lactoferrin gene expression is estrogen responsive in human and

Molecular Human Reproduction Vol.8, No.1 pp. 58–67, 2002

Lactoferrin gene expression is estrogen responsive inhuman and rhesus monkey endometrium

Christina T.Teng1,5, Wesley Gladwell1, Clara Beard1, David Walmer1,2, Ching S.Teng3 andRobert Brenner4

1Gene Regulation Group, Laboratory of Reproductive and Developmental Toxicology, National Institute of Environmental HealthSciences, National Institutes of Health, Research Triangle Park, NC 27709, 2Department of Obstetrics and Gynecology, Division ofReproductive Endocrinology and Infertility, Duke University Medical Center, Durham, NC 27710, 3Department of APR, School ofVeterinary Medicine, North Carolina State University, Raleigh, North Carolina and 4Division of Reproductive Sciences, OregonRegional Primate Research Center, Beaverton, OR 97006, USA

5To whom correspondence should be addressed at: P.O.Box 12233, MD E2-01, RTP, NC 27709, USA. E-mail: [email protected]

We have previously shown that the estrogen responsiveness of the human lactoferrin gene in a transient transfectionsystem is mediated through an imperfect estrogen response element (ERE) and a steroidogenic factor 1 bindingelement (SFRE) 26 bp upstream from ERE. Reporter constructs containing SFRE and ERE respond to estrogenstimulation in a dose-dependent manner, whereas mutations at either one of the response elements severely impairedthe estrogen responsiveness. In this study, we demonstrated that estrogen receptor (ERα) binds to the humanlactoferrin gene ERE and forms two complexes in an electrophoresis mobility shift assay (EMSA). These complexescould be supershifted by an antibody to ERα. We also showed that in normal cycling women, lactoferrin geneexpression in the endometrium increases during the proliferative phase and diminishes during the luteal phase.This in-vivo study thus supported the finding from transient transfection experiments that the human lactoferringene expression is elevated in an environment with a high level of estrogen. The estrogen effect on lactoferrin geneexpression in the rhesus monkey endometrium was studied by Western blotting and immunohistochemistry. Theimmunohistochemistry results showed that immunoreactive lactoferrin protein was not detectable in the untreatedovariectomized monkey endometrium, was elevated by estrogen treatment, and was suppressed by sequential,combined estrogen plus progesterone treatment. In conclusion, this study has shown that lactoferrin gene expressionis responsive to estrogen in primate endometrium.

Key words: endometrium/estrogen response element/human and monkey/lactoferrin/lactoferrin gene promoter

IntroductionLactoferrin, a non-haem iron-binding glycoprotein, was firstdiscovered in milk and later found in the wet surface mucosaepithelium. It is a major protein in the secondary granules ofneutrophils and is present in many biological secretions includ-ing the saliva, tears and semen (Sanchez et al., 1992; Levayand Viljoen, 1995; Lonnerdal and Iyer, 1995; Nuijens et al.,1996). The protein has been shown to kill bacteria, playan immunomodulatory role and participate in inflammatoryresponse (Sanchez et al., 1992; Levay and Viljoen, 1995;Lonnerdal and Iyer, 1995; Nuijens et al., 1996). However, itsmechanism of action is presently unknown. In the mouseuterus, lactoferrin expression is up-regulated by both estrogensand epidermal growth factor (EGF) (Teng, 1999). The mRNAand protein expression of lactoferrin can be induced severalhundred-fold in the uterus of immature mice by estrogentreatment (Teng et al., 1986; Pentecost and Teng, 1987). DNA

58 © European Society of Human Reproduction and Embryology

elements of the mouse lactoferrin gene responsible for estrogenand EGF stimulation have been identified and characterizedin a transiently transfected cell culture system (Liu and Teng,1991, 1992; Liu et al., 1993; Shi and Teng, 1994, 1996). Theestrogen response element (ERE) of the mouse lactoferringene consists of an imperfect palindromic AGGTCA motif,which overlaps with a chicken ovalbumin upstream promoter(COUP)-transcription factor (TF) binding site (Liu and Teng,1992). Estrogen receptor alpha (ERα) binds COUP/ERE andforms stable complexes with a half-life of 30 min, whereas aninteraction between COUP-TF and the DNA elements lastsonly 5 min (Liu et al., 1993). Thus, estrogen-bound ERαactivates the mouse lactoferrin gene readily in the transfectionsystem and in the mouse uterus. During the mouse estruscycle, lactoferrin content in the uterine epithelium fluctuateswith the level of circulating estrogen (Newbold et al., 1992;Walmer et al., 1992). For example, the expression of the

Lactoferrin gene is regulated by estrogen in primate endometrium

uterine lactoferrin gene increases during proestrus, becomeshighest at estrus, declines during metestrus, and is undetectableat diestrus. These changes suggest that the lactoferrin proteinplays an important role in the physiology of the mouse uterus.

