Download - Steroidal Regulation of Uterine Edema and Tissue Inhibitors ...

500

BIOLOGY OF REPRODUCTION 70, 500–508 (2004)Published online before print 20 October 2003.DOI 10.1095/biolreprod.103.020834

Steroidal Regulation of Uterine Edema and Tissue Inhibitors of Metalloproteinase(TIMP)-3 Messenger RNA Expression Is Altered in TIMP-1-Deficient Mice1

Warren B. Nothnick,2,3,4 Xuan Zhang,3 and Han-E Zhou3

Departments of Obstetrics and Gynecology, Division of Basic and Clinical Women’s Research,3 and Molecular andIntegrative Physiology,4 University of Kansas Medical Center, Kansas City, Kansas 66160

ABSTRACT

Tissue inhibitors of metalloproteinases (TIMPs) are expressedwithin the uteri of virtually all species where they are postulatedto control extracellular matrix turnover, cellular apoptosis, andproliferation. The objective of the current study was to examinethe steroidal regulation of uterine TIMP expression and to de-termine the potential role of the TIMP-1 gene product in thisregulation. To accomplish these goals, ovariectomized femaleTIMP-1 wild-type and null mice were treated with estradiol, pro-gesterone, or estradiol and progesterone and killed at varioustimes after steroid administration. Estradiol induced a significantreduction in uterine TIMP-3 expression in wild-type mice at 8and 24 h post-steroid administration, but the ability of this ste-roid to decrease TIMP-3 expression was impaired in the uteri ofTIMP-1 null mice. Further, estrogen-induced uterine wet-weightgain/edema was enhanced in the TIMP-1 null mice, and the an-tiestrogen compound ICI 182,780 or progesterone could onlypartially block this estrogenic effect. It is concluded from thisstudy that steroidal modulation of uterine TIMP-3 expressionand regulation of wet-weight gain/edema are altered in TIMP-1null mice. These observations suggest that steroids induce uter-ine TIMP-1 expression and, in turn, that TIMP-1 influences TIMP-3 mRNA expression and uterine edema.

estradiol, female reproductive tract, progesterone, uterus

INTRODUCTION

Within the uterus, the female sex steroids estrogen andprogesterone play pivotal roles in the establishment of asuitable environment for embryo implantation and preg-nancy. More specifically, these steroids regulate a multitudeof cellular processes that include cell proliferation and dif-ferentiation as well as regulation of vascular permeability,angiogenesis, and adenogenesis [1–7]. To bring about thesechanges, estrogen and progesterone must appropriatelymodulate a variety of factors that include growth factors,cytokines, extracellular matrix proteins, and adhesion mol-ecules [1–7]. One such family of factors that has been pos-tulated to play a role in uterine physiology and the estab-lishment of pregnancy is the tissue inhibitors of metallo-proteinases, or TIMPs [8–16].

TIMPs are a multifunctional family of inhibitors of ma-trix metalloproteinases (MMPs) that include four distinct

1This research was supported by NIH grant award HD39765 to W.B.N.2Correspondence: Warren B. Nothnick, University of Kansas MedicalCenter, Department of Obstetrics and Gynecology, 3901 Rainbow Blvd.,Kansas City, KS 66160. FAX: 913 588 6271; e-mail: [email protected]

Received: 9 July 2003.First decision: 25 July 2003.Accepted: 14 October 2003.Q 2004 by the Society for the Study of Reproduction, Inc.ISSN: 0006-3363. http://www.biolreprod.org

members: TIMP-1, TIMP-2, TIMP-3, and TIMP-4 [8, 17,18]. All but TIMP-4 are expressed within the uterus of avariety of species that include mice [19], sheep [20], hu-mans [21–25], and nonhuman primates [26, 27]. TIMPs areknown modulators of cell proliferation, differentiation, andapoptosis [8, 17] and elicit these activities through pro-cesses that are either dependent on or independent of theirability to regulate MMP activity [17]. In the human, TIMPsare expressed within the uterus during the course of themenstrual cycle and have been postulated to play a role inthe maintenance of endometrial integrity [21]. Studies usingTIMP-1 null mice have suggested a similar role of thisTIMP during the course of the rodent estrous cycle [19,28], and loss of function of this TIMP is associated withsubfertility [19, 28], reduction in reproductive life span[28], and altered uterine MMP [29] and TIMP [19] patternsof expression.

Uterine TIMPs expression appears to be regulated bysteroids in a species-specific fashion. For example, in vivostudies in the cycling sheep uterus suggest that expressionof TIMP-1 is down-regulated by estrogen, while that ofTIMP-2 may be up-regulated by progesterone [20]. In con-trast, in nonhuman primates, progesterone withdrawal [25,26] is associated with a rapid increase in uterine TIMP-1expression followed by a reduction in expression. Resultsfrom in vitro studies that incorporated human endometrialstromal cells are conflicting. Initial studies by Salamonsenand colleagues [27] suggested that withdrawal of proges-terone did not influence TIMP-1 or TIMP-2 expression. Amore recent study by this group [30] using a long-termculture of human endometrial stromal cells and progester-one analogs indicated that TIMP-1 expression was in-creased. A similar up-regulation of TIMP-3 by progesteronewas found in vitro [23]. While the majority of the datawould indicate that uterine TIMPs are regulated by steroids,data are conflicting. These deviations are most likely theresult of differences in experimental designs (in vitro vs. invivo studies, short-term culture vs. long-term culture) aswell as possible species differences.

