Expression and Biological Activity of ParathyroidHormone-Related Peptide in Pregnant Rat UterineArtery: Any Role for 8-Iso-Prostaglandin F2�?

Ferhat Meziani, Angela Tesse, Sandra Welsch, Helene Kremer, Mariette Barthelmebs,Ramaroson Andriantsitohaina, Francis Schneider, and Alexis Gairard

Universite Louis Pasteur-Strasbourg I, Institut Gilbert-Laustriat and Centre National de la Recherche Scientifique (CNRS)Unite Mixte de Recherche (UMR) 7175 (F.M., H.K., A.G.), Departement Pharmacologie et Physicochimie, Faculte dePharmacie, 67401 Illkirch, France; Institut National de la Sante et de la Recherche Medicale (INSERM) UMR 771 (F.M.,A.T., R.A.); CNRS UMR 6214; Faculte de Medecine, Universite d’Angers, 49045 Angers, France; INSERM Unite 727 (S.W.,M.B.), Universite Louis Pasteur (Strasbourg I), 67070 Strasbourg, France; Departement de Reanimation medicale et demedecine hyperbare (F.M.), Centre Hospitalier Universitaire, 49933 Angers, France; and Reanimation Medicale (H.K., F.S.),Hopital de Hautepierre, Hopitaux Universitaires de Strasbourg, 67000 Strasbourg, France

PTHrP is produced in vessels and acts as a local modulator oftone. We recently reported that PTHrP(1–34) is able to inducevasorelaxation in rat uterine arteries, but in pregnancy, thisresponse is blunted and becomes strictly endothelium depen-dent. The present study aimed to get insights into the mech-anisms involved in these changes because the adaptation ofuterine blood flow is essential for fetal development. On d 20of gestation, RT-PCR analysis of uterine arteries showed thatPTH/PTHrP receptor (PTH1R) mRNA expression was de-creased, whereas that of PTHrP mRNA was increased. Thiswas associated with a redistribution of the PTHrP/PTH1Rsystem, with both PTH1R protein and PTHrP peptide becom-ing concentrated in the intimal layer of arteries from preg-nant rats. On the other hand, the blunted vasorelaxation in-

duced by PTHrP(1–34) in uterine arteries from pregnant ratswas specifically restored by indomethacin and a specific cy-clooxygenase-2 inhibitor, NS 398. This was associated with anincrease in cyclooxygenase-2 expression and in 8-iso-prosta-glandin F2� release when uterine arteries from pregnant ratswere exposed to high levels of PTHrP(1–34). Most interest-ingly, 8-iso-prostaglandin F2� itself was able to increasePTHrP expression and reduce PTH1R expression in culturedrat aortic smooth muscle cells. These results suggest a localregulation of uterine artery functions by PTHrP during preg-nancy resulting from PTH1R redistribution. Moreover, theyshed light on a potential role of 8-iso-prostaglandin F2�.(Endocrinology 149: 626–633, 2008)

THE PTHrP WAS FIRST identified as a factor causingmalignant humoral hypercalcemia. In contrast to PTH,

the expression of PTHrP is observed in a wide range oftissues where it is involved in the control of calcium ho-meostasis, cartilage resorption, and cell proliferation/apoptosis and in vasodilatation (1). PTHrP is considered asa regulator of vascular tone through both paracrine andautocrine pathways (2). PTHrP has also been reported to playa key role in the pathophysiology of several diseases (3, 4).In addition, the expression of PTHrP and its cognate recep-tor, the PTH/PTHrP receptor (PTH1R), are known to bedifferentially regulated during development (5) and inflam-matory states (6).

In a previous study, we reported that exogenous PTHrP(1–34) acts on vascular smooth muscle via the cAMP pathwayto promote relaxation in uterine artery from nonpregnant rat,whereas the same agonist induces almost exclusively an en-dothelium-dependent relaxation during pregnancy acting

via nitric oxide and endothelial-derived hyperpolarizing fac-tor pathways (7). The reasons for these changes still remainunknown. Furthermore, we reported that the relaxation inresponse to PTHrP(1–34) is significantly reduced in uterineartery from pregnant rats. Vasoconstrictor metabolites releasedfrom vascular cells, originating namely from cyclooxygenases(COXs) are known to participate in the impairment of agonist-induced relaxation. The COX metabolites involved are not onlythose that can activate the thromboxane A2/endoperoxide re-ceptor such as prostaglandin (PG)E2 or thromboxane A2 butalso 8-iso-PGF2�. On the other hand, changes in the uterinevascular reactivity to PTHrP could also be linked to a changein PTH1R expression. Indeed, PTHrP-induced renal vasodila-tation was shown to closely correlate with PTH1R expressionunder various experimental settings (8, 9).

The aim of this study was to better understand whyPTHrP-induced relaxation of uterine arteries became endo-thelium dependent and was reduced in late pregnant rats. Inthis context, we investigated both 1) the expression of thePTHrP/PTH1R system and 2) the modulation of the vascularresponse to PTHrP during pregnancy by COX-1/COX-2 me-tabolites, with particular interest in 8-iso-PGF2�. The mostoriginal finding of our study is that PTHrP expression isenhanced in late pregnant uterine arteries, although PTH1Rexpression is decreased, but both concentrate to endothelial

First Published Online November 29, 2007Abbreviations: Ang II, Angiotensin II; AoSMC, aortic smooth muscle

cells; COX, cyclooxygenase; PG, prostaglandin; PSS, physiological saltsolution; PTH1R, PTH/PTHrP receptor.Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving theendocrine community.

0013-7227/08/$15.00/0 Endocrinology 149(2):626–633Printed in U.S.A. Copyright © 2008 by The Endocrine Society

doi: 10.1210/en.2007-0568

626

and subendothelial layers of the vascular wall. Moreover,exogenous PTHrP(1–34) is able to overexpress COX-2 in thesame layers and enhance 8-iso-PGF2� release when the pro-stanoid itself amplifies PTHrP expression in cultured vas-cular smooth muscle cells. These data provide novel evi-dence for a possible contribution of the PTHrP/PTH1Rsystem in uterine artery physiology.

Materials and Methods

The experiments were conducted in compliance with statutory re-quirements by accredited research scientists (Authorization No. 67-125from the French Ministry of Agriculture).

