C4s(iI4mI3(taCao

emittbeassdi

RS

R

CORRESPONDENCE

1

Human endometrium expresses urocortin II and IIImessenger RNA and peptides

Urocortin II and urocortin III are recently identified corticotropin-releasing factor–related neuropeptides, andtheir endometrial expression throughout the menstrual cycle and in early pregnancy was evaluated in thepresent study by semiquantitative reverse transcriptase polymerase chain reaction and immunohistochemistry.The endometrial expression of urocortin II mRNA was significantly (P�.01) higher in the early proliferativephase of the menstrual cycle than in other phases and maternal decidua (MD), whereas that of urocortin IIImRNA was higher (P�.01) in MD than in all other endometrial samples. Both peptides were immunolocalizedin epithelial, stromal, and endothelial cells. (Fertil Steril� 2006;86:1766–70. ©2006 by American Society for

Reproductive Medicine.)

tpl

Ippe

MShwamfbcneetfit

npota(spB4

orticotropin-releasing factor (CRF) is a neuropeptide of1 amino acids and has a key role in diverse aspects of thetress regulation of the hypothalamus-pituitary-adrenalHPA) axis (1). The family of CRF comprises recentlydentified neuropeptides, namely urocortin (UCN) I, II, andII (2–4). Urocortin I is a 40-amino-acid peptide that shares5% sequence homology with CRF (2); UCN II showsoderate homology to human CRF (34%) and human UCN(43%) and UCN III (37%–40%), which in turn shares

2% homology to human CRF and 26% to human UCN I3, 4). The UCNs and CRF bind with different affinity towo distinct receptors, respectively, CRF-type 1 (CRF-R1)nd -type 2 (CRF-R2). Urocortin I displays high affinity forRF-R2, whereas CRF has approximately ten-fold lowerffinity to this receptor; UCN II and III specifically bindnly to CRF-R2 (5).

Corticotropin-releasing factor (6, 7) and UCN I (8) arexpressed by the human endometrium (epithelial and stro-al cells) (6, 7), and immunoreactive CRF concentrations

n endometrial biopsies are higher in the secretory than inhe proliferative phase (9). These findings, together withhe evidence that the human endometrium also expressesoth CRF-R1 (10) and CRF-R2 (11), suggest that thendometrium is the source but also the target of both CRFnd UCNs. In this regard, the activation of CRF receptorubtypes has been demonstrated to have important roles ineveral aspects of endometrial physiology, as in the case ofecidualization of endometrial stromal cells (12, 13), thenduction of blastocyst implantation and early maternal

eceived August 3, 2005; revised and accepted May 9, 2006.upported by grants from the Italian Ministry of University and ScientificResearch (MURST) and the University of Siena.

eprint requests: Felice Petraglia, M.D., Chair of Obstetrics and Gyne-cology, Department of Pediatrics, Obstetrics, and ReproductiveMedicine, University of Siena, Policlinico “Le Scotte” Viale Bracci,

p53100 Siena, Italy (FAX: �39 0577 233.454; E-mail: petraglia@unisi.it).

766 Fertility and Sterility� Vol. 86, No. 6, December 2006Copyright ©2006 American Society for Reproductive Medicine,

olerance (14), and the inhibition of tumor cell growth androliferation of endometrial cells derived from a tumor celline (15).

Therefore, the present study evaluated whether UCNI and III messenger (m) RNA and peptides are ex-ressed by human endometrium and whether their ex-ression changes through the endometrial cycle and inarly pregnant decidua.

ATERIALS AND METHODSamples of nonpregnant endometrium were obtained fromealthy fertile subjects (n � 34; age range 37 to 43 years)ith normal menstrual cycle and actually ovulatory, as

ssessed by transvaginal ultrasonography (16), who wereedical students or laboratory staff and who volunteered

or the study, underwent hysteroscopy, and were found toe free of endometrial pathologies. Subjects who had re-eived steroid treatment during the previous 6 months wereot included in the study. Specimens were classified asarly (EP; n � 10) and late (LP; n � 9) proliferative, andarly (ES; n � 8) and late (LS; n � 7) secretory endome-rium according to the last day of menstruation and con-rmed by both transvaginal ultrasonography (16) and by

he histologic criteria of Noyes et al. (17).

