Human Reproduction vol.12 no.6 pp.1325–1337, 1997

Local relaxin biosynthesis in the ovary and uterusthrough the oestrous cycle and early pregnancy in thefemale marmoset monkey (Callithrix jacchus)

Almuth Einspanier1,4, M.Rasoul Zarreh-Hoshyari- and the recently described relaxin-like factor. Because relaxinwas originally described as a hormone of the perinatal phaseKhah2, Marga Balvers2, Lorraine Kerr 3,

Kerstin Fuhrmann 1 and Richard Ivell 2,4 of pregnancy, most research until recently has concentrated onthis role. However, this perinatal function, in which relaxin is1Department of Reproductive Endocrinology, German Primateresponsiblein vivo for widening of the pubic symphysis,Centre, Kellnerweg 4, 37077 Go¨ttingen,2Division of Reproductivesoftening of the cervix and, in the rat, in the development ofSciences, IHF Institute for Hormone and Fertility Research,

University of Hamburg, Grandweg 64, 22529 Hamburg, Germany, the teats prior to suckling, has been largely elucidated on theand3MRC Reproductive Biology Unit, 37 Chalmers Street, basis of studies in certain animal models: namely the pig, theEdinburgh, Scotland, UK rat and the guinea pig (reviewed in Sherwood, 1994). In these4To whom correspondence should be addressed species there is a large increase in circulating relaxin in the

third trimester of pregnancy, to reach hormone concentrationsThe pattern of peripheral serum concentration for theat least 10-fold higher than the known dissociation constantpeptide hormone relaxin in women points to the possibility(Kd) of the relaxin receptor. In the human and anthropoidof an interesting paracrine function in the cycle and earlyprimates this is not the case: maximal circulating concentrationspregnancy. In order to investigate this physiology in detail,of relaxin are found in the first trimester of pregnancy andit was decided to examine local relaxin biosynthesis in anthey barely exceed 1–2 ng/ml (Eddieet al., 1986; Steinetzestablished primate model for human female reproductiveet al., 1992), which is below the 0.44 nM (~3 ng/ml) Kdfunction, the marmoset monkey (Callithrix jacchus). In thisreported for binding of the hormone to its receptor (Osheroffinitial study relaxin biosynthesis was assessed using aand King, 1995). However, concentrations as low as 0.03 nMcombination of molecular and immunological techniqueshave been shown to be effective in cell culture (Croninet al.,through the oestrous cycle in the marmoset monkey. The1987; Sortinoet al., 1989). This would suggest that, at leastnucleotide sequence of the full-length relaxin gene tran-in the human and other primates, a systemic relaxin physiologyscript was cloned from the marmoset ovary and found toin the perinatal phase, as in other species, is unlikely; serumbe closely homologous to that of the human H2 relaxin.relaxin concentrations are more likely to reflect local paracrineUsing gene specific probes derived from this sequence,functions which predominate in early pregnancy, possibly inRNase protection assays, reverse transcription–polymeraserelation to implantation. Support for a role in implantation ischain reaction (RT–PCR) assays and in-situ hybridization,provided by Stewartet al. (1993), who showed significantlyshowed relaxin gene expression within the ovary in thecaincreased concentrations of relaxin in women undergoingcells and corpora lutea in the oestrous cycle, increasing inspontaneous early abortion. Relaxin-like immunoreactivity hasearly pregnancy. Relaxin gene expression was also identifiedbeen detected by immunoassay in the uterus, ovary andat a low level in the uterus and placenta, and at a highercervix of non-pregnant women (Bongers-Binderet al., 1991).level in the prostate in the male marmoset monkey. UsingAdditionally, relaxin has been shownin vitro to be a majortwo different relaxin-specific antisera, relaxin-like immuno-stimulator of cAMP in endometrial cells from the human non-reactivity was observed in the ovary with a pattern ofpregnant uterus (Feiet al., 1990), and is evidently involved indistribution coincident with that obtained by in-situ hybrid-the decidualization of endometrial stromal cells (Zhuet al.,ization. Immunoreactivity was also found in the non-1990; Bellet al., 1991; Tabanelliet al., 1992).pregnant uterus, within the endometrial epithelium of the

Little is known about the expression of relaxin or its functionlate proliferative phase and increasing within the glandsin either the non-pregnant or pregnant cycles, or in earlythrough the secretory phase. Taken together, the patternpregnancy, especially in primates. In order to examine such aof relaxin peptide and mRNA expression show there is thephysiology we have first assessed the biosynthesis of relaxinbasis for local relaxin physiology within the ovarian follicleat both mRNA and protein levels in the female reproductiveand corpus luteum, and within the uterus during thetissues of a new world monkey, the marmoset (Callithrixoestrous cycle in this new world monkey.jacchus). This model was chosen because of its demonstratedKey words:corpus luteum/endometrium/placenta/primate/prostatesuitability as a model of primate/human reproductive physi-ology, with well-defined follicular and luteal phases of endo-metrial and ovarian function (Hillieret al., 1987). Additionally,

Introduction the 28-day oestrous cycle can be controlled by a luteolyticprostaglandin-F2α injection to provide precisely-regulated tis-The peptide hormone relaxin belongs to a family of related

peptides which includes insulin, the insulin-like growth factors sues (Summerset al., 1985). Finally, the marmoset has been

© European Society for Human Reproduction and Embryology 1325

A.Einspanier et al.

demonstrated to have circulating relaxin-like immunoreactivity(Steinetz et al., 1995). Local biosynthesis of relaxin wasdemonstrated using a combination of mRNA techniquesemploying species-specific cloned relaxin cDNA probes, aswell as immunohistochemistry using two different antisera toindicate the precise cellular distribution of the hormone.

Materials and methods

Animals and tissues

Adult female marmoset monkeys (Callithrix jacchus) from a captivebreeding colony at the German Primate Center in Go¨ttingen, Germany,were housed in pairs under controlled environmental conditions, andcycle status was monitored by twice weekly assessment of plasmaprogesterone concentrations (Einspanieret al., 1994). The length ofthe oestrous cycle in the marmoset is 28 days, ovulation normallyoccurring around day 10. In certain experiments animals weresynchronized by application of a luteolytic dose of prostaglandin F2αon day 12 of the luteal phase (Einspanieret al., 1994). For non-conceptive cycles, females were housed with castrated males. Tissueswere removed from 14 animals under anaesthesia as describedpreviously (Einspanieret al., 1994). Ovaries and uteri were collectedfrom seven different stages: early (n 5 2) or late [before endogenousluteinizing hormone (LH) increase,n 5 2; after exogenous HCGapplication,n 5 2] follicular phase, early (day 3–4 after ovulation;n 5 2), mid- (day 8–9 after ovulation;n 5 2) and late (day 12–13after ovulation;n 5 2) luteal phase (equivalent to days 17–18, days21–22, and days 26–27 of the luteal phase respectively, in the human),as well as early pregnancy (,40 days,n 5 2). Placental tissue wasobtained immediately following a spontaneous or assisted birthrespectively. Prostate tissue was obtained from a sexually maturemale culled as part of a management programme. Tissues from eachanimal were separately portioned, one part being frozen immediatelyin liquid nitrogen and stored until use at –80°C, and one partimmersion-fixed for immunohistochemistry in phosphate-buffered 4%paraformaldehyde for 4 h at 4°C, prior to paraffin-embedding (seebelow). For the preparation of the cDNA library, ovaries werecollected from two adult female marmosets from the colony at theMRC Reproductive Biology Unit in Edinburgh. These animals werein the follicular phase of the non-pregnant cycle.