Animals

Female Wistar rats (Nico Iffa strain), 12–15 wk of age, supplied byCharles River (L’Arbresle, France) were housed separately in temper-ature- and humidity-controlled quarters with constant 12-h light, 12-hdark cycles and were provided with standard food (A04 diet; Safe,Villemoisson, France) and water ad libitum. The rats were mated, and theday on which the sperm plug was observed was denoted as d 1 ofgestation. Nonpregnant rats (n � 18) or rats on d 20 of pregnancy (n �15) were anesthetized with pentobarbital (60 mg/kg ip; Sanofi, Paris,France) and killed by cervical dislocation. After laparotomy, the uterustogether with both ovaries and uterine arteries were excised and im-mediately placed in cold (4 C) physiological salt solution (PSS; see belowfor composition).

Aortic smooth muscle cells (AoSMC) in culture

Primary cultures of rat AoSMC were obtained by the explant methodas reported previously (10) and used at passages 15–17. AoSMC werecultured in DMEM supplemented with 10% fetal bovine serum, peni-cillin (100 U/ml), and streptomycin (0.1 mg/ml) at 37 C in 10% CO2.Cells at 80% confluence were grown for a 24-h period of quiescence inserum-free medium supplemented with 0.1% BSA, before being exposedfor 8 h to 8-iso-PGF2� (10 �m; Cayman Chemical, SPI-BIO, Montigny leBretonneux, France), or angiotensin II (Ang II, 100 nm; Sigma-Aldrich,Saint Quentin Fallavier, France).

RNA extraction and RT-PCR analysis in arteries and cells

The uterine arteries from pregnant and nonpregnant rats were dis-sected under a magnifier. Endothelial cells were removed from somepreparations as described previously (7), by a bolus injection of 0.3%CHAPS,3-[(3-cholamidopropyl) dimethyl- ammonio]-1-propanesulfon-ate, so that arteries with (E�) and without (E�) endothelium could beanalyzed. Total RNA was rapidly extracted from homogenized uterinearteries with the QIAGEN kit for fibrous tissue (QIAGEN S.A., Courta-boeuf, France) or from cells in culture by TRIzol according to the man-ufacturer’s protocol (Invitrogen, Cergy-Pontoise, France) and quantifiedby absorbance at 260 nm. RT was performed on 5 �g RNA in the presenceof 400 U reverse transcriptase (leukemia virus reverse transcriptase,Fermentas, St. Leon-Rot, Germany). Quantitative real-time PCR analysiswas used to amplify cDNA with the LightCycler-FastStart DNA MasterSYBR Green kit (Roche Diagnostics, Meylan, France) as described pre-viously (11). A standard curve was obtained by serial dilutions of mixedcDNA samples. Amplifications were performed in a 20-�l mix, con-taining 4 mm MgCl2, 0.5 �m of each primer set, reaction buffer (Taq DNApolymerase, dNTP mix, 0.2 �m SYBR Green I dye) and 50 ng cDNAsample. The following sets of primers were used: rat PTH1R, sense5�-GGG CAC AAG AAG TGG ATC AT-3� and antisense 5�-GGC CATGAA GAC GGT GTA GT-3�; rat PTHrP, sense 5�-CAG CCG AAA TCAGAG CTA CC-3� and antisense 5�-CTC CTG TTC TCT GCG TTT CC-3�;and rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH), sense5�-CAT GGA GAA GGC TGG GGC TC-3� and antisense 5�-AAC GGATAC ATT GGG GGT AG-3�. GAPDH was used as a housekeeping geneto normalize the expression of other genes. PCR were run at 95 C for 10min, followed by 45 cycles of 10 sec at 95 C, 5 sec at 60 C, and 16 sec at72 C (PTH1R and PTHrP) or 20 sec at 72 C (GAPDH). As a negativecontrol, cDNA was replaced by PCR-grade water. Each sample was run

twice. Data were quantified with the LightCycler analysis software(Roche Diagnostics) and mRNA signals normalized for GAPDH (PTH1Ror PTHrP/GAPDH mRNA ratio). PCR products were identified on 1.5%of agarose gels containing 0.5 �g/ml ethidium bromide by their ex-pected size of 210 bp (PTH1R), 206 bp (PTHrP), and 415 bp (GAPDH)(not shown).

In vitro vascular reactivity

The uterine arteries from pregnant and nonpregnant rats were dis-sected under a magnifier, and uterine arterial rings were mounted on awire myograph (EMKA Technologies, Paris, France) for the in vitro studyof vascular reactivity exactly as previously described (7). Briefly, ex-periments were performed in prewarmed (37 C), oxygenated (95% O2and 5% CO2) PSS medium of the following composition (in mm): 119NaCl; 4.7 KCl; 14.9 NaHCO3; 1.2 MgSO47H2O; 2.5 CaCl2; 1.18 KH2PO4;5.5 glucose. The force of vessels was recorded with an isometric forcetransducer, and data acquisition was performed using Datanalyst ver-sion 1.4 and playback software (EMKA Technologies). After an equil-ibration period under an optimal passive tension for at least 20 min, thephenylephrine concentration giving 80% of the maximal contraction(PE80) was determined on each vessel and chosen for preconstriction inthe following relaxation experiments. All the studies were performed inarteries with functional endothelium as assessed by the ability of ace-tylcholine to induce relaxation. The effects of PTHrP on the vessels werequantified as the percentage of relaxation in preparations previouslycontracted with phenylephrine. Concentration-response curves wereconstructed by cumulative application of PTHrP(1–34) (1–100 nm; Neo-system, Strasbourg, France) in the absence or the presence of COXinhibitors indomethacin (10 �m/liter), the selective COX-2 inhibitorN-(2-cyclohexyloxy-4-nitrophenyl) methanesulfonamide (NS-398, 10�m/liter), and the selective COX-1 inhibitor valeryl salicylate (3 mm/liter). The inhibitors were added in the bath 30 min before addition ofPTHrP(1–34). The drugs used in the experiments were purchased fromSigma Chemical Co. (St. Louis, MO) unless otherwise stated.

