Endometrial epithelial integrity and subepithelial reticular fibre

Human Reproduction, Vol. 15, (Suppl. 3), pp. 189-196, 2000

Endometrial epithelial integrity and subepithelialreticular fibre expression in progestin contraceptive

acceptors

Sugito Wonodirekso1'5, Biran Affandi2, Budiningsih Siregar3,Atikah C.Barasila1, Lia Damayanti1 and Peter A.W.Rogers4

1 Department of Histology, 2Department of Obstetrics and Gynaecology and 3 Department of AnatomicPathology, Faculty of Medicine, University of Indonesia, Jakarta and 4Monash University Department of

Obstetrics and Gynaecology, Monash Medical Centre, Clayton, Victoria, Australia5To whom correspondence should be addressed at: Department of Histology, Faculty of Medicine,

University of Indonesia, Jakarta. E-mail: [email protected]

Long-acting progestin contraceptives have beenavailable in many countries for a number ofyears with a large number of women now usingthem. Although some improvements in deliverysystems have been made, the major problemwith progestin-only contraceptives remainsunpredictable endometrial breakthrough bleed-ing (BTB), which is responsible for more than50% of drop-outs from this form of contracep-tion. Using hysteroscopy, endometrial petechiaeand ecchymoses are a common finding amongNorplant users, although these features do notalways correlate with BTB. It has been postu-lated that epithelial and subepithelial tissuesmay provide a barrier to BTB, as long asepithelial integrity is maintained. The aim ofthis pilot study is to explore structural changesin the endometrial surface epithelium, andsubepithelial collagen III fibres. Endometrialbiopsies from noresthisterone-enanthate (Net-En) users (n = 6) and controls (n = 6) wereassessed using routine haematoxylin and eosinstaining and immunohistochemical staining forcytokeratins 8, 18 and 19, and collagen III. Aconventional silver impregnation method wasalso used to identify subepithelial collagen IIIfibres. Most of the Net-En tissues showedreduced surface epithelial cell height comparedto controls (P = 0.002). Cytokeratin stainingwas weaker (P = 0.04) and distributed evenlybetween basal and apical parts of the cell inNet-En tissue, compared to more apically incontrols. Both immunohistochemical and con-

ventional silver staining methods revealed thatthe subepithelial collagen III meshworkremained unchanged in Net-En compared tocontrol endometrium. Both staining methodsidentified collagen fibres with equal sensitivity.In conclusion, atrophic changes remain the dom-inant appearance for progestin-exposed endo-metrium, with reduced cytokeratin staining, butapparently there is little change in subepithelialcollagen III expression.Key words: breakthrough bleeding/collagen III/endometrial epithelium/norethi sterone-enanthate/progestin-only contraceptives

Introduction

Progestin contraceptives are safe, effective, eco-nomical and have been used worldwide for anumber of years. They are available as oral pills,long-acting injections, and implants. Unfortunately,all of them cause some unpredictable endometrialbreakthrough bleeding (BTB) that is responsiblefor more than 50% of drop-outs from this form ofcontraception. To date, no definitive mechanismresponsible for BTB has been identified and con-sequently no rational treatment approach is avail-able. Several investigations have exploredendometrial structural changes due to the progestineffects, including vascular density (Rogers et al.,1993; Rogers, 1996), oestrogen and progesteronereceptors (Critchley et al., 1993), and basementmembrane components (Palmer et al., 1996). How-

© European Society of Human Reproduction and Embryology 189

Downloaded from https://academic.oup.com/humrep/article-abstract/15/suppl_3/189/652023by gueston 20 March 2018

S.Wonodirekso et al.

ever, despite endometrial structural changes beingevident with progestin-only use, none of thesecorrelated with the occurrence of BTB.

