Human Myometrial Genes Are Differentially Expressed in Labor: A ...

0013-7227/02/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 87(6):2435–2441 Copyright © 2002 by The Endocrine Society

Human Myometrial Genes Are Differentially Expressed in Labor: A Suppression Subtractive Hybridization Study E. C. CHAN, S. FRASER, S. YIN, G. YEO, K. KWEK, R. J. FAIRCLOUGH,

AND

R. SMITH

Mothers and Babies Research Centre (E.C.C., S.Y., R.S.), Endocrine Unit, John Hunter Hospital, Hunter Region Mail Centre, Newcastle, New South Wales 2310, Australia; Victoria University of Technology (S.F., R.J.F.), Werribee, Victoria 3030, Australia; and Singapore KK Women’s and Children’s Hospital (G.Y., K.K.), Singapore 229899 Human parturition is effected by a cascade of factors, of which many are unknown. We aim to identify the genes that are changed by labor in the human myometrium by suppression subtractive hybridization . We also seek to ascertain whether these genes are differentially expressed in the myometrium at the upper or fundal and lower segments of the uterus. Term myometrial tissues were obtained from laboring and nonlaboring women undergoing cesarean section after obtaining informed consent. Total RNA was used in suppression subtractive hybridization (CLONTECH PCR Select) to produce two subtracted cDNA libraries enriched for genes expressed during or before labor, labor and not-in-labor libraries, respectively. Dot blot screening of 400 positive clones, constituting 20% of the two subtracted libraries, revealed 30 differentially expressed clones, 14 of which were up-regulated by labor. Among the 10 known genes that were up-regulated in labor, 6 had apparent immune regulatory and inflammatory roles. Three are well-known inflammatory mediators and modulators that were previously linked with parturition: IL-8, manganese superoxide dismutase (MnSOD), and metalloproteinase-9. Three others, interferon-inducible 1-8d gene, elongation factor 1␣, and nucleophosmin, have not been previously linked with labor. Constitutively expressed genes, including cyclophilin and ␣-actin, were found to be altered by labor. Quantitative real-time RT-PCR using Taqman probes further confirmed the up-regulation of some of these genes.

N

ORMAL HUMAN PARTURITION is the culmination of four intertwined events: fetal maturation, coordinated uterine contractions, cervical ripening/remodeling, and placental separation. The changes in the uterus, both contractions and cervical ripening, are instrumental for the delivery process because babies may be born with different degrees of maturity, although placental separation alone does not trigger normal delivery. The transformation of the uterus from quiescence to active contractile synchrony remains a puzzle. The prevailing model for labor suggests that the uterus, predominantly the myometrium or the smooth muscle core, is kept relatively quiescent throughout most of pregnancy by uterotonic agents including progesterone (1, 2). When these agents are supplanted by uterolytic agents or contraction-associated

Abbreviations: B23, Nucleophosmin; DIG, digoxigenin; EF1␣, elongation factor 1␣; IFN 1-8d, interferon-inducible gene 1– 8d; IR, immunoreactivity; McSOD, manganese superoxide dismutase; MMP9, matrix metalloproteinase 9; NIL, not in labor; QRT-RT-PCR, quantitative realtime RT-PCR; SSH, suppression subtractive hybridization.

The amounts of the specific genes assayed were standardized to 18S ribosomal RNA and are expressed as mean ⴞ SEM. Quantitative real-time RT-PCR showed that IL-8 mRNA rose from 0.003 ⴞ 0.002 in nonlaboring samples (n ⴝ 38) to 0.24 ⴞ 0.11 (n ⴝ 20) in gestational-age-matched spontaneously laboring women (P ⴝ 0.035). Similarly, MnSOD rose from 0.11 ⴞ 0.02 (n ⴝ 24) to 1.23 ⴞ 0.56 (n ⴝ 24) in gestational-age-matched women (P ⴝ 0.047). Additionally, cyclophilin, often used as a constitutive or housekeeping gene marker, increased from 0.0008 ⴞ 0.0002 (n ⴝ 6) to 0.002 ⴞ 0.0004 (n ⴝ 6; P ⴝ 0.008) during labor. Notably, MnSOD mRNA was differentially distributed between the upper (0.63 ⴞ 0.18) and lower (0.15 ⴞ 0.05; n ⴝ 15; P ⴝ 0.022) segments of the uterus, but IL-8 was not (n ⴝ 17; P ⴝ 0.97). Induced labor further showed significantly higher levels of IL-8 (0.63 ⴞ 0.21; n ⴝ 14) than spontaneous labor (0.22 ⴞ 0.11; n ⴝ 20; P ⴝ 0.046), but not MnSOD (P ⴝ 0.1). This work identifies novel as well as known genes that were not previously associated with parturition. It extends previous data indicating that there is differential expression of some, but not all genes within the gravid human uterus. Inflammatory genes constitute a major proportion of the known genes found to be up-regulated in labor, lending support to the hypothesis of an inflammatory mechanism for human parturition. This work further indicates that many factors associated with human labor and their complex interactions remain to be elucidated. (J Clin Endocrinol Metab 87: 2435–2441, 2002)