In vitro prostanoids release by uterine arteries

Uterine arteries were dissected from pregnant and nonpregnant ratsas described above, put in plate wells with 5 ml PSS, and incubated at37 C for 4 h in the absence or the presence of 100 nm PTHrP(1–34).Thromboxane B2, 8-iso-PGF2�, and PGE2 metabolites were measured inthe incubation medium by enzymatic immunoassay kits (CaymanChemical).

Staining and imaging by confocal microscopy

Freshly dissected uterine arteries from pregnant and nonpregnantrats were used for the double immunostaining with endothelial anti-PECAM-1 (C-20) antibody (1:50; Santa Cruz Biotechnology, Heidelberg,Germany) and either anti-PTH1R or anti-PTHrP antibodies. Vesselswere frozen and cut in 10-�m sections, fixed (5 min at �20 C) in 100%methanol, and treated (1 h at room temperature) in blocking buffer (5%nonfat dry milk in PBS and 0.05% Tween 20) to block nonspecific bindingsites. Fixed and blocked tissue sections were coincubated (overnight at4 C) with a polyclonal goat anti-PECAM-1 antibody and a monoclonalmouse anti-PTH1R antibody (1:50; Acris, Hiddenhausen, Germany) ora rabbit anti-PTHrP(1–34) antibody (1:20; Bachem, Voisins-le-Breton-neux, France). After three washes, tissue sections were incubated (1 h atroom temperature) with Alexa fluor-488-labeled antigoat IgG (1:100) tostain endothelial PECAM-1, Alexa fluor-546-labeled antimouse IgG (1:100) to stain PTH1R or Alexa fluor-546-labeled antirabbit IgG (1:100) tostain PTHrP (all from Invitrogen Molecular Probes, Leiden, The Neth-erlands). After final washes, vessel sections were mounted on glassslides.

In another set of experiments, freshly dissected uterine arteries frompregnant and nonpregnant rats exposed to vehicle or 100 nm PTHrP(1–34) for 4 h were used. Vessels were frozen, cut in 10-�m sections, andprocessed as described above. Fixed and blocked tissue sections wereincubated (overnight at 4 C) for the double immunostaining with apolyclonal rabbit anti-COX-2 antibody (1:10; Transduction Laboratories,Heidelberg, Germany) and a polyclonal goat anti-PECAM-1 antibody tostain COX-2 and the endothelium, respectively. After three washes,

Meziani et al. • PTHrP Biological Activity in Rat Uterine Artery Endocrinology, February 2008, 149(2):626–633 627

tissue sections were incubated (1 h at room temperature) with Alexafluor-546-labeled antirabbit IgG (1:100; Invitrogen Molecular Probes)and Alexa fluor-488-labeled antigoat IgG (1:100; Invitrogen Molecularprobes).

Slides were examined with an Olympus light microscope FluoviewFU 300 Laser Scanning Confocal Imaging System (Paris, France),equipped with an argon ion laser (EM 488 nm) and helium-neon ion laser(excitation wavelength 543 nm). Pictures were taken with a �40 objec-tive (oil immersion). The laser was adjusted in the green/red fluores-cence mode, which yielded two excitation/emission wavelengths at488/520 nm for Alexa 488 and at 543/572 nm for Alexa 546. Z-series werecollected in 1-�m steps, and final images were obtained after stacking.Green and red images were obtained from two separated channels anda third superposition of green and red pictures. Red staining corre-sponded to COX-2, PTHrP, or PTH1R detections, whereas green stainingcorresponded to endothelial PECAM-1 detection in the same section.

Data analysis

One-way ANOVA, Kruskal-Wallis, and Mann-Whitney U tests ortwo-way ANOVA for repeated measurements and subsequent post hoctests were performed with the Statview version 5.0 software (SAS In-stitute, Cary, NC). P � 0.05 was considered statistically significant. Allvalues are presented as mean � sem for n experiments, and n representsthe number of animals.

ResultsPregnancy decreases PTH1R mRNA but increases PTHrPmRNA in uterine arteries

The expression of PTH1R and PTHrP were analyzed byquantitative real-time RT-PCR on uterine arteries with(E�) and without (E�) endothelium, freshly dissectedfrom pregnant and nonpregnant rats. The expression ofPTH1R mRNA is reduced by 50% in the E� arteries frompregnant rats (Fig. 1A) compared with those from non-pregnant rats. In endothelium-denuded (E�) pregnant ar-teries, this decrease becomes even more pronounced (75%)(Fig. 1B). In contrast, PTHrP mRNA expression seems tobe increased overall by a factor of three in uterine E�arteries from pregnant rats (Fig. 1C). However, in endo-thelium-denuded pregnant arteries, PTHrP expression isno longer significantly increased in the remaining medial

layer (Fig. 1D). These results show that changes in PTHrPand PTH1R mRNA expression are heterogeneous withinthe vascular wall of pregnant arteries.

Uterine artery expression of PTH1R and PTHrP is shiftedfrom medial layer to intimal layer during pregnancy

To analyze the influence of pregnancy on the expressionand distribution of PTH1R and PTHrP in the vascular wallof uterine arteries, double staining was performed forendothelial PECAM-1 and either PTH1R protein or PTHrPpeptide. Representative immunographs of uterine arteriesfrom one pregnant and one nonpregnant rat are given onFig. 2, with similar scale showing the increase in size ofarteries from pregnant rats. The most remarkable obser-vation is the changes in the distribution of PTH1R andPTHrP in arteries from pregnant and nonpregnant rats.Indeed, PTH1R and PTHrP immunostaining is uniformlydistributed in the vascular wall of uterine arteries fromnonpregnant rats (Fig. 2, E and K), whereas only somePTH1R colocalization with endothelial PECAM-1 is as-sessed by the yellow staining in the merged figure (Fig.2F). In contrast, in pregnant uterine arteries, marked la-beling for PTH1R or PTHrP is found in the intima (i.e.endothelial and subendothelial layers), whereas labeling isvery weak in the medial layer of the arteries (Fig. 2, B andH). Colocalization of PTH1R or PTHrP and endothelialPECAM-1 is observed in endothelial layer of pregnantarteries (Fig. 2, C and I, respectively). Negative controls,obtained by incubation of arteries only with the secondaryfluorescence-labeled IgG, display no staining (not shown).Taken together, these results show that pregnancy is as-sociated with a redistribution of PTH1R protein andPTHrP peptide within the vascular wall, staining beingmarkedly increased in endothelial and subendothelial lay-ers as demonstrated by the colocalized staining of PTH1Ror PTHrP with the endothelial protein PECAM-1.