It has been noted that petechiae and ecchymosesare common in the endometrium of progestincontraceptive users, and that the endometriumbleeds easily following minor trauma (Hickeyet al, 1996) Despite these observations, this studyfound that not all of the progestin contraceptiveusers had BTB. A possible explanation for thisfinding could be that BTB was prevented bythe integrity of the endometrial epithelial andsubepithelial tissues. Several components play arole in maintaining epithelial integrity, includingcytokeratins, intercellular adhesion molecules, andbasement membrane. To maintain their protectivefunction, all of these components need to beanchored firmly to the stromal tissue through theunderlying reticular fibre meshwork.

Endometrial epithelium contains cytokeratins 7,8, 18 and 19, from the intracellular intermediatefilament family. Most of these cytoskeletal fibresare meshed around the nucleus in both surface andglandular epithelium and determine the epithelialshape. Some bundles of fibres are anchored todesmosomes to maintain epithelial cell shape andadhesion to neighbouring cells, and also to hemi-desmosomes through specific proteins to anchorthe epithelium to the basement membrane (Quinlanet al, 1985; Nagle et al, 1989, 1990; Skalli et al,1994). These intermediate filaments are dynamicstructures and are subject to changes induced byseveral external influences (Gijbels et al, 1992).Such influences may include cyclical hormonalchanges during the menstrual cycle and long-termprogestin exposure.

Human endometrium contains collagen type Iwhich is predominantly distributed in the interstitialspaces and considered to be more involved instromal integrity than epithelial integrity. Basementmembrane contains mainly laminin and collagenIV, whereas collagen III (reticular fibres) is mainlyfound beneath the basement membrane of surfaceand glandular epithelium, and around the vascularwall. Collagen III fibres also occupy, but at alesser density, the extracellular space of the stroma(Stoval et al, 1992). Collagen III fibres are alsoknown as argyrophilic fibres due to their affinity

for silver staining. Epithelial integrity, and theunderlying collagen fibres to which the epitheliumis anchored, protect the underlying stromal tissueand may act to prevent small haemorrhages fromleaking into the uterine lumen. In normal menstru-ation it has been postulated that bleeding wasinitiated by tissue degradation due to matrixmetalloproteinase (MMP) activation because ofsudden progesterone deprivation (Marbaix et al,1996; Salamonsen and Woolley, 1996). This mech-anism for BTB seems less likely in long-actingprogestin acceptors since the systemic progestinconcentration remains relatively constant.

Norethisterone-enanthate (Net-En) is a progestincontraceptive injected i.m., and has been usedworldwide. (Howard et al, 1982, 1986; Joshiet al, 1988). Chemically, Net-En is a C-18 steroidbelonging to the 19-nortestosterone group (17-ethiny 1- P-heptanoiloxy-4-esterene-one) dispensedin an oily solvent mixture of benzyl benzoateand castor oil. Net-En has been shown to arrestendometrial growth in the rat (Bhowmik et al,1988).

The aim of this investigation was to observe, innormal control and Net-En endometrium, changesin: the surface epithelium, cytokeratin expressionand distribution, cellular shape and arrangementand the underlying reticular fibre expression andarrangement.

Materials and methods

Subjects

In this study we examined 12 endometrial biopsysamples collected from women of reproductive age(range 25^4-0, mean 32.33 years). The controlgroup consisted of six normal endometria fromearly and late proliferative and secretory stages,recruited from infertility subjects attending MonashMedical Centre gynaecological clinic where theinfertility was shown to be due to male factorproblems. The treatment group consisted of sixNet-En (Schering AG, Germany) endometriacollected from women who had received Net-Enfor more than 2 years, recruited from Raden SalehClinic, Jakarta.

Bleeding patterns

Bleeding patterns among Net-En users wererecorded daily by the subject during a 90 day

190


Progestin action on endometrial epithelium

reference period prior to endometrial biopsy. Dueto small numbers of subjects, bleeding patternswere classified into two categories: heavy bleeding(bleeding/spotting >10 days) (n = 1) and lightbleeding (bleeding/spotting <10 days) (n = 5).