Our recent observation [19] that the pattern of murineuterine TIMPs fluctuate with the estrous cycle suggested tous, that like in other species, uterine TIMPs in the mousemight be regulated by steroids. This observation, coupledwith the finding that uterine TIMP expression appears tobe altered in TIMP-1 null mice [19], led to the followingseries of experiments that were designed to examine thesteroidal regulation of uterine TIMPs expression and thecontribution of TIMP-1 to this process. In the current study,wild-type and TIMP-1-deficient mice were utilized, and therole of estrogen and progesterone in regulation of uterineTIMPs expression was assessed. Unexpectedly, we foundthat although estrogen and progesterone regulated TIMP-1,TIMP-2, and TIMP-3 mRNA expression in wild-type uteri

Dow

nloaded from https://academ

ic.oup.com/biolreprod/article/70/2/500/2712788 by guest on 08 M

arch 2022

501ALTERATION IN STEROIDOGENIC RESPONSES IN TIMP-1 NULL UTERI

FIG. 1. Steroidal regulation of uterine TIMP-1 mRNA expression in TIMP-1 wild-type mice. Mice were treated with either estrogen (A) or estrogen andprogesterone (B) and killed at 0, 2, 4, 6, 8, or 24 h after steroid administration. Total mRNA was extracted and subsequently analyzed for TIMP-1expression as described in Materials and Methods using 10 mg of total RNA/lane. TIMP-1 mRNA was detected as a single transcript of approximately0.9 kb consistent with previous reports. TIMP-1 mRNA expression is expressed as the mean ratio of TIMP-1/18S transcript 6 SEM and is reported asoptical density (OD) units for four separate observations (n 5 4 mice/treatment group). Autoradiographic exposures for TIMP-1 were for 24 h at 2758C,while that of 18S rRNA was for 1–3 h at 2758C. Different letters indicate statistical significance (P , 0.05) among treatment groups as determined byone-way ANOVA. A transcript consistent with that of TIMP-1 was not detected in the null mice (data not shown).

through what appeared to be classical steroid receptor-me-diated mechanisms, regulation of TIMP-3 mRNA was al-tered in TIMP-1 null mice. Similarly, estrogenic regulationof uterine edema also appeared to be altered in the nullmice. As such, these findings may be interpreted to suggestthat steroidal induction of TIMP-1 within the uterus impartsa subsequent role for this TIMP in maintaining selectivesteroidal responses within this organ.

MATERIALS AND METHODS

Animals

TIMP-1 null and wild-type mice were utilized for all studies. TIMP-1-deficient animals (SVTER 129 background) were generated by homol-ogous recombination of a neocontaining gene-targeting vector in mouseembryonic stem cells. Transmission of the mutant allele and the genotypeof mice were determined by polymerase chain reaction analysis of theneosequences in genomic tail DNA. TIMP-1 deficiency was confirmed atthe transcript and protein level by Northern analysis and protease inhibitorassays, respectively [19].

A breeding colony of both genotypes was generated at the Universityof Kansas Medical Center. Mice were housed within environmentally con-trolled conditions under the supervision of a licensed veterinarian. Allanimal procedures for these experiments were approved by the Universityof Kansas Medical Center Institutional Animal Care and Use Committee(IACUC). Mice were maintained on a 14L:10D cycle and provided waterand mice chow ad libitum. Eight- to 12-wk-old female mice of both ge-notypes were ovariectomized and rested for 14 days, after which animalswere randomly assigned to the respective treatment groups.

Administration of Treatments and Tissue HarvestingFourteen days after ovariectomy, wild-type and TIMP-1 null mice were

injected s.c. with either vehicle (0.1 ml sesame oil), estradiol-17b (E2, 10mg/kg body weight [BW]), progesterone (P4, 100 mg/kg BW), or E2 1P4 (previous doses). Animals were then killed at 2, 4, 6, 8, and 24 h aftersteroid treatment by cervical dislocation, and uteri were removed, trimmedof fat and connection, and weighed. Uteri were then either snap frozen inliquid nitrogen or stored in RNAlater (Ambion Inc., Austin, TX) untilutilized for RNA extraction.

To verify that E2 and P4 regulation of uterine TIMPs expression oc-curred via activation of their specific receptors, the E2 receptor antagonistICI-182,780 (ICI; Tocris Cookson Inc., Ellisville, MO) and the P4 receptorantagonist RU-486 (Mifepristone; Dr. A.F. Parlow, NIDDK’s NationalHormone and Pituitary Program) were used. Mice were injected s.c. witheither ICI (20 mg/kg dissolved in 100% ethanol and resuspended in sesameoil), RU-486 (20 mg/kg dissolved in 100% ethanol and resuspended insesame oil), or vehicle (ethanol 1 oil). Thirty minutes later, mice receivedeither oil vehicle, E2, P4, or E2 1 P4 (previous doses), and mice were thenkilled at 6 (for TIMP-2), 8 (for TIMP-1), and 8 and 24 h (for TIMP-3)after steroid administration. These time points were chosen, as these weredetermined to be the times for maximal changes in uterine TIMP-1, TIMP-2, and TIMP-3 mRNA expression in the first experiment.

RNA IsolationTotal RNA was isolated from uteri by separately homogenizing the

tissue in 1 ml of TRIZOL reagent (Invitrogen Life Technologies, Carlsbad,CA) per 100 mg of tissue wet weight. RNA was then extracted withchloroform and precipitated with isopropyl alcohol according to the rec-ommendations of the manufacturer. Total RNA samples were then electro-phoresed through 1.0% agarose gels containing 2.2 M formaldehyde andwere transferred to nylon membranes (Nytran, Schleicher and Schuell,

Dow



arch 2022

502 NOTHNICK ET AL.

FIG. 2. The effects of estrogen and progesterone receptor antagonists onuterine TIMP-1 expression. Animals were treated the estrogen receptor(ICI) or progesterone receptor (RU-486) antagonists followed 30 min laterby treatment with either E2 or E2 1 P4 (as indicated in the figure). Micewere killed at 8 h after steroid administration, and TIMP-1 mRNA ex-pression was analyzed as described in Materials and Methods using 10mg of total RNA/lane. TIMP-1 mRNA expression is expressed as the meanratio of TIMP-1/18S transcript 6 SEM and is reported as optical density(OD) units for five separate observations (n 5 5 mice/treatment group).Autoradiographic exposures for TIMP-1 were for 24 h at 2758C. Differentletters indicate statistical significance (P , 0.05) among treatment groupsas determined by one-way ANOVA.