P NP0.0

0.5

1.0

1.5

PT

H1R

mR

NA

(GA

PD

H r

atio

)

P NP0

1

2

3

4

5

PT

HrP

mR

NA

(GA

PD

H r

atio

)

P NP0.0

0.5

1.0

1.5

PT

H1R

mR

NA

(GA

PD

H r

atio

)

P NP0

1

2

3

4

5

PT

HrP

mR

NA

(GA

PD

H r

atio

)

A

B

C

D

**

*

(E+)

(E-)

(E+)

(E-)

P NP0.0

0.5

1.0

1.5

PT

H1R

mR

NA

(GA

PD

H r

atio

)

P NP0

1

2

3

4

5

PT

HrP

mR

NA

(GA

PD

H r

atio

)

P NP0.0

0.5

1.0

1.5

PT

H1R

mR

NA

(GA

PD

H r

atio

)

P NP0

1

2

3

4

5

PT

HrP

mR

NA

(GA

PD

H r

atio

)

**

*

(E+)

(E-)

(E+)

(E-)

FIG. 1. PTH1R and PTHrP mRNA expression in uter-ine arteries from pregnant and nonpregnant rats.Quantitative real-time RT-PCR analysis was per-formed for the evaluation of PTH1R (A and B) andPTHrP (C and D) mRNA in uterine arteries freshlyprepared from pregnant (P) and nonpregnant (NP) ratswith (E�) and without (E�) endothelium. Expressionof target transcripts was compared with the expressionof GAPDH used as a housekeeping gene. Normalizedratios were calculated (PTH1R/GAPDH mRNA andPTHrP/GAPDH mRNA) and set at 1 for nonpregnantrats. Note that PTHrP expression is spontaneouslyup-regulated and PTH1R expression down-regulatedin pregnant uterine arteries. Of particular interest, inthe E� uterine arteries from pregnant rats, the down-regulation of PTH1R expression becomes more pro-nounced, whereas PTHrP expression is no longer sig-nificantly increased. Results are given as mean � SEMfor three to six independent experiments.

628 Endocrinology, February 2008, 149(2):626–633 Meziani et al. • PTHrP Biological Activity in Rat Uterine Artery

COX-2 inhibitors enhance relaxation to PTHrP in arteriesfrom pregnant rat

As previously reported (7), PTHrP(1–34) produces a con-centration-dependent relaxation in uterine arteries that wasblunted in arteries from pregnant rats (Fig. 3, A and B).Indeed, the maximal relaxation to PTHrP(1–34) is then de-creased by about half (28 � 6% in pregnant vs. 46 � 10% innonpregnant; n � 6; P � 0.05). To investigate whether COXderivatives could explain this difference, we evaluatedPTHrP(1–34)-induced uterine relaxation in the presence ofCOX inhibitors with specificity for COX-1 (valeryl salicylate)or COX-2 (NS-398) as well as in the presence of a nonspecificCOX inhibitor (indomethacin).

Indomethacin does not modify the relaxation induced byPTHrP(1–34) in arteries from nonpregnant rats (Fig. 3B) butsignificantly enhances the response in arteries from pregnantrats (Fig. 3A). The selective COX-1 inhibitor, valeryl salicy-late, does not affect the relaxation to PTHrP(1–34) whateverthe origin of the arteries is. In contrast, the selective COX-2inhibitor, NS-398, markedly enhances PTHrP-induced vaso-dilatation in uterine arteries from pregnant rats (Fig. 3A).Taken together, these results show that vasoconstrictor me-

tabolite(s) derived from COX-2 partially inhibit PTHrP-induced uterine relaxation in pregnant rats only.

PTHrP(1–34) enhances COX-2 expression in uterine arteriesfrom pregnant rats

To assess whether COX-2 expression is changed duringpregnancy and to localize the protein within the vascularwall, we performed double immunohistochemical detectionof both endothelial PECAM-1 and COX-2 protein. No stain-ing or weak staining of COX-2 is found in naive vessels fromnonpregnant rats (Fig. 4H), whereas weak staining is foundon the adventitial layer of vessels from pregnant rats (Fig.4B). However, marked COX-2 labeling is observed in theendothelial, subendothelial, and adventitial layers of uterinearteries from pregnant rats exposed for 4 h to 100 nmPTHrP(1–34) (Fig. 4E). Colocalization with endothelialPECAM-1 is assessed by the yellow staining in the mergedfigure (Fig. 4F). In contrast, no labeling or only weak labelingof COX-2 is found after PTHrP treatment of arteries fromnonpregnant rats (Fig. 4K). Negative controls, obtained byincubation of arteries only with the secondary fluorescence-labeled IgG, display no staining (not shown).

PECAM-1

PTH1R

Merge

PTH1R PTHrP

P NPP NP

PECAM-1

PTHrP

Merge

A D

B E

C F

G

I

J

L

H K

PECAM-1

PTH1R

Merge

PTH1R PTHrP

P NPP NP

PECAM-1

PTHrP

Merge

A D

B E

C F

G

I

J

L

H K

PTH1R PTHrP

P NPP NP

PECAM-1

PTHrP

Merge

A

B E

C

GG

II

JJ

LL

HH KK

F

D

FIG. 2. Immunohistochemical staining for PTH1R and PTHrP in pregnant and nonpregnant uterine arteries. Representative pictures are shownof immunohistochemical red staining for PTH1R in uterine arteries from pregnant (P) (B) or nonpregnant rats (NP) (E) and green staining forendothelial PECAM-1 in uterine arteries from pregnant (A) and nonpregnant rats (D). Colocalization of PTH1R and PECAM-1 (yellow staining)is pronounced in vessels from pregnant rats (C) but weak in vessels from nonpregnant rats (F). Representative pictures are shown ofimmunohistochemical red staining for PTHrP in uterine arteries from pregnant (P) (H) or nonpregnant rats (NP) (K) and green staining forendothelial PECAM-1 in uterine arteries from pregnant (G) and nonpregnant rats (J). Colocalization of PTHrP and PECAM-1 (yellow staining)is pronounced in vessels from pregnant rats (I) but very weak in vessels from nonpregnant rats (L). Note that the diffuse staining present innonpregnant rats for both PTH1R and PTHrP becomes restricted to the intimal layer in pregnant rats. Bars, 50 �m.