RNA preparation and analysis

RNA was prepared from frozen tissue fragments according to themethod of Chomczynski and Sacchi (1987). In order to obtain aspecies-specific cDNA fragment, reverse transcription–polymerasechain reaction (RT–PCR) was performed as previously described(Hartung et al., 1995), using oligonucleotide primer combinations Figure 1. Nucleotide sequence of the full-length cDNA andcorresponding to conserved regions of the relaxin precursor sequencededuced amino acid sequence of the marmoset relaxin precursor(Hartung et al., 1995). The resulting ~440 bp PCR fragment was polypeptide. The amino acid sequence is shown as a comparisonsubcloned into the plasmid pCRII (Invitrogen, San Diego, CA, USA)with the human H2 relaxin sequence with which it bears close

homology. Asterisks indicate amino acid identity. The putativeand sequenced.receptor binding domain within the B-domain is boxed.An RNase protection assay was established with the help of a

commercial kit (Ambion, Austin, TX, USA). For the probe, a 271bp DNA fragment was generated from the marmoset relaxin-cDNA inserted PCR sequence, included 87 nucleotides of vector sequence. As

control for the integrity of the RNA a second RNase protection(see below) by RT–PCR using as forward and reverse primersoligonucleotides from positions 260–279 and 512–531 (see Figure 1) assay was established for the marmoset glyceraldehyde 3-phosphate

dehydrogenase (GAPDH) enzyme. A 197 bp PCR product wasrespectively. This PCR product was cloned into the pGEM-T vector(Promega, Madison, WI, USA) and the sequence checked by direct obtained from a longer cloned cDNA fragment of the marmoset

GAPDH–cDNA (Einspanier and Ivell, 1997), using for the PCRdouble-stranded sequencing. This plasmid was linearized withNcoIand transcribed in the presence of [α32P]-CTP (Amersham-Buchler, reactions the oligonucleotides 59-GTCTTCACCACCATGGAG-39

(forward primer) and 59-ATGGATGACCTTGGC-39 (reverse primer).Braunschweig, Germany) using SP6 RNA polymerase (Stratagene,La Jolla, CA, USA) to provide a cRNA probe which, in addition to the This PCR product was cloned into the pCRII vector (Invitrogen).

1326

Relaxin biosynthesis in the marmoset monkey

Figure 2. RNase protection assay for specific relaxin gene transcripts (a) and for glyceraldehyde 3-phosphate dehydrogenase (GAPDH)transcripts as control of RNA integrity (b). M 5 markers; P5 undigested cRNA probe; C5 negative control substituting yeast tRNA forsample RNA. RNA samples: 1, CL day 5 luteal phase; 2, CL day 5 luteal phase; 3, CL day 8/9 luteal phase; 4, CL day 8/9 luteal phase; 5,CL early pregnancy (~2 months); 6, CL early pregnancy (~2 months); 7, CL mid-pregnancy (~3 months); 8, residual ovary (minus CL)early pregnancy; 9, residual ovary (minus CL) mid-pregnancy; 10, granulosa cells before the luteinizing hormone (LH) surge; 11, granulosacells following exogenous human chorionic gonadotrophin (HCG) application; 12, theca cells before the LH surge; 13, uterus, proliferative(day 7 follicle) phase; 14, uterus, secretory (day 9/10 luteal) phase; 15, placenta (Caesarean section); 16, placenta (spontaneous birth); 17,pituitary; 18, prostate; 19, liver. Note that for some samples there was insufficient extracted RNA to perform the GAPDH control in additionto the assessment of relaxin (RLX) mRNA.

Figure 3. Reverse transcription–polymerase chain reaction (RT–PCR) analysis for specific relaxin (RLX) gene transcripts (a,b) and forglyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA as control (c) for the integrity of the extracted RNA. Each sample represents adifferent animal/tissue. Panelsb andc show autoradiograms exposed for 12 h. Panela is similar to panelb, but exposed for only 30 min.Samples: C5 water negative control; 1, uterus, proliferative (day 4 follicle) phase; 2, uterus, proliferative (day 4 follicle) phase; 3, uterus,proliferative (day 4 follicle) phase; 4, uterus, secretory (day 14 luteal) phase; 5, uterus, secretory (day 12 luteal) phase; 6, uterus, secretory(day 9 luteal) phase; 7, uterus, proliferative (day 7 follicle) phase; 8, uterus, secretory (day 9/10 luteal) phase; 9, CL day 5 luteal phase; 10,CL day 5 luteal phase; 11, CL day 8/9 luteal phase; 12, CL day 8/9 luteal phase; 13, CL early pregnancy (~2 months); 14, CL earlypregnancy (~2 months); 15, CL mid-pregnancy (~3 months); 16, prostate.

The plasmid was linearized withHindIII and transcribed using T7 derived from the cloned full-length cDNA sequence (forward primer:nucleotides 131–149; reverse primer: nucleotides 512–531; seeRNA polymerase (Stratagene) to provide a cRNA probe of 323 bp,

which additionally included 126 bp of vector sequence. RNase Figure 1). This assay yields a PCR product of 400 bp. Total RNA(2 µg) was used as template in a standard 50µl reverse transcriptionprotection assays were evaluated by electrophoresis on 6.5% poly-

acrylamide sequencing gels followed by autoradiography. reaction. 1µl of the resultant single-strand (ss)-cDNA was then usedas template for the PCR reaction which included a touch-downSince in some tissue samples the specific RLX-mRNA was either

very low or undetectable using the RNase protection assay, a specific annealing temperature of 55–45°C, an elongation temperature of 72°Cand a denaturing step at 95°C. All steps were for 1 min each, andRT–PCR assay was established using new oligonucleotide primers

1327

A.Einspanier et al.