Keene, NH) as recommended by the manufacturer. The murine TIMP-1,TIMP-2, TIMP-3, and TIMP-4 cDNA probes (kindly provided by Dr. Dy-lan Edwards, University of East Anglia, Norwich, UK) were excised fromtheir respective plasmids with the appropriate restriction endonucleases,and the resulting inserts were labeled using Ready-To-Go DNA labelingbeads (Amersham Pharmacia Biotech, Inc., Piscataway, NJ). Probes werelabeled to a specific activity of 5 3 108 to 1 3 1010 dpm/mg of DNAusing [a-32P] dCTP (Perkin-Elmer Life Sciences, Boston, MA). Filterswere hybridized overnight, washed, and exposed to Blue Sensitive auto-radiography film (Midwest Scientific, St. Louis, MO) for 24 h at 2758C.Hybridization signals were allowed to decay, and filters were subsequentlyhybridized for the 18S transcript using a rat cDNA probe (kindly providedby Dr. Michael Melner, Vanderbilt University, Nashville, TN) that crosshybridizes with the mouse transcript. In all experiments, TIMP data werenormalized to the relative expression of the 18S transcript for each of thestudy groups. Data were digitized and quantitated using the GDS-8000System (Ultra Violet Products, Upland, CA).

Statistical Analysis

All data were analyzed across treatment regimes within genotype byone-way ANOVA. When an F-test indicated statistical significance, posthoc analysis was done using the Student-Newman-Keuls procedure.Planned comparisons between genotypes within each specific treatmentgroup were made using unpaired t-tests. Significance was set at P , 0.05for all comparisons [31].

RESULTS

Steroidal Regulation of Uterine TIMP-1 TranscriptExpression in Wild-Type Mice

TIMP-1 mRNA was not detected in the uteri of ovari-ectomized mice by Northern analysis using 10 mg of totalRNA per lane. However, administration of exogenous E2significantly increased TIMP-1 steady-state mRNA levelsin a time-dependent manner. TIMP-1 transcript expressionwas detectable as early as 2 h post-E2 administration, peak-ed between 4 and 8 h post-E2 administration, and then de-creased by 24 h after E2 administration (Fig. 1A). Admin-istration of E2 plus P4 also induced a time-dependent in-crease in TIMP-1 steady-state mRNA expression, but thisexpression peaked earlier (6 h post-steroid administration)compared to E2 administration alone (Fig. 1B). Progester-one administered alone had no effect on uterine TIMP-1expression, as levels remained undetectable by Northernanalysis at all time points (data not shown). These datawere interpreted to suggest that E2 but not P4 could increaseuterine TIMP-1 steady-state mRNA levels. To verify thatE2 action was mediated through its receptor proteins, a sec-ond experiment was conducted using the E2 receptor an-tagonist, ICI 182,780 (ICI). As expected, pretreatment ofmice with ICI resulted in a significant reduction in uterineTIMP-1 steady-state mRNA expression in response to E2treatment (Fig. 2). This suppression of TIMP-1 expressionwas also detected in mice pretreated with ICI followed byE2 1 P4. In contrast, mice pretreated with the P4 antagonistRU-486 then challenged with E2 1 P4 expressed similarlevels of TIMP-1 transcript compared to E2 and E2 1 P4treated mice (Fig. 2). Finally, treatment with either ICI orRU-486 alone had no effect on uterine TIMP-1 transcrip-tion. Transcript expression was similar to vehicle-treatedanimals, as TIMP-1 mRNA steady-state levels were unde-tectable by Northern analysis (data not shown).

Steroidal Regulation of Uterine TIMP-2 TranscriptExpression in Wild-Type and TIMP-1 Null Mice

Both the 3.5- and the 1.0-kilobase (kb) transcripts forTIMP-2 were detected in the uteri of both wild-type andTIMP-1 null mice (Fig. 3, A and B). Levels of both tran-scripts were initially analyzed individually and did not sig-nificantly differ among treatment groups or between ge-notypes. As such, TIMP-2 expression is reported as themRNA steady-state levels for both transcripts combined. Incontrast to TIMP-1, TIMP-2 expression was highest in theuteri of wild-type ovariectomized (0 h, oil-treated) mice,and E2 administration decreased TIMP-2 expression. Spe-cifically, E2 significantly reduced uterine TIMP-2 expres-sion at 4, 6, and 8 h post-steroid administration comparedto 0-h time points (Fig. 3A) with maximal suppression at6 h after E2 treatment. By 24 h post-E2 administration,uterine TIMP-2 transcript expression returned to controllevels (Fig. 3A). A similar pattern of TIMP-2 regulationwas induced by E2 1 P4 treatment, but the suppressiveeffect of both steroids was more evident at 4 h post-steroidadministration compared to 6 h for E2 alone. In TIMP-1null mice, the pattern of E2 and E2 1 P4 regulation ofuterine TIMP-2 mRNA expression was similar to that de-tected in wild-type mice with no significant differences inTIMP-2 transcript expression between genotypes withintime points for either steroid treatments (Fig. 3, A and B).Finally, P4 alone did not influence TIMP-2 expression in

Dow



arch 2022


FIG. 3. Steroidal regulation of uterine TIMP-2 mRNA expression in TIMP-1 wild-type and null mice. Mice were treated with either estrogen (A) orestrogen and progesterone (B) and killed at 0, 2, 4, 6, 8, or 24 h after steroid administration. Total mRNA was extracted and subsequently analyzed forTIMP-2 expression as described in Materials and Methods using 10 mg of total RNA/lane. TIMP-2 mRNA was detected as two transcripts of approximately3.5 kb (upper band) and 1.0 kb (lower band) consistent with previous reports. TIMP-2 mRNA expression is expressed as the mean ratio of TIMP-2/18Stranscript 6 SEM for both transcripts combined (3.5 and 1.0 kb) and is reported as optical density (OD) units for four separate observations (n 5 4mice/genotype/treatment group). Autoradiographic exposures for TIMP-2 were for 24 h at 2758C. Different letters indicate statistical significance amongthe treatment groups as determined by one-way ANOVA (block letters indicate comparisons within wild-type mice, while italics indicate comparisonswithin null mice). For all comparisons, P , 0.05 was considered statistically significant.

mice of either genotype at any time point assessed (datanot shown).

To verify that steroidal modulation of uterine TIMP-2steady-state mRNA expression was mediated through therespective steroid receptor proteins, mice were treated withICI 182,780 and RU-486 and killed 6 h after steroid ad-ministration. Pretreatment with ICI 182,780 blocked the E2-induced down-regulation of TIMP-2 transcript expressionin mice of both genotypes, as expression levels were similarto those of vehicle controls (Fig. 4). The E2 1 P4 reductionin steady-state levels of TIMP-2 mRNA expression ap-peared to be due to E2 action but not P4. This suggestionis based on the observation that pretreatment with ICIblocked the E2 1 P4 down-regulation of TIMP-2, but pre-treatment with RU486 failed to restore TIMP-2 expressionto those similar to the 0-h time point (Fig. 4). Finally, treat-ment with either ICI or RU-486 alone had no effect onuterine TIMP-2 steady-state mRNA levels (data notshown).