We also collected samples of maternal decidua (MD;� 7, gestational age ranging from 8 to 12 weeks from

regnant women who underwent a voluntary terminationf pregnancy) and pregnant myometrium (n � 3) ob-ained from women undergoing elective cesarean sectiont term, which were used as positive control for UCN II18). Informed written consent was obtained from allubjects before inclusion in the study, for which ap-roval was obtained from the local Institutional Reviewoard. Frozen kidney tissues (about 1 g) obtained withinh postmortem from two male patients were used as

ositive control for UCN III (19).

0015-0282/06/$32.00Published by Elsevier Inc. doi:10.1016/j.fertnstert.2006.05.041

MRtiqtRppTHm(tgga�eap(ra

Ucj(asPcUmf7cafGssfommtgmI

it

o3TmmiitwbLpmoimha

wagdswwiorl

sgvmw

RTR(awq(thes(

w

Fe

Samples were disrupted and homogenized usingixer Mill MM 300 (Qiagen, Milan, Italy), and totalNA was extracted with RNeasy Protect Mini Kit and

hen treated with RNase-free DNase according to thenstructions of the manufacturer (Qiagen). RNA wasuantified by UV absorption, and 1 �g was reverseranscribed to prepare complementary DNA (cDNA).eaction conditions for reverse transcriptions were asreviously described (20). Two microliters of reactionroduct was used for polymerase chain reaction (PCR).he PCR conditions were as follows: 20 mmol/L Tris-Cl (pH 8.4), 50 mmol/L KCl, 1.5 mmol/L MgCl2, 0.25mol/L dNTPs, 1 U Taq DNA polymerase recombinant

Invitrogen, Milan, Italy), and 0.4 �mol/L (final concen-ration) UCN II (sense 5=-gtgtcggccactgctgagcct-agaga-3= and antisense 5=-atctgatatgacctgcatgacagtg-ct-3=) and UCN III (sense 5=-tgctgctcctgctgctgctc-3=nd antisense 5=-gtgtcctggcgtggctttccc-3=) primers in 50L total volume. Sequence homology among the differ-nt oligomers used in the present study was alsovoided, excluding possible cross-reactions. The ex-ected size of the amplified fragment was 195 base pairsbp) for UCN II and 310 bp for UCN III. Blank for eacheaction, consisting of amplifications performed in thebsence of reverse transcriptase enzyme (RT), was done.

To estimate differences in mRNA expression ofCN II and UCN III, different phases of the menstrual

ycle were compared by semiquantitative RT-PCR, ad-usted with glyceraldehyde-3-phosphate dehydrogenaseGAPDH; primers: sense 5=-gaaggtgaaggtcggagtca-3=nd antisense 5=-ctgagaacgggaagcttgtc-3=) as the internaltandard. The number of cycles needed to perform theCR in linear phase was selected after a number ofalibration experiments in the range of 28 –36 runs. TheCN II amplification was carried out at 94°C for 1inute, at 60°C for 1 minute, and at 72°C for 1 minute

or 30 thermed cycle steps, followed by a final step at2°C for 10 minutes. The UCN III amplification wasarried out at 94°C for 1 minute, at 63°C for 1 minute,nd at 72°C for 1 minute for 32 thermed cycle steps,ollowed by a final step at 72°C for 10 minutes. TheAPDH amplification was carried out at 94°C for 30

econds, at 52 °C for 30 seconds, and at 72°C for 30econds for 28 cycles, followed by a final step at 72°Cor 10 minutes. Amplification products were visualizedn 2% agarose gel and stained with 3% ethidium bro-ide. The expected bands were quantified by densito-etric analysis performed using Image J software (Na-

ional Institutes of Health) by an employee of the federalovernment. Relative amount of UCN II and UCN IIIRNA was calculated as UCN II/GAPDH and UCN

II/GAPDH mRNA ratios.