Meziani et al. • PTHrP Biological Activity in Rat Uterine Artery Endocrinology, February 2008, 149(2):626–633 629

PTHrP(1–34) stimulates the release of 8-iso-PGF2� only inuterine arteries from pregnant rats

To further document the identity of COX-2 vasoconstrictorproducts released by uterine arteries, we measure throm-boxane B2, the stable metabolite of thromboxane A2, as wellas PGE2 and 8-iso-PGF2� in arteries freshly dissected fromboth pregnant and nonpregnant rats. As shown on Fig. 5, thebasal release of PGE2 is 2.5-fold higher in arteries from preg-nant rats (Fig. 5A), whereas there is no difference in the basalrelease of 8-iso-PGF2� (Fig. 5B) or thromboxane B2 (Fig. 5C).In contrast, when arteries are in vitro exposed for 4 h to 100nm PTHrP(1–34), no change in the release of prostanoids isobserved in arteries from nonpregnant rats, whereas therelease of 8-iso-PGF2� is enhanced by a factor of two in uterinearteries from pregnant rats (Fig. 5B).

8-Iso-PGF2� regulates PTHrP and PTH1R mRNAexpression in AoSMC

Because Ang II, as well as other vasoconstrictors, havebeen reported to modify PTHrP and PTH1R expressions invascular smooth muscle cells (12, 13), we tested whether8-iso-PGF2� may be an additional factor able to modulatePTHrP and/or PTH1R expression in vessels. For this study,we used AoSMC that were incubated for 8 h with 8-iso-PGF2�

(10 �m) or Ang II (100 nm), which is used here as a referencedrug. As expected, the RT-PCR analysis shows that Ang IIinduces a 2-fold increase in PTHrP mRNA expression (Fig.6B) and markedly down-regulates the expression of PTH1RmRNA (Fig. 6A). More interestingly, incubation of the cellsin the presence of 8-iso-PGF2� elicits a quite similar effect,inducing overexpression of PTHrP and down-regulation ofPTH1R expression (Fig. 6, B and A, respectively). These re-sults establish now that 8-iso-PGF2� belongs to the extensivefamily of factors regulating the vascular expression of thePTHrP/PTH1R system.

Discussion

We recently reported that PTHrP(1–34) is able to inducevasorelaxation in rat uterine arteries, but this response wasblunted in arteries from pregnant rats and became strictlyendothelium dependent. The present study was performedto better understand the mechanisms involved in PTHrP(1–34)-induced uterine vasodilatation changes during preg-nancy. The main findings were that the overall expression ofPTH1R was decreased, whereas that of PTHrP was increasedin uterine arteries from pregnant rats. This was associatedwith a redistribution of the PTHrP/PTH1R system in thesevessels. Thus, both PTH1R protein and PTHrP peptide ex-pression became concentrated in the intimal layer of arteriesfrom pregnant rats. These data were in agreement with RT-PCR results in E� and E� arteries, suggesting that PTH1Rand PTHrP mRNA expression are high in the endotheliumof pregnant arteries. Besides, relaxation in response toPTHrP(1–34) was restored in the presence of COX-2 inhib-itors in uterine arteries from pregnant rats. This was asso-ciated with an increased COX-2 expression within the vesselwall and an enhanced 8-iso-PGF2� release. Most interestingly,8-iso-PGF2� was able to reduce PTH1R expression and in-creased PTHrP expression in cultured vascular smooth mus-cle cells.

PTH1R and PTHrP are widely expressed in the cardio-vascular system, including the heart, vascular smooth mus-cle, and endothelial cells. The effectiveness of the endoge-nous PTHrP/PTH1R system as a modulator of regionalhemodynamics mainly depends on local expression of bothPTH1R and PTHrP (2, 9). In the present study, we clearlydocument the presence of a local PTHrP/PTH1R system inuterine arteries. Indeed, PTH1R and PTHrP transcripts aswell as immunoreactive proteins were detected. Of interest,the vascular uterine PTHrP/PTH1R system adapts duringpregnancy with a down-regulation of PTH1R and an overalloverexpression of PTHrP as shown by our results from RT-PCR studies. However, PTHrP-induced vasorelaxation per-sists in arteries from pregnant rats, although at a lower level,and becomes strongly endothelium dependent as we re-

Pregnant

Non Pregnant

-10 -9 -8 -7 -6

0

25

50

75

log [PTHrP] (M)

Rel

axat

ion

%

-10 -9 -8 -7 -6

0

25

50

75

**

log [PTHrP] (M)

Rel

axat

ion

%

Pregnant

Non Pregnant

A

-10 -9 -8 -7 -6

0

25

50

75

log [PTHrP] (M)

Rel

axat

ion

%

-10 -9 -8 -7 -6

0

25

50

75

**

log [PTHrP] (M)

Rel

axat

ion

%

B

FIG. 3. Effects of COX inhibitors on PTHrP(1–34)-induced relaxationin pregnant and nonpregnant uterine arteries. A, Concentration-response curves on uterine arteries from pregnant rats (n � 6) ex-posed to increasing concentrations of PTHrP(1–34) in the absence (�)or in the presence of indomethacin (f), NS-398 (�), or valeryl salic-ylate (Œ). B, Concentration-response curves on uterine arteries fromnonpregnant rats (n � 6) exposed to increasing concentrations ofPTHrP in the absence (�) or in the presence of indomethacin (f),NS-398 (�), or valeryl salicylate (Œ). Please note that PTHrP(1–34)produces a vasorelaxant effect that is more pronounced in vesselsfrom pregnant rats when indomethacin or a COX-2 inhibitor is addedbut not when the study is performed in the presence of valeryl sa-licylate, a COX-1 inhibitor. Results are given as means � SEM andwere compared by a two-way ANOVA on repeated measurements,followed by Tukey test for multiple comparisons.*, P � 0.05 vs. control(absence of inhibitor).