proteins, labor ensues. Various agents, including oxytocin (3) and prostaglandins (4), have been proposed as parturition triggers in the human; however, no agent as yet fills the role as the trigger as definitively as the obligate cortisol surge that precedes labor in the sheep (5) and wallaby (6, 7). A lack of a suitable animal model, the ethics of working with the human model, and technological limitations are the factors restricting advances in our understanding of human parturition. Using modern techniques of discovery-based RNA subtraction and quantitative real-time RT-PCR (QRT-RTPCR), we seek to address this issue by determining the differences in the pattern of gene expression in the laboring compared with the nonlaboring myometrium. The appropriate site for study of the human uterus has been under critical scrutiny over the past few decades (8, 9). Of major concern is whether lower segment tissue, which is the site of incision of most cesareans and thus most accessible for study, reflects the state of the cervix or the fundus. Multistrain measurements of uterine contractions have shown that contractions start in the fundus and travel in waves

2435

2436

J Clin Endocrinol Metab, June 2002, 87(6):2435–2441

Chan et al. • Human Myometrial Genes

downward toward the cervix, such that the longitudinal muscle fibers changed from extension in the fundus to compression in the lower segment (10). As such, the fundal region may be more relevant to studies into contractile processes of parturition, whereas the lower segment may be more akin to cervical changes required for full effacement and delivery. Recent studies have tended to confirm that the upper and lower segments of the uterus express different genes (8, 9) and thereby have different functions. We have also addressed the question of site-specific expression of genes by comparing expression levels in paired upper and lower segment uterine specimens from pregnant women.

subtractions to produce two libraries, labor and not-in-labor (NIL), according to the SSH protocol (PCR-Select cDNA Subtraction Kit, CLONTECH Laboratories, Inc.). Briefly, the two cDNA pools were digested with RsaI and ligated to two different adaptors. The subtractions were performed by stoichiometrically mixing the cDNAs in two rounds of controlled hybridizations. The resulting cDNA molecules were then subjected to two rounds of PCR to amplify and enrich the desired differentially expressed sequences. The primers used in the first or primary PCR were based on the adaptor sequence, whereas the forward and reverse primers for the secondary PCR were based on sequences nested into the adaptors. The PCR products were then cloned into pCR 2.1 vector using a TA cloning kit (Invitrogen, San Diego, CA). Blue-white selection using ␤-galactosidase gene ␣-complementation showed that there were about 2,000 white colonies containing inserts obtained for each library.

Subjects and Methods Experimental subjects

SSH screening

Informed consent was obtained from women before elective or emergency cesarean sections in accordance with the institutional guidelines of the Singapore KK Women’s and Children’s Hospital, the Newcastle John Hunter Hospital, and the University of Newcastle Ethics committees. Tissue slivers (0.5 ⫻ 1 cm) were obtained from the upper margins of the lower uterine segments of consenting patients (n ⫽ 88). Tissue biopsies were obtained from both the lower and fundal segments of women undergoing hysterectomies during the cesarean section (n ⫽ 4). In addition, needle biopsies were taken from the upper uterine segment from 10 women requiring cesareans. Lower segment samples were also obtained for this group to allow paired comparisons. The samples were obtained by a research obstetrician who was not the patient’s obstetrician immediately on delivery of the baby. The tissues were freed of connective tissue and decidua and immediately snap-frozen in liquid nitrogen for storage at ⫺80 C until analysis. Women undergoing cesarean sections were examined for overt clinical signs of infection. High vaginal swabs for microbiological culture were performed. The placenta and membranes were also examined microscopically after delivery for infection such as chorioamnionitis. Women opting for elective cesarean sections were examined for signs of labor onset including uterine contractions and cervical changes. The indications included previous cesarean sections, breech and transverse lie. The indications for emergency cesarean sections included fetal distress, previous cesarean sections, breech, placental abruption, macrosomia, and prolonged labor or failure to progress. Both spontaneous and induced labors were sampled, and the duration of labor before surgery varied from 2–16 h. The women were matched for gestational and maternal ages.

RNA extraction Myometrial samples were pulverized under liquid nitrogen, and RNA was extracted by the guanidinium thiocyanate-acid phenol-chloroform method (11). Total RNA was treated with DNase (DNA free, Ambion, Geneworks Pty. Ltd., Thebarton, Australia). The RNA integrity and quality were checked on agarose gel to confirm the presence of strong and intact ribosomal 28 S and 18 S bands and the absence of large molecular weight DNA-sized material. Treated total RNA (1.5 ␮g) was reverse-transcribed with random primers using Superscript II kit (Life Technologies, Inc., Mulgrave, Australia).