TIMP-3 was detected as a major transcript of 4.5 kb inthe uteri of both wild-type and TIMP-1 null mice (Fig. 5).In wild-type mice, E2 administration induced a slight but

significant increase in uterine TIMP-3 expression at 2, 4,and 6 h after steroid treatment. At 8 and 24 h, E2 signifi-cantly reduced uterine TIMP-3 steady-state mRNA levelsto at or below the 0-h levels (Fig. 5A). In TIMP-1 nullmice, E2 again induced increases at 2, 4, and 6 h post-steroid administration. However, in contrast to TIMP-3 ex-pression in wild-type mice, E2 treatment did not suppressuterine TIMP-3 steady-state mRNA levels at 8 h post-ste-roid administration, but a reduction was noted by 24 h post-E2 treatment (Fig. 5). Administration of E2 1 P4 resultedin a similar pattern of TIMP-3 expression with increases intranscript steady-state levels at 2, 4, and 6 h after admin-istration. Again at 8 h posttreatment, TIMP-3 expressionwas not suppressed by E2 1 P4 in the null mice as it wasin the wild-type mice. By 24 h post-steroid administration,TIMP-3 expression was reduced compared to the 2-, 4-, or6-h groups in mice of both genotypes. Further, the reductionof TIMP-3 expression by E2 1 P4 was not as marked as itwas for E2 alone in mice of either genotype, indicating thatat 24 h, P4 may be capable of blocking E2 down-regulationof TIMP-3 steady-state mRNA levels. Finally, P4 alone didnot influence TIMP-3 steady-state mRNA levels in mice ofeither genotype at any time point assessed (data not shown).

To verify that steroidal modulation of uterine TIMP-3was mediated through the respective steroid receptor pro-

Dow



arch 2022

504 NOTHNICK ET AL.

FIG. 4. The effects of estrogen and progesterone receptor antagonists onuterine TIMP-2 expression. Animals were treated the estrogen receptor(ICI) or progesterone receptor (RU-486) antagonists followed 30 min laterby treatment with either E2 or E2 1 P4 (as indicated in the figure). Micewere killed at 6 h after steroid administration, and TIMP-2 mRNA ex-pression was analyzed as described in Materials and Methods using 10mg of total RNA/lane. TIMP-2 mRNA expression is expressed as the meanratio of TIMP-2/18S transcript 6 SEM and is reported as optical density(OD) units for four separate observations (n 5 4 mice/treatment group).Autoradiographic exposures for TIMP-2 were for 24 h at 2758C. Differentletters indicate statistical significance (P , 0.05) among treatment groupsas determined by one-way ANOVA.

teins, mice of both genotypes were treated with ICI 182,780or RU-486, challenged with steroids, and killed 8 and 24 hfollowing steroid administration. In wild-type mice, E2 in-duced an approximate 35% decrease (P . 0.05) in TIMP-3 mRNA expression at 8 h posttreatment (Fig. 6A), whichwas in agreement with time course studies (Fig. 5). Pre-treatment with ICI 182,780 blocked this E2-induced de-crease in wild-type mice. In contrast, E2 did not influenceTIMP-3 expression in null mice at 8 h posttreatment, astranscript levels were similar to vehicle treated mice (Fig.6A). When compared within treatment between genotypes,E2 induced a significant reduction in TIMP-3 mRNA ex-pression in the wild-type mice (Fig. 6A; compare 1/1 to2/2 E2 groups). Finally, treatment with E2 1 P4 alone orwith antagonists did not influence TIMP-3 mRNA steady-state levels in mice of either genotype (Fig. 6A).

At 24 h, E2 induced a significant reduction in TIMP-3steady-state mRNA levels in mice of both genotypes, andthis level of expression did not differ between genotypes(Fig. 5). Pretreatment with ICI confirmed that this reductionwas via an E2 receptor-specific pathway in both wild-typeand null mice (Fig. 6B). Also in accord with the time coursestudies (Fig. 5), E2 1 P4 treatment restored TIMP-3 steady-state mRNA levels similar to those of vehicle-treated mice

and greater than those levels detected in mice treated withE2 alone (Fig. 6B). This regulation was similar in mice ofboth genotypes and suggests that P4 could block the E2-induced suppression. This was confirmed by RU-486 pre-treatment, as TIMP-3 steady-state mRNA levels were sig-nificantly reduced in the E2 1 P4 1 RU486 group com-pared to the E2 1 P4 group in both wild-type and null mice(Fig. 6B).


TIMP-4 was not expressed within the uterus of eitherwild-type or TIMP-1 null mice under control or any hor-monal treatment regime (data not shown). In accord withour previous findings [32], positive ovarian control tissuedid express TIMP-4 mRNA (data not shown), verifying infact that uterine TIMP-4 transcript is undetectable byNorthern analysis using 10 mg of total RNA.

Estrogen Induction of Uterine Wet-Weight Gain IsPartially Independent of Estrogen Receptor-SpecificPathways in TIMP-1 Null Mice

As reproductively cycling TIMP-1 null mice have largeruteri compared to wild-type counterparts [19], the effect ofsteroid treatment on uterine wet-weight gain/edema was as-sessed. E2 significantly increased uterine wet weight abovevehicle levels in mice of both genotypes (Fig. 7). Compar-ison between genotypes within treatment group revealedthat this E2-induced increase in wet weight was signifi-cantly greater in the TIMP-1 null mice compared to wild-type counterparts. In addition, the antiestrogen ICI 182,780blocked the E2-induced increase in uterine wet weight inwild-type mice, reducing uterine weight wet to similar val-ues compared to vehicle-treated mice. In the null mice, ICIreduced uterine wet weight, but these weights were stillsignificantly greater compared to vehicle-treated null mice(Fig. 7). Comparison between genotypes within treatmentgroup revealed that ICI could only partially block the E2induction of uterine wet-weight gain.