As with the mouse lactoferrin gene, an imperfect ERE,located at similar position in the human lactoferrin gene, canbe activated by estrogen in transiently transfected cells (Tenget al., 1992). However, in the human lactoferrin gene, anadditional element, Steroid Factor 1 Response Element (SFRE)(Yang and Teng, 1994), located 26 bp upstream from the ERE,is involved in estrogen responsiveness (Yang et al., 1996).Mutations at SFRE reduce the activity of ERα to 50% althoughthe ERE remains intact. Also, an estrogen receptor-relatedreceptor (ERRα1) binds to SFRE and is responsible formodulating the transcriptional activity mediated by ERα (Yanget al., 1996; Zhang and Teng, 2000). Therefore, estrogenregulation of the human lactoferrin gene appears to be subjectto an additional level of fine-tuning as compared to the mouse(Teng et al., 1992; Yang et al., 1996).

Lactoferrin was observed in the human endometrium atvarious stages of menstrual cycle more than 30 years ago(Masson et al., 1968; Tourville et al., 1970). However, examina-tion of lactoferrin expression in the human endometrium duringthe menstrual cycle by immunohistochemistry has yieldedinconsistent results. It has been found to be highly expressedin the endometrium of both proliferative (Kelver et al., 1996)or secretory phase (Tourville et al., 1970; Walmer et al., 1995).However, most of the studies were conducted with polyclonalrabbit anti-lactoferrin serum that may also react with othermembers of the transferrin gene family since they share 60–100% sequence identity at certain regions of the protein(Pentecost and Teng, 1987). The human studies have also beenconstrained by limited samples and variations among thehuman subjects. It has also been reported that lactoferrin geneexpression in normal and pathological human endometrium isup-regulated by estrogen (Walmer et al., 1995; Kelver et al.,1996). The goal of the present study was to investigate humanendometrium lactoferrin expression at various stages of thenormal cycle in carefully staged tissue samples with a well-characterized antiserum to lactoferrin. In addition, we examinedlactoferrin expression in the endometrium of rhesus monkeyunder experimentally controlled hormonal conditions.

Materials and methods

Patient population and sample preparation

The inclusion criteria of women in the study and preparation of theendometrium samples have been described previously (Bush et al.,1998). The samples were collected in accordance with the guidelinesof the Internal Review Board committee at Duke University MedicalCenter and informed consent was obtained from the patients. Briefly,two sets of endometrium samples (100297 as Set 1 and 102897 asSet 2) were collected from women of reproductive age (28–48 yearsold and with spontaneous menstrual cycles occurring every 26–35days), whose uterus was removed during hysterectomy. All uteri withsevere pathology, excessive bleeding, fibroids and endometrosis wereexcluded from the study. Screening for uterine pathology included aclinical history, physical examination, documenting regular monthlymenses at a normal interval and three independent blinded readings

59

by a single board-certified gynaecological pathologist to rule outunsuspected histological abnormalities. The endometrium was scrapedfrom the fundus in a uniform fashion with a scalpel and wasimmediately placed into solubilization buffer [1% Triton X-100,2 mmol/l EDTA, 2 mmol/l EGTA, 1 mmol/l Na3VO4, 20 mmol/lNaF, 50 µmol/l Na2Mo4O4, aprotinin and leupeptin (20 µg/ml each),and 15 µmol/l (4-amidinophenyl)-methansulphonyl fluoride in 20mmol/l HEPES, pH 7.4] and homogenized with a Virtishear (VirtisCorp., Gardiner, NY, USA) in the cold. After centrifugation at21 000 g for 1 min, the supernatant was divided into small aliquotsand stored at –70°C until use. The protein concentration was deter-mined with the Pierce BCA Protein Assay (Pierce, Rockford, IL,USA). Human milk was obtained from a lactating mother with amilk pump. Human prostate tissue (from a normal 67 year old whitemale) was obtained from the Sourthern Division of CooperativeHuman Tissue Network (Birmingham, AL, USA).

Rhesus monkey sample preparation

Animal treatments and endometrial collection have been describedpreviously (Slayden et al., 1993). Briefly, sexually mature rhesusmacaques (Macaca mulatta) were housed and cared for by theveterinary staff of the Oregon Regional Primate Research Center inaccordance with the NIH Guide for the Care and Use of LaboratoryAnimals. Ovariectomy was performed by standard techniques. Afterovariectomy, the monkeys were either left untreated to serve ascontrols, treated for 4 weeks with estradiol-17β (E2) to provideestrogenized samples, or treated with E2 for 2 weeks, then E2 �progesterone for 2 subsequent weeks to provide artificial luteal phasesamples. Tissues collected for Western blotting were cut freehandwith a razor blade (2 mm thick), frozen in liquid propane and storedat –70°C until use. The frozen tissue samples were pulverized inliquid N2, then immediately homogenized in 5 volumes of RIPA lysisbuffer [150 mmol/l NaCl, 0.1% (sodium dodecyl sulphate) SDS, 1%CHAPS, 1 mmol/l EDTA and 10 mmol/l Tris–HCl at pH 7.4]containing a cocktail of protease inhibitors (0.5 mmol/l PMSF, 10µg/ml aprotinin, 2 µg/ml pepstatin, 2 µg/ml leupeptin; Sigma, St Louis,MO, USA). After sonication with a Branson Sonifier (Ultrasonics Inc.,Plain View, NY, USA), the samples were cleared by centrifugationat 10 000 g for 20 min. Aliquots of the clear supernatant werestored at –70°C until use. Tissues collected for histology andimmunohistochemistry were fixed in 4% paraformaldehyde, embeddedin paraffin, sectioned at 6 µm thickness and mounted on slides.Monkey prostate tissues were obtained from normal adult malesduring regularly scheduled Primate Center necropsies. Monkey milkwas obtained from lactating females with a milk pump by standardprocedures.