Figure 4. In-situ hybridization for specific relaxin (RLX gene transcripts in cryostat sections from different marmoset ovaries. (A–H)Sections probed with the anti-sense cRNA. (I–P) Sections probed with the sense cRNA as control. (A, B, I , J) Preovulatory follicle, gc5granulosa cells; ti5 theca interna. (C, D, K , L ) Early luteal phase, cl5 corpus luteum. (E, F, M , N) Mid/late luteal phase. (G, H, O, P)early pregnancy, sections illustrated from within a corpus luteum.A, C, E, G, I , K , M andO are photographed using darkfield reflectancemicroscopy.B, D, F, H, J, L , N andP are the equivalent transmitted light images (original magnifications:A–F, I–N 375; G, H, O, P3150).

1328

Relaxin biosynthesis in the marmoset monkey

following 30 cycles the reaction products were elongated for 10 min antibody solution. Two different rabbit polyclonal anti-porcine relaxinantibodies were used (serum 258; Sherwood and Rutherford, 1981;at 72°C. The PCR products were analysed by 1.2% agarose gel

electrophoresis followed by transfer to Hybond N1 (Amersham- and R6; O’Bryne and Steinetz, 1976; courtesy of Dr O.DavidSherwood, University of Illinois, and Dr Bernard Steinetz, New YorkBuchler, Braunschweig, Germany) membranes and hybridization with

an internal relaxin-specific probe. This probe was derived by PCR University Medical Center respectively), diluted 1:1000 for ovariansections and 1:2000 for uterine sections. A rabbit polyclonal 3β-HSDusing the cloned relaxin cDNA (Figure 1) as template and as primers

the oligonucleotides 193–211 (forward primer) and 372–391 (reverse antibody (S683; Keeneyet al., 1993; courtesy Professor Ian Mason,University of Edinburgh) was used at a dilution of 1:500, and a mouseprimer). This 198 bp PCR product was labelled with [α32P]-dCTP

(Amersham-Buchler) by random primer labelling and the hybridizing monoclonal anti-progesterone receptor (PR; Affinity Bioreagents Inc,Golden, CO, USA) was diluted 1:50 before use. All antibody dilutionsDNA bands visualized by conventional autoradiography. As control

for the quality of the ss-cDNA prepared, parallel PCR reactions were were in PBS. Immunological epitopes were detected using avidin–biotin–peroxidase (ABC) and PAP visualization techniques, asperformed using oligonucleotide primers specific for the marmoset

GAPDH gene sequence as described previously (Einspanier and described previously (Einspanier and Ivell, 1997). Some sectionswere additionally counterstained with haematoxylin, as indicated inIvell, 1997). After transfer of the electrophoresed products to nylon

membranes (Hybond N1), these were visualized by hybridization to the figure legends.For the detection of both progesterone receptor and relaxin in thean internal 25mer oligonucleotide (59-CCTTCGATGCTGGGGCTG-

GCATTGC) radiolabelled by T4 polynucleotidyl kinase (New Eng- same sections, the following double-staining technique was used.After rehydration in a descending series of ethanol, sections wereland Biolabs, Schwalbach, Germany) in the presence of [γ32P]-ATP

(Amersham-Buchler). incubated in 10 mM citrate buffer, pH 6.0, at 120°C for 10 min,followed by cooling for 30 min to room temperature. Sections were

Molecular cloning of the marmoset relaxin cDNA then rinsed in PBS, incubated in 0.0125% trypsin (Gibco-BRL,Glasgow, UK) at 37°C, then rinsed again in PBS. Slides were thenIn order to obtain a cDNA sequence corresponding to the completetreated with 3% H2O2 in a humidified chamber for 20 min, rinsedrelaxin gene transcript of the marmoset ovary, and hence also theagain and then incubated with 1:10 diluted mouse IgG (Dianova) forfull-length encoded amino acid sequence of the relaxin precursor30 min, followed by an overnight incubation at 4°C with the firstpolypeptide, a cDNA library was constructed using the lambdaprimary antibody, against the progesterone receptor, also at a dilutionbacteriophage vector Uni-Zap-XR (Stratagene) from mRNA derivedof 1:10 in PBS. After further rinsing in PBS, biotinylated anti-mousefrom four ovaries from two non-pregnant adult female marmosets,IgG (Vector Laboratories, Burlingame, MA, USA) was applied at aexactly as described elsewhere for marmoset testis (Saunderset al.,dilution of 1:300 for 30 min. After rinsing in PBS, sections were1996). The library, which had an estimated complexity of 23105

incubated in a humidified chamber for 40 min at 4°C in the ABCindependent clones and an average insert size of 870 bp (range 300–complex reagent (10µl avidin, 10µl biotinylated horseradish peroxid-2000 bp), was screened using the original ~440 bp relaxin-specificase in 1 ml PBS; Vector Laboratories), followed after further briefPCR product (see above). Two independent clones were selected andrinsing in PBS by DAB/nickel incubation in a humified chamber incDNA isolated by phagemid rescue followed by conventional cDNAthe dark at room temperature for 8 min. Following this step, thepreparation procedures (Sambrooket al., 1989). The nucleotidesections were washed successively in water and PBS each for 5 minsequence of the two clones was obtained in both orientations by thebefore incubating in normal rabbit serum (diluted 1:100 in PBS) inconventional dideoxynucleotide termination protocol for double-a humidified chamber for 15 min. After draining off the serum, thestranded DNA.second primary antibody, anti-porcine relaxin, diluted 1:500 in PBS,was then applied to the sections for 1 h in a humidified chamber.In situ hybridizationSections were then washed twice in PBS and the secondary antibodyIn-situ hybridization was performed on 10µm cryofixed sections of(biotinylated anti-rabbit IgG; Vector Laboratories) applied. This wasfrozen marmoset ovaries from the various cycle stages as indicated,followed by a 40 min incubation in ABC complex reagent (see above)using [α35S]-CTP-labelled cRNA probes as previously described byat room temperature. Sections were then again washed twice inIvell et al. (1990). The antisense and sense cRNA probes werePBS before applying the second chromogen substrate, AEC (Vectortranscribed from the 441 bp PCR fragment which is specific forLaboratories) for 15 min. The sections were then washed for 5 minmarmoset relaxin and had been cloned into the pCRII vector (seein water before mounting in Mowiol (Hoechst, Frankfurt, Germany).above). For antisense probes the vector was linearized withNotI and

incubated with SP6 RNA polymerase; for the sense probes used asnegative control the same vector was linearized withHindIII andincubated with T7 RNA polymerase. Results

Cloning and sequence analysis of marmoset relaxinImmunohistochemistry

Immunohistochemistry was carried out on 3µm paraffin sections Using a combination of oligonucleotide primers shown tomounted on 0.01% poly-L-lysine (Sigma, Deisenhofen, Germany)- exhibit high homology between the relaxin molecules alreadycoated slides. Sections were dried for 1 h at 45°C, dewaxed in xyloldescribed from various species, a 441 bp PCR fragment wasfor 6 min, followed by descending ethanols, and then rinsed inisolated using luteal RNA from early pregnancy as templatephosphate-buffered saline (PBS; Instamed Dulbecco’s without Ca21