Administration of E2 1 P4 reduced uterine wet weightcompared to E2 treatment alone in wild-type mice (Fig. 7).However, compared to E2 treatment alone in the null mice,the combination of these steroids did not significantly re-duce uterine wet weights (Fig. 7). Comparison between ge-notypes within treatment group further revealed that E2 1P4 regulation of uterine wet weight was altered, as null micehad significantly greater uterine wet weights (Fig. 7). Ad-ministration of ICI in combination with E2 1 P4 reduceduterine wet weight compared to E2 1 P4 in mice of bothgenotypes, but again there was a significant alteration inregulation of uterine wet weights between genotypes. Fi-nally, administration of the P4 receptor antagonist RU486(Mifepristone) in combination with E2 1 P4 resulted in anincrease in uterine wet weight in both wild-type and TIMP-1 null mice, and this increase was significantly greater inthe null mice (Fig. 7). Overall, it appears that E2, in theface of either P4 or the E2 receptor antagonist ICI, has amore profound effect on uterine wet weight in the TIMP-1 null mice.

DISCUSSION

The pattern of expression of uterine TIMPs during themenstrual [21, 22, 24–26] and estrous [19] cycles suggeststhat these factors might be under the transcriptional regu-

Dow



arch 2022


FIG. 5. Steroidal regulation of uterine TIMP-3 mRNA expression in TIMP-1 wild-type and null mice. Mice were treated with either estrogen (A) orestrogen and progesterone (B) and killed at 0, 2, 4, 6, 8, or 24 h after steroid administration. Total mRNA was extracted and subsequently analyzed forTIMP-3 expression as described in Materials and Methods using 10 mg of total RNA/lane. TIMP-3 mRNA was detected as a major transcript ofapproximately 4.5 kb consistent with previous reports. TIMP-3 mRNA expression is expressed as the mean ratio of TIMP-3/18S transcript 6 SEM andis reported as optical density (OD) units for six separate observations (n 5 6 mice/genotype/treatment group). Autoradiographic exposures for TIMP-3were for 8 to 12 h at 2758C. Different letters indicate statistical significance among the treatment groups as determined by one-way ANOVA (blockletters indicate comparisons within wild-type mice, while italics indicate comparisons within null mice). Asterisks (*) indicate statistically significantdifferences between genotypes within treatment group by planned comparisons using unpaired t-tests. For all comparisons, P , 0.05 was consideredstatistically significant.

lation of E2 and or P4. In vivo and in vitro studies suggestthat this is true, but there appears to be species-specificdifferences in the effects of these steroids on TIMP ex-pression [20–28]. In the current study, we support our pre-vious postulation [19] that, in the mouse, uterine TIMP ex-pression is regulated by the steroids E2 and P4. In additionto providing further evidence for steroidal regulation ofuterine TIMPs, this study makes several novel and impor-tant observations that include distinct patterns of steroidalregulation for TIMP-1, TIMP-2, and TIMP-3 and a poten-tial role for TIMP-1 in the modulation of subsequent ste-roidal effects within the uterus.

In the current study, it was demonstrated that uterineexpression of TIMP-1 is up-regulated by E2 alone or incombination with P4. The ability of the estrogen receptorantagonist ICI 182,780 but not the progesterone receptorantagonist RU-486 to block E2 (or E2 1 P4) induction ofTIMP-1 confirms the specificity of the E2 induction of thisTIMP. The precise role of TIMP-1 within the uterus of non-menstruating species is largely unknown, but disruption ofthe gene is associated with reproductive abnormalities [19,28]. Determination of the temporal pattern of uterine TIMP-1 expression may provide insight into the role of this TIMPwithin the uterus. In the current study, the temporal patternof E2-induced uterine TIMP-1 expression parallels that ofthe immediate early responses induced by E2, such as waterimbibition/edema and macromolecular uptake [33, 34].

When TIMP-1 null mice were treated with E2, the increasein uterine edema (assessed at 8 h post-E2 administration asuterine wet weight) was approximately 50% greater com-pared to wild-type counterparts, suggesting that TIMP-1may functionally control the extent of uterine edema.

The mechanisms by which TIMP-1 may regulate uterineedema may rely on the ability of this TIMP to regulateMMP activity. Compared to wild-type mice, uterine MMPactivity is elevated in TIMP-1 null mice [29; unpublishedresults], and this elevation in active MMP may play a rolein the induction of uterine edema. Recent evidence suggeststhat estrogen induction of uterine edema is mediated at leastin part through the action of vascular endothelial growthfactors (VEGF) [35, 36]. Further evidence indicates thatVEGF up-regulates MMP expression in vascular smoothmuscle [37] and endothelial cells [38], and it is well estab-lished that MMPs themselves are potent inducers of vas-cular permeability [39, 40]. As such, in the current study,the absence of TIMP-1 may lead to elevated MMP activity(compared to wild-type mice), which in turn leads to en-hanced uterine edema via the action of these MMPs.

One of the more interesting findings of the current studywas the observation concerning the regulation of uterineedema. More specifically, E2-induced uterine edema couldbe blocked in the wild-type mice by either ICI or P4, butthis inhibition was only partially blocked in the null mice.This observation may suggest that disruption of the TIMP-

Dow



arch 2022

506 NOTHNICK ET AL.

FIG. 6. The effects of estrogen and progesterone receptor antagonists on uterine TIMP-3 expression. Animals were treated the estrogen receptor (ICI)or progesterone receptor (RU-486) antagonists followed 30 min later by treatment with either E2 or E2 1 P4 (as indicated in the figure). Mice werekilled at 8 h (A) and 24 h (B) after steroid administration, and TIMP-3 mRNA expression was analyzed as described in Materials and Methods using 10mg of total RNA/lane. TIMP-3 mRNA expression is expressed as the mean ratio of TIMP-3/18S transcript 6 SEM and is reported as optical density (OD)units for five separate observations (n 5 5 mice/treatment group). Autoradiographic exposures for TIMP-3 were for 8 h at 2758C. Different letters indicatestatistical significance among the treatment groups as determined by one-way ANOVA (block letters indicate comparisons within wild-type mice, whileitalics indicate comparisons within null mice). For all comparisons, P , 0.05 was considered statistically significant.