Nuclear protein preparation and EMSA

Nuclear protein enriched with ERα from human endometrial carcin-oma RL95-2 cells (RL95-2) was prepared as previously described(Yang et al., 1996). ERα was overexpressed in RL95-2 cells withexpression vector (HEGO, a gift from Pierre Chambon, Paris, France)for 48 h before harvesting the cells and extracting the nuclear protein.Baculovirus expressed ERα (BV-ERα) was a gift from MalcomParker (London, UK). The probes that were used in the electrophoresismobility shift assay (EMSA) have been characterized (Teng et al.,1992; Yang and Teng, 1994) and the EMSA was performed aspreviously described (Liu and Teng, 1992). The antibody to ERαwas obtained from Abbott Laboratories (Abbott ER-ICA monoclonalH222 kit, Chicago, IL, USA).

Western blot analysis

Tissue extracts containing 25 µg of protein (BioRad Protein AssayKit) from the prostate or endometrial samples of the human and

C.T.Teng et al.

monkey were examined in Western blot analysis. For antibodycharacterization, mouse lactoferrin (mLF) was isolated from theuterine fluid of an estrogen-treated mouse (Teng et al., 1986). Bovinemilk lactoferrin (bLF), human milk lactoferrin (hLF), and mousetransferrin (mTF) were obtained commercially (Sigma). The rabbitpolyclonal antibodies, raised against mouse lactoferrin (mLF 8344),human lactoferrin (hLF 12484) and mouse transferrin (mTF), havebeen previously described (Teng et al., 1986; Panella et al., 1991).Proteins were loaded onto precast 4–12% Bis–Tris NuPAGE gels andelectrophoresed in a 1� MOPS buffer at 120 mA and 200 V for1.25 h in a Novex XCell II Mini-Cell system (Novex, San Diego,CA). Proteins were transferred from the gel onto an Immobilon-Ppolyvinylidene difluoride (PDVF) membrane (Millipore) at 200 mAand 25 V for 2 h in a 1� NuPAGE transferring buffer containing20% methanol. Immunodetections were carried out with an enhancedchemiluminescence (ECL) kit (Amersham) according to the manufac-turer’s specifications. Primary antibodies were diluted �10 000 andincubated with the protein blots overnight in the cold and with constantshaking. (For the transferrin antibody neutralization experiments,10 µg/ml of transferrin from either human or rat was included in theprimary antibody interaction step.) After washing, the blots wereincubated with diluted (�10 000) second antibody (donkey anti-rabbitIgG linked with horseradish peroxidase) for another hour before ECLdetection. X-ray film was exposed to the blot for 1 min and developedin Konica SRX-101 developer (Tokyo, Japan).

IgG purification and immunohistochemistry staining

IgG fractions of the hLF 12484 and mLF 8344 were purified fromthe rabbit antiserum according to the IgG purification protocol fromImmunoPure™ IgG (Protein A) Purification Kit (Pierce, Rockford,IL. USA). The immunohistochemistry protocol was derived from theHistostain SP Kit for rabbit polyclonal antibody (Zymed LaboratoriesInc., San Francisco, CA, USA). Briefly, the deparaffinized tissuesections were first circled with a PAP to define the area and allsubsequent steps were carried out in a dark and humidified box. Thetissue sections were blocked by two different blocking solutions: firstthe Pierce Peroxidase Suppressor to block the endogenous peroxidaseactivity and then another blocking solution supplied with the Kit toblock non-specific interactions. The primary antibody, mLF 8344 IgGat a concentration of 10 µg/ml in phosphate-buffered saline containingeither 1 mg/ml of bovine serum albumin (BSA) or rat transferrin,was applied to the tissue sections overnight at room temperature. Thesecondary antibody, Pharmingen goat anti-rabbit biotin, was diluted1:1500 before use. The colour reaction was then developed withhorseradish peroxidase–steptavidin and aminoethyl carbazole, andcounter-stained with a modified Harris haematoxylin (Sigma).

Results

ERα binding to the imperfect ERE of human lactoferrin gene

Multiple DNA elements in the human lactoferrin gene promoterare involved in estrogen regulation (Figure 1). The imperfectpalindromic ERE (–362 and –336) has been shown to mediatethe estrogen action (Teng et al., 1992) and a stronger estrogenresponse can be obtained if the SFRE (–390 and –379) is alsoincluded (Yang et al., 1996). Whether ERα binds the ERE inthe context of human lactoferrin gene was investigated in thecurrent study. By EMSA we demonstrated that the BV- ERαbinds to the ERE in vitro and forms two complexes (Figure2A, lanes 2 and 3). To authenticate the binding, we showedthat both complexes were supershifted by antibody specific to

60

the receptor (lane 4). Furthermore, the receptor expressed inthe baculovirus and in RL95-2 cells (NPE) binds ERE inthe context of the human lactoferrin gene (hLF –417/–336)promoter (Figure 2B). The specific binding of ERα (lane 1and 5) could be competed off by 50� excess cold EREoligonucleotides (lane 3), and the ERα antibody (H222)retarded the mobility of the complexes in EMSA (lanes 2 and 4).