To evaluate the localization of UCN II and UCN III,mmunohistochemistry was carried out on 5-�m thick sec-

ions, obtained from paraffin-embedded samples, mounted e

rtility and Sterility�

n electrostatically charged slides, and dried overnight at7°C. Sections were dewaxed, rehydrated, and washed inris-buffered saline [TBS: 20 mmol/L Tris-HCl, 150mol/L NaCl (pH 7.6)]. Tissue solutions were heated in aicrowave oven twice for 5 minutes at 750 W and rinsed

n 3% H2O2 to block endogenous peroxidase. Slides werencubated overnight at room temperature with primary an-ibody. Antibodies used for UCN II (21) and UCN III (22)ere rabbit antihuman polyclonal diluted 1:1000, providedy Professor W. Vale (Salk Institute for Peptide Biology,a Jolla, CA). Anti-UCN II and anti-UCN III are affinity-urified goat polyclonal antibodies, raised against peptideappings at the carboxy terminus of UCN II and UCN III

f human origin. These antibodies, identical to correspond-ng mouse and rat sequences, recognize UCN II and III ofouse, rat, and human origin with no cross-reactivity, and

uman or rat CRF, sheep CRF, PACAP38, human ACTH,nd sauvagine (21, 22).

The reactions were developed by successive incubationsith antirabbit immunoglobins labeled with biotin, the

vidin-biotin peroxidase complex (Vector Lab, Burlin-ame, CA), and 1 mg/mL 3,3=-diaminobenzidine tetrahy-rochloride (Sigma, St. Louis, MO) as chromogen sub-trate in TBS containing 0.3% H2O2. Harris hematoxylinas used for nuclear counterstaining. A positive reactionas characterized by the presence of granular brown stain-

ng in the cytoplasm. For each case, a negative control wasbtained by using the antibody preadsorbed with the cor-esponding peptide at the concentration of 20 �g/mL di-uted antibody.

After confirming a normal distribution, the data wereummarized as mean � standard error (SE). Between-roup differences were evaluated by using the analysis ofariance test, followed by the post-hoc Tukey test forultiple comparisons. Statistical significance was assumedhen P�.05.

ESULTSotal RNA extracted from endometrium was analyzed byT-PCR. Predicted bands corresponding in size to UCN II

196 bp) and to UCN III (310 bp) products were obtained,nd no amplified fragment caused by DNA contaminationas detected in any experiment. When evaluated by semi-uantitative RT-PCR, the expression of UCN II mRNAexpressed as UCN II/GAPDH mRNA ratio) was highest inhe early proliferative endometrium, significantly (P�.01)igher than in other samples and decidua, and the highestxpression of UCN III mRNA was in maternal decidua,ignificantly (P�.01) higher than in endometrial phasesdata not shown).

In proliferative endometrium, immunoreactive UCN IIas found in luminal and glandular epithelial and vascular

ndothelial cells, and UCN II-positive stromal cells were

1767

mssgper

gce1a

D

TUdmIta(rcm

F

1

ost abundant close to the superficial layers (Fig. 1A). Inecretory endometrium, luminal epithelial cells weretrongly immunostained, whereas staining was weaker inlandular cells (Fig. 1B). Maternal decidual cells showedositive UCN II stain, with immunoreactivity found inpithelial stromal and endothelial cells from spiral arte-ioles (Figs. 1C and 1D).

Urocortin III intensely immunostained superficial andlandular cells as well as vascular endothelial and stromalells of the proliferative endometrium (Fig. 1F). Secretoryndometrium (Fig. 1G) and maternal decidua (Figs. 1H andI) showed intense UCN III immunostaining in superficial

IGURE 1

CN II (A–E) and III (F–J) immunolocalization in humanndometrium. Intense UCN II immunostaining was mainlandular (white arrows) cells, whereas a weaker staininessel walls (dotted circles) showed an intense stainingrrows) and luminal (black arrows) cells showed intense

ntense in stromal cells. (C, D) Early pregnant endometrrrows) and endothelial (dotted circles) cells from spiralntibody preadsorbed with UCN II (20 �g). (F) Proliferatainly localized in epithelial luminal (black arrows) andas found in stromal cells. Endothelial cells of the vessecretory endometrium. Epithelial glandular (white arrow

II staining. Again, UCN III immunostaining was less inteCN III was localized in epithelial (black arrows), stromaontrol obtained by using the antibody preadsorbed wi

lorio. Urocortin 2 and 3 in human endometrium. Fertil Steril 2006.