1 gene product alters the mechanism by which E2 regulatesuterine edema and shifts the mechanism from an estrogenreceptor-dependent mechanism (blocked by ICI or P4) tothat of an estrogen receptor-independent mechanism (not oronly partially blocked by ICI or P4). It is well establishedthat E2 actions are mediated through the nuclear estrogenreceptors ER-a and/or ER-b [41] and that ICI and P4 canblock E2 action via inhibition of signaling through E2 re-ceptors [42, 43]. However, recent accumulating evidencesuggests that steroid hormones can elicit nongenomic ac-tions within reproductive tissues [44]. Within the uterus, E2has been shown to stimulate calcium entry into endometrialcells [45] and tissue [46, 47], and this mechanism is thoughtto play a role in normal uterine physiology. Pertinent to thecurrent study is the observation that an increase in intra-cellular calcium has been associated with increases in MMPactivity [48–50] and that MMPs themselves [51] can furtherincrease calcium influx into cells. Other potential mecha-nisms that may stimulate the transcriptional activity E2 re-ceptors independent of E2 itself may include cross talk be-tween growth factors and the E2 receptor [52]. Elevateduterine MMP activity, which is characteristic of the TIMP-1 null mice [29], may lead to the liberation of free biolog-ically active growth factors that could stimulate E2 recep-

tor-dependent pathways independent of the presence of thissteroid. We propose that in the TIMP-1 null mice, elevatedMMP activity may lead to an enhanced action of calciumor growth factors (whose biological activity can be in-creased by MMP actions [53, 54]) that in turn may regulateE2 induction of uterine edema independent of the classicalnuclear E2-dependent E2 receptor pathway.

Another interesting observation was the regulation ofuterine TIMP-2 and TIMP-3 expression between the wild-type and TIMP-1 knockout mice. In mice of both geno-types, TIMP-2 expression was high in 0-h control mice;decreased to the lowest levels in response to either E2 orE2 1 P4 at 4 and 6 h post-steroid administration, respec-tively; and then returned to baseline (0 h) levels at 24 h.Furthermore, this steroidal regulation could be blocked bythe E2 receptor antagonist ICI in mice of both genotypes.Collectively, these data suggest that steroidal regulation ofuterine TIMP-2 is similar in wild-type and null mice andthat the presence or absence of TIMP-1 does not influencethe expression of TIMP-2. E2 appears to down-regulateTIMP-2 expression, while P4 has no effect on its expres-sion. While we interpret these data to suggest that E2 de-creases steady-state levels of TIMP-2 mRNA, we are awarethat there could actually be an increase in TIMP-2 trans-

Dow



arch 2022


FIG. 7. Effects of steroids and steroid receptor antagonists on uterinewet weights between genotypes. Mice were killed 8 h after treatmentadministration, and uterine wet weights were determined. Uterine wetweights are expressed as percentage of body weight and are expressedas the mean 6 SEM for six mice/genotype/treatment group (n 5 6). Dif-ferent letters indicate statistical significance among the treatment groupsas determined by one-way ANOVA (block letters indicate comparisonswithin wild-type mice, while italics indicate comparisons within nullmice). Asterisks (*) indicate statistically significant differences betweengenotypes within treatment group by planned comparisons using un-paired t-tests (* 5 P , 0.05, ** 5 P , 0.01, *** 5 P , 0.001). For allcomparisons, P , 0.05 was considered statistically significant.

lation and protein expression. This would suggest that theend result of E2 treatment would be an increase in TIMP-2 activity.

In contrast to TIMP-2 regulation, steroidal regulation ofuterine TIMP-3 expression appears to differ between wild-type and TIMP-1 null mice, but only after 8 h post-steroidadministration were these differences in regulation betweengenotypes noted. In wild-type mice, E2 induced a decreasein TIMP-3 expression at 8 and 24 h post-steroid adminis-tration, but in the null mice a reduction in TIMP-3 expres-sion did not occur until 24 h post-E2 administration. Oneinterpretation of this observation may be that the regulatorymechanism by which uterine TIMP-3 expression is modu-lated is out of phase or delayed in the TIMP-1 null mice.This postulate is supported by the observation in the currentstudy that at 8 h post-E2 treatment, TIMP-3 mRNA steady-state levels begin to decrease in the wild-type mice, andthis decrease can be blocked by pretreatment with the E2receptor antagonist ICI. In contrast, E2 does not suppressTIMP-3 expression at 8 h in the null mice, and ICI pre-treatment has no effect on steady-state mRNA levels of thisTIMP. However, by 24 h post-E2 treatment, TIMP-3 mRNAsteady-state levels are significantly reduced in mice of bothgenotypes, and this reduction can be blocked by ICI. Whilethe mechanism(s) for this ‘‘delayed’’ regulation of TIMP-3expression is unknown, alterations in calcium or growthfactor signaling that result from the absence of TIMP-1 maybe responsible.

In summary, uterine TIMP-1 is regulated by both E2 andE2 1 P4. It appears that while disruption of the TIMP-1gene is not associated with an alteration in steroidal regu-lation of uterine TIMP-2 expression, it is associated withan altered regulation of uterine TIMP-3 expression. Estro-gen up-regulation of TIMP-1 and down-regulation ofTIMP-2 (and to a lesser extent that of TIMP-3) occur con-current to the induction of uterine wet-weight gain/edema,

indicating that these TIMPs may play a role in regulatingthis process. ICI blocks estrogen regulation of TIMP-1,TIMP-2 expression, and uterine edema, indicating a similarestrogen pathway via classical ER. Precise roles of TIMP-1 and TIMP-2 in regulating uterine edema are uncertain butmay depend on their MMP-dependent or MMP-indepen-dent actions. Based on data with the TIMP-1 null mice, wepostulate that TIMP-1 controls the extent of edema, as ab-sence of this inhibitor was associated with a more pro-nounced increase in uterine edema. The fact that this edemawas only partially blocked by ICI or P4 in the null micesuggests a complicated role for TIMP-1 that occurs in re-sponse to estrogen treatment. Possibilities include elevationof MMP activity, which in turn could alter bioavailability/bioactivity of growth factors/cytokines, which could influ-ence edema.

REFERENCES

1. Fazleabas AT, Strakova Z. Endometrial function: cell specific changesin the uterine environment. Mol Cell Endocrinol 2002; 186:143–147.