(Hartung et al., 1995). This was subcloned, sequenced andand Mg21, and made up at 9.55 g/l to pH 7.5; Biochrom, Berlin,

used as a probe to isolate full-length clones from a cDNAGermany), before incubating for 30 min at room temperature inlibrary derived from marmoset ovary. Altogether, there were0.03% H2O2. After rinsing again in PBS, sections were then blocked19 positive clones identified from 13106 screened, implyingfor 30 min in mouse (for progesterone receptor detection), or rabbit [fora frequency of gene expression of 0.002% for whole ovary ofrelaxin and 3β-hydroxysteroid dehydrogenase (3β-HSD) detection]the oestrous cycle.immunoglobulin (Ig)G, both diluted 1:10 in PBS. After further

washing in PBS, sections were then incubated in the primary The final cDNA sequence of 744 nucleotides (Figure 1),

1329

A.Einspanier et al.

Figure 5.

1330

Relaxin biosynthesis in the marmoset monkey

determined by complete sequencing of two independent full- after longer autoradiographic exposure (not shown). Granulosacells were negative, as were the uterine samples. Because thelength clones, encoded the complete relaxin precursor polypep-

tide of 185 amino acids from the marmoset. This polypeptide GAPDH controls of the illustrated uterine samples suggestedonly low concentrations of mRNA, six further uterine samplesconformed to the common A-B-C domain structure of all the

relaxin molecules so far described, and includes a classic from independent animals, three from the luteal phase andthree from the follicular phase were additionally analysed. Insignal peptide at theN-terminus, with which the molecule is

directed to the secretory pathway. The probable signal peptidase spite of strong GAPDH-mRNA signals, all six tissues remainednegative for relaxin-mRNA (not shown). Of other tissuescleavage site, following the rules of Von Heijne (1986), is at

Ala24. The marmoset prorelaxin shows greatest homology to tested, the prostate (lane18) was positive, but liver and pituitary,used as negative controls, were not. Because of the verythe human H2 prorelaxin (71%) with the homology being

equally distributed through A-, B- and C-domains. Possible limited amounts of RNA available it was not possible toperform GAPDH controls on all samples. Thus for samples 5,cleavage sites within the precursor to form an A–B heterodimer

are indicated in Figure 1 based on the human homology, though 7, and 10–12, this assessment of RNA integrity was notperformed. However, small aliquots of all RNA samples werewhether the marmoset prorelaxin is indeed cleaved must await

further studies. checked prior to all analyses by gel electrophoresis to verifythe quantity and quality especially of the 18S and 28S ribosomalThe full-length cDNA sequence predicts a mRNA of ~0.9 kb,

allowing ~150 residues for the polyadenylate tail. This is RNA (not shown). According to these criteria all RNA samplesappeared to be intact and equivalent amounts were used forsimilar to what has been observed in Northern hybridization

of human luteal RNA (Ivellet al., 1989). Unfortunately, it was the subsequent assays.not possible to extract marmoset tissues to obtain sufficient

Reverse transcription–polymerase chain reactionRNA for northern hybridization. However, tissue samples wereanalysed using an RNase protection assay (see below) which RT–PCR assays were established for relaxin and GAPDH

using species-specific oligonucleotide primers. RT-PCR assaysconfirmed the homology in the region of the subcloned andprotected probe. are much more sensitive than either Northern hybridization or

RNase protection assays, and were therefore used especiallyRNase protection analysis to detect possible low level expression of relaxin gene tran-

scripts in those tissue samples which had proved negative inBecause of the small amounts of RNA which could be extractedfrom marmoset tissues, specific relaxin-mRNA concentrations the less sensitive assays.

All the samples which were positive in the RNase protectionwere assessed using an RNase protection assay (Figure 2).RNase protection assays, by depending upon the protection assay gave strong positive signals in the RT–PCR analysis

(Figure 3). Additionally, weak positive signals were nowfrom RNase A digestion by duplex formation of a single-stranded cRNA probe with a mRNA molecule, are absolutely obtained for most of the luteal RNA samples from the

non-pregnant cycle (Figure 3, lanes 9–12). Furthermore, onstoichiometric and provide a reliable quantitative estimate ofspecific mRNA concentrations in a given tissue. The presence prolonged autoradiographic exposure of the Southern blotted

PCR products, weak positive signals could now be obtainedof relaxin-mRNA in a sample gave rise to a protected band atthe expected 271 b (Figure 2). The appearance of a doublet at for all of the uterine samples (Figure 3, lanes 1–8). Unfortu-

nately, there was insufficient RNA from the granulosa andthis position is probably due to a single base heterogeneity inthe cloned probe construct. The tissues with the highest levels theca cell preparations to perform RT–PCR on these samples.of expression were the corpus luteum and stromal tissue

In-situ hybridization(including luteal islets and some preantral follicles, but exclud-ing corpora lutea) from ovaries of early to mid-pregnancy The 441 bp RT–PCR product (Hartunget al., 1995) was

subcloned into the pBS plasmid and after linearization with(Figure 2, lanes 5–9), and in the male, the prostate (Figure 2,lane 18). Even after long exposure, no specific signals were eitherHindIII or NotI used to direct in-vitro transcription

using either T7 (sense) or SP6 (antisense) RNA polymerasedetectable from luteal RNA of the non-pregnant cycle (lanes1–4). Additionally, RNA from theca cells (lane 12), and from respectively, in the presence of [α-35S]-CTP. The resulting

cRNA probes were then used to detect relaxin gene transcriptsplacental tissue (lane 15) showed a positive signal. Althoughin Figure 2 only the sample of placental RNA from the by in-situ hybridization to cryostat sections of marmoset

ovaries at various stages of the oestrous cycle (Figure 4).spontaneous birth provides a visible signal, a weaker positivesignal is also evident for the placenta of the induced birth Specific signals (white silver grains) are already evident in the

Figure 5. Relaxin immunoreactivity in follicles and corpora lutea from different stages of the marmoset oestrous cycle and pregnancy.(A) Negative control indicating parts of both a corpus luteum (right) and a large antral follicle (left), substituting pre-immune serum for thespecific primary immunoglobulin (Ig)G. (B) Early cycle corpus luteum. (C) Mid-cycle corpus luteum. (D) Late cycle corpus luteum.(E) Corpus luteum of early pregnancy. (F) Antral follicle from day 7 of the follicular phase. (G) Preovulatory follicle on day 8/9, 20 h afteran exogenous human chorionic gonadotrophin (HCG) application. The sections inB–G made use of the anti-relaxin antiserum 258(Sherwood and Rutherford, 1981). (H) Preovulatory follicle exactly as in (G), but using the R6 anti-relaxin antibody (O’Byrne and Steinetz,1976). Immunostaining is localized within the cytoplasm; nuclei are counterstained with haematoxylin (original magnification3375).