nd glandular epithelial, stromal, and vascular cells. m

768 Florio et al. Correspondence

ISCUSSION

he present study is the first to refer to the expression ofCN II and III into human endometrium and maternalecidua and to their immunolocalization in epithelial, stro-al and endothelial cells. The mRNA expression of UCN

I was higher in the early proliferative endometrium, andhat of UCN III in pregnant decidua. The facts that UCN IInd UCN III are members of the CRF family of peptides2–4) and that CRF (6, 7), UCN I (8, 23) and CRFeceptors (10,11) are expressed throughout the endometrialycle together reinforce the notion that the human endo-etrium possesses the ability to synthesize peptides/hor-

ometrium and pregnant decidua. (A) Proliferativecalized in epithelial luminal (black arrows) and

as found in stromal cells. Endothelial cells of theSecretory endometrium. Epithelial glandular (whiteN II staining. Again, UCN II immunostaining was less. UCN II was localized in epithelial stromal (blackrioles. (E) Negative control obtained by using thendometrium. Intense UCN III immunostaining was

dular (white arrows) cells, whereas a weaker stainingalls (dotted circles) showed an intense staining. (G)nd luminal (black arrows) cells showed intense UCNin stromal cells. (H, I) Early pregnant endometrium.d endothelial (dotted circles) cells. (J) Negative

CN III (20 �g).

F

U ende ly log g wv . (B)a UCi iuma artea ive em glanw el wS s) aI nseU l, anc th U

ones that are normally expressed into the brain. More-

Vol. 86, No. 6, December 2006

oUt

Tdfdtw

baewUmTeagamzPmv

pdCvrsti

epsp

R

1

1

1

1

1

1

1

1

1

Fe

ver, the findings of the local expression of UCN II andCN III open the question of the possible roles played by

hese neuropeptides in the human endometrium.

Urocortins II and III specifically bind to CRF-R2 (5).he fact that the coupling of CRF-R1 and -R2 activatesifferent intracellular signalling cascades that mediate dif-erent effects of CRF and UCNs leads us to suggest thatistinct signaling networks exist in the human endome-rium, involving UCN II and III and CRF-R2 receptors asell as involving CRF and UCN I with CRF-R1.

Vascular endothelial cells express CRF-R2 (24, 25), theinding of UCNs to CRF-R2 mediates vasodilatation (26),nd both UCN II and III have shown to exert cytoprotectiveffect in myocardiocytes (27). The endometrium is a tissueith strong remodeling, and a potential involvement of bothCNs in endometrial cell survival and apoptosis during theenstrual cycle and early in pregnancy may be hypothesized.herefore, we may propose that UCN II and III regulatendometrial angiogenesis and/or vascular endothelial tone thatre key events in the mechanisms of menstruation. Prosta-landins (PGs) are involved in menstruation (28), and thectivation of CRF-Rs modulates its synthesis (12) through theodulation of cyclooxygenase-2 (COX-2) activity, the en-

yme that locally synthesizes PGs (29, 30). The inhibition ofG release is also fundamental for pregnancy to occur, and aarked inhibition of decidual PGs synthetase activity occurs

ery soon during early pregnancy (27).

These findings and the evidence that CRF suppresses theroduction of PGs from human endometrial cells in time- andose-dependent fashion, probably through the inhibition ofOX-2 (12, 31), lead us to suggest that UCNs may be in-olved in regulating the endometrial PG pathway. Finally, theecent studies on the potent effects of UCNs on the immuneystem (32), together with the effect of CRF on early maternalolerance (14), lead us to suggest that UCNs may also havemportant local roles on early maternal tolerance.

In conclusion, UCN II and III are expressed by humanndometrium and early decidua. Their different secretoryatterns, together with their selective binding to CRF-R2,uggest that they may have different actions in endometrialhysiology.