Histological and immunohistochemical stainingmethods

Routine haematoxylin and eosin (H&E) stainingwas applied to formalin-fixed, paraffin-embeddedsections of all the specimens to assess the appear-ance of surface epithelial cell shape. To identifycytokeratin 8, 18 and 19 expression, biotin-strepta-vidin amplified detection immunohistochemistrystaining methods were applied to paraffin blockspecimens using monoclonal primary antibody tocytokeratins 8, 18 and 19 (BioGenex, San Ramon,CA, USA; No. AM131-5M). Antigen retrieval withpepsin (BioGenex; EK000-05) was applied priorto the primary antibody. For labelling, the super-sensitive secondary antibody immunodetection sys-tem (QP900-9L) and DAB (BioGenex; HK093-5K) which stains a brown colour, were used.

Immunohistochemical methods were also usedto identify collagen III expression using a similarantigen retrieval and labelling system as appliedto the cytokeratins. Supersensitive anticollagen IIIprimary antibody (BioGenex; AM167-5M) wasused. A conventional histochemical silver stainingmethod was also employed to identify collagen IIIas this has been shown to give good results inrecent studies (Marbaix et al., 1996). Thisstaining is specific for collagen III fibres whichare also called reticular fibres or argyrophilic fibres(Culling, 1980).

Scoring

The epithelial cells were classified as high colum-nar and given a score of 3 when their height wasequal to or higher than twice their width. A scoreof 2 was given to low columnar cell shapes whenheight was less than twice the width. A score of1 was given to cuboidal cell shapes when heightwas equal to or less than width. Ocular andobjective micrometer calibration based on the 20(objective) X10 (ocular) magnification was used tomeasure the epithelial height.

The endometrium was considered atrophic when

the surface epithelium is low (score 1) and nostromal proliferative profiles could be found.

Epithelial damage was defined as epithelial dis-continuity covered by blood or mucus with leuko-cyte infiltration within the underlying stroma. Ascore of 0 was given for no damage and 1 fordamage.

For cytokeratin expression, 0 = negative stain-ing, 1 = weak staining, 2 = moderate staining,3 = strong staining, and 4 = intense staining(Wonodirekso et al., 1996). Results for cytokeratindistribution within the cell were descriptive only.

Collagen III expression was given a score of 2when it appeared continuous and 1 if not continu-ous. In addition, score 2 was given when it appearedthick (the dense plus the less dense collagen layersdirectly under the surface epithelium were S IO|im) and 1 when it was thin (the denser plus theless dense collagen layers directly under the surfaceepithelium were <10 |im). This scoring was alsoused to compare immunostaining and silver stain-ing results for collagen III. Ocular and objectivemicrometer calibration based on the 20(objective) X10 (ocular) magnification was used tomeasure the collagen layer thickness.

Normal endometrial tissues were classified intoearly proliferative (EP), late proliferative (LP),early secretory (ES) and late secretory (LS). Resultsrepresent the dominant findings in each specimen.Only specimens found to contain surface epithe-lium using routine H&E staining were included inthis study. A number of specimens were excludeddue to lack of surface epithelium in the biopsyspecimen.

Results

Endometrial stages and bleeding patterns

By histopathological analysis the control endomet-rium was divided into one EP, one LP, two ESand two LS specimens. The Net-En endometriummostly showed atrophic change (five of six), withfive scored as EP and showing lower surfaceepithelium and/or scarce glands. One specimenwas LP. Most of the Net-En tissues came fromlight bleeders (five of six) with only one experien-cing heavy bleeding during the 90 day referenceperiod (Table I).

191



Epithelial shape and damage

In all of the assessments, the results recorded inthe tables represent the dominant or most commonappearance within each specimen. All control endo-metria showed simple high columnar epithelium(SHCol) with no epithelial damage evident (Figure1A). By contrast the Net-En endometria werevaried and overall the epithelium was significantlylower (P = 0.002). One of the Net-En specimensshowed SHCol and this was also the specimen thatwas classified as LP. Three Net-En specimensshowed simple low columnar (SLCol) epitheliumand the other two showed simple cuboidal (SCub)epithelium with some squamous cells (Figure IBand 1C, and Table II). These last five Net-Enspecimens were all classified as EP. The heavy

Table I. Menstrual cycle12 subjects in this study

Endometrialtissue

ControlNet-En

Stages

EP

15

stages and bleeding patterns

and bleeding patterns

LP ES LS LB

1 2 2 -1 0 0 5

for the

HB

1

EP = early proliferative; LP = late proliferative; ES = earlysecretory; LS = late secretory; LB = light bleeding; HB =heavy bleeding; Net-En = noresthisterone-enanthate.