1331

A.Einspanier et al.

theca cell layer of preovulatory follicles (Figure 4A). This a third potentially steroidogenic tissue component capable ofproducing relaxin.specific signal persists after ovulation with the formation of

The physiological role of relaxin in the cycle and earlythe corpus luteum. The luteal signal is relatively weak in thepregnancy is unclear, but has been linked to implantation ofearly luteal phase (Figure 4C) of the oestrous cycle, butthe blastocyst and endometrial function. Additionally, in someincreases in intensity through the mid (Figure 4E) to latespecies, there is evidence to suggest a local production of(Figure 4G) luteal phases. Control sections hybridized withrelaxin within the uterus. In order to characterize the presencethe sense cRNA probe (Figures 4I–P), indicated no significantof relaxin in the marmoset uterus, tissue sections were analyseddifferences in signal intensity from background levels. Inimmunohistochemically through the follicular and luteal phasesaddition to the specific signals illustrated in Figure 4, relaxinof the cycle and in pregnancy. In the non-pregnant cycle, theretranscripts were also sporadically detected within the stromalis staining for relaxin-like epitopes in the epithelial cells duringtissue of the ovary, in small islands of luteal-like cells (lutealthe proliferative phase (Figure 7B). This staining increases inislets; see later).the secretory phase (Figure 7C), where there is in additionstaining in some stromal cells. Maximal staining is seen in theImmunohistochemical detection of relaxin-like epitopesglandular epithelium during the late secretory phase. It should

Expression of the relaxin peptide was assessed immunohisto-be noted that in the early to mid-proliferative phase, the

chemically using two different polyclonal antibodies raisedstaining for relaxin is not homogeneous in all epithelial cells,

against porcine relaxin, and shown to cross-react with relaxinsome cells being heavily stained, others showing no staining

from several species. Within the ovary, relaxin epitopes were(Figure 7B). During pregnancy, there is enhanced staining in

detected using both antibodies within the theca cell layer ofthe deep stromal cells (Figure 7F), though the epithelial cells

antral and preovulatory follicles and within corpora lutea. Inin this region are negative. Closer to the luminal surface ofthe follicular phase there is an increase in staining intensity inthe uterus (Figure 7G) there is an increase in immunoreactivethe theca cells, which attains a maximum following the artificialstaining not only in the stromal cells, but also now in thegonadotrophin surge (Figure 5F,G). Similar intensity ofepithelial cells of the glands. All decidual cells (Figure 7H)immunostaining is obtained irrespective of which polyclonalstain strongly for relaxin immunoreactivity.anti-relaxin antibody is used (Figure 5G,H). Granulosa cellsremain negative throughout the cycle. Once formed the corpusluteum shows steadily increasing relaxin immunoreactivityDiscussionthrough the luteal phase (Figure 5B–D), though there is aLike a number of other mammalian groups, the female marmo-decline in intensity on approaching luteal regressionset exhibits the expression of a relaxin gene within tissues of(Figure 5D). Note that not all luteal cells are positive. Highestthe reproductive tract. As in most species, the corpus luteumsignal intensity is observed in corpora lutea of early pregnancyof pregnancy appears to produce the highest concentrations of(Figure 5E), followed by mid luteal phase, then early or latethis hormone. This probably accounts for the maximum levelsluteal phase. In preliminary double-labelling experiments, itof relaxin immunoreactivity detected for this species in peri-could be shown that some but not all of the relaxin-positivepheral serum in mid pregnancy by Steinetzet al. (1995).luteal cells also stained for 3β-HSD and PR (not shown), Indeed, in women becoming pregnant after ovarian stimulationthough there did not appear to be a clear-cut correlation of thethe serum concentrations of relaxin appear to reflect the numberepitopes in terms of either distribution or intensity of staining.of follicles ovulating, and hence the number of resultingStromal cells were consistently negative for relaxin immuno-corpora lutea (Kristianssonet al., 1996). In the present studyreactivity. there is good agreement between the in-situ hybridization

In addition to the follicles and corpora lutea, small islandsdata, the RT–PCR and RNase protection assessment of RNAof luteal-like cells are apparent within the stroma, whichconcentrations and the immunohistochemical findings, all con-also react positively to the relaxin antisera (Figure 6E–G),sistently showing the corpus luteum and the theca cells of thethroughout the entire follicular and early/mid-luteal phases.preovulatory follicle to be major sites of relaxin biosynthesis,Nearly all of the luteal-like cells in these clusters stainand equally validating the reliability of the antibodies to detectpositively for relaxin epitopes. These luteal islets also showtrue relaxin epitopes.marked in-situ hybridization to the relaxin-specific cRNA Although highest levels of relaxin expression are evidentprobes (Figure 6B), showing that there is true relaxin biosyn-in pregnancy, there nevertheless appears to be significantthesis in these ectopic structures. A subsequent, more detailedexpression also in the non-pregnant cycle, within both theimmunohistochemical analysis indicates that these luteal isletstheca cells and the corpus luteum. This pattern of expressionare also positive for steroidogenic components, such as 3β- is similar to that reported also for the pig (Fields and Fields,HSD (not shown) and the progesterone receptor (Figure 6D).1985; Bagnellet al., 1987, 1990a,b), and human (Ivellet al.,Whereas the majority of the luteal-like cells also stain for 3β- 1989; Stoelket al., 1991). It has been suggested that relaxinHSD, only about half stain for PR. Thus in double-labelling may be involved in the rat in stimulating the production byexperiments (Figures 6F,G) luteal-like cells are evident whichgranulosa cells of collagen metabolizing enzymes, such as tPAstain positively for both relaxin and PR (arrows), and some(Tooet al., 1984), and hence having a role in tissue remodelling,for relaxin only. Taken together, it would appear that thepossibly associated with ovulation and formation of the corpus

luteum (Hwanget al., 1996). Relaxin also appears to influencemarmoset ovary contains besides the corpora lutea and follicles

1332

Relaxin biosynthesis in the marmoset monkey

Figure 6. In-situ transcript hybridization for relaxin mRNA and immunohistochemistry for relaxin peptide and progesterone receptor (PR) inluteal islets. (A) In-situ hybridization using a sense probe as negative control. (B) In-situ hybridization using an anti-sense probe.A andBare darkfield reflectance micrographs showing positive signals as white silver grains. (C) Transmitted light micrograph of the same field ofview as inB. (D) PR immunoreactivity (dark nuclear stain), without counterstaining. (E) Relaxin immunoreactivity (cytoplasmic stain),without counterstaining. (F) Double-labelling for PR and relaxin. Typical double-labelled cells are arrowed. (G) Overview of a completeluteal islet double-stained for PR and relaxin, as above. Typical double-stained cells are marked by arrows. Negative controls forimmunohistochemistry are shown as inserts inD–G. Tissue was paraffin-embedded for immunohistochemistry and cryopreserved for the in-situ transcript hybridization (original magnifications:A–C, 3250; D–F, 3375; G, 3300).

