Identification of -1, endothelin-2 and endothelin-3 in human endometrium I. T. Cameron, C. Plumpton, R. Champeney, C. van Papendorp, M. J. Ashby and A. P. Davenport department of Obstetrics and Gynaecology and 1 Clinical Pharmacology Unit, University of Cambridge Clinical School, Cambridge, UK

This study determined the presence of specific endothelin isoforms in human endometrium using high performance liquid chromatography (HPLC) combined with radioimmunoassay, and immunocytochemistry to detect the endothelin precursors (proendothelin-1, proendo- thelin-2 and proendothelin-3). Endothelin-like immunoreactivity was measured in HPLC eluates using antisera raised in rabbits against the carboxy-terminal heptapeptide of endo- thelin-1, which is common to the three endothelin isoforms. Of eight endometrial samples analysed by HPLC, three were in the proliferative phase of the cycle, and five in the secretory phase. Endothelin-1 was detected in seven samples, whereas endothelin-2 and endothelin-3 were seen in four and five specimens, respectively. No relationship was seen between endothelin isoforms and the stage of the cycle. Immunocytochemistry was performed on five proliferative and three secretory phase tissues. When present, staining for the precursor proendothelins was localized to endometrial glandular and luminal epithelium (proendo- thelin-1, 5 of 8; proendothelin-2, 5 of 8; proendothelin-3, 6 of 8). In two sections, staining was also seen in vascular endothelium using antibody raised against proendothelin-1 (n = 1) and proendothelin-3 (n = 1). These data provide evidence that the three endothelin isoforms are present in human endometrium, and suggest that these potent vasoactive agents may play a role in the paracrine control of the uterine vascular bed.

Introduction thelin-like immunoreactivity has also been detected in con¬ ditioned medium from human endometrial glandular and stromal cells short-term The comprise a family of 21 amino acid vasoactive in culture (Ecónomos et al, 1992). However, compounds, originally thought to be exclusively of endothelial it has not been possible to elucidate the precise endothelin isoform because of origin, and now known to be produced by a number of epi¬ in endometrium of crossreactivity anti¬ thelial tissues (Yanagisawa et al, 1988; MacCumber et al, 1989; bodies directed against either the whole endothelin molecule Baley et al, 1990; Marciniak et al, 1992). Human genomic or its carboxy terminus with the three endothelin isoforms. sequencing has predicted three distinct isoforms, endothelin-1, The aim of the present study was to determine the endo¬ endothelin-2 and endothelin-3 (Inoue et al, 1989), of which thelin isoforms present in normal human endometrium using endothelial-derived endothelin-1 has been described as a potent both high performance liquid chromatography (HPLC) and vasoconstrictor. The sites of synthesis and actions of endo¬ radioimmunoassay, and immunocytochemistry to detect the thelin-2 and endothelin-3 have been less extensively docu¬ endothelin precursors, proendothelin-1, proendothelin-2 and mented; endothelin-3 may act as a (Shinmi et al, proendothelin-3. 1989), but the role of endothelin-2 remains to be determined. Demonstration of the endothelins in human endometrium has suggested that these may play a role in the para¬ Materials and Methods crine control of the uterine vascular bed (Cameron and Davenport, 1992). Immunoreactive endothelin and specific binding sites for Materials iodinated peptides have been localized to endometrial glandular epithelium (Davenport et al, 1991; Cameron et al, 1992) and Endothelin peptides, proendothelin-1, and sarafatoxins S6b mRNA for the three endothelin peptides and their two known and S6c were supplied by Peninsula (St Helens, Merseyside, receptor subtypes, ETA and ETg, has been demonstrated UK) and Novabiochem (Switzerland). Other reagents were from throughout the menstrual cycle (O'Reilly et al, 1992). Endo- Cambridge Research Biochemicals, Cheshire (), Peninsula intestinal mouse "Address correspondence to: Professor I. T. Cameron, The Queen Mother's (vasoactive (VIP), vaso- Hospital, Yorkhill, Glasgow GJ 8SH, UK. active intestinal constrictor (VIC), , met-enkephalin, Received J August 1992. ), Bachern, Saffron Waiden, Essex (human Downloaded from Bioscientifica.com at 09/24/2021 06:23:28PM via free access gene-related peptide-1 (CGRP-1), atrial natriuretic factor (ANF), performed on one column using 10_6moll_1 injections of ET (CCK), , somatostatin, substance P, isoforms and online UV detection (Waters 455 UV spectro- dynorphin A (1-6)), and Sigma Chemicals, Poole (epidermal photometer, 214 nm). After extensive system flushing, includ¬ growth factor (EGF), angiotensin II (A II) and other reagents). ing a change of guard column, a separate but batch-matched Commercial rabbit antisera were from Peninsula. The antisera column was used, and the 1 min fractions subsequently collected raised in rabbits against the endothelin-1 carboxy-terminal hepta- were assayed for endothelin-like immunoreactivity by radio¬ peptide, endothelin-l,15_21) have been described elsewhere immunoassay. This column was standardized with endothelin (Cameron et al, 1992). Secondary swine anti-rabbit sera and isoforms (10-9 mol l-1). The selected gradient was A, 100% for rabbit were obtained from Dako Ltd 10 min, increasing to B, 100% at 50 min (solvent A and peroxidase—antiperoxidase - (High Wycombe, Bucks). acetonitrile:50 mmol dibasic sodium phosphate 1_1:TFA (24: 75.9:0.1 and 40:59.9:0.1, v:v:v, respectively)). Fractions 5 to 44 inclusive were evaporated to dryness overnight and either Tissues assayed immediately or sealed and stored at 70°C. The sys¬ — tem was equilibrated with at least 30 ml solvent. A between Endometrium was obtained at curettage or hysterectomy each standard or sample run, and after every four to six samples from 16 women undergoing surgery for benign disease. Six or standard runs, a blank run was performed and assayed by women had leiomyomata, two had pelvic pain, three had a pro¬ radioimmunoassay to exclude carry over of immunoreactivity and five of heavy Informed consent lapse complained periods. from previous separations. was sought, and the study was approved by the District Ethical Committee of the Cambridge Health Authority. Samples were immediately snap-frozen in liquid nitrogen and stored at Radioimmunoassay 70°C until extraction for HPLC (n = 8), or sectioning for