Lactoferrin gene expression in human endometrium duringthe menstrual cycle

Endometrial samples collected from women during earlyproliferative (EP), mid proliferative (MP), mid luteal (days20–22), and late luteal (days 26–28) phases of the cycle wereexamined by Western blotting analysis with human lactoferrinantiserum, hLF 12484 (Figure 3A). Two sets of humanendometrial samples were evaluated. With Set 1 (100297;lanes 1–4), lactoferrin was found in the early proliferativephase (lane 1) but not in other phases (lanes 2–4). With Set 2(102897; lanes 6–9), lactoferrin was mainly found in early(lane 6) and mid (lane 7) proliferative.

The initial characterization of the antibody indicated thatthis lot of anti-hLF antibody cross-reacted weakly with themouse transferrin (Figure 4, top panel). Therefore, the faintband that ran faster than the milk lactoferrin (Figure 3A, lane10) and was detected by the antibody in all endometrialsamples could be transferrin or a related family member. Toverify the presence of transferrin in the endometrial samples,we stripped the lactoferrin antibody from the membrane andre-blotted with antibody to the mTF. Indeed, a transferrin bandwas detected in every sample (Figure 3A, lanes 11–14 and16–19) except the milk sample that contains a high level oflactoferrin but no transferrin (lane 15). For this reason, werepeated the Western blotting experiment of Set 2 (102897)and included 10 µg/ml of human transferrin in the primaryantibody to neutralize the immunoreactivity of transferrin(Figure 3B). As expected, the transferrin antibody activity wascompletely blocked (lanes 1–4) whereas lactoferrin immuno-reactive protein was detected in the early and mid proliferativephase endometrial samples (lanes 5 and 6). As a loadingcontrol, a duplicated gel was stained with colloidal blue (Figure3C, lanes 1–4) to verify the equal loading. These resultsdemonstrated that human lactoferrin gene in the endometriumis highly expressed during the early and mid proliferativephases of the menstrual cycle, when estrogen is the dominanthormone.

Estrogen effect on lactoferrin expression in rhesus monkeyendometrium

To explore whether lactoferrin in monkey endometrium isregulated by estrogen, we first evaluated the species specificityof the rabbit polyclonal anti-lactoferrin sera. Among the tenrabbit antisera that we analysed, the hLF 12484 and the mLF8344 antisera were used in the present study (Figure 4). ThehLF 12484 was very specific to the hLF and did not cross-react with lactoferrin of other species (Figure 4, top panel),whereas the mLF 8344 cross-reacted with lactoferrin fromother species as well as with the mTF although with lessintensity (middle panel). The antibody to mTF reacted to all


Figure 1. Diagram of estrogen response modules of the human lactoferrin gene promoter. The numbers indicate the position upstreamfrom the start site. COUP/SFRE � overlapping COUP-TF (cross-hatched box) and SF-1 binding elements (shaded box); COUP/ERE � overlapping COUP-TF (cross-hatched box) and ERα binding elements (shaded box); GATA � the consensus GATA1 bindingelement (clear box).

Figure 2. Electrophoresis mobility shift assay (EMSA) detection of ERα binding to the human lactoferrin (hLF) estrogen responseelement (ERE). (A) Baculovirus expressed ERα (BV-ERα) binds the ERE. Concentrations of the BV-ERα used to bind ERE are 0.6and 3.2 ng protein. (B) BV-ERα (3.2 ng) and ERα in the nuclear extract of RL95-2 cells (NPE, 4 µg) bind the ERE in the contextof lactoferrin estrogen response module (–417/–336) H222, antibody to ERα. Double arrows: the receptor–DNA complexes; singlearrow: ERα antibody supershifted complex.

61

C.T.Teng et al.

Figure 3. Lactoferrin expression in human endometrium during the proliferative phase of the menstrual cycle. (A) Western blotting ofendometrium extract with human anti-lactoferrin 12484 (hLF) or mouse transferrin (mTF) antiserum. Two sets of samples wereevaluated, Set 1 (100297) and Set 2 (102897). Equal amounts of human endometrium extract (25 µg protein) were loaded andanalysed by Western blotting with enhanced chemiluminescence detection. EP � early proliferative phase; MP � mid proliferativephase; L20–22 � luteal phase at day 20–22 of the cycle; L26–28 � luteal phase at day 26–28 of the cycle; ST � standard mol. wtmarkers; h milk � human milk.(B) Samples from Set 2 study were Western-blotted with hLF or mTF in the presence of 10 µg/ml human transferrin. The arrowindicates the lactoferrin and the mol. wts are marked. (C) Colloidal blue staining of the protein samples from B.

transferrin tested (data not shown) and to the bLF slightly(bottom panel).