Pasquale Florio, M.D., Ph.D.a

Paulo B. Torres, M.D.a

Michela Torricelli, M.D.a

Paolo Toti, M.D.b

Wylie Vale, Ph.D.c

Felice Petraglia, M.D.aa Department of Pediatrics, Obstetrics, and

Reproductive Medicine and b Department of HumanPathology and Oncology, University of Siena, Siena,Italy; and c Clayton Foundation Laboratories forPeptide Biology, The Salk Institute,

La Jolla, California

rtility and Sterility�

EFERENCES1. Florio P, Vale W, Petraglia F. Urocortins in human reproduction.

Peptides 2004;25:1751–7.2. Vaughan J, Donaldson C, Bittencourt J, Perrin MH, Lewis K, Sutton

S, et al. Urocortin, a mammalian neuropeptide related to fish uroten-sin I and to corticotropin-releasing factor. Nature 1995;378:287–92.

3. Lewis K, Li C, Perrin MH, Blount A, Kunitake K, Donaldson C,et al. Identification of urocortin III, an additional member of thecorticotropin-releasing factor (CRF) family with high affinity for theCRF2 receptor. Proc Natl Acad Sci U S A 2001;98:7570–5.

4. Reyes TM, Lewis K, Perrin MH, Kunitake KS, Vaughan J, Arias CA,et al. Urocortin II: a member of the corticotropin-releasing factor(CRF) neuropeptide family that is selectively bound by type 2 CRFreceptors. Proc Natl Acad Sci U S A 2001;98:2843–8.

5. Aguilera G, Nikodemova M, Wynn PC, Catt KJ. Corticotropinreleasing hormone receptors: two decades later. Peptides 2004;25:319 –29.

6. Makrigiannakis A, Zoumakis E, Margioris AN, Theodoropoulos P,Stournaras C, Gravanis A. The corticotropin-releasing hormone(CRH) in normal and tumoral epithelial cells of human endometrium.J Clin Endocrinol Metab 1995;80:185–9.

7. Mastorakos G, Scopa CD, Kao LC, Vryonidou A, Friedman TC,Kattis D, et al. Presence of immunoreactive corticotropin-releasinghormone in human endometrium. J Clin Endocrinol Metab 1996;81:1046–50.

8. Florio P, Arcuri F, Ciarmela P, Runci Y, Romagnoli R, Cintorino M,et al. Identification of urocortin mRNA and peptide in the humanendometrium. J Endocrinol 2002;173:R9–14.

9. Zoumakis E, Chatzaki E, Charalampopoulos I, Margioris AN, An-gelakis E, Koumantakis E, et al. Cycle and age-related changes incorticotropin-releasing hormone levels in human endometrium andovaries. Gynecol Endocrinol 2001;15:98–102.

0. Di Blasio AM, Giraldi FP, Vigano P, Petraglia F, Vignali M, Cav-agnini F. Expression of corticotropin-releasing hormone and its R1receptor in human endometrial stromal cells. J Clin Endocrinol Metab1997;82:1594–7.

1. Karteris E, Papadopoulou N, Grammatopoulos DK, Hillhouse EW.Expression and signalling characteristics of the corticotrophin-releasing hormone receptors during the implantation phase in thehuman endometrium. J Mol Endocrinol 2004;32:21–32.

2. Zoumakis E, Margioris AN, Stournaras C, Dermitzaki E, AngelakisE, Makrigiannakis A, et al. Corticotrophin-releasing hormone (CRH)interacts with inflammatory prostaglandins and interleukins and af-fects the decidualization of human endometrial stroma. Mol HumReprod 2000;6:344–51.

3. Ferrari A, Petraglia F, Gurpide E. Corticotropin releasing factordecidualizes human endometrial stromal cells in vitro. Interactionwith progestin. J Steroid Biochem Mol Biol 1995;54:251–5.

4. Makrigiannakis A, Zoumakis E, Kalantaridou S, Coutifaris C, Mar-gioris AN, Coukos G, et al. Corticotropin-releasing hormone pro-motes blastocyst implantation and early maternal tolerance. Nat Im-munol 2001;2:1018–24.

5. Graziani G, Tentori L, Portarena I, Barbarino M, Tringali G, PozzoliG, et al. CRH inhibits cell growth of human endometrial adenocar-cinoma cells via CRH-receptor 1–mediated activation of cAMP-PKApathway. Endocrinology 2002;143:807–13.

