation was associated with a cessation of cell proliferation in both the functionalis and basalis zones as evident by the absence of mitotic and Ki-67 positive cells on day 3 after P withdrawal (Fig. 5). By late menses on day 4 after P withdrawal, Ki-67-positive cells were evident in the healing glands in the functionalis zone. At this time there was a striking increase in the number of apoptotic cells noted in the basalis zone (Fig. 4c). Proliferative phase By day 8 of the natural follicular phase, the vervet endometrium was fully healed (Fig. 3c) and the glands displayed abundant mitotic cells (Fig. 4d). This morphological state was also evident after 7 and 14 days of E2 in the proliferative phase of the artificial cycle. Like other nonhuman primates and women, the proliferative-phase endometrium exhibited tubular non-sacculated glands with abundant mitotic cells in the functionalis zone. Analysis by Ki-67 ICC revealed that on days 7 and 14 of the proliferative phase the abundance of Ki-67-positive cells in both the glands and stroma was greatly increased compared with the secretory and menstrual phases of the cycle (Fig. 5). Counts of proliferative phase cells (days 7 and 14 combined) revealed that 4878% of the epithelial cells were Ki-67 positive. In contrast, E2 did not stimulate cell proliferation (assessed by mitotic cells or Ki-67 staining) in the basalis zone. This is similar to the regulation observed in macaques, where E2 fails to stimulate basalis zone growth [Padykula et al., 1989]. Regulation of ER and PR Figure 6 shows ICC staining for ER and PR in the vervet endometrium. In general, this pattern of ER and PR staining was very similar to that seen in other nonhuman primates and women. Photomicrographs of ER and PR staining during the artificial cycle are shown in Figure 6. Staining from vervets in the natural cycle was identical to staining in the artificial cycle. ER ER was most strikingly regulated in the functionalis zone during the cycle. For instance, ER staining was very strong in functionalis zone during the proliferative phase of the cycle (Fig. 6a and d). By day 21 of the cycle (day 7 of the secretory phase), P suppressed ER in the stroma of the functionalis zone (Fig. 6b). Progesterone-induced downregulation of ER in the functionalis stroma and glands was more pronounced at the end of the artificial cycle (Fig. 6c). Withdrawal of P during the menstrual period resulted in a gradual increase in ER staining in functionalis zone stroma and ER expression was clearly upregulated in the glands of the functionalis zone by late mense (not shown). In contrast, in the basalis zone, the glands maintained ER expression throughout the cycle. PR staining Like ER, PR staining was strongly regulated in the functionalis zone during the cycle (Fig. 6g–i), the abundance of strong PR-positive cells was greatest in both the stroma and glands of the upper zones (Fig. 6g and j) during the late proliferative phase. By day 21 of the cycle, P treatment greatly suppressed PR in the stroma and glandular epithelium (Fig. 6h). This downregulation of PR was even more evident at the end of the secretory phase (Fig. 6i) After P withdrawal at the end of the cycle, there was a striking increase in stromal PR in the functionalis zone stroma (not shown) that was essentially the same as observed at mid-cycle. Am. J. Primatol. DOI 10.1002/ajp Endometrial Cycles in the Vervet / 911 Fig. 6. Photomicrographs showing estrogen receptor (ER), progesterone receptor (PR) immunostaining in the vervet. In the functionalis zone both ER (a) and PR (g) were strong in the glandular epithelium and stroma. During the early secretory phase (d 21) stromal ER (b) and PR (h) were strikingly suppressed. P-induced suppression of ER and PR in both the glands and stroma was evident by the end of the cycle (c and i). In the basalis zone, ER and PR staining were maintained in the glandular epithelium. Stromal ER was suppressed by P in the basalis zone. PR staining in the basalis zone was variable. h—glandular looks very similar to i, more so than to g. As seen with ER, PR was maintained in the glands of the basalis zone. Stromal PR staining in the basalis zone was more variable and appeared to be suppressed by P during the secretory phase (Fig. 6j–l). MMPs During Menstruation In rhesus macaques and women, menstruation is associated with an abrupt increase in the endometrial levels of several MMP enzymes after P withdrawal during the artificial cycle. To