— immunocytochemistry (n = 8). A portion of each tissue was Fractions were assessed by radioimmunoassay using antibody also fixed in formalin for histological assessment (Noyes et al, raised against the endothelin-1 carboxy terminus. The limit of 1950). detection of the assay, determined by the lowest amount of unlabelled peptide that could be detected above zero, was < 1.25 fmol per tube. Crossreactivity was tested against a High performance liquid chromatography variety of peptides including endothelin-1, endothelin-2 and endothelin-3 (100%), A II (< 0.005%), ANF (< 0.005%), human After the tissue was cut into small a thawing, pieces using CGRP-1 (< 0.005%), (<0.01%), mouse clean razor and with mol acid 1~ /0.1% proendothelin-l(22_38) homogenized 0.5 acetic VIC (46%) and the sarafatoxins S6b (1.7%) and S6c (1.2%). trifluoroacetic acid (TFA, lOmlg-1 wet weight) in a 45 ml polycarbonate centrifuge tube (Sorval, Du Pont, Stevenage). The homogenates were placed in a boiling water bath for Immunocytochemistry 15 min, and then allowed to cool to room temperature. After centrifugation at 48 000 g for 20 min at 4°C, the resulting Cryostat sections through endometrium and adjacent myo¬ supernatant was stored in polypropylene tubes at 70°C until metrium, 30 µ thick, were thaw-mounted onto gelatin-subbed — assayed. slides, and allowed to warm to room temperature. Two or three Supernatants were solid phase extracted using Amprep C2 sections were examined from tissues from each of the eight minicolumns (Amersham) and a vacuum manifold apparatus women. Sections were fixed in 4% paraformaldehyde, incubated (Spe-ed, Lab Impex, Twickenham). The minicolumns were with 10% goat serum to block nonspecific staining, and then activated with methanol and washed with water. Samples cor¬ incubated with primary antisera directed against proendothelin- responding to up to 2 g wet weight of original tissue were then 1, proendothelin-2 and proendothelin-3 for 48 to 60 h at a applied under low vacuum at < 1ml min-1. After washing dilution of 1:100 and 1:300 (v/v). Binding of primary antibody twice with 2.8 ml 0.1% TFA to remove loosely bound material, was visualized using an anti-rabbit second antibody (diluted bound peptides were eluted with 2 ml 40% acetonitrile/0.1% 1:200 (v/v) in 0.05 mol PBS \ pH 7.4) followed by rabbit TFA and collected in polypropylene tubes (Sarstedt, Leicester). peroxidase-antiperoxidase (1:400, v/v). The resulting com¬ Eluates were evaporated to dryness overnight using a centri¬ plexes were detected with diaminobenzidine and hydrogen fugal evaporator connected to a freeze-drier (Speedvac, Savant, peroxide in 0.05 mol Tris-HCI 1_1 (pH 7.4) in sections and Life Sciences, Luton). The dried samples were then stored at counterstained with haemalum (Sternberger et al, 1970). 70°C. Positive controls included the detection of staining in the