DNA synthesis by granulosa cells (Zhang and Bagnell, 1993). species, such as the human (Gagliardiet al., 1992). Here,synthesis only appears to occur once these cells have luteinizedThe finding of relaxin biosynthesisin vivo within the theca

cells and luteal cells of the cycle, but not in granulosa cells,in vitro, and thus might be comparable to the relaxin-positivecells of the corpus luteumin vivo, although whether such lutealcould nevertheless be consistent with the observations of

relaxin biosynthesis in cultured granulosa cells from other cells are of theca or granulosa cell origin, or both, is not clear.

1333

A.Einspanier et al.

Figure 7.

1334

Relaxin biosynthesis in the marmoset monkey

An exciting finding from the present study was the relaxin immunoreactivity in the endometrium of earlypregnancy, especially in those regions close to the sites ofpresence of so-called luteal islets within the stromal tissue

of the ovary and discrete from luteal and follicular structures. implantation, but not, for example, in the mesometrialendometrium or in the non-conceptus-bearing uterus. RelaxinThe cells comprising these islets exhibit nuclear progesterone

receptor immunoreactivity as well as the 3β-HSD enzyme, also appears to influence myometrial contractility, acting,for example, in the rat non-pregnant uterus to dampenand can thus be considered as true, albeit ectopic luteal

cells. That these cells also express high concentrations of oxytocin- or endothelin-induced contractions (Goldsmithet al., 1989; McGovernet al., 1992) as well as reducingrelaxin mRNA and peptide suggests that they are served by

similar regulatory mechanisms as the other steroidogenic spontaneous contractions in a number of species (Downingand Hollingsworth, 1993).tissues of the ovary at this time. Whether these luteal islets

also contribute to the circulating concentrations of relaxin Systemic relaxin in the late cycle and early pregnancy ofthe primate, including the human (Eddieet al., 1986; Stewartis not known, though their mass relative to that of the

corpus luteum would suggest the latter still to be the majoret al., 1990; Johnsonet al., 1993), chimpanzee (Steinetzet al., 1992), rhesus macaque (Stewartet al., 1993; Duffysource of serum relaxin. The source of these luteal islets is

obscure; they might be stromal in origin, or derive from et al., 1995) and marmoset (Steinetzet al., 1995) neverexceeds ~2 ng/ml, i.e lower than the determined Kd for theexisting corpora lutea, or alternatively be remnants of

luteinized but non-ovulated follicles. Morphologically, how- rat relaxin receptor (Osheroff and King, 1995). In agreementwith this, relaxin only influences porcine granulosa cells orever, they would appear to be different from the pseudodecid-

ual cells described for the pregnant human ovary by human endometrial stromal cells at concentrations 1,10 ng/ml, with a half-maximal response at 4–5 ng/ml (Feiet al.,Blankenshipet al. (1994).

Specific relaxin-mRNA in total uterine extracts is undetect- 1990; Zhang and Bagnell, 1993), though anterior pituitarycells can respond to as little as 0.2 ng/ml (Croninet al.,able using the RNase protection assay, but can be detected

at low values by RT–PCR. Using two different antibodies, 1987). Thus, serum relaxin concentrations in primates duringthe cycle and in early pregnancy may not in themselves beimmunohistochemistry indicates clear positive signals in

endometrial tissues of both the cycle and pregnancy. Relaxin physiologically relevant, but rather reflect local biosynthesiswithin the ovary and uterus. Local concentrations withinwould appear to be present in both stromal as well as

epithelial cells, though especially the latter indicate significant the follicle, corpus luteum and uterus are likely to be muchhigher, and thus certainly able to activate the relaxinrelaxin immunoreactivity first in the late secretory phase

and especially in the decidua. Thus it would appear that, receptors which have been shown for several species to bein the immediate neighbourhood of these relaxin sources.as in other species such as the rabbit (Lee and Fields,

1990), pig (Knox et al., 1994) human (Bryant-Greenwood Specific relaxin-mRNA signals were also detected in theterm placenta and the prostate gland, showing that theet al., 1993) and guinea pig (Larkin and Renegar, 1986),

relaxin can be produced locally within the endometrium, marmoset conforms closely with the tissue-specific patternof gene expression demonstrated for the human. Also inparticularly in the secretory phase of the cycle and in

pregnancy. the human, maximal concentrations are expressed in thecorpus luteum, followed in order of magnitude by theThere are few data discussing a function for relaxin in

the uterus of the menstrual cycle. The most significant of placenta, and prostate. Although a new world monkey, andthus not closely related to the anthropoid primates, thethese is the observation that relaxin appears to be an

important stimulator of cAMP (Feiet al., 1990) and together marmoset thus appears to present a suitable model withwhich to assess the role of relaxin in the physiology of thewith progesterone can induce decidualization in in-vitro

cultures of human endometrial stromal cells from the cycle female primate, especially during the cycle and peri-implantation period. The cloning of the specific marmoset(Zhu et al., 1990; Bell et al., 1991; Tabanelliet al., 1992;

Laneet al., 1994). Taken together these findings would imply transcript for relaxin allows not only the development ofsuch sequence specific assays as in-situ hybridization andan important role for ovarian relaxin in the differentiation of

the endometrium at the end of the cycle.In vivo, this is the RNase protection assay, but also offers the opportunityto produce biotechnologically a bioactive marmoset relaxinan essential prerequisite for implantation. It is interesting to

note that both in the rat (Fieldset al., 1992) and the rabbit by expression in appropriate cell systems. The developmentof such tools would then allow optimal use of the(Fields and Lee, 1991), there appears to be expression of

Figure 7. Immunohistochemistry for relaxin in the non-pregnant (A–C) and early pregnant (E–G) uterus, and in decidua (D, H) of themarmoset monkey. PanelsA, D andE are negative controls where the primary antibody was replaced by an equivalent amount ofimmunoglobulin (Ig)G. PanelsA–C show non-pregnant uterus from the proliferative (B) and secretory (A, C) phases, indicating weak orstrong immunostaining, respectively, in the glandular epithelial cells. PanelsE andF represent early pregnant uterus, showing deep glandsstaining specifically only in stromal cells (F), but in more luminal areas (G) there is strong immunostaining in both stromal and epithelialcells. PanelsD andH shows a section of the decidua from an early pregnant uterus, with specific staining in nearly all cells. All illustratedsections made use of anti-relaxin 258 (Sherwood and Rutherford, 1981). The R6 antibody gave similar results (not shown). Positiveimmunostaining for relaxin in these paraformaldehyde-fixed and paraffin-embedded sections is represented by a red colour. Sections arecounterstained for cell nuclei using haematoxylin (original magnifications:A–C andE–G, 3375; D andH, 3625).