2. Tseng L, Mazella J. Endometrial cell specific gene activation duringimplantation and early pregnancy. Front Biosci 2002; 7:d1566–d1574.

3. Bagchi IC, Li Q, Cheon YP. Role of steroid-hormone-regulated genesin implantation. Ann N Y Acad Sci 2001; 943:68–76.

4. Lessey BA. Adhesion molecules and implantation. J Reprod Immunol2002; 55:101–112.

5. Ma W, Tan J, Matsumoto H, Robert B, Abrahamson DR, Das SK,Dey SK. Adult tissue angiogenesis: evidence for negative regulationby estrogen in the uterus. Mol Endocrinol 2001; 15:1983–1992.

6. Fazleabas AT, Bell SC, Fleming S, Sun J, Lessey BA. Distribution ofintegrins and the extracellular matrix proteins in the baboon endo-metrium during the menstrual cycle and early pregnancy. Biol Reprod1997; 56:348–356.

7. Gray CA, Bartol FF, Tarleton BJ, Wiley AA, Johnson GA, Bazer FW,Spencer TE. Developmental biology of uterine glands. Biol Reprod2001; 65:1311–1323.

8. Fata JE, Ho AT, Leco KJ, Moorehead RA, Khokha R. Cellular turn-over and extracellular matrix remodeling in female reproductive tis-sues: functions of metalloproteinases and their inhibitors. Cell MolLife Sci 2000; 57:77–95.

9. Zhao YG, Xiao AZ, Cao XM, Zhu C. Expression of matrix metallo-proteinase-2, -9 and tissue inhibitors of metalloproteinase-1, -2, -3mRNAs in rat uterus during early pregnancy. Mol Reprod Dev 2002;62:149–158.

10. Salamonsen LA, Nagase H, Woolley DE. Matrix metalloproteinasesand their tissue inhibitors at the ovine trophoblast-uterine interface. JReprod Fertil 1995; 49:29–37.

11. Bass KE, Li H, Hawkes SP, Howard E, Bullen E, Vu TK, McMasterM, Janatpour M, Fisher SJ. Tissue inhibitor of metalloproteinase-3expression is upregulated during human cytotrophoblast invasion invitro. Developmental Genetics 1997; 21:61–67.

12. Das SK, Yano S, Wang J, Edwards DR, Nagase H, Dey SK. Expres-sion of matrix metalloproteinases and tissue inhibitors of metallopro-teinases in the mouse uterus during the periimplantation period. De-velopmental Genetics 1997; 21:44–54.

13. Harvey MB, Leco KJ, Arcellana-Panlilio MY, Zhang X, Edwards DR,Schultz GA. Proteinase expression in early mouse embryos is regu-lated by leukemia inhibitory factor and epidermal growth factor. De-velopment 1995; 121:1005–1014.

14. Reponen P, Leivo I, Sahlberg C, Apte SS, Olsen BR, Thesleff I, Try-ggvason K. 92-kDA type IV collagenase and TIMP-3, but not 72-kDatype IV collagenase or TIMP-1 or TIMP-2, are highly expressed dur-ing mouse embryo implantation. Dev Dyn 1995; 202:388–396.

15. Waterhouse P, Denhardt DT, Khokha R. Temporal expression of tissueinhibitors of metalloproteinases in mouse reproductive tissues duringgestation. Mol Reprod Dev 1993; 35:219–226.

16. Menino AR Jr, Hogan A, Schultz GA, Novak S, Dixon W, FoxcroftGH. Expression of proteinases and proteinase inhibitors during em-bryo-uterine contact in the pig. Developmental Genetics 1997; 21:68–74.

17. Gomez DE, Alonso DF, Yoshiji H, Thorgeirsson UP. Tissue inhibitorsof metalloproteinases: structure, regulation and biological functions.Eur J Cell Biol 1997; 74:111–122.

Dow



arch 2022

508 NOTHNICK ET AL.

18. Leco KJ, Apte SS, Taniguchi GT, Hawkes SP, Khokha R, Schultz GA,Edwards DR. Murine tissue inhibitor of metalloproteinase-4 (TIMP-4): cDNA isolation and expression in adult mouse tissues. FEBS Lett1997; 401:213–217.

19. Nothnick WB. Disruption of the tissue inhibitor of metalloproteinase-1 gene results in altered reproductive cyclicity and uterine morphologyin reproductive-age female mice. Biol Reprod 2000; 63:905–912.

20. Hampton AL, Butt AR, Riley SC, Salamonsen LA. Tissue inhibitorsof metalloproteinases in endometrium of ovariectomized steroid-treat-ed ewes during the estrous cycle and early pregnancy. Biol Reprod1995; 53:302–311.

21. Zhang J, Salamonsen LA. Tissue inhibitor of metalloproteinases(TIMP)-1, -2, and -3 in human endometrium during the menstrualcycle. Mol Hum Reprod 1997; 3:735–741.

22. Hampton AL, Salamonsen LA. Expression of messenger ribonucleicacid encoding matrix metalloproteinases and their tissue inhibitors isrelated to menstruation. J Endocrinol 1994; 141:R1–R3.

23. Higuchi T, Kanzaki H, Nakayama H, Fujimoto M, Hatayama H, Ko-jima K, Iwai M, Mori T, Fujita J. Induction of tissue inhibitor ofmetalloproteinase 3 gene expression during in vitro decidualization ofhuman endometrial stromal cells. Endocrinology 1995; 136:4973–4981.

24. Rodgers W, Matrisian LM, Giudice LC, Dsupin B, Cannon P, SvitekC, Gorstein F, Osteen KG. Patterns of matrix metalloproteinase ex-pression in cycling endometrium imply differential functions and reg-ulation by steroid hormones. J Clin Invest 1994; 94:946–953.

25. Brenner RM, Rudolph L, Matrisian L, Slayden OD. Non-human pri-mate models: artificial menstrual cycles, endometrial matrix metallo-proteinases and s.c. endometrial grafts. Hum Reprod 1996; 11:150–164.

26. Rudolph-Owen LA, Slayden OD, Matrisian LM, Brenner RM. Matrixmetalloproteinase expression in Macaca mulatta endometrium: evi-dence for zone-specific regulatory tissue gradients. Biol Reprod 1998;59:1349–1359.