Suppression subtractive hybridization (SSH) Lower segment myometrial samples from term (40 wk) nonlaboring and laboring women were used to produce two subtracted cDNA libraries. The nonlaboring sample was donated from an elective cesarean section with breech presentation, whereas the emergency/nonelective cesarean section resulted from a failure to progress after laboring for nearly 21 h. Total RNA was extracted and treated as above and validated by PCR for the housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase, to be free of genomic material. Total RNA (1 ␮g) was reversetranscribed using Superscript II (Life Technologies, Inc.) and SMART PCR cDNA synthesis kit (CLONTECH Laboratories, Inc., Palo Alto, CA) to produce two pools of full-length cDNA. These were used in two

The two subtracted cDNA libraries produced were enriched for genes up-regulated 1) in labor or 2) for quiescence (NIL). Initial screening of the white positive bacterial colonies obtained was performed by dot blot hybridizations. Random positive colonies of Escherichia coli from our subtracted libraries were grown in 96-well plates with 100 ␮l Luria Broth containing 50 mg/liter ampicillin at 37 C for 2 h with shaking. The bacterial cultures were then directly used for PCR with nested primers. The PCR was performed in a Perkin-Elmer GeneAmp PCR System 9600 machine (PE Applied Biosystems, Scoresby, Australia) using the following cycling parameters: 94 C for 30 sec, and 23 cycles of two-step PCR (95 C for 30 sec, 68 C for 3 min). The PCR products were first checked on a 2% agarose gel in Tris, borate, EDTA buffer to ensure that there was amplification of products. Then, the products that should correspond to the cDNA insert were denatured with NaOH before manually spotting onto four identical nylon membranes (Hybond N⫹, Amersham Pharmacia Biotech, Castlehill, Australia). To minimize false positive results, negative controls were used, and all spots were made in duplicate. Two negative controls, a testes-specific protein (GenBank accession no. X52128) and human semenogelin II mRNA (GB accession no. AN M81652) were also included. The dot blots were UV-cross-linked on a Spectrolinker SL-1000 UV cross-linker (Spectronics Corporation, Medos Co. Pty. Ltd., Lidcombe, Australia) before hybridization with four different cDNA probes. The probes included both the labor- and NIL-enriched cDNAs as well as the nonsubtracted cDNAs. The subtracted probes were made using the primary PCR products. Briefly, diluted primary PCR products were used as templates for limited cycles (10) of secondary PCR using nested primers. The products were purified using Wizard PCR Preps (Promega Corp., Annandale, Australia) before digestion with RsaI, EagI, and SmaI to remove the adaptors from the probes that would introduce high background. The adaptors were separated from the probes by using the Wizard Preps. The resulting adaptor-free subtracted cDNAs as well as the unsubtracted cDNAs are labeled by random primer method with digoxigenin (DIG) according to the manufacturer’s protocol (DIG High Prime DNA Labeling Kit, Roche Molecular Biochemicals, Castle Hill, Australia). For each dot blot screening, four identical blots were prehybridized in a proprietary mix (DIG Easy Hyb, Roche) for 2 h at 42 C in a hybridization oven (Hybaid Mini 10, Integrated Sciences). The probes were denatured by boiling and added to fresh hybridization mixes for overnight hybridizations at 42 C with continuous agitation. After overnight hybridizations, the blots were washed with low-stringency solutions (2⫻ SSC, 0.1% SDS) and warm (68 C) highstringency solutions (0.2⫻ SSD, 0.1% SDS). The blots were then detected by the protocol for DIG detection (DIG Wash and Block Buffer Set, Roche). Briefly, the membranes were treated with a blocking solution for 1 h before reaction with an anti-DIG-alkaline phosphatase antibody (1/10,000 dilution) for 30 min at room temperature with shaking. The blots were then washed before the chemiluminescent substrate disodium 3-(4-methoxyspiro[1,2-dioxetane-3,2⬘-(5⬘chloro)tricyclo[3.3.1.1(3,7)]decan]-4-yl) phenyl phosphate was added. After incubation at 37 C for 15 min, the blots were exposed to x-ray film (Biomax ML, Kodak, Amersham Pharmacia Biotech, Castlehill, Australia) for 15–30 min. The films were developed by an automated processor.

Chan et al. • Human Myometrial Genes

Clones identified as differentially expressed by dot blots were used as probes in Southern blots with RT-PCR products (virtual Northern blots) to confirm the differential expression before further investments of sequencing and analyses. The full-length cDNAs made from reverse transcribing the two initial RNAs were diluted and amplified by PCR. The PCR products from the labor-enriched and nonlabor-enriched cDNAs were electrophoresed on a 1% agarose gel and transferred onto nylon membrane (Hybond N⫹) using neutral (10⫻ SSC) downward transfer with a turboblotter (Schleicher & Schuell, Inc., Medos Co. Pty. Ltd., Lidcombe, Australia). Identical hybridization and stringency wash conditions were used for the Southern blots with RT-PCR products as for the dot blot analysis, except that each blot was screened with the specific probe identified by dot blot screening. Detection of the signals was also identical with the dot blots. Most of the Southern blots confirmed the differential expression of the specific genes in the two initial samples used for SSH. Those clones that passed the dot blot and Southern screenings were sent for sequencing in forward and reverse orientations using nested primers and M13 primers. Automated sequencing using fluorescent dye terminator chemistry on an ABI 377 automatic sequencer was performed by the Australian Genome Research Facility (Brisbane, Australia). Sequence analyses were performed using Blastn and FASTA for GenBank queries. Sequence alignments using ClustalW and Line-UP (Australian National Genomic Information Service) were used to compare our clones with published sequences to ascertain identities.