Figure 1. Haematoxylin and eosin staining of controlendometrium and endometrium from noresthisterone-enanthate (Net-En) users. (A) Proliferative endometrium fromcontrol showing simple high columnar epithelium. (B)Cuboidal epithelium from a Net-En user. (C) Squamous (seearrows) epithelium from a Net-En user. Originalmagnification X200.

bleeder showed SLCol epithelium. Histologicalevidence of subepithelial and deeper stromal bleed-ing was found in almost all the Net-En specimens.Some of this bleeding may have occurred duringthe biopsy procedure, or could have intratissueBTB from before the biopsy (Figure 2). Epithelialdamage of different severity and an underlyingstromal inflammatory reaction (Figure 2), wasfound in all of the Net-En endometria.

Cytokeratin distribution and staining intensity

Three of the control endometria with apical cyto-keratin distribution were from EP, LP and ESstages. The other three had equal cytokeratin distri-bution and were from ES and LS stages. In contrast,all Net-En endometria showed equal distributionof cytokeratin (Table III). It should be noted thatthe specimens were usually not uniformly stained(Figure 3) and the most typical appearance isrepresented in Table III.

Table II. Epithelial shapes and damage for the 12 subjects inthis study

Endometrialtissue

Epithelial cell shape and damage

SHCol SLCol SCub Damage

ControlNet-En

SHCol = simple high columnar; SLCol = simple lowcolumnar; SCub = simple cuboidal; epithelial damage =damage; Net-En = noresthisterone-enanthate.

Figure 2. Collagen III immunostaining of noresthisterone-enanthate endometrium. Note the relatively intact collagen IIImeshwork under the damaged epithelium (arrow). It isprobable that the damage occurred prior to biopsy, sinceleukocyte infiltration is visible beneath the stroma. Originalmagnification X200.

192



Cytokeratin immunostaining intensity resultswere quite variable but the control endometriatended to have stronger staining and the Net-Entissue tended to be weaker (P = 0.04). All ofthe specimens showed some immunostaining forcytokeratin (Table IV). Cytokeratin was clearlyseen intracytoplasmically in both surface andglandular epithelium as light to dark fine brownfibres meshed around the nucleus. Stromal cellsand extracellular spaces were negative.

Table III. Epithelial cytoplasmic cytokeratin distributionfound by immunohistochemical staining for cytokeratins 8,18 and 19

Endometrialtissue

ControlNet-En

Cytokeratin distribution

Apical Equal

3 30 6

Basal

00

Distribution is either predominantly apical or basal to the cellnucleus.

Figure 3. Immunostaining for cytokeratins 8, 18 and 19 innoresthisterone-enanthate (Net-En) (A and B) and normalsecretory endometrium (C). (A) Weak staining; (B) moderatestaining; (C) strong staining. Original magnification X200.

Collagen III immunostaining and silverstaining

Collagen III immunostaining demonstrated a brownto dark brown subepithelial meshwork whichappeared continuous and thick in both control andNet-En tissues (Table V). Fine individual fibresand a looser network occupied interstitial spacesbetween stromal cells (Figure 4A). Only one ofthe Net-En tissues showed a thinner arrangement ofthe collagen III network under surface epithelium.There was no significant difference in immuno-staining results between normal and Net-En tissue.