demonstrate that MMPs are involved in menses in the vervet, we analyzed samples of RNA for MMP transcripts by focused gene array. Results of this analysis are shown in Figure 7. Compared to the secretory phase of the cycle, withdrawal of P during the artificial menstrual phase resulted Am. J. Primatol. DOI 10.1002/ajp 912 / Carroll et al. Fig. 7. Endometrial matrix metalloproteinase (MMP) transcript expression in artificially cycled vervets detected by focused gene array. Compared with the end of the secretory phase, expression of MMP-1, -2, -3, -7, -10, -11, -13, was significantly increased during menses. Tissue inhibitor of metalloproteinase-1 (Timp-1) but not Timp-2 was also elevated during menses. The levels of these enzymes declined to baseline after menses during the proliferative phase of the cycle. in a 30–100-fold increase in expression of MMP-1, 2, 3, 7, 10 and 11. Several other MMP transcripts including MMP-12, 13 and 26 were also elevated, but to a lesser extent. Expression of these enzymes then declined back to baseline during the proliferative phase of the artificial cycle. This pattern of MMP expression was very similar to that previously published for macaques [Rudolph-Owen et al., 1998]. Immunostaining for MMP-1, -2 and -7 during the artificial menstrual cycle is shown in Figure 8. MMP-1 and -2 were localized in the stroma and the fragmenting zone during menstruation on day 3 after P withdrawal. MMP-1 and -2 were greatly decreased by day 5. Similar strong staining was observed during early menstruation in the natural cycle (not shown). No specific MMP staining was observed in the non-menstruating animals in either the natural or artificial cycle. MMP-7 showed a strong signal in the glandular epithelium and not the stroma during late menses. Glandular expression of MMP-7 is unique compared with other MMPs, which are primarily stromal. Epithelial MMP-7 staining was also noted in macaques [Rudolph-Owen et al., 1998] and may correlate with the beginning of glandular remodeling at the end of the menstrual phase. DISCUSSION This study is the first assessment of experimentally induced, hormonal control of menstruation in the vervet. Compared to other primates including women [Jabbour et al., 2006], macaques [Slayden et al., 2001] and baboons, [D’Hooghe et al., 1994], vervets display exceptionally light menstrual bleeding that requires daily testing by vaginal swab to identify menstrual periods [Hess et al., 1979]. However, in most other respects, the vervet is very similar to other nonhuman primates. Am. J. Primatol. DOI 10.1002/ajp Endometrial Cycles in the Vervet / 913 Fig. 8. Photomicrographs showing immunostaining for matrix metalloproteinase (MMP)-1, -2 and -7 in the vervet endometrium. Stromal staining was clearly evident for MMP-1 and 2 during menses. Staining for MMP-7 increased during late menses but was localized the glandular epithelium. For instance, as in other nonhuman primates, E2 increased staining for ER and PR, both well-recognized endpoints of estrogen action. P during the luteal phase inhibited this action of E2 in the functionalis zone and suppressed ER and PR. This basic regulation by E2 in the vervet endometrium is identical to the wellrecognized pattern in other nonhuman primates and women. One limitation was the availability of animals for evaluation in this study. For instance, minimal sample size prevented a quantitative analysis of ER and PR staining. To reduce the variability among animals in this study, we deliberately treated animals with artificial cycles to control progestin levels as well as the timing of P withdrawal. This resulted in a robust effect on cell proliferation and MMP expression. For instance, E2 during the proliferative phase of the cycle stimulated epithelial cell proliferation only in the functionalis zone of the endometrium. During the secretory (luteal) phase of the cycle P stimulated cell proliferation that was restricted to the basalis zone. During the luteal phase P-suppresses proliferation in the functionalis zone. This cyclic pattern of functionalis–basalis cell proliferation was identical to that seen in macaques [Brenner et al., 2003] but quite different from the pattern seen in women, where P is not reported to stimulate basalis zone proliferation. Another striking change is the expression of MMPs, which increased as much as 200-fold in some cases during menses. Like other menstruating primates, menstrual bleeding and sloughing in the vervets was