6. Severi FM, Bocchi C, Florio P, Cobellis L, Ignacchiti E, Petraglia F.Transvaginal ultrasonography in women receiving emergency contra-ception. Fertil Steril 2003;79:1074–7.

7. Noyes RW, Hertig AT, Rock J. Dating the endometrial biopsy. Am JObstet Gynecol 1975;122:262–3.

8. Karteris E, Hillhouse EW, Grammatopoulos D. Urocortin II is ex-pressed in human pregnant myometrial cells and regulates myosinlight chain phosphorylation: potential role of the type-2 corticotropin-releasing hormone receptor in the control of myometrial contractility.

Endocrinology 2004;145:890–900.

1769

1

2

2

2

2

2

2

2

2

2

2

3

3

3

1

9. Takahashi K, Totsune K, Murakami O, Saruta M, Nakabayashi M,Suzuki T, et al. Expression of urocortin III/stresscopin in human heartand kidney. J Clin Endocrinol Metab 2004;89:1897–903.

0. Petraglia F, Tabanelli S, Galassi MC, Garuti GC, Mancini AC,Genazzani AR, et al. Human decidua and in vitro decidualized endo-metrial stromal cells at term contain immunoreactive corticotropin-releasing factor (CRF) and CRF messenger ribonucleic acid. J ClinEndocrinol Metab 1992;74:1427–31.

1. Chen A, Vaughan J, Vale WW. Glucocorticoids regulate the expres-sion of the mouse urocortin II gene: a putative connection between thecorticotropin-releasing factor receptor pathways. Mol Endocrinol2003;17:1622–39.

2. Li C, Chen P, Vaughan J, Blount A, Chen A, Jamieson PM, et al.Urocortin III is expressed in pancreatic beta-cells and stimulatesinsulin and glucagon secretion. Endocrinology 2003;144:3216–24.

3. Petraglia F, Florio P, Gallo R, Simoncini T, Saviozzi M, Di Blasio AM,et al. Human placenta and fetal membranes express human urocortinmRNA and peptide. J Clin Endocrinol Metab 1996;81:3807–10.

4. Simoncini T, Apa R, Reis FM, Miceli F, Stomati M, Driul L, et al.Human umbilical vein endothelial cells: a new source and potentialtarget for corticotropin-releasing factor. J Clin Endocrinol Metab1999;84:2802–6.

5. Jain V, Longo M, Ali M, Saade GR, Chwalisz K, Garfield RE.Expression of receptors for corticotropin-releasing factor in the vas-

culature of pregnant rats. J Soc Gynecol Invest 2000;7:153–60.

770 Florio et al. Correspondence

6. Wiley KE, Davenport AP. CRF2 receptors are highly expressed inthe human cardiovascular system and their cognate ligands uro-cortins 2 and 3 are potent vasodilators. Br J Pharmacol 2004;143:508 –14.

7. Brar BK, Jonassen AK, Egorina EM, Chen A, Negro A, Perrin MH,et al. Urocortin-II and urocortin-III are cardioprotective against isch-emia reperfusion injury: an essential endogenous cardioprotectiverole for corticotropin releasing factor receptor type 2 in the murineheart. Endocrinology 2004;145:24–35.

8. Jabbour HN, Sales KJ. Prostaglandin receptor signalling and functionin human endometrial pathology. Trends Endocrinol Metab 2004;15:398–404.

9. Jones RL, Kelly RW, Critchley HO. Chemokine and cyclooxygenase-2expression in human endometrium coincides with leukocyte accumu-lation. Hum Reprod 1997;12:1300–6.

0. Ota H, Igarashi S, Sasaki M, Tanaka T. Distribution of cyclooxygenase-2in eutopic and ectopic endometrium in endometriosis and adenomyosis.Hum Reprod 2001;16:561–6.

1. Fleisher-Berkovich S, Danon A. Effect of corticotropin-releasingfactor on prostaglandin synthesis in endothelial cells and fibroblasts.Endocrinology 1995;136:4068–72.

2. Suda T, Kageyama K, Sakihara S, Nigawara T. Physiological roles ofurocortins, human homologues of fish urotensin I, and their receptors.

Peptides 2004;25:1689–701.

Vol. 86, No. 6, December 2006