630 Endocrinology, February 2008, 149(2):626–633 Meziani et al. • PTHrP Biological Activity in Rat Uterine Artery

ported previously (7). One of the novel findings of thepresent study is the redistribution of the PTHrP/PTH1Rsystem within the uterine arteries during pregnancy thatmay explain the obligatory role of the endothelium in theresponse to PTHrP in vessels from pregnant rats. Indeed,PTH1R mRNA expression is strongly reduced in endothe-lium-denuded uterine arteries from pregnant rats, and im-munostaining is almost absent in the medial layer of preg-nant arteries, whereas strong expression of PTH1R protein isfound in the endothelium. To our knowledge, no such vas-cular redistribution of PTH1R has been described before.Taken together, our results suggest that PTHrP can be con-sidered as one of the factors regulating uterine hemodynam-ics, even during pregnancy, inasmuch as the local expressionof PTHrP itself is increased in the endothelium.

The blood supply to the uterus is provided in major partby the uterine arteries. During pregnancy, they are trans-formed into large dilated vessels undergoing dramatic struc-tural changes in their wall to maximize the delivery of ma-ternal blood to the intervillous space. Their arterial lumenbecomes wider and responsiveness to vasoconstrictors isreduced (14). Hyporesponsiveness to Ang II is at least par-tially mediated by prostacyclin, the potent vasodilator whosesynthesis is enhanced during pregnancy (15). PTHrP mayalso play a role in conjunction with other well-known factorssuch as NO (16, 17) and PGI2 (18). Direct impairment ofuterine blood flow results in fetoplacental abnormalities (19).

The present results highlight the interplay between PTHrPand COX-2 in uterine arteries from pregnant rats. To someextent, COX-2 derivatives blunt PTHrP-induced uterine re-

laxation as shown by the increased response in the presenceof NS-398 or indomethacin. It is, however, not clear whatkind of COX-2 derivative is involved in this blunting. PGE2

may account for this blunting, because it is the only onewhose release is enhanced in naive pregnant uterine arteries.On the other hand, 8-iso-PGF2� constitutes an alternativepossibility because it is the only one whose release is in-creased in uterine arteries after induction of COX-2 expres-sion. Overexpression of COX-2 is specifically observed in thepresent study in pregnant uterine arteries exposed for aperiod of 4 h to high concentration of PTHrP(1–34). Of in-terest, COX-2 expression is then restricted to the endothelialand subendothelial layer where the PTHrP/PTH1R systemis expressed in the pregnant uterine arteries. We cannotexclude that such an effect occurs in vitro during the courseof the construction of the concentration-response curve toPTHrP(1–34) in the uterine arteries. 8-iso-PGF2� can thencontribute to blunting PTHrP-induced vasorelaxation byeliciting vasoconstrictor and functional antagonism. How-ever, 8-iso-PGF2� is probably not responsible for the redis-tribution of the PTH1R during normal pregnancy. Indeed,COX-2 is not expressed in naive pregnant arteries, and 8-iso-PGF2� release is not enhanced in the same arteries. Becausebasal PGE2 release was increased in pregnant arteries, welooked at the effects of PGE2 on the expression of the PTHrP/PTH1R system in AoSMC but found it inactive (Barthelmebs,M., personal results). The mechanism of the PTHrP/PTH1Rredistribution in uterine arteries during pregnancy remainstherefore unknown and needs further investigation, as is also

P-vehicle P-PTHrP NP-vehicle NP-PTHrP

PECAM-1

COX-2

Merge

A D

B E

C F

G J

H K

I L

P-vehicle P-PTHrP NP-vehicle NP-PTHrP

PECAM-1

COX-2

Merge

A D

B E

C F

G J

H K

I L

FIG. 4. PTHrP(1–34)-enhanced COX-2 expression in uterine arteries from pregnant rats. A, D, G, and J, Representative pictures of greenimmunohistochemical staining for endothelial PECAM-1 in vessels from pregnant (A and D) or nonpregnant rats (G and J); B and H,representative pictures of red immunohistochemical staining for COX-2 in uterine arteries from pregnant (B) and nonpregnant rats (H) treatedwith vehicle; E and K, representative pictures of immunohistochemical staining for COX-2 after exposure of uterine arteries from pregnant(E) or nonpregnant rats (K) to PTHrP(1–34) (100 nM for 4 h). F, Note that colocalization of COX-2 and PECAM-1 (yellow staining) in vesselsfrom pregnant rats incubated with PTHrP(1–34) (100 nM for 4 h). Bars, 50 �m. Note also that the COX-2 labeling is localized in endothelialand subendothelial layers of vessels from pregnant rats.

Meziani et al. • PTHrP Biological Activity in Rat Uterine Artery Endocrinology, February 2008, 149(2):626–633 631

the case for the specific effect of PTHrP on COX-2 overex-pression in pregnant uterine arteries.

Although the mechanism of this particular effect of PTHrPremains to be determined, its physiopathological signifi-cance may be of particular value. PTHrP is considered as amember of the cascade of cytokines induced during the in-flammatory response (6). PTHrP expression is increased in anumber of vascular diseases characterized by inflammatory

processes and oxidative stress, such as atherosclerotic lesions(3, 20), hypertension (2), and ischemic cardiac injury (21). Thepossibility of COX-2 induction by high local concentrationsof PTHrP adds further support for an active contribution ofthis peptide in these inflammatory processes. Moreover, inthe same experimental settings, PTH1R expression is usuallyfound to be down-regulated. Our data now add 8-iso-PGF2�

as a possible factor responsible for both PTHrP up-regulationand PTH1R down-regulation, a profile shared with Ang II(13). In this context, it is interesting to note that AngII-induced PTHrP overexpression in AoSMC was inhibited byindomethacin (12). During normal rat pregnancy, deleteriousvascular effects of 8-iso-PGF2� are probably blunted by en-dothelial NO (7). 8-Iso-PGF2� is considered as a valuablemarker for vascular oxidative stress (22–24).