1335

A.Einspanier et al.

of relaxin in the pregnant rat: immunolocalization in the corpora lutea andmarmoset model with homologous assays and the appropriateendometrium.Endocrinology, 130,2985–2990.

physiological peptide. Gagliardi, C.L., Goldsmith, L.T., Daketos, M.et al. (1992) Human chorionicgonadotrophin stimulation of relaxin secretion by luteinized humangranulosa cells.Fertil. Steril., 58, 314–320.

Acknowledgements Goldsmith, L.T., Skurnick, J.H., Wojtczuk, A.S.et al. (1989) The antagonisticeffect of oxytocin and relaxin on rat uterine segment contractility.Am.We should like to thank Drs David Sherwood, Bernard Steinetz andJ. Obstet. Gynecol., 161,1644–1649.Ian Mason for supplying the anti-relaxin and anti-3β-HSD antibodies,

Hartung, S., Kondo, S., Abend, N.et al. (1995) The search for ruminantand Dr David Sherwood also for generously providing the purifiedrelaxin. In McLennan, A.H., Tregear, G. and Bryant-Geenwood, G.D. (eds),porcine relaxin used for controls. We are also grateful to ProfessorsProgress in Relaxin Research.World Scientific Publishing, Singapore, pp.

Keith Hodges, Freimut Leidenberger, and Heinrich Schulte for provi- 439–456.sion of excellent research facilities, as well as to Werner RustHillier, S.G., Harlow, C.R., Shaw, H.J.et al. (1987) Granulosa celland Alexandra Marten for some very helpful technical advice and differentiation in primate ovaries: the marmoset monkey (Callithrix jacchus)assistance. Aspects of this study form part of the doctoral thesis by as a laboratory model. In Stouffer, R. (ed.),The Primate Ovary.PlenumM.R.Z-H-K. to be submitted to the Faculty of Biology, Hamburg Press, New York, pp. 61–73.University. Research was generously supported by the DeutscheHwang, J-J., Lin, S-W., Teng, C-H.et al. (1996) Relaxin modulates theForschungsgemeinschaft (grants Ei333/4–1 and Iv7/7–1). ovulatory process and increases secretion of different gelatinases from

granulosa and theca–interstitial cells in rats.Biol. Reprod., 55, 1276–1283.Ivell, R., Furuya, K., Nollmeyer, D.et al. (1990) The expression of

neuropeptide genes in the mammalian testis. In Isidori, A., Fabbri, A. andReferencesDufau, M.L. (eds),Hormonal Communicating Events in the Testis.SeronoBagnell, C.A., Frando, L.B., Downey, B.R.et al. (1987) Localization ofSymposium Publications. Vol. 70. Raven Press, New York, pp. 45–56.relaxin in the pig follicle during preovulatory development.Biol. Reprod.,

Ivell, R., Hunt, N., Khan-Dawood, F. and Dawood, M.Y. (1989) Expression37, 235–240.of the human relaxin gene in the corpus luteum of the menstrual cycle andBagnell, C.A., Tashima, L., Tsark, W.et al. (1990a) Relaxin gene expressionin the prostate.Mol. Cell. Endocrinol., 66, 251–255.in the sow corpus luteum during the cycle, pregnancy, and lactation.

Johnson, M.R., Carter, G., Grint, C. and Lightman, S.L. (1993) RelationshipEndocrinology, 126,2514–2520.between ovarian steroids, gonadotrophins and relaxin during the menstrualBagnell, C.A., Tsark, W., Tashima, L.,et al. (1990b) Relaxin gene expressioncycle.Acta Endocrinol., 129,121–125.in the porcine follicle during preovulatory development induced by

Keeney, D.S., Naville, D., Milewich, L.et al. (1993) Multiple isoforms ofgonadotrophins.J. Mol. Endocrinol., 5, 211–219.3β-hydroxysteroid dehydrogenase/∆5–4-isomerase in mouse tissues: male-Bell, S.C., Jackson, J.A., Ashmore, J.et al. (1991) Regulation of insulin-likespecific isoforms are expressed in the gonads and liver.Endocrinology,growth factor-binding protein-1 synthesis and secretion by progestin and133,39–45.relaxin in long term cultures of human endometrial stromal cells.J. Clin.

Knox, R.V., Zhang, Z., Day, B.N. and Anthony, R.V. (1994) Identification ofEndocrinol. Metab., 72, 1014–1024.relaxin gene expression and protein localization in the uterine endometriumBlankenship, T., Stewart, D.R., Benirschke, K.et al. (1994) Immunocyto-during early pregnancy in the pig.Endocrinology, 135,2517–2525.chemical localization of nonluteal ovarian relaxin.J. Reprod. Med., 39,

Kristiansson, P., Sva¨rdsudd, K., Von Schoultz, B. and Wramsby, H. (1996)235–240.Supraphysiological serum relaxin concentration during pregnancy achievedBongers-Binder, S., Burgardt, A., Seeger, H.et al. (1991) Distribution ofby in-vitro fertilization is strongly correlated to the number of growingimmunoreactive relaxin in the genital tract and in the mammary gland offollicles in the treatment cycle.Hum. Reprod., 11, 2036–2040.non-pregnant women.Clin. Exp. Obstet. Gynecol., 18, 161–164.