- Immediately before chromatography, the samples were cytoplasm of human umbilical artery and vein endothelial cells reconstituted in 200 µ solvent A (see below) and centrifuged at with antisera raised against proendothelin-1. No staining was seen 11 000 £ for 5 min at room temperature. Of the resulting super¬ when primary antibody was omitted, or substituted with pre¬ natant, 100 µ was injected onto the HPLC column for separation immune serum. Pre-absorption with synthetic proendothelin-1, and subsequent radioimmunoassay. proendothelin-2 and proendothelin-3 (10~ moll-1) abolished Reverse-phase HPLC was performed on a Waters 600 Multi- staining with the corresponding proendothelin antisera. Speci¬ solvent delivery system, Rheodyne 7125 sample injector, Phase ficity was determined by incubating primary antibodies with separations ODS II C18 250 4.6 mm column. Flow rate was peptides unrelated to the endothelins (10-6 mol 1_1), but which 1.0 ml min-1 and 1 ml fractions were collected by an LKB 2212 may be present in the uterus or have a similar cyclic structure, Helirac fraction collector. The initial gradient selected was including human CGRP-1, ANF, VIP, CCK-4, pentagastrin, Downloaded from Bioscientifica.com at 09/24/2021 06:23:28PM via free access 500 (a) Table 1. Detection of endothelin-1, endothelin-2 and endo¬ thelin-3 in homogenates of human endometrium by HPLC/ radioimmunoassay

-2 Day Endothelin-1 Endothelin-2 Endothelin-3

ö 25°r -3 9 ++ + + -1 12 + + 13 ++ ++++— ++ 19 ++ 20 ++ — ++— I Ln_r-cnnljnnnlillilnnn„nirri7?7 22 + —+ 25 45 22 26 +++— ++++— ++++—

Indistinguishable from baseline values; + equivalent to <50 fmol per tube -1 -2 —when compared with standard; -f— +++ and ++-1—t- equivalent to 500.- 50-100, 100-500 and > 500 fmol tube. (b) per