J Clin Endocrinol Metab, June 2002, 87(6):2435–2441 2437

sive reference dye, Rox. All RNA samples and controls were assayed in triplicate. Appropriate controls (no reverse-transcriptase, no template controls) were included. The cyclophilin and 18S rRNA assays were performed on the same samples in separate tubes. The IL-8 and MnSOD assays were multiplexed with 18S rRNA, i.e. simultaneous detection of both the target gene (either IL-8 or MnSOD) and 18S rRNA as the internal reference/control were performed in the same tube. This allows standardization of the amount of target gene to the internal reference gene to control for different amounts of cDNA used. The standardized target gene was then compared with an external reference or calibrator sample,

Northern blots To further confirm the identities of the clones obtained and their differential expression, Northern blots of gestational age-matched myometrial samples from women in labor (n ⫽ 3) and NIL (n ⫽ 3) were performed. The chemiluminescent DIG method was used. Total RNA (20 ␮g) was electrophoresed in denaturing formaldehyde gel (1.2% agarose) using 3[N-morholino]propanesulfonic acid buffer. The gel was blotted onto nylon membrane (Hybond N⫹) as for the Southern blots. The membranes were UV-cross-linked as for the dot blots, prehybridized with continuous agitation in Easy Hyb at 50 C for 2 h, then hybridized overnight at 50 C with specific probes. The probes were made by PCR of the specific clones from the subtracted library. The cDNA fragments were labeled by random priming (DIG High Prime) as for the dot blots. After overnight hybridizations, the blots were washed using low-stringency solutions (2⫻ SSC, 0.1% SDS) and warm high-stringency solutions (0.1⫻ SSC, 0.1% SDS) at 60 C. Chemiluminescent DIG detection was performed as for the dot blots. On completion of the first blot, the membranes were stripped using boiling 0.1% SDS and reprobed with DIG-labeled 18S probe (pTRI RNA 18S, Ambion, Inc.) to check for equivalent sample loading and integrity.

QRT-RT-PCR The extent of differential gene expression between labor and NIL tissues, as well as between upper and lower uterine segments, was investigated using QRT-RT-PCR to enable accurate quantification. Total RNA (1.5 ␮g) was treated with DNase to remove genomic contamination and reverse-transcribed as described under RNA extraction. Some RNA samples termed No RT Controls had no reverse transcriptase enzyme added as a further check for the absence of genomic DNA contamination in the samples. The cDNAs were diluted 10-fold, and specific mRNAs were measured in QRT-RT-PCR assays using Taqman probes 5⬘-fluorescent labeled with either 6-FAM or VIC in a thermal cycler (ABI Prism 7700 Sequence Detector system, PE Applied Biosystems) linked to a MacIntosh G4 (Apple Computer Inc., PE Applied Biosystems). Three predeveloped Taqman assays were used to quantify IL-8, cyclophilin, and 18 S rRNA as reference or control gene (PDAR, PE Applied Biosystems). In addition, we designed the following primers and probes for detection of manganese superoxide dismutase (MnSOD): Forward primer: 5⬘- AAGGGAGATGTTACAGCCCAGATA-3⬘ Reverse primer: 5⬘-TCCAGAAAATGCTATGATTGATATGAC-3⬘ Probe: 5⬘-6-FAM-CCACCATTGAACTTCAGTGCAGGCTG-TAMRA-3⬘ PCR was performed in 96-well plates using the Universal Master Mix (PE Applied Biosystems) comprising the thermo-stable polymerase AmpliTaq Gold, uracil-DNA-N-glycosylase, nucleotides, and an internal pas-

FIG. 1. Dot blot hybridization of putative cDNA clones from subtracted SSH libraries. The cDNA inserts in clones from the subtracted libraries were amplified by PCR and blotted onto four identical membranes. Each clone was spotted in duplicate. The dot blots shown are a representative hybridization with four probes (see SSH screening). The results with only two probes, from labor (top) and NIL (bottom) subtractions, are shown. The clones (circled) that reacted with the labor probe but not the NIL probe were investigated further. The two negative controls that were included (boxed) consistently showed no difference, regardless of probe used. TABLE 1. Genes identified by SSH to be up-regulated in labor GenBank accession no.

Known genes Oxytocin receptor X64878 MMP9 NM004994 Fibronectin U60068 IL-8 M28130 Genes not previously linked with labor MnSOD S77127 B23 M23613 IFN 1-8d X57351 EF1␣ J04617 Cyclophilin Y00052 ␣-Actin X13839 Four novel genes were identified with no matching sequences in databases.

2438

J Clin Endocrinol Metab, June 2002, 87(6):2435–2441

Chan et al. • Human Myometrial Genes

a placental cDNA pool that was used in every assay. The result of this comparative method of quantification is transformed to an exponential value, 2⫺⌬⌬Ct, where Ct refers to the cycle threshold, or the cycle when the product was first detected by the software (ABI Prism 7700 Sequence Detection System User Bulletin 2, PE Applied Biosystems). Standard cycling conditions for QRT-RT-PCR were used in all assays (ABI Prism 7700 Sequence Detection System User Bulletin 2, PE Applied Biosystems).

Statistical analyses The differences in RNA expressed between the upper and lower segments were analyzed using t test (paired). Correlation and regression analyses and ANOVA were performed on the RNA and clinical data.