To detect the immunoreactive protein by Western blotting,monkey endometrial tissue extracts were prepared and analysed(Figure 5A). Milk, which contains a high level of lactoferrinbut no transferrin, and prostate tissue extracts, which containboth lactoferrin and transferrin, were used as controls. Bothhuman and monkey samples were used. The tissue extractsand the control samples were examined by antisera to mLF

62

(lanes 1–6), hLF (lanes 7 and 8) and mTF (lanes 9–14). The84 kDa immunoreactive band detected in the samples ofmonkey milk and prostate (lane 1 and lane 2, upper band) islikely to represent a form of monkey lactoferrin, based on thelack of reactivity to the anti-mTF antiserum. The size of thisprotein contrasts with the 76 kDa protein detected in humanmilk (lanes 5 and 8) and prostate (lane 6) lactoferrin. Interes-tingly, the monkey prostate and endometrial tissue extractscontained a 70 kDa lactoferrin-immunoreactive protein, which


Figure 4. Characterization of the rabbit polyclonal antisera by Western blotting. Sources of the purified protein are indicated in theMaterials and methods section. The antibody, hLF 12484, mLF 8344 or mTF used in the Western blots is indicated at the bottom ofeach panel. Concentrations of the protein loaded in each lane are indicated at the top and the mol. wt marker is shown. The primaryantibody was diluted 10 000-fold and the secondary antibody was diluted 10 000-fold. mTF � mouse transferrin; bLF � bovinelactoferrin;hLF � human lactoferrin; mLF � mouse lactoferrin.

Figure 5. Detection of lactoferrin in the monkey endometrium. (A) Western blotting of tissue extracts from monkey (MK) and human(HU) prostate (pros) and monkey endometrium (endo) treated with estrogen alone (endo E) or estrogen plus progesterone (endo E �P). Milk was included as the control. The mouse anti-lactoferrin 8344 (mLF), human anti-lactoferrin 12484 (hLF) and mouse anti-transferrin (mTF) antisera were examined. (B) Western blotting of the monkey samples with mLF in the presence of 10 µg/ml of rattransferrin. The double arrow indicates the location of the lactoferrin.

was distinct from the 84 kDa milk lactoferrin (lanes 2–4, lowerband). Transferrin ran slightly ahead of the lactoferrin ona SDS–polyacrylamide gel electrophoresis as shown earlier(Figure 3). Therefore, the size of monkey transferrin was

63

~78 kDa (Figure 5A, lanes 10–12, the upper band) and thehTF was ~74 kDa (lane 14). As expected, the transferrin wasnot detected in either monkey (lane 9) or human (lane 13)milk, and the hLF antiserum only reacted to hLF (lane 8) and

C.T.Teng et al.

not with monkey lactoferrin (lane 7). There was some cross-reactivity between the mTF and the monkey lactoferrin-reactive70 kDa protein (lanes 10–12, lower band). The monkeyendometrial extracts of E2 and E2 � progesterone consistedof heterogenous cell populations from various zones of theendometrium and infiltrating neutrophils which have a highcontent of lactoferrin. For this reason, Western blot experimentsof the monkey endometrium can only provide qualitative butnot quantitative information on lactoferrin. Therefore, thelactoferrin-reactive protein from E2 and E2 � progesteronesamples showed equal intensity (Figure 5A, lanes 3 and 4).We also performed Western blotting on samples of monkeyprostate and endometrium with human and rhesus monkeymilk as control, in which 10 µg/ml rat transferrin was addedto the mLF 8344 antibody. Comparing the amino acid sequenceof the transferrin protein from mouse and rat (GeneBank, rat-d38380 and mouse tf- bco12310), a 90% sequence homologywas found. Therefore, the presence of neutralizing rattransferrin with the mLF antibody could block most of theantibody epitopes that recognize mTF. Under these conditionsthe mLF antibody detected only the 70 kDa lactoferrin-likeprotein in monkey endometrium (Figure 5B). The fast-movingbands apparent in the monkey milk and prostate samples maybe degradation products of lactoferrin.

To immunolocalize the lactoferrin protein in the monkeyendometrium, we isolated the IgG fractions from the mLF8344 antiserum and performed immunohistochemistry withthe mLF 8344 IgG purified fraction (Figure 6). The endometrialsamples examined in this study were from ovariectomizedmonkeys either untreated or treated with E2 or E2 � progester-one as described in Materials and methods. In preliminarywork, high background was seen in the stroma of E2 �progesterone samples. This background persisted even withthe purified mLF 8344 IgG fraction whereas the hLF 12484antiserum or IgG fraction was negative on all samples (datanot shown). It is possible that the high level of serum transferrinin the tissue could be recognized by the mTF epitopes of themLF. To minimize the interference of monkey transferrin onthe immunohistochemistry study, we added 1 mg/ml of rattransferrin to the mLF 8344 IgG during the overnight primaryantibody-staining step. Inclusion of transferrin greatly reducedall background staining and revealed specific lactoferrin stain-ing in the endometrial glands. The residual positive stainingin the stroma could be due to lactoferrin synthesized by anunknown cell type or released from the secondary granules ofthe infiltrating neutrophils (Masson et al., 1969) during hor-monal treatment. The staining was absent in the untreatedanimals, strong in the glands of E2-treated animals, and greatlysuppressed in the glands of the E2 � progesterone-treatedrhesus macaque monkeys (Figure 6, compare A, B andC). These results demonstrated that estrogen induces, andsequential progesterone suppresses, lactoferrin gene expressionin the glands of monkey endometrium (Figure 6D–F).