Furthermore, it was negatively implicated in trophoblastinvasion where it reduces metalloproteinase activity (25).Besides, preeclampsia has been associated with increasedoxidative stress and subsequent increase in 8-iso-PGF2� levelsin fetoplacental membranes and amniotic fluids (26). In arecent review, Maioli et al. (27) suggested that reduced localproduction of PTHrP could be a major causative factor inpreeclampsia for defective placentation. Our present datashed a different light on the possible implication of PTHrP,suggesting that overexpression of PTHrP during patholog-ical states might be a harmful factor because it can furtherincrease vascular 8-iso-PGF2� release and local oxidativestress. Of interest, preeclampsia occurred in a pregnantwoman with a PTHrP-secreting pancreatic neuroendocrinetumor and high circulating levels of PTHrP (28).

In conclusion, the present results suggest that PTHrP par-ticipates in the regulation of local uterine hemodynamicsduring gestation. Besides, the present study points out a localfeedback control loop involving the interplay between

P

P-PTHrP NP

NP-PTHrP

0

10

20

30

pg/m

g W

ds

Thromboxane-B2

C

P

P-PTHrP NP

NP-PTHrP

0

1

2

3 *

pg/m

g W

dsProstaglandin E2

P

P-PTHrP NP

NP-PTHrP

0.0

2.5

5.0

7.5

10.0 #

pg/m

g W

ds

8-iso Prostaglandin F2αααα

A

B

FIG. 5. Effects of pregnancy and PTHrP(1–34) on prostanoids pro-duction by pregnant and nonpregnant uterine arteries. Arachidonicacid derivatives related to COX-2 pathway released from pregnant (P)or nonpregnant (NP) uterine arteries, exposed or not to PTHrP(1–34)(100 nM/liter for 4 h), was assessed. A, Basal PGE2 release is higherfrom pregnant vs. nonpregnant arteries (#, P � 0.05) but is not mod-ified by exposure to PTHrP(1–34); B, basal 8-iso-PGF2� release is notdifferent between pregnant and nonpregnant arteries but is enhancedfrom pregnant arteries exposed to PTHrP(1–34) (*, P � 0.05); C,thromboxane B2 release is no different among the four groups. Resultsare given as means � SEM (n � 3).

Contro

l

8-iso

P 8h

AngII

8h

PT

HrP

mR

NA

(GA

PD

H r

atio

)

0

1

2

3

4

5

Contro

l

8-iso

P 8h

AngII

8h

PT

H1R

mR

NA

(GA

PD

H r

atio

)

0.0

0.5

1.0

1.5

*#

A

*

B

#

Contro

l

8-iso

P 8h

AngII

8h

PT

HrP

mR

NA

(GA

PD

H r

atio

)

0

1

2

3

4

5

Contro

l

8-iso

P 8h

AngII

8h

PT

H1R

mR

NA

(GA

PD

H r

atio

)

0.0

0.5

1.0

1.5

*#

* #

FIG. 6. Modulation by 8-iso-PGF2� of PTHrP and PTH1R mRNA ex-pression in rat aortic vascular smooth muscle cells. Quantitativereal-time RT-PCR analysis was performed for the evaluation ofPTH1R (A) and PTHrP (B) mRNA in aortic vascular smooth musclecells, under control conditions and after incubation for 8 h with 8-iso-PGF2� (10 �M) or Ang II (100 nM). Expression of target transcripts wascompared with the expression of GAPDH used as a housekeepinggene. Normalized ratios were calculated (PTHrP/GAPDH mRNA andPTH1R/GAPDH mRNA) and set at 1 for control. Of note, 8-iso-PGF2�

induces down-regulation of PTH1R mRNA and up-regulation ofPTHrP mRNA (*, P � 0.05) exactly as Ang II does (#, P � 0.05).Results are given as mean � SEM for three to six independent exper-iments.

632 Endocrinology, February 2008, 149(2):626–633 Meziani et al. • PTHrP Biological Activity in Rat Uterine Artery

PTHrP/PTH1R and 8-iso-PGF2� release from COX-2. Be-cause the above pathway could have implications in utero-placental perfusion abnormalities in pathological pregnancy,the quantitative importance of PTHrP in maternal physiol-ogy needs to be further clarified.

Acknowledgments

We appreciate the valuable technical assistance of J. F. Poirier and V.Roques (Universite Louis Pasteur de Strasbourg), L. Preisser fromIFR132, M. H. Guilleux from Service Commun de Cytometrie etd’Analyses Nucleotidiques, and R. Mallet of the Service commund’imageries et d’analyses microscopiques (Univesite d’Angers). Finally,we thank A. Bernard for careful review of the English editing of themanuscript.

Received April 30, 2007. Accepted November 20, 2007.Address all correspondence and requests for reprints to: Pr. Alexis

Gairard, Institut Gilbert-Laustriat, Centre National de la RechercheScientifique Unite Mixte de Recherche 7175, Faculte de Pharmacie, 74,route du Rhin, 67401 Illkirch, France. E-mail: alexis.gairard@pharma.u-strasbg.fr.

This work was supported by the Medical School of Louis PasteurUniversity from Strasbourg (Grant 2002-5196) for financial support.

Disclosure Statement: The authors have nothing to disclose.

References

1. Clemens TL, Cormier S, Eichinger A, Endlich K, Fiaschi-Taesch N, FischerE, Friedman PA, Karaplis AC, Massfelder T, Rossert J, Schluter KD, SilveC, Stewart AF, Takane K, Helwig JJ 2001 Parathyroid hormone-related proteinand its receptors: nuclear functions and roles in the renal and cardiovascularsystems, the placental trophoblasts and the pancreatic islets. Br J Pharmacol134:1113–1136

2. Massfelder T, Helwig JJ 1999 Parathyroid hormone-related protein in car-diovascular development and blood pressure regulation. Endocrinology 140:1507–1510

3. Martin-Ventura J, Ortego M, Esbrit P, Hernandez-Presa M, Ortega L, EgidoJ 2003 Possible role of parathyroid hormone-related protein as a proinflam-matory cytokine in atherosclerosis. Stroke 34:1783–1789

4. Godler DE, Stein AN, Bakharevski O, Lindsay MM, Ryan PF 2005 Parathy-roid hormone-related peptide expression in rat collagen-induced arthritis.Rheumatology (Oxford) 44:1122–1131