Results

Dot blot screening of 400 clones revealed 30 differentially expressed clones. Figure 1 is a representative dot blot, showing the results with the subtracted probes. Only clones that reacted strongly with the labor-enriched probe (top panel) but not the nonlabor enriched probe (bottom panel) were followed up to minimize false positive screenings. The negative controls did not hybridize in all blots, demonstrating the specificity of hybridization. Sequence analyses of the positive clones revealed the identities of genes obtained, some of which were previously linked with parturition. Others, including novel sequences with no known function, had not been linked with labor (Table 1). Southern blots with the RT-PCR products confirmed the differential expression patterns evident in dot blot screening. In addition, the Southern blots revealed the sizes of the reacting cDNA species. This provided added useful confirmatory data when the identity of the genes was known from the DNA sequencing. Final confirmation of the results of the subtractions was obtained by Northern blot analyses using RNA extracted from three pairs of gestational-age-matched women either in labor or NIL (Fig. 2). Importantly, the data showed the expected sizes of transcripts for the genes examined including IL-8, MnSOD, cyclophilin, nucleophosmin (B23), elongation factor 1␣ (EF1␣), and interferon-inducible gene 1-8d (IFN 1-8d). Reprobing these blots with DIG-labeled 18S rRNA probe showed that equivalent amounts of RNA were loaded across all lanes so that the differences observed were due to the differential expression of the specific genes (data not shown). QRT-RT-PCR analyses showed that IL-8 was significantly elevated during labor, rising from 0.003 ⫾ 0.002 in nonlaboring samples (n ⫽ 38) to 0.24 ⫾ 0.11 in gestational-agematched spontaneously laboring women (n ⫽ 20; P ⫽ 0.035) (Fig. 3A). Similar significant elevations of cyclophilin [0.0008 ⫾ 0.0002 (n ⫽ 6) to 0.002 ⫾ 0.0004 (n ⫽ 6); P ⫽ 0.008] and MnSOD [0.11 ⫾ 0.02 (n ⫽ 24) to 1.23 ⫾ 0.56 (n ⫽ 24); P ⫽ 0.047] with labor were also observed (Fig. 3, B and C). MnSOD mRNA was differentially distributed between the upper (0.63 ⫾ 0.18) and lower (0.15 ⫾ 0.05; n ⫽ 15; P ⫽ 0.022) segments of the uterus (Fig. 4). Of the 15 paired samples from

FIG. 2. Northern blotting of human myometrial tissues. Lower segment myometrial samples from gestational age-matched women who were NIL (n ⫽ 3; lanes 1–3, from left to right, of each blot) or in labor (n ⫽ 3; lanes 4 – 6 in each blot) were probed with the specific genes as indicated: A, EF1␣ (1.1 kb mRNA); B, MnSOD (two splice forms of 4

and 0.9 kb, respectively); and C, B23 (1.3 kb) and IFN1-8d (900 bp). Blot C was stripped after hybridization with IFN1-8d and reprobed with B23. All blots were stripped and reprobed with DIG-18S rRNA to check that sample loading was similar for all samples (data not shown).

Chan et al. • Human Myometrial Genes

J Clin Endocrinol Metab, June 2002, 87(6):2435–2441 2439

observed for MnSOD mRNA (P ⫽ 0.1; data not shown). The data also showed that all myometrial samples from women who were not in labor had lower IL-8 mRNA compared with those obtained from laboring women at all gestational ages. Discussion

FIG. 3. QRT-RT-PCR shows that IL-8, cyclophilin, and MnSOD are significantly elevated during labor. The specific mRNAs were standardized against 18S rRNA. Individual values for each specific gene (spots) for patients who were clinically NIL or in labor as well as the concentration mean ⫾ SD data (bars) are shown for as follows: A, IL-8; B, cyclophilin; and C, MnSOD. The semilog plots are used to show individual sample values. The t test analyses of the results are represented by *, P ⫽ 0.035; ⫹, P ⫽ 0.008; and #, P ⫽ 0.047, respectively.

the subjects included in the analysis, 10 had higher fundal MnSOD, whereas the remaining 5 women had similar fundal and lower segment MnSOD. In contrast, IL-8 was not differentially distributed between the uterine segments (n ⫽ 17; P ⫽ 0.97; data not shown). IL-8 mRNA was significantly higher in induced labor (0.63 ⫾ 0.21; n ⫽ 14) than spontaneous labor (0.22 ⫾ 0.11; n ⫽ 20; P ⫽ 0.046) (Fig. 5), but there were no such differences