DiscussionEstrogen regulation of lactoferrin gene expression in the mouseuterus is well established (Teng, 1999). However, there are

64

inconsistent reports on estrogen regulation of the lactoferringene in human endometrium (Tourville et al., 1970; Tenget al., 1992; Walmer et al., 1995; Kelver et al., 1996). Thepresent study demonstrated that the estrogen receptor binds tothe imperfect ERE of human lactoferrin gene both in isolationand in the context of the gene promoter region. This findingsupported our previous observations that the human lactoferringene is activated in transiently transfected human endometrialcarcinoma cells (RL95-2) through an ER-mediated processand that the imperfect ERE of the gene is required (Teng et al.,1992). In addition to the in vitro physical interaction betweenER and the ERE of the human lactoferrin gene, we demon-strated that lactoferrin gene expression in the endometriumfluctuates during the menstrual cycle. In Western blottingexperiments, lactoferrin was detected in the proliferative phasebut not in the secretory phase, suggesting that estrogen inducesand progesterone suppresses lactoferrin gene expression invivo. It is unlikely that lactoferrin from neutrophils confoundedthese results, as more are present in the endometrium duringthe secretory phase when lactoferrin is marginally detectableon the Western blots. There are two major differences betweenthe current study and the previous studies on the same subjectmatter (Tourville et al., 1970; Teng et al., 1992; Walmer et al.,1995; Kelver et al., 1996). First, the tissue samples werecollected by carefully scraping the functionalis endometriumoff the basalis zone and the myometrium. This provides arelatively homogeneous cell population. Second, the epitopesof the antiserum that recognizes transferrin were neutralizedby a high level of transferrin protein incorporated in theblotting procedure. This step significantly reduced any interfer-ence caused by close family members of lactoferrin.

In the present study, we also investigated the use of a non-human primate model to study estrogen regulation of lactoferringene expression in the endometrium. The lactoferrin of rhesusmonkey has high similarity to the human protein in both aminoacid composition and carbohydrate moiety (Davidson andLonnerdal, 1986). Moreover, monkey lactoferrin, like humanlactoferrin, has an unusual amino acid sequence at the N-terminus which is essential for binding to bacterial lipopoly-saccharide and to the mammalian lactoferrin receptor (Wuet al., 1995; van Berkel et al., 1997), indicating that monkeylactoferrin could function similarly to the human protein. Thereproductive physiology of the female rhesus macaque monkeyis also highly comparable to that of women (Slayden et al.,1993). However, the polyclonal antibody produced againsthuman lactoferrin did not cross-react with the monkey milklactoferrin nor did it detect any immunoreactive protein in theendometrial tissue extracts or in the immunostaining study(data not shown). However, the rabbit anti-mouse lactoferrinserum, mLF8344, reacted to an 84 kDa lactoferrin in themonkey milk and to a 70 kDa protein, presumed to belactoferrin, in the endometrium. With this antibody, we wereable to show that lactoferrin-reactive protein in monkey endo-metrium is up-regulated by estrogen. This finding is in agree-ment with the human study showing that lactoferrin geneexpression is increased under the influence of estrogen.

A reduced immunostaining for lactoferrin in the endometrialglands after E2 � progesterone treatment suggests that a high


Figure 6. Effect of estrogen on lactoferrin expression in the monkey endometrium. Ovariectomized monkeys were either untreated ortreated with estrogen only or estrogen plus progesterone (E � P). (A) Untreated. (B) Treated with estrogen for 14 days. (C) Treatedwith estrogen for 14 days plus another 14 days of estrogen and progesterone. The original magnification of untreated endometriumwas �100 (A) and the other two samples were �40 (B and C). (D–F) Higher magnifications (original magnification �400) of theglands from each sample are shown. Rat transferrin (1 mg/ml) was included with the mLF 8344 IgG (10 µg/ml) in theimmunostaining procedure. The pink stain indicates the presence of lactoferrin. The tissue sections were counter-stained withhaematoxylin. Scale bars � 100 µm. En � endometrium; My � myometrium.

level of progesterone may block the affects of E2 on lactoferrin,probably by suppressing glandular estrogen receptors, aspreviously reported (West and Brenner, 1985). Nonetheless,in the basal glands of the endometrium, the intensity oflactoferrin staining was similar in both estrogen- and progester-one-dominant stages. This observation is consistent with thefinding that 2–56% of the basal glands in the human endomet-

65

rium at any give time during the menstrual cycle are lactoferrinpositive (Walmer et al., 1995). Whether lactoferrin geneexpression in the human and monkey endometrium exhibitregional specificity is under investigation. Based on the organ-ization of the estrogen response module and the transcriptionfactors that are involved in estrogen action (Teng et al., 1992;Yang et al., 1996; Teng, 1999), the molecular mechanisms

C.T.Teng et al.

of estrogen action in the human and non-human primateendometrium could be very different from that in the mouseuterus. Multiple levels of control may be required to fine-tunethe estrogenic effect in human endometrial cells.