associated with the expression of MMP enzymes (e.g., MMP-1, 2, 3, 7 10, 11) during the menstrual phase of the cycle. Immunostaining revealed that MMP-1 and -2 proteins were strictly localized to the stromal cells in the sloughing zone of the endometrial functionalis. This is consistent with reports on other primates, including women and macaques, that the decline of P at the end of the cycle triggers the expression of MMP enzymes [Matrisian et al., 1994; Rudolph-Owen et al., 1998]. Am. J. Primatol. DOI 10.1002/ajp 914 / Carroll et al. The MMPs are capable of degrading the extracellular matrix and play a role in the menstrual dissociation of endometrial tissue in women [Matrisian et al., 1994; Osteen et al., 2003; Zhang & Salamonsen, 2002] and macaques. Expression of MMP-1, -2, -3, -10, -11, and -14 in the rhesus macaque endometrium peaks on days 2–4 of the luteal–follicular transition (LFT), then declines spontaneously after menstruation during the proliferative phase of the cycle even though P is absent [Rudolph-Owen et al., 1998]. Similar MMP enzymes are upregulated in the vervet. Paradoxically, PR consensus sequences have not been identified on the genes for these menstruation-associated MMPs. Therefore, P mediated regulation of endometrial MMPs appears to be indirect. Our studies show that similar MMP enzymes are expressed in both macaques [Rudolph-Owen et al., 1998] and vervets, therefore, differences in MMP expression do not explain light menstruation in the vervet. Our studies further indicate that ovarian function is not responsible for attenuated menstruation in this species. Similar light menses was observed during artificial cycles where ovarian steroid hormone levels were tightly controlled. Why, therefore, do vervets display exceptionally light menstrual bleeding? It is possible that vervet-specific anatomy may play a role in light menses observed in this species compared with the macaque. Our observation that the upper cervix was frequently cut in half during hysterectomy suggests that it is longer relative to the macaque. Some of the menstrual material may seep into the vaults of the cervical mucosa and never reach the vulva. Although difficult to quantify, the level of fragmentation in the vervet endometrium appeared to be restricted to the upper fourth or less of the functionalis zone. Moreover, at each time point in the artificial cycle there was less tissue sloughing than seen in other macaques or menstruating species. It is currently not known what physiological factors limit the expression of MMPs and menstruation to the upper functionalis zone. Like other menstruating nonhuman primates, vervets possess endometrial spiral arteries that vascularize the upper endometrial zones. These arteries appear to be critical to normal menstruation in primates [Markee, 1950]. Additional studies are needed to extensively analyze arterial patterns in vervets and to determine if differences in arterial function are responsible for reduced tissue sloughing in this species. However, reduced tissue volume and an elongated cervix may combine to be the reason that the external bleeding associated with menstruation is attenuated in this species. Despite minimal menstrual discharge in this species, strong similarities between the vervet and macaque endometrium indicate that vervets can be a useful animal model for studies of menstruation, including analysis of the role of MMPs in endometrial remodeling, and identification of zone-specific factors that regulate proliferation in the nonhuman primate endometrium. CONCLUSIONS Cyclic regulation of the vervet endometrium is very similar to that seen in macaques and women. P withdrawal at the end of the cycle induces menstrual breakdown marked by upregulation of the same MMP family typically observed in other nonhuman primates. The lack of external bleeding appears primarily because of the reduced amount of tissue that is sloughed off during the menstrual process. Therefore, this species could provide a viable animal model in which to study mechanisms that underlie the zonal differences in MMP expression, steroid receptor regulation and control of proliferation. Am. J. Primatol. DOI 10.1002/ajp Endometrial Cycles in the Vervet / 915 ACKNOWLEDGMENTS We gratefully acknowledge the Animal Care Technicians in the Division of Laboratory Animal Medicine for care of the animals during this study. REFERENCES Bartelmez GW. 1951. Cyclic changes in the endometrium of the rhesus monkey (Macaca