-3 Immunocytochemical localization of the proendothelins was determined in five proliferative phase (days 2, 6, 7, 13, 14) and three secretory phase tissues (days 15, 20, 23). Immunoreactive 250 proendothelin-1, proendothelin-2 and proendothelin-3 were detected in endometrium throughout the cycle. The pattern of distribution of proendothelin-like immunoreactivity was similar in all tissues examined. Staining was localized to the cytoplasm of endometrial glandular and luminal epithelium (Fig. 2), and less often to the cytoplasm of vascular endothelium, although staining n^.,nnnnllllíinnnnJ was ensure that 25 45 of parallel sections not performed to endothelial cells had not been lost slide preparation (Table 2). Pro¬ Fraction number (min) during endothelin-like immunoreactivity did not appear to be present in 1. of Fig. Representative HPLC/radioimmunoassay profiles (a) pro¬ endometrial stroma, and there was no relationship between the liferative and (b) secretory endometrium. Peaks have been labelled specific isoform detected and the pattern of localization, the phase endothelin(ET)-l, endothelin(ET)-2 and endothelin(ET)-3 to correspond of the menstrual or the to retention times of endothelin standards. ET-IR: endothelin- cycle, underlying pathology. Staining synthetic was not affected with like immunoreactivity. by pre-absorption non-proendothelin-like peptides, but was abolished after pre-absorption with the corresponding synthetic proendothelins. somatostatin, substance P, eledoisin, dynorphin A(l-6), met- enkephalin, bradykinin, neuropeptide Y and EGF. Proendothelin antisera did not crossreact with their respective cleaved 21 Discussion amino acid endothelin peptides (Howard et al, 1992). This study suggests that the peptides endothelin-1, endothelin- 2 and endothelin-3 are present in human endometrium. The Results detection of endothelin-like immunoreactivity in the endo¬ metrium is in accord with previous work, although the relative Of the eight endometrial samples analysed by HPLC, three contribution of the endometrial epithelial and stromal compo¬ were in the proliferative phase (days 9, 12, 13) and five in the nents is debated, and may differ among species (Orlando et al, secretory phase of the cycle (days 19, 20, 22, 22, 26), and this 1990; Maggi et al, 1991; Cameron et al, 1992; Ecónomos et al, was confirmed by histological assessment. Representative 1992). However, previous studies have been unable to detect HPLC/radioimmunoassay profiles are shown (Fig. 1) and the specific endothelins because all antibodies used, whether raised distribution of endothelin-like immunoreactivity in the eluates against the whole 21 amino acid endothelin peptides or their corresponding to the three endothelin isoforms is given (Table common carboxy terminus, crossreact with all three ). Endothelin-1 was present in seven of the eight tissues; isoforms. The use of HPLC in the study reported here has endothelin-2 and endothelin-3 were detected in four and five permitted separation of the endothelin peptides, which were samples, respectively. Detection of the different ET isoforms did subsequently measured in the appropriate fractions by radio¬ not appear to be related to the stage of the menstrual cycle, and immunoassay, using an antibody raised against the carboxy although results (Table 1) might suggest an increase in endo¬ terminal heptapeptide, endothelin-l:5_2r Detection of the dif¬ thelin-like immunoreactivity premenstrually, loss of peptide ferent isoforms was not considered to be the result of carry over during the extraction process precludes precise quantification. from previous separations, as immunoreactive endothelin was Downloaded from Bioscientifica.com at 09/24/2021 06:23:28PM via free access Table 2. Identification of proendothelin-like immunoreactivity in human endometrium

Proendothelin-1 Proendothelin-2 Proendothelin-3

Day

2 + + + 6 + + + 7 + + + - + 13 - +

- 14 +- - - 15 - 20 — — +

- 23 +- +

( + ) The presence or ( ) absence of specific immunoreactivity in endometrial glandular epithelium (G)— and vascular endothelium (E).