The 70 kDa immunoreactive band detected in the monkeyendometrium and presumed to be lactoferrin, is smaller thanthe milk protein. This 70 kDa smaller form of lactoferrin wasalso detected in the prostate. Interestingly, lactoferrin in thehuman prostate is also smaller than that in the milk. Proteinmodification and presence of isoforms could contribute to thesize differences in various tissues. Isoforms of lactoferrin havebeen identified in the human milk (Furmanski and Fortuna,1989), granulocytes (Furmanski and Li, 1990) and seminalplasma (Sorrentino et al., 1999). These isoforms share physical,chemical and antigenic properties with lactoferrin, yet differin functions. Recently, a subset of cytotrophoblasts of thehuman placenta was reported to be recognized by a lactoferrinmonoclonal antibody but not by several other polyclonal andmonoclonal antibodies to human lactoferrin, suggesting thatthe cytotrophoblasts express a unique epitope of lactoferrin(Thaler et al., 1999). In addition, an alternative form of humanlactoferrin mRNA that is expressed in adult and fetal tissuesbut not tumour-derived cell lines has been described (Siebertand Huang, 1997). The biological significance of lactoferrinand its isoforms in these tissues is not known. It is welldocumented that estrogen regulates the mucosal immunesystem of the female rodent reproductive tract (Wira and Stern,1992; Wira and Kaushic, 1996) and it is therefore likely thatlactoferrin, an immunomodulator, could participate in themucosal immunity of the primate endometrium.

AcknowledgementsWe thank Malcom Parker, Pierre Chambon and Nenyu Yang for thereagents used in this study and we appreciate Barbara Davis andSylvia Hewitt for reading the manuscript.

ReferencesBush, M.R., Mele, J.M., Couchman, G.M. and Walmer, D.K. (1998) Evidence

of juxtacrine signaling for transforming growth factor α in humanendometrium. Biol. Reprod., 59, 1522–1529.

Davidson, L.A. and Lonnerdal, B. (1986) Isolation and characterization ofrhesus monkey milk lactoferrin. Pediatr. Res., 20, 197–201.

Furmanski, P., Li, Z.P., Fortuna, M.B. Fortuna, M.B., Swamy, C.V. and Das,M.T. (1989) Multiple molecular forms of human lactoferrin. J. Exp. Med.,170, 415–429.

Furmanski, P. and Li, Z.P. (1990) Multiple forms of lactoferrin in normal andleukemic human granulocytes. Exp. Hematol., 18, 932–935.

Kelver, M.E., Kaul, A., Nowicki, B. Findley, W.E., Hutchens, T.W. andNagamani, M. (1996) Estrogen regulation of lactoferrin expression inhuman endometrium. Am. J. Reprod. Immunol., 36, 243–247.

Levay, P.E. and Viljoen, M. (1995) Lactoferrin: a general review.Haematologica, 80, 252–267.

Liu, Y.H. and Teng, C.T. (1991) Characterization of estrogen-responsivemouse lactoferrin promoter. J. Biol. Chem., 266, 21880–21885.

Liu, Y.H. and Teng, C.T. (1992) Estrogen response module of the mouselactoferrin gene contains overlapping chicken ovalbumin up-streampromoter transcription factor and estrogen receptor-binding elements. Mol.Endocrinol., 6, 355–364.

Liu, Y.H., Yang, N.Y. and Teng, C.T. (1993) COUP-TF acts as a competitiverepressor for estrogen receptor-mediated activation of the mouse lactoferringene. Mol. Cell. Biol., 13, 1836–1846.

66

Lonnerdal, B. and Iyer, S. (1995) Lactoferrin molecular structure and biologicalfunction. Annu. Rev. Nutr., 15, 93–110.

Masson, P.L., Heremans, J.F. and Ferin, J. (1968) Presence of an iron bindingprotein (lactoferrin) in the genital tract of the human female. Fertil. Steril.,19, 679–689.

Masson, P.L., Heremans, J.F. and Schonne, E. (1969) Lactoferrin, an iron-binding protein in neutrophilic leukocytes. J. Exp. Med., 130, 643–658.

Newbold, R.R., Teng, C.T., Beckman, W.C. Jr, Jefferson, W.N., Hanson, R.B.,Miller, J.V. and McLachlan, J.A. (1992) Fluctuations of lactoferrin proteinand messenger ribonucleic acid in the reproductive tract of the mouseduring the estrous cycle. Biol. Reprod., 47, 903–915.

Nuijens, J.H., van Berkel, P.H.C. and Schanbacher, F.L. (1996) Structure andbiological action of lactoferrin. J. Mam. Gland Biol. Neopl., 1,285–295.

Panella, T.J., Liu, Y.H., Huang, A.T. and Teng, C.T. (1991) Polymorphismand altered methylation of the lactoferrin gene in normal leukocytes,leukemic cells, and breast cancer. Cancer Res., 51, 3037–3043.

Pentecost, B.T. and Teng, C.T. (1987) Lactotransferrin is the major estrogeninducible protein of mouse uterine secretions. J. Biol. Chem., 262,10134–10139.

Sanchez, L., Calvo, M. and Brock, J.H. (1992) Biological role of lactoferrin.Arch. Dis. Child., 67, 657–661.

Shi, H.P. and Teng, C.T. (1994) Characterization of a mitogen reponse unitin the mouse lactoferrin gene promoter. J. Biol. Chem., 269, 12973–12980.

Shi, H.P. and Teng, C.T. (1996) Promoter specific activation of mouselactoferin gene by EGF involves two adjacent regulatory elements. Mol.Enodcrinol., 10, 732–741.