The screening of 20% of the SSH libraries showed that the subtractions were successful because a group of genes long associated with labor were found. These include oxytocin receptor, matrix metalloproteinase 9 (MMP9), fibronectin, and IL-8. A larger cohort of genes identified included new and known genes not previously linked with labor. This report also confirms previous data showing site-specific expression of genes, specifically that MnSOD, but not IL-8, is differentially expressed at the fundal and lower uterine segments. Among the 10 known labor up-regulated genes, there is extensive literature on oxytocin (3) and fibronectin (12) as having important roles in myometrial contractility and cervical remodeling, respectively. The role of metalloproteinases, especially MMP9, in parturition is also well documented. Labor, whether term or preterm, is associated with a significant increase in MMP9 enzyme activity and immunoreactivity (IR) in the human amnion, placenta, choriodecidua, amniotic fluid, and lower uterine segment (13–17). Our pilot data showing a significant MMP mRNA increase during labor thus parallels the reported protein expression and activity in the myometrium. Similarly, much data has been gathered on IL-8. IL-8 IR has been previously identified in human amnion, amniotic fluid, chorion, decidua, and villous placenta, and its level increases during labor (18 –21). Levels of IL-8 IR correlate with the release of collagenases, a crucial step that regulates the process of cervical extracellular matrix remodeling (17). Recently, IL-8 IR has also been found to be increased with labor in the myometrium (22). Interestingly, Thomson et al. (9) and Winkler et al. (22) reported that the lower segment of the myometrium from laboring women had massive infiltration of leukocytes and macrophages compared with nonlaboring gestational-age matched controls. This is consistent with the chemoattractant activity of IL-8, and it is likely that the myometrial IL-8 induced infiltration of the cervical stroma by neutrophils and a subsequent release of proteinases may play a key role in parturition. It has also been reported recently that IL-8 release from cervical explants was stimulated by prostaglandin E2 and nitric oxide and inhibited by progesterone and dexamethasone (23), data consistent with IL-8 being involved in parturition and with our finding of significantly higher IL-8 mRNA levels in induced labors compared with spontaneous labor. Our finding that IL-8 is not differentially expressed in the upper compared with the lower segment agrees with findings of similar distribution of leukocytes in both segments of the myometrium because the IL-8 is mainly derived from these cells (9). MnSOD, a mitochondrial antioxidant, metabolizes superoxide. Superoxide destructively combines with vasodilator nitric oxide (24). As such, MnSOD has been postulated to play a role in uterine quiescence. Its enzymic activity and immunostaining have been described in myometrium, pla-

2440

J Clin Endocrinol Metab, June 2002, 87(6):2435–2441

Chan et al. • Human Myometrial Genes

FIG. 4. MnSOD mRNA is differentially expressed at the upper and lower segments of the uterus during pregnancy. The specific mRNAs were standardized against 18S rRNA. Paired upper segment (US) and lower segment (LS) uterine samples from the same women are joined by lines. The t test analysis indicates that the difference is significant (P ⫽ 0.022).

FIG. 5. Myometrial IL-8 mRNA is significantly higher in induced labor compared with spontaneous labor. The specific mRNAs were standardized against 18S rRNA. The plot of relative IL-8 mRNA vs. gestational ages show that induced labor (circles) was associated with higher IL-8 mRNA than spontaneous labor (triangles); P ⫽ 0.046. Both laboring groups had significantly higher IL-8 mRNA than nonlaboring women (diamonds).

centa, decidua, and fetal membranes (24 –26). In the myometrium, it has been found in both myocytes and endothelial cells and some vascular smooth muscle cells, although Telfer et al. (25) did not find any changes in SOD enzyme activity or staining intensity in the tissues obtained before or during labor. This contrasts with our data indicating increased MnSOD mRNA levels during labor. The divergence of data may be due to differing methodological sensitivities. Importantly, we found MnSOD to be significantly higher in the upper segment of the uterus. This is consistent with a homeostatic role in response to reactive oxygen species produced as a by-product of the increased metabolic demands for parturition. The three other genes that we found to be up-regulated at labor were IFN 1-8d, EF1␣, and B23. They have not been previously linked with parturition. IFN 1-8d, found to be up-regulated in radiation-induced cell death (27), is a transmembrane mediator of interferon that exerts its effects through the induction of more than 30 different proteins with antiviral, antiproliferative, and immunomodulatory functions. Significantly, the presence of IFN in cervicovaginal fluid from late second trimester is an important risk factor for preterm labor (28). EF1␣ is a multifunctional protein well characterized as a cellular mediator of the transfer of aminoacyl-tRNA to ribosomes (29). Its expression has been found to decrease toward the end of the life span of fibroblasts (30), but in-

creased in tumors from pancreas, colon, breast, lung, and gastric mucosa (31). EF1␣ has another less well cited function, its activity in severing stable microtubules (32), and is thought to be involved in cytoskeletal reorganization. Its activity is stimulated by phorbol esters (33) and inhibited by didemnin B (34), a marine natural product with immuosuppressive activities in animals. B23 is a nucleolar phosphoprotein with a well characterized integral role in cell growth and signaling function. It is abundant in tumor cells and has been well studied in its role in the control of cellular apoptosis and immortalization (35). It is involved in ribosome synthesis and dissociates from nucleoli of cells after treatments with various anticancer drugs (36). Apart from its role as a molecular chaperone, B23 is one of two proteins found to endogenously immunize multiparous women against antigens purified from ovarian tumors (37). Sera from multiparous women were found to react consistently with B23 and EF1␣, which were among the antigens isolated from ovarian tumors. Thus, B23 has been postulated to have a role in protecting multiparous women from developing ovarian cancer, a suggestion that fits in with epidemiological data indicating reduced ovarian cancer risk with increased parity. Interestingly, there are two reports that suggest that B23 is an estrogen-regulated gene in human vascular smooth muscle cells (38) as well as an estrogenregulated protein associated with acquired estrogen-independence in human breast cancer cells (39). Because the