in human endometrium throughout the menstrual cycle (O'Reilly et al, 1992). Detection of the three endothelin peptides in human endo¬ metrium is further borne out by the detection of their precursor proendothelins by immunocytochemistry. The pattern of distri¬ bution of the proendothelins was similar to that of the mature 21 amino acid peptides, being confined to endometrial glandu¬ lar and luminal epithelium, and vascular endothelium (Cameron et al, 1992), but differed in that unlike the mature endothelin peptides, proendothelin-like immunoreactivity was not seen in all sections studied. This may be a reflection of the efficiency of the antibodies used, or the state of proendothelin cleavage; it is thought that the endothelins may be stored in tissues in pre¬ cursor form and released when required by cleavage using a putative endothelin converting enzyme (Yanagisawa et al, 1988). The precise role of the different endothelin isoforms in endo¬ metrium is not known. In other tissues, including vascular smooth muscle and anterior pituitary cells, endothelin-1 is thought to act on the ETA receptor to effect intense and long- lasting vasoconstriction (Fried and Samuelson, 1991), or the release of gonadotrophins (Stojilkovic et al, 1992). Endothelin-3 has been demonstrated in the central nervous system, and may a role as a neurotransmitter el Fig. 2. Light field photomicrographs showing the localization of pro¬ play brain—gut peptide (Shinmi al, endothelin-1-like immunoreactivity in human endometrium. (a) Low 1989). This isoform binds to the ETB receptor with similar affinity power view of proliferative endometrium showing localization of to both endothelin-1 and endothelin-2. Activation of the ETB immunoreactivity to luminal and glandular epithelium, (b) High power receptor in endothelial cells may result in both the release of view of the same tissue as in (a), (c) Negative control of the same tissue prostacyclin and nitric oxide, and upregulation of endothelin-1 as in (a) and (b) with omission of primary antibody. Cryostat sections (Yokokawaefß/., 1991). Messenger RNAforETA and ETB receptor were fixed and incubated with directed (30 µ thick) primary antibody subtypes has been demonstrated in human endometrium through¬ against and for 48 proendothelin-1, proendothelin-2 proendothelin-3 out the menstrual cycle, with an increase in the ratio of to 60 h at 4°C at a dilution of 1:100 and 1:300 (v/v). Antibody binding ETB:ETA evident in the mid-luteal (O'Reilly et al, 1992). was visualized using the unlabelled antibody-enzyme method in subtype phase The of endothelin-1, endothelin-3 and both endothelin sections counterstained with haemalum. In two sections, staining was presence in would that the also seen in vascular endothelium using antibody raised against pro¬ receptor subtypes endometrium suggest of the endothelin-1 or proendothelin-3 (see Table 2). L: uterine cavity lumen; endothelins could contribute to the paracrine control G: gland; S: stroma. Scale bars: (a) and (c) = 100 µ , (b) = 50 µ . uterine vascular bed to permit vasoconstriction and haemostasis at menstruation, when endometrial sloughing would allow access of glandular endothelins to the abluminal aspect of the spiral not found in equivalent eluate fractions from blank runs. arterioles. Alternatively, activation of the ETB receptor might Demonstration of the three isoforms is supported by the pres¬ permit the release of vasodilatory mediators such as prostacyclin ence of mRNA for endothelin-1, endothelin-2 and endothelin-3 and nitric oxide at the time of implantation (De Nucci et al, 1988). Downloaded from Bioscientifica.com at 09/24/2021 06:23:28PM via free access The by endothelin-2 in local vascular control has not Howard PG, Plumpton C and Davenport AP (1992) Anatomical localisation and part played of mature endothelins and their in human yet been defined, and its tissue of has not been demon¬ pharmacological activity precursors origin vascular tissue Journal of Hypertension 10 1379—1386 to of strated. However, endothelin-2 will bind both the endothelin Inoue A, Yanagisawa M, Kimura S, Kasuya Y, Miyauchi T, Goto and Masaki receptor subtypes, and although it binds to the ETA receptor with (1989) The human endothelin family: three structurally and pharmacologi¬ lower affinity than does endothelin-1, it may be more potent than cally distinct isopeptides predicted by three separate genes Proceedings of the endothelin-1 in pharmacological studies (Morton and Davenport, National Academy of Sciences USA 86 2863-2867 1992). MacCumber MW, Ross CA, Glaser BM and Snyder SH (1989) Endothelin: visual¬ ization of mRNAs by in situ hybridization evidence for local action In this has evidence that