Collagen III, which was formerly identifiedas reticular fibres, can also be identified by itsargyrophilic properties using a silver impregnationstaining procedure. This technique gave goodresults as shown in Table VI. All of the controltissues and five of the Net-En tissues showed acontinuous, thick, meshwork arrangement (Figure4B), while one of the Net-En tissues showed athinner, discontinuous meshwork (Figure 4C).There were no significant differences in stainingresults between control and Net-En tissues usingsilver staining (not significant), although the indi-

Table V. Collagen III immunostaining resultscontinuity and thickness of the fibre network

Endometrialtissue

ControlNet-En

Collagen III

Continuous

66

showing

immunostaining pattern

Notcontinuous

00

Thick

65

the

Thin

01

'Thick' is defined as both the denser layer directly under theepithelium and the thinner network underneath it being 3^10|j.m thick, and 'thin' indicates that they are <10|J.m.Net-En = noresthisterone-enanthate.

Table IV. Endometrial surface epithelium cytokeratinEn) subjects

Endometrial tissue

ControlNet-En

Cytokeratin staining intensity

Intense Strong

1 30 2

immunostaining intensity

Moderate

21

for control and noresthisterone-enanthate (Net-

Weak

03

Negative

00

Immunohistochemical staining was with an antibody to cytokeratins 8, 18 and 19.

193



Figure 4. (A) Collagen III immunostaining; (B, C) collagenIII silver staining. (A) and (B) show thick continuous layers;(C) shows a thinner broken layer. (A) and (B) were takenfrom normal proliferative endometrium; (C) was taken fromnoresthisterone-enanthate endometrium. Originalmagnification X200.

Table VI. Collagen III silver staining results showing thecontinuity and arrangement of the fibres

Endometnaltissue

ControlNet-En

Collagen III silver staining pattern

Continuous Not Thick Thincontinuous

Net-En = noresthisterone-enanthate.

vidual collagen III fibres looked thicker with silvercompared to immunohistochemical staining. Bothsilver and immunostaining for collagen III gavesimilar results (not significant).

Discussion

The main finding of this study was lowersurface epithelium in Net-En users. These findingsagree with the results of earlier studies that showedthat long-term progestin contraceptive exposureresults in endometnal atrophy with reduced glandnumber and size, stromal thickness, and epithelialcell height (Bhowmik and Mukherjea, 1988;

Johannissonefa/., 1991; Wonodireksoe?a/., 1993).Whereas all of the control tissues showed simplehigh columnar epithelium, only one of the Net-Entissues had this appearance (Table II). In addition,almost all of the Net-En tissues were classified asat the EP stage of the cycle. Only one of them wasLP, possibly as a result of endogenous oestrogenproduction by a growing ovarian follicle.

In control endometria, cytokeratin immuno-staining was stronger than in Net-En tissues andthe cells were significantly higher. In control tissuesthe cytokeratin distribution was either apical orequal. These two different patterns could be dueto changes during the menstrual cycle since endo-metrial epithelium develops secretory phasemucus-filled vacuoles whose contents could dis-place cytoplasmic cytokeratin.

Variable results were found with Net-En tissueimmunostaining for cytokeratins. Four specimensshowed moderate or weaker staining. Net-En epi-thelial cells also tended to be lower in height. Allof the Net-En tissues showed equal cytokeratindistribution, possibly indicating that the epithelialcells were not fully active or developed (Gijbelset al., 1992). Similar findings have also beenpublished for Norplant exposed endometriumwhere the epithelium is usually lower with roundedcells and weaker cytokeratin immunostaining(Wonodirekso et al, 1993, 1996). These findingsmay also be due to inhibition of the ovarian cycleleading to endometrial atrophy.

Stromal bleeding in superficial and/or deeperregions was found in all Net-En specimens despitethe fact that only one of these subjects experiencedheavy bleeding. This finding supports the resultsof a hysteroscopic investigation which found thatpetechiae and ecchymoses were common in Norpl-ant-exposed endometrium without necessarilybeing accompanied by discernable BTB (Hickeyet al., 1996). It is possible that overt BTB may beprevented by an intact endometrial epithelium.Even if the epithelium is denuded, it is still possiblethat the underlying reticular fibres, which werefound in this study to be relatively unchanged inNet-En users, may act to stop the bleeding byinitiating blood clotting. Thus epithelial integrityand the underlying collagen III fibres may becrucial in preventing blood leaking into the uterine

194



lumen following endometrial microvascularrupture.