Chan et al. • Human Myometrial Genes

uterus is a major target organ of gestational estrogens, B23 may have important effects in parturition. Interestingly, the remaining two labor up-regulated genes, cyclophilin and ␣-actin, are often measured as constitutive or housekeeping gene markers. It is perhaps not surprising that they are up-regulated in the myometrium during labor because the laboring uterus is in a highly metabolically active state, requiring cyclophilin to accelerate folding of enzymes and protein into working structures, and more ␣-actin for sustained contractions. SSH has been shown to be a productive technique for studies into parturition in sheep myometrium (40), human amnion (19), and human placenta (41). Together with these studies, our data demonstrate the power of SSH as a discovery tool. Our data indicate that a major proportion of genes up-regulated during human labor are associated with inflammatory-immune pathways and suggest an inflammatory mechanism for parturition. They also suggest that many factors associated with human labor and their complex interactions remain to be elucidated. Acknowledgments Received September 20, 2001. Accepted January 13, 2002. Address all correspondence and requests for reprints to: E.-Cheng Chan, Mothers and Babies Research Center, Endocrine Unit, John Hunter Hospital, Hunter Region Mail Centre, Locked Bag 1, New South Wales 2310, Australia. E-mail: [email protected]. This work was supported by a grant from the Australian New South Wales Department of Health. The Mothers and Babies Research Center is part of the Hunter Medical Research Institute.

References 1. Lo´pez Bernal A, Rivera J, Europe-Finner GN, Phaneuf S, Asbo´th G 1995 Parturition: activation of stimulatory pathways or loss of uterine quiescence? Adv Exp Med Biol 395:435– 451 2. Smith R 1999 The timing of birth. Sci Am 280:68 –75 3. Giraldi A, Enevoldsen AS, Wagner G 1990 Oxytocin and the initiation of parturition: a review. Dan Med Bull 37:377–383 4. Gibb W 1998 The role of prostaglandins in human parturition. Ann Med 30:235–241 5. Challis JRG, Matthews SG, Gibb W, Lye SJ 2000 Endocrine and paracrine regulation of birth at term and preterm. Endocr Rev 21:514 –550 6. Ingram JN, Shaw G, Renfree MB 1999 Cortisol in fetal fluids and the fetal adrenal at parturition in the tammar wallaby (Macropus eugenii) Biol Reprod 60:651– 655 7. Shaw G, Renfree MB, Fletcher TP 1996 A role for glucocorticoids in parturition in a marsupial, Macropus eugenii. Biol Reprod 54:728 –733 8. Sparey C, Robson SC, Bailey J, Lyall F, Europe-Finner GN 1999 The differential expression of myometrial connexin-43, cyclooxygenase-1 and -2, and Gs ␣ proteins in the upper and lower segments of the human uterus during pregnancy and labor. J Clin Endocrinol Metab 84:1705–1710 9. Thomson AJ, Telfer JF, Young A, Campbell S, Stewart CJ, Cameron IT, Greer IA, Norman JE 1999 Leukocytes infiltrate the myometrium during human parturition: further evidence that labour is an inflammatory process. Hum Reprod 14:229 –236 10. Mizrahi J, Karni Z, Polishuk WZ 1977 Strain uterography in labour. Br J Obstet Gynaecol 84:930 –936 11. Chomczynski P, Sacchi N 1987 Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162: 156 –159 12. Leppert PC 1995 Anatomy and physiology of cervical ripening. Clin Obstet Gynecol 38:267–279 13. Tsatas D, Baker MS, Rice GE 1999 Differential expression of proteases in human gestational tissues before, during and after spontaneous-onset labour at term. J Reprod Fertil 116:43– 49 14. Athayde N, Romero R, Gomez R, Maymon E, Pacora P, Mazor M, Yoon BH, Fortunato S, Menon R, Ghezzi F, Edwin SS 1999 Matrix metalloproteinases-9 in preterm and term human parturition. J Matern Fetal Med 8:213–219 15. Osmers RG, Bla¨ser J, Kuhn W, Tschesche H 1995 Interleukin-8 synthesis and the onset of labor. Obstet Gynecol 86:223–229