27. Salamonsen LA, Butt AR, Hammond FR, Garcia S, Zhang J. Produc-tion of endometrial matrix metalloproteinases, but not their tissue in-hibitors, is modulated by progesterone withdrawal in an in vitro modelfor menstruation. J Clin Endocrinol Metab 1997; 82:1409–1415.

28. Nothnick WB. Reduction in reproductive lifespan of tissue inhibitorof metalloproteinase 1 (TIMP-1)-deficient female mice. Reproduction2001; 122:923–927.

29. Nothnick WB. Disruption of the tissue inhibitor of metalloproteinase-1 gene in reproductive-age female mice is associated with estrouscycle stage-specific increases in stromelysin messenger RNA expres-sion and activity. Biol Reprod 2001; 65:1780–1788.

30. Hampton AL, Nie G, Salamonsen LA. Progesterone analogues simi-larly modulate endometrial matrix metalloproteinase-1 and matrix me-talloproteinase-3 and their inhibitor in a model for long-term contra-ceptive effects. Mol Hum Reprod 1999; 5:365–371.

31. Rosner B. Fundamentals of Biostatistics, 3rd ed. Boston: PWS-KentPublishing; 1990.

32. Nothnick WB. Tissue inhibitor of metalloproteinase-1 (TIMP-1) de-ficient mice display reduced serum progesterone levels during corpusluteum development. Endocrinology 2003; 144:5–8.

33. Kirkland JL, Gardner RN, Ireland JS, Stancel GM. The effect of hy-pophysectomy on the uterine response to estradiol. Endocrinology1977; 101:403–410.

34. Harris J, Gorski J. Evidence for a discontinuous requirement for es-trogen in stimulation of deoxyribonucleic acid synthesis in the im-mature rat uterus. Endocrinology 1978; 103:204–245.

35. Rockwell LC, Pillai S, Olson CE, Koos RD. Inhibition of vascular

endothelial growth factor/vascular permeability factor action blocksestrogen-induced uterine edema and implantation in rodents. Biol Re-prod 2002; 67:1804–1810.

36. Hastings JM, Licence DR, Burton GJ, Charnock-Jones DS, Smith SK.Soluble vascular endothelial growth factor receptor 1 inhibits edemaand epithelial proliferation induced by 17b-estradiol in the mouseuterus. Endocrinology 2003; 144:326–334.

37. Wang H, Keiser JA. Vascular endothelial growth factor upregulatesthe expression of matrix metalloproteinases in vascular smooth musclecells. Role of flt-1. Circ Res 1998; 83:832–840.

38. Unemori EN, Ferrara N, Bauer EA, Amento EP. Vascular endothelialgrowth factor induces interstitial collagenase expression in human en-dothelial cells. J Cell Physiol 1992; 153:557–562.

39. Partridge CA, Jeffrey JJ, Malik AB. A 96-kDa gelatinase induced byTNF-a contributes to increased microvascular endothelial permeabil-ity. Am J Physiol 1993; 265:438–447.

40. Zucker S, Mirza H, Conner CE, Lorenz AI, Drews MH, Bahov WF,Jesty J. Vascular endothelial growth factor induces tissue factor andmatrix metalloproteinase production in endothelial cells: conversionof prothrombin to thrombin results in progelatinase A activation andcell proliferation. Int J Cancer 1998; 75:780–786.

41. Truss M, Beato M. Steroid hormone receptors: interaction with deox-yribonucleic acid and transcription factors. Endocr Rev 1993; 14:459–479.

42. Wakeling AE, Dules M, Bowler J. A potent specific pure antiestrogenwith clinical potential. Cancer Res 1991; 51:3867–3873.

43. Clark JH, Mani SK. Actions of steroid hormone receptors. In KnobilE, Neill J (eds.), The Physiology of Reproduction. New York: RavenPress; 1994:1011–1059.

44. Revelli A, Massobrio M, Tesarik J. Nongenomic actions of steroidhormones in reproductive tissues. Endocr Rev 1998; 19:3–17.

45. Perret S, Dockery P, Harvey BJ. 17b-Oestradiol stimulates capacita-tive Ca21 entry in human endometrial cells. Mol Cell Endocrinol2001; 176:77–84.

46. Ishii K, Kano T, Ando J. Calcium channel, Ca11 mobilization, andmechanical reactivity of estrogen- and progesterone-treated rat uterus.Jpn J Pharmacol 1986; 41:47–54.

47. Batra S. Increase by oestrogen of calcium entry and calcium channeldensity in uterine smooth muscle. Br J Pharmacol 1987; 92:389–392.

48. Kohn EC, Jacobs W, Kim YS, Alessandro R, Stetler-Stevenson WG,Liotta LA. Calcium influx modulates expression of matrix metallo-proteinase-2 (72-kDa type IV collagenase, gelatinase A). J Biol Chem1994; 269:21505–21511.

49. Lohi J, Keski-Oja J. Calcium ionophores decrease pericellular gela-tinolytic activity via inhibition of 92-kDa gelatinase expression anddecrease of 72-kDa gelatinase activation. J Biol Chem 1995; 270:17602–17609.

50. Kobayashi T, Hattori S, Nagai Y, Tajima S, Nishikawa T. Differentialregulation of MMP-2 and MMP-9 gelatinases in cultured human ke-ratinocytes. Dermatology 1998; 197:1–5.

51. Conant K, Haughey N, Nath A, St Hillaire C, Gary DS, Pardo CA,Wahl LM, Bilak M, Milward E, Mattson MP. Matrix metalloprotei-nase-1 activates a pertussis toxin-sensitive signaling pathway thatstimulates the release of matrix metalloproteinase-9. J Neurochem2002; 82:885–893.

52. Smith CL. Cross-talk between peptide growth factor and estrogen re-ceptor signaling pathways. Biol Reprod 1998; 58:627–632.

53. Sternlicht MD, Werb Z. How matrix metalloproteinases regulate cellbehavior. Annu Rev Cell Dev Biol 2001; 17:463–516.

54. Fowlkes JL, Winkler MK. Exploring the interface between metallo-proteinase activity and growth factor and cytokine bioavailability. Cy-tokine Growth Factor Rev 2002; 13:277–287.

Dow



arch 2022