mulatta). Contrib Embryol 34:99–144. Bartelmez GW. 1957. The phases of the menstrual cycle and their interpretation in terms of the pregnancy cycle. Am J Obstet Gynecol 74:931–955. Bennett HS, Wyrick AD, Lee SW, McNeil Jr JJ. 1976. Science and art in preparing tissues embedded in plastic for light microscopy, with special reference to glycol methacrylate, glass knives and simple stains. Stain Technol 51:71–97. Brenner RM, Slayden OD. 1994. Cyclic changes in the primate oviduct and endometrium. In: Knobil E, Neill JD, editors. The physiology of reproduction. New York: Raven Press. p 541–569. Brenner RM, Slayden OD. 2004. Steroid receptors in blood vessels of the rhesus macaque endometrium: a review. Arch Histol Cytol 67:411–416. Brenner RM, Slayden OD, Rodgers WH, Critchley HOD, Carroll R, Nie XJ, Mah K. 2003. Immunocytochemical assessment of mitotic activity with an antibody to phosphorylated histone H3 in the macaque and human endometrium. Hum Reprod 18: 1185–1193. Brenner RM, Nayak NR, Slayden OD, Critchley HOD, Kelly RW. 2002. Premenstrual and menstrual changes in the macaque endometrium: relevance to endometriosis. Ann NY Acad Sci 955:60–74. D’Hooghe TM, Bambra CS, Suleman MA, Dunselman GA, Evers HL, Koninckx PR. 1994. Development of a model of retrograde menstruation in baboons (Papio anubis). Fertil Steril 62:635–638. Eley RM, Tarara RP, Worthman CM, Else JC. 1989. Reproduction in the vervet monkey (Cercopithecus aethiops): III. The menstrual cycle. Am J Primatol 17:1–10. Gerdes J, Lemke H, Baisch H, Wacker HH, Schwab U, Stein H. 1984. Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J Immunol 133:1710–1715. Hess DL, Hendrickx AG, Stabenfeldt GH. 1979. Reproductive and hormonal patterns in the African green monkey (Cercopithecus aethiops). J Med Primatol 8:273–281. Hodgen GD. 1983. Surrogate embryo transfer combined with estrogen–progesterone therapy in monkeys. JAMA 250:2167–2171. Jabbour HN, Kelly RW, Fraser HM, Critchley HO. 2006. Endocrine regulation of menstruation. Endocr Rev 27:17–46. Kudolo GB, Mbai FN, Eley RM. 1986. Reproduction in the vervet monkey (Cercopithecus aethiops): endometrial oestrogen and progestin receptor dynamics during normal and prolonged menstrual cycles. J Endocrinol 110:429–439. Markee JE. 1950. The morphological and endocrine basis for menstrual bleeding. In: Meigs JV, Surgis SH, editors. Progress in gynecology. New York: Grune and Stratton. p 63–74. Matrisian LM, Gaire M, Rodgers WH, Osteen KG. 1994. Metalloproteinase expression and hormonal regulation during tissue remodeling in the cycling human endometrium. In: Koide H, Hayashi T, editors. Extracellular matrix in the kidney. Basel: Karger. p 94–100. Okulicz WC, Balsamo M, Tast J. 1993. Progesterone regulation of endometrial estrogen receptor and cell proliferation during the late proliferative and secretory phase in artificial menstrual cycles in the rhesus monkey. Biol Reprod 49:24–32. Osteen KG, Yeaman GR, Bruner-Tran KL. 2003. Matrix metalloproteinases and endometriosis. Semin Reprod Med 21:155–164. Padykula HA, Coles LG, Okulicz WC, Rapaport SI, McCracken JA, King NW Jr, Longcope C, Kaiserman-Abramof IR. 1989. The basalis of the primate endometrium: a bifunctional germinal compartment. Biol Reprod 40:681–690. Rudolph-Owen LA, Slayden OD, Matrisian LM, Brenner RM. 1998. Matrix metalloproteinase expression in Macaca mulatta endometrium: evidence for zone-specific regulatory tissue gradients. Biol Reprod 59: 1349–1359. Slayden OD, Brenner RM. 2004. Hormonal regulation and localization of estrogen, progestin and androgen receptors in the endometrium of nonhuman primates: effects of progesterone receptor antagonists. Arch Histol Cytol 67:393–409. Slayden OD, Chwalisz K, Brenner RM. 2001. Reversible suppression of menstruation with progesterone antagonists in rhesus macaques. Hum Reprod 16:1562–1574. Slayden OD, Hirst JJ, Brenner RM. 1993. Estrogen action in the reproductive tract of rhesus monkeys during antiprogestin treatment. Endocrinology 132:1845–1856. Am. J. Primatol. DOI 10.1002/ajp 916 / Carroll et al. Slayden OD, Koji T, Brenner RM. 1995. Microwave stabilization enhances immunocytochemical detection of estrogen receptor in frozen sections of macaque oviduct. Endocrinology 136:4012–4021. Am. J. Primatol. DOI 10.1002/ajp Zhang J, Salamonsen LA. 2002. In vivo evidence for active matrix metalloproteinases in human endometrium supports their role in tissue breakdown at menstruation. J Clin Endocrinol Metab 87:2346–2351.