Siebert, P.D. and Huang, B.C.B. (1997) Identification of an alternative formof human lactoferrin mRNA that is expressed differentially in normaltissues and tumor-derived cell lines. Proc. Natl. Acad. Sci. USA, 94,2189–2203.

Slayden, O.D., Hirst, J.J. and Brenner, R.M. (1993) Estrogen action inthe reproductive tract of rhesus monkeys during antiprogestin treatment.Endocrinology, 132, 1845–1856.

Sorrentino, S., D’Alessandro, A.M., Maras, B. Di Ciccio, L., D’Andrea, G.,De Prisco, R., Bossa, F., Libonati, M. and Oratore, A. (1999) Purificationof a 76-kDa iron-binding protein from human seminal plasma by affinitychromatography specific for ribonuclease: structural and functional identitywith milk lactoferrin. Biochim. Biophys. Acta, 1430, 103–110.

Teng, C.T. (1999) Regulation of lactoferrin gene expression by estrogen andepidermal growth factor. Cell. Biochem. Biophys., 31, 49–64.

Teng, C.T., Walker, M.P., Bhattacharya, S.N. Klapper, D.G., DiAugustine,R.P. and McLachlan, J.A. (1986) Purification and properties of an estrogen-stimulated mouse uterine glycoprotein (approx. 70 kDa). Biochem. J., 240,413–422.

Teng, C.T., Liu, Y.H., Yang, N.Y. Walmer, D. and Panella, T. (1992) Differentialmolecular mechanism of the estrogen action that regulates lactoferrin genein human and mouse. Mol. Endocrinol., 6, 1969–1981.

Thaler, C.J., Labarrere, C.A., Hunt, J.S., McIntyre, J.A. and Faulk, W.P.(1999) Unique epitopes of lactoferrin expressed in human cytotrophoblastsinvolved in immunologic reactions. Am. J. Obstet. Cynecol., 181, 460–467.

Tourville, D.R., Ogra, S.S., Lippes, J. and Tomasi, T.B. Jr (1970) The humanfemale reproductive tract: immunohistological localization of γA, γG. γM,secretory ‘piece,’ and lactoferrin. Am. J. Obstet. Gynecol., 108, 1102–1108.

van Berkel, P.H.C., Geerts, M.E.J., van Veen, H.A. Mericskay, M., de Boer,H.A. and Nuijens, J.H. (1997) N-terminal stretch Arg2, Arg3, Arg4 andArg5 of human lactoferrin is essential for binding to heparin, bacteriallipopolysaccharide, human lysozyme and DNA. Biochem. J., 328, 145–151.

Walmer, D.K., Wrona, M.A., Hughes, C.L. and Nelson, K.G. (1992) Lactoferrinexpression in the mouse reproductive tract during the natural estrous cycle:correlation with circulating estradiol and progesterone. Endocrinology, 131,1458–1466.

Walmer, D.K., Padin, C.J., Wrona, M.A., Healy, B.E., Bentley, R.C., Tsao,M.S., Kohler, M.F., McLachlan, J.A. and Gray, K.D. (1995) Malignanttransformation of the human endometrium is associated with overexpressionof lactoferrin messenger RNA and protein. Cancer Res., 55, 1168–1175.

West, N. and Brenner, R. (1985) Progesterone mediated suppression ofestradiol receptors in cynomolgus macaque cervix, endometrium and oviductduring sequential estradiol-progesterone treatment. J. Steroid Biochem., 22,29–37.

Wira, C.R. and Kaushic, C. (1996) Mucosal immunity in the femalereproductive tract: effect of sex hormones on immune recognition andresponses. In Kiyono, H., Ogra, P.L. and McGhee, J.R. (eds), Mucosal


Vaccines: New Trends in Immunization. Academic Press, New York,pp. 375–388.

Wira, C.R. and Stern, J. (1992) Endocrine regulation of the mucosalimmune system in the female reproductive tract: control of IgA, IgG,and secretory component during the reproductive cycle, at implantationand throughout pregnancy. In Pasqualini, J.R. and Scholler, R. (eds),Hormones and Fetal Pathophysiology. Marcel Dekker, New York,pp. 343–368.

Wu, H.F., Monroe, D.M. and Church, F.C. (1995) Characterization of theglycosaminoglycan-binding region of lactoferrin. Arch. Biochem. Biophys.,317, 85–92.

67

Yang, N.Y. and Teng, C.T. (1994) Identification of COUP-TF bindingelement in the human lactoferrin promoter. Endocrine J., 2, 241–248.

Yang, N.Y., Shigeta, H., Shi, H.P. and Teng, C.T. (1996) Estrogen relatedreceptor, hERR1, modulates estrogen receptor mediated response ofhuman lactoferrin gene promoter. J. Biol. Chem., 271, 5795–5804.

Zhang, Z.P. and Teng, C.T. (2000) Estrogen receptor-related receptor α1interacts with coactivator and constitutively activates the estrogenresponse elements of the human lactoferrin gene. J. Biol. Chem., 275,20837–20846.

Received on May 8, 2001; accepted on October 10, 2001

lactoferrin gene expression is estrogen responsive in human and

Documents