the three provides conclusion, study provided Proceedings of the National Academy of Sciences USA 86 7285-7289 endothelin isoforms, endothelin-1, endothelin-2 and endothelin- Maggi M, Vannelli GB, Peri A, Brandi ML, Fantoni G, Giannini S, Torrisi C, 3, are present in human endometrium, and suggests that these Guardabasso V, Barni T, Toscano V, Massi G and Serio M (1991) Immuno- and of endothelin in rabbit uterus: vasoactive may a role in the local control of localization, binding, biological activity potent agents play effect of ovarian steroids American 260 E292—E305 the uterine vascular bed. Journal of Physiology Marciniak SJ, Plumpton C, Barker PJ, Huskisson NS and Davenport AP (1992) Localization of immunoreactive endothelin and proendothelin in human lung The authors thank R. Kuc for invaluable technical assistance. The Pulmonary Pharmacology 5 175—182 was from The East Health study supported by grants Anglian Regional Morton AJ and Davenport AP (1992) Cerebellar neurones and glia respond Authority (ITC), Birthright, Royal College of Obstetricians and differentially to endothelins and sarafotoxin S6b Brain Research 581 299—306 Gynaecologists (ITC), the British Heart Foundation (APD) and the Noyes RW, Hertig AT and Rock J (1950) Dating the endometrial biopsy Fertility Isaac Newton Trust (APD). and Sterility 1 3—25 O'Reilly G, Charnock-Jones DS, Davenport AP, Cameron IT and Smith SK (1992) Presence of mRNA for endothelin-1, endothelin-2 and endothelin-3 in References human endometrium, and a change in the ratio of ETA and ETa receptor subtype across the menstrual cycle Journal of Clinical Endocrinology and Metabolism 75 1545-1549 Baley PA, Resink TJ, Eppenberger U and Hahn WA (1990) Endothelin messenger Orlando C, Brandi ML, Peri A, Giannini S, Fantoni G, Calabresi E, Serio M RNA and receptors are differentially expressed in cultured human breast and stromal cells Journal of Clinical Investigation 85 1320—1323 and Maggi M (1990) Neurohypophyseal hormone regulation of endothelin epithelial secretion from rabbit endometrial cells culture Cameron IT and AP (1992) Endothelins in in primary Endocrinology 126 Davenport reproduction Reproductive 1780-1782 Medicine Review 1 99-113 Shinmi Kimura Sawamura T, Yoshizawa Y, Cameron IT, Davenport AP, van Papendorp CL, Huskisson NS, Gilmour RS, O, S, Sugita Y, T, Uchiyama Brown MJ and Smith SK (1992) Endothelin-like immunoreactivity in human Yanagisawa M, Goto K, Masaki and Kanazawa I (1989) Endothelin-3 is a of endometrium Journal of Reproduction and Fertility 95 623—628 novel neuropeptide: isolation and sequence determination endothelin-1 Davenport AP, Cameron IT, Smith SK and Brown MJ (1991) Binding sites for and endothelin-3 in porcine brain Biochemical and Biophysical Research 164 iodinated endothelin-1, endothelin-2 and endothelin-3 demonstrated on Communications 587—593 human uterine glandular epithelial cells by quantitative high-resolution Sternberger LA, Hardy PH Jr, Cuculis JJ and Meyer HG (1970) The unlabeled autoradiography Journal of Endocrinology 129 149—154 antibody—enzyme method of immunocytochemistry. Preparation and De Nucci G, Thomas R, D'Orleans-Juste P, Antunes E, Walder C, Warner TD and properties of soluble antigen—antibody complex (horseradish peroxidase— and its use in identification of Vane JR (1988) Pressor effects of circulating endothelin are limited by its antihorseradishperoxidase) spirochetes Journal removal in the pulmonary circulation and by the release of prostacyclin and of Histochemistry and Cytochemistry 18 315—333 endothelium-derived relaxing factor Proceedings of the National Academy of Stojilkovic SS, Balla , Fukuda S, Cesnjaj M, Merelli F, Krsmanovic LZ and Catt Sciences USA 85 9797-9800 KJ (1992) Endothelin ETA receptors mediate the signaling and secretory Ecónomos K, MacDonald PC and Casey ML (1992) Endothelin-1 gene expression actions of endothelins in pituitary gonadotrophs Endocrinology 130 465—474 and protein biosynthesis in human endometrium: potential modulator of Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobayashi M, Mitsui Y, endometrial blood flow Journal of Clinical Endocrinology and Metabolism 74 Yazaki Y, Goto T and Masaki (1988) A novel potent vasoconstrictor 14-19 peptide produced by vascular endothelial cells Nature 332 411—415 Fried G and Samuelson U (1991) Endothelin and neuropeptide Y are vasocon¬ Yokokawa K, Kohno M, Yasunari K, Murakawa and Takeda (1991) Endo¬ strictors in human uterine blood vessels American Journal of Obstetrics and thelin-3 regulates endothelin-1 production in cultured human endothelial Gynecology 164 1330-1336 cells Hypertension 18 304-315

Downloaded from Bioscientifica.com at 09/24/2021 06:23:28PM via free access