J Clin Endocrinol Metab, June 2002, 87(6):2435–2441 2441

16. Osmers RG, Adelmann-Grill BC, Rath W, Stuhlsatz HW, Tschesche H, Kuhn W 1995 Biochemical events in cervical ripening dilatation during pregnancy and parturition. J Obstet Gynaecol 21:185–194 17. Winkler M, Oberpichler A, Tschesche H, Ruck P, Fischer DC, Rath W 1999 Collagenolysis in the lower uterine segment during parturition at term: correlations with stage of cervical dilatation and duration of labor. Am J Obstet Gynecol 181:153–158 18. Keelan JA, Marvin KW, Sato TA, Coleman M, McCowan LM, Mitchell MD 1999 Cytokine abundance in placental tissues: evidence of inflammatory activation in gestational membranes with term and preterm parturition. Am J Obstet Gynecol 181:1530 –1536 19. Tashima LS, Millar LK, Bryant-Greenwood GD 1999 Genes upregulated in human fetal membranes by infection or labor. Obstet Gynecol 94:441– 449 20. Laham N, Rice GE, Bishop GJ, Ransome C, Brennecke SP 1993 Interleukin 8 concentrations in amniotic fluid and peripheral venous plasma during human pregnancy and parturition. Acta Endocrinol (Copenh) 129:220 –224 21. Romero R, Ceska M, Avila C, Mazor M, Behnke E, Lindley I 1991 Neutrophil attractant/activating peptide-1/interleukin-8 in term and preterm parturition. Am J Obstet Gynecol 165:813– 820 22. Winkler M, Fischer DC, Ruck P, Marx T, Kaiserling E, Oberpichler A, Tschesche H, Rath W 1999 Parturition at term: parallel increases in interleukin-8 and proteinase concentrations and neutrophil count in the lower uterine segment. Hum Reprod 14:1096 –1100 23. Denison FC, Calder AA, Kelly RW 1999 The action of prostaglandin E2 on the human cervix: stimulation of interleukin 8 and inhibition of secretory leukocyte protease inhibitor. Am J Obstet Gynecol 180:614 – 620 24. Boggess KA, Oury TD, Kay HH, Crapo JD 1998 Extracellular superoxide dismutase localization and activity within the human placenta. Placenta 19:417– 422 25. Telfer JF, Thomson AJ, Cameron IT, Greer IA, Norman JE 1997 Expression of superoxide dismutase and xanthine oxidase in myometrium, fetal membranes and placenta during normal human pregnancy and parturition. Hum Reprod 12:2306 –2312 26. Myatt L, Eis AL, Brockman DE, Kossenjans W, Greer IA, Lyall F 1997 Differential localization of superoxide dismutase isoforms in placental villous tissue of normotensive, pre-eclamptic, and intrauterine growth-restricted pregnancies. J Histochem Cytochem 45:1433–1438 27. Clave E, Carosella ED, Gluckman E, Socie´ G 1997 Radiation-enhanced expression of interferon-inducible genes in the KG1a primitive hematopoietic cell line. Leukemia 11:114 –119 28. Chuileanna´in FN, Brennecke S 1998 Prediction of preterm labour in multiple pregnancies. Baillieres Clin Obstet Gynaecol 12:53– 66 29. Reed VS, Wastney ME, Yang DC 1994 Mechanisms of the transfer of aminoacyl-tRNA from aminoacyl-tRNA synthetase to the elongation factor 1 ␣. J Biol Chem 269:32932–32936 30. Cavallius J, Rattan SI, Clark BF 1986 Changes in activity and amount of active elongation factor 1 ␣ in aging and immortal human fibroblast cultures. Exp Gerontol 21:149 –157 31. Grant AG, Flomen RM, Tizard ML, Grant DA 1992 Differential screening of a human pancreatic adenocarcinoma ␭ gt11 expression library has identified increased transcription of elongation factor EF-1 ␣ in tumour cells. Int J Cancer 50:740 –745 32. Shiina N, Gotoh Y, Kubomura N, Iwamatsu A, Nishida E 1994 Microtubule severing by elongation factor 1 ␣. Science 266:282–285 33. Derventzi A, Rattan SI, Clark BF 1993 Phorbol ester PMA stimulates protein synthesis and increases the levels of active elongation factors EF-1 ␣ and EF-2 in ageing human fibroblasts. Mech Ageing Dev 69:193–205 34. Crews CM, Collins JL, Lane WS, Snapper ML, Schreiber SL 1994 GTPdependent binding of the antiproliferative agent didemnin to elongation factor 1 ␣. J Biol Chem 269:15411–15414 35. Finch RA, Chan PK 1996 ATP depletion affects NPM translocation and exportation of rRNA from nuclei. Biochem Biophys Res Commun 222:553–558 36. Chan PK, Bloom DA, Hoang TT 1999 The N-terminal half of NPM dissociates from nucleoli of HeLa cells after anticancer drug treatments. Biochem Biophys Res Commun 264:305–309 37. Shields LB, Gercel-Taylor C, Yashar CM, Wan TC, Katsanis WA, Spinnato JA, Taylor DD 1997 Induction of immune responses to ovarian tumor antigens by multiparity. J Soc Gynecol Investig 4:298 –304 38. Koike H, Karas RH, Baur WE, O’Donnell Jr TF, Mendelsohn ME 1996 Differential-display polymerase chain reaction identifies nucleophosmin as an estrogenregulated gene in human vascular smooth muscle cells. J Vasc Surg 23:477– 482 39. Skaar TC, Prasad SC, Sharareh S, Lippman ME, Brunner N, Clarke R 1998 Two-dimensional gel electrophoresis analyses identify nucleophosmin as an estrogen regulated protein associated with acquired estrogen-independence in human breast cancer cells. J Steroid Biochem Mol Biol 67:391– 402 40. Wu WX, Zhang Q, Ma XH, Unno N, Nathanielsz PW 1999 Suppression subtractive hybridization identified a marked increase in thrombospondin-1 associated with parturition in pregnant sheep myometrium. Endocrinology 140:2364 –2371 41. Morrish DW, Linetsky E, Bhardwaj D, Li H, Dakour J, Marsh RG, Paterson MC, Godbout R 1996 Identification by subtractive hybridization of a spectrum of novel and unexpected genes associated with in vitro differentiation of human cytotrophoblast cells. Placenta 17:431– 441