5. Ferguson 2nd JE, Seaner RM, Bruns DE, Iezzoni JC, Bruns ME 1998 Expres-sion and specific immunolocalization of the human parathyroid hormone/parathyroid hormone-related protein receptor in the uteroplacental unit. Am JObstet Gynecol 179:321–329

6. Funk JL 2001 A role for parathyroid hormone-related protein in the patho-genesis of inflammatory/autoimmune diseases. Int Immunopharmacol1:1101–1121

7. Meziani F, Van Overloop B, Schneider F, Gairard A 2005 Parathyroid hor-mone-related protein-induced relaxation of rat uterine arteries: influence of theendothelium during gestation. J Soc Gynecol Investig 12:14–19

8. Fritsch S, Lindner V, Welsch S, Massfelder T, Grima M, Rothhut S, Bar-thelmebs M, Helwig JJ 2004 Intravenous delivery of PTH/PTHrP type 1receptor cDNA to rats decreases heart rate, blood pressure, renal tone, reninangiotensin system, and stress-induced cardiovascular responses. J Am SocNephrol 15:2588–2600

9. Welsch S, Schordan E, Coquard C, Massfelder T, Fiaschi-Taesch N, Helwig

JJ, Barthelmebs M 2006 Abnormal renovascular parathyroid hormone-1 re-ceptor in hypertension: primary defect or secondary to angiotensin II type 1receptor activation? Endocrinology 147:4384–4391

10. Endlich N, Endlich K, Taesch N, Helwig JJ 2000 Culture of vascular smoothmuscle cells from small arteries of the rat kidney. Kidney Int 57:2468–2475

11. Schordan E, Welsch S, Rothhut S, Lambert A, Barthelmebs M, Helwig JJ,Massfelder T 2004 Role of parathyroid hormone-related protein in the regu-lation of stretch-induced renal vascular smooth muscle cell proliferation. J AmSoc Nephrol 15:3016–3025

12. Pirola CJ, Wang HM, Kamyar A, Wu S, Enomoto H, Sharifi B, Forrester JS,Clemens TL, Fagin JA 1993 Angiotensin II regulates parathyroid hormone-related protein expression in cultured rat aortic smooth muscle cells throughtranscriptional and post-transcriptional mechanisms. J Biol Chem 268:1987–1994

13. Okano K, Wu S, Huang X, Pirola CJ, Juppner H, Abou-Samra AB, Segre GV,Iwasaki K, Fagin JA, Clemens TL 1994 Parathyroid hormone (PTH)/PTH-related protein (PTHrP) receptor and its messenger ribonucleic acid in rataortic vascular smooth muscle cells and UMR osteoblast-like cells: cell-specificregulation by angiotensin-II and PTHrP. Endocrinology 135:1093–1099

14. Espinoza J, Romero R, Mee Kim Y, Kusanovic JP, Hassan S, Erez O, GotschF, Than NG, Papp Z, Jai Kim C 2006 Normal and abnormal transformationof the spiral arteries during pregnancy. J Perinat Med 34:447–458

15. Habermehl DA, Janowiak MA, Vagnoni KE, Bird IM, Magness RR 2000Endothelial vasodilator production by uterine and systemic arteries. IV. Cy-clooxygenase isoform expression during the ovarian cycle and pregnancy insheep. Biol Reprod 62:781–788

16. Nelson SH, Steinsland OS, Wang Y, Yallampalli C, Dong YL, Sanchez JM2000 Increased nitric oxide synthase activity and expression in the humanuterine artery during pregnancy. Circ Res 87:406–411

17. Bird IM, Sullivan JA, Di T, Cale JM, Zhang L, Zheng J, Magness RR 2000Pregnancy-dependent changes in cell signaling underlie changes in differentialcontrol of vasodilator production in uterine artery endothelial cells. Endocri-nology 141:1107–1117

18. Poston L, McCarthy AL, Ritter JM 1995 Control of vascular resistance in thematernal and feto-placental arterial beds. Pharmacol Ther 65:215–239

19. Svedas E, Nisell H, Vanwijk MJ, Nikas Y, Kublickiene KR 2002 Endothelialdysfunction in uterine circulation in preeclampsia: can estrogens improve it?Am J Obstet Gynecol 187:1608–1616

20. Ishikawa M, Akishita M, Kozaki K, Toba K, Namiki A, Yamaguchi T, OrimoH, Ouchi Y 2000 Expression of parathyroid hormone-related protein in humanand experimental atherosclerotic lesions: functional role in arterial intimalthickening. Atherosclerosis 152:97–105

21. Ross G, Schluter KD 2005 Cardiac-specific effects of parathyroid hormone-related peptide: modification by aging and hypertension. Cardiovasc Res66:334–344

22. Morrow JD, Roberts LJ 2002 The isoprostanes: their role as an index of oxidantstress status in human pulmonary disease. Am J Respir Crit Care Med 166:S25–S30

23. Habib A, Badr KF 2004 Molecular pharmacology of isoprostanes in vascularsmooth muscle. Chem Phys Lipids 128:69–73

24. Cracowski JL, Ormezzano O 2004 Isoprostanes, emerging biomarkers andpotential mediators in cardiovascular diseases. Eur Heart J 25:1675–1678

25. Staff AC, Ranheim T, Henriksen T, B. H 2000 8-Iso-prostaglandin F2� reducestrophoblast invasion and matrix metalloproteinase activity. Hypertension 35:1307–1313

26. Walsh SW, Vaughan JE, Wang Y, Roberts 2nd LJ 2000 Placental isoprostaneis significantly increased in preeclampsia. FASEB J 14:1289–1296

27. Maioli E, Fortino V, Pacini A 2004 Parathyroid hormone-related protein inpreeclampsia: a linkage between maternal and fetal failures. Biol Reprod71:1779–1784

28. Abraham P, Ralston SH, Hewison M, Fraser WD, Bevan JS 2002 Presentationof a PTHrP-secreting pancreatic neuroendocrine tumour, with hypercalcaemiccrisis, pre-eclampsia, and renal failure. Postgrad Med J 78:752–753

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving theendocrine community.

Meziani et al. • PTHrP Biological Activity in Rat Uterine Artery Endocrinology, February 2008, 149(2):626–633 633