Lane, B., Oxberry, W., Mazella, J. and Tseng, L. (1994) Decidualization ofBryant-Greenwood, G.D., Rutanen, E.M., Partanen, S.et al. (1993) Sequentialhuman endometrial stromal cellsin vitro: effects of progestin and relaxinappearance of relaxin, prolactin and IGFBP-1 during growth andon the ultrastructure and production of decidual secretory proteins.Hum.differentiation of the human endometrium.Mol. Cell. Endocrinol., 95,23–29.Reprod., 9, 259–266.Chomszinsky, P. and Sacchi, N. (1987) Single step method of RNA isolation

Larkin, L.H. and Renegar, R.H. (1986) Immunochemical and cytochemicalby acid guanidinium thiocyanate–phenol–chloroform extraction.Anal.studies of relaxin-containing cells in the guinea pig uterus.Am. J. Anat.,Biochem., 162,156–159.176,353–365.Cronin, M.J., Malaska, T. and Bakhit, C. (1987) Human relaxin increases

Lee, V.H. and Fields, P.A. (1990) Rabbit endometrial relaxin:cyclic AMP levels in cultured anterior pituitary cells.Biochem. Biophys.immunohistochemical localization during preimplantation, pregnancy, andRes. Commun., 148,1246–1251.lactation.Biol. Reprod., 42, 737–745.Downing, S.J. and Hollingsworth, M. (1993) Action of relaxin on uterine

McGovern, P.G., Goldsmith, L.T., Schmidt, C.L.et al. (1992) Effects ofcontractions – a review.J. Reprod. Fertil., 99, 275–282.endothelin and relaxin on rat uterine segment contractility.Biol. Reprod.,Duffy, D.M., Stouffer, R.L. and Stewart, D.R. (1995) Dissociation of relaxin46, 680–685.and progesterone secretion from the primate corpus luteum by acute

O’Byrne, E.M. and Steinetz, B.G. (1976) Radioimmunoassay of relaxin inadministration of a 3β-hydroxysteroid dehydrogenase inhibitor during thesera of various species using an antiserum to porcine relaxin.Proc. Soc.menstrual cycle.Biol. Reprod., 53, 447–453.Exp. Biol. Med., 152,272–276.Eddie, L.W., Bell, R.J., Lester, A.et al. (1986) Radioimmunoassay of relaxin

Osheroff, P.L. and King, K.L. (1995) Binding and cross-linking of32P-labelledin pregnancy with an analogue of human relaxin.Lancet, i, 1344–1346.human relaxin to human uterine cells and primary rat atrial cardiomyocytes.Einspanier, A. and Ivell R. (1997) Oxytocin and oxytocin receptor expressionEndocrinology, 136,4377–4381.in reproductive tissues of the male marmoset monkey.Biol. Reprod., 56,

Sambrook, J., Fritsch, E. and Maniatis, T. (1989)Molecular Cloning, A416–422.Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold SpringEinspanier, A., Ivell, R., Rune, G. and Hodges, J.K. (1994) Oxytocin geneHarbor, USA.expression in the ovary of the common marmoset monkey.Biol. Reprod.,

Saunders, P.T.K., Gaughan, J., Saxty, B.A.et al. (1996) Expression of50, 1216–1222.protamine P2 in the testis of the common marmoset and man visualizedFei, D.T.W., Gross, M.C., Lofgren, J.L.et al. (1990) Cyclic AMP response tousing non-radioactivein situ hybridization.Int. J. Androl., 19, 212–219.recombinant human relaxin by cultured human endometrial cells – a specific

Sherwood, O.D. (1994) Relaxin. In Knobil, E. and Neill, J.D. (eds),Theand high throughputin vitro bioassay.Biochem. Biophys. Res. Commun.,Physiology of Reproduction.2nd edn. Raven Press, New York, pp. 861–1009.170,214–222.

Sherwood, O.D. and Rutherford, J.E. (1981) Relaxin immunoactivity levelsFields, P.A. and Fields, M.J. (1985) Ultrastructural localization of relaxin inin ovarian extracts obtained from rats during various reproductive statesthe corpus luteum of the nonpregnant, pseudopregnant, and pregnant pig.and from adult cycling pigs.Endocrinology, 108,1171–1177.Biol. Reprod., 32, 1169–1179.

Sortino, M.A., Cronin, M.J. and Wise, P.M. (1989) Relaxin stimulates prolactinFields, P.A. and Lee, V.H. (1991) Conceptus-mediated integrity of endometrialsecretion from anterior pituitary cells.Endocrinology, 124,2013–2015.epithelial cells and maintenance of relaxin synthesis in pregnant rabbits:

effects of unilateral oviduct ligation.Biol. Reprod., 44, 364–374. Steinetz, B.G., Randolph, C. and Mahoney, C.J. (1992) Serum concentrationsof relaxin, chorionic gonadotropin, estradiol-17β, and progesterone duringFields, P.A., Lee, A.B., Haab, L.M.et al. (1992) Evidence for a dual source

1336

Relaxin biosynthesis in the marmoset monkey

the reproductive cycle of the chimpanzee (Pan troglodytes). Endocrinology,130,3601–3607.

Steinetz, B.G., Randolph, C. and Mahoney, C.J. (1995) Patterns of relaxinand steroids in the reproductive cycle of the common marmoset (Callithrixjacchus): effects of prostaglandin F2α on relaxin and progesterone secretionduring pregnancy.Biol. Reprod., 53, 834–839.

Stewart, D.R., Celniker, A.C., Taylor, C.A.et al. (1990) Relaxin in the peri-implantation period.J. Clin. Endocrinol. Metab., 70, 1771–1773.

Stewart, D.R., Overstreet, J.W., Celniker, A.C.et al. (1993a) The relationshipbetween hCG and relaxin secretion in normal pregnancies vs peri-implantation spontaneous abortions.Clin. Endocrinol., 38, 379–385.

Stewart, D.R., Stouffer, R., Overstreet, J.W.et al. (1993b) Measurement ofperiimplantational relaxin concentrations in the macaque using a homologousassay.Endocrinology, 132,6–12.

Stoelk, E., Chegini, N., Lei, Z.M.et al. (1991) Immunocytochemicallocalization of relaxin in human corpora lutea: cellular and subcellulardistribution and dependence on reproductive state.Biol. Reprod., 44,1140–1147.

Summers, P.M., Wenninck, C.J. and Hodges, J.K. (1985) Cloprostenol-inducedluteolysis in the marmoset monkey (Callithrix jacchus). J. Reprod. Fertil.,73, 133–138.

Tabanelli, S., Tang, B. and Gurpide, E. (1992)In vitro decidualization ofhuman endometrial stromal cells.J. Steroid Biochem. Mol. Biol.,42,337–344.

Too, C.K.L., Bryant-Greenwood, G.D. and Greenwood, F. (1994) Relaxinincreases the release of plasminogen activator, collagenase, andproteogycanase from rat granulosa cellsin vitro. Endocrinology, 115,1043–1050.

Von Heijne, G. (1986) Patterns of amino acids near signal-sequence cleavagesites.Eur. J. Biochem., 133,17–21.

Zhang, Q. and Bagnell, C.A. (1993) Relaxin stimulation of porcine granulosacell deoxyribonucleic acid sysnthesisin vitro: interactions with insulin andinsulin-like growth factor I.Endocrinology, 132,1643–1650.

Zhu, H.H., Huang, J.R., Mazella, J.et al. (1990) Differential effects ofprogestin and relaxin on the synthesis and secretion of immunoreactiveprolactin in long term culture of human endometrial stromal cells.J. Clin.Endocrinol. Metab., 71, 889–899.

Received on February 3, 1997; accepted on April 18, 1997

1337