In control cycles, endometrial epithelial develop-ment is regulated in part by the influence ofoestrogen, with oestrogen receptor concentrationsbeing higher in the proliferative phase of thecycle (Coppens et al, 1993; Lessey et al, 1998).Endometrial epithelial oestrogen receptor expres-sion is down-regulated by Norplant and a levonor-gestrel releasing intrauterine device (Critchleyet al, 1993, 1998). Since oestrogen can initiateendometrial development through proliferation ofboth stroma and epithelium, it has been used bysome investigators to combat BTB problems. Todate, the effectiveness of this approach has beenvaried (Archer et al, 1996; Boonkasemsanti et al.,1996; WHO, Special Programme of Research,1996; Witjaksono et al, 1996).

Collagen III fibres are found in several tissues,including endometrium, in which they serve as asupporting meshwork under the basement mem-brane of epithelia and around blood vessel walls.(Stoval et al, 1992; Lodish et al, 1995). As shownin Figure 4, in both control and Net-En tissuesthe fibres are clearly visible under the surfaceepithelium. Similar features can also be seen usingsilver staining (Figure 4B,C). The fibres remaincontinuous and thick in almost all of the specimensin both control and Net-En tissues. Interestinglyunder the damaged or absent surface epithelium inNet-En tissues, the collagen fibre meshwork isfound to be intact (Figure 2). Thus the collagenIII fibre meshwork under the surface epitheliumcould also serve as a mechanical barrier to preventblood escaping into the endometrial cavity.

In summary, and in the light of the limitednumber of samples in this study, it is concluded thatatrophic changes are apparent in Net-En exposedendometrium, and cytokeratin immunostaining isweaker in Net-En tissues compared with controls.Despite these differences, there are minimalchanges in the endometrial collagen III fibre mesh-work in Net-En tissue compared with controls,suggesting that the collagen III meshwork couldact as a secondary barrier to the escape of bloodinto the uterine lumen.

AcknowledgementsThis research was funded by WHO grant 96908. Thanksare due to Sister Rosminah for subject recruitment.

ReferencesArcher, D.F., Philput, C.A. and Weber, M.E. (1996)

Management of irregular uterine bleeding and spottingassociated with Norplant®. Hum. Reprod., 11, 24-30.

Boonkasemsanti, W., Reinprayoon, D., Pruksananonda,K. et al. (1996) The effect of transdermal estradiolon bleeding pattern, hormonal profiles and sex steroidreceptor distribution in the endometrium of Norplantusers. Hum. Reprod., 11, 115-123.

Bhowmik, T. and Mukherjea, M. (1988) Histologicalchanges in the ovary and uterus of rat after injectablecontraceptive therapy. Contraception, 37, 529-539.

Coppens, M.T., Dhont, M.A., De Boever, J.G. et al.(1993) The distribution of estrogen and progesteronereceptors in the human endometrial basal andfunctional layer during the normal menstrual cycle.Histochemistry, 99, 121-126.

Critchley, H.O.D., Bailey, D.A., Au, C.L. et al. (1993)Immunohistochemical sex steroid receptor distributionin endometrium from long-term subdermallevonorgestrel users and during the normal menstrualcycle. Hum. Reprod., 8,1632-1639.

Critchley, H.O.D., Wang, H., Kelly, R.W. et al. (1998)Progestin receptor isoforms and prostaglandindehydrogenase in the endometrium of women usinglevonorgestrel-releasing system. Hum. Reprod., 13,1210-1217.

Culling, C.F.A. (1980) Connective tissue. In Handbookof Histopathological and Histochemical Techniques,3rd edn. Butterworth, Toronto, pp. 427^28.

Gijbels, M.J.J., van der Ham, R, van Bennekum, A.M.and Roholl, P.J.M. (1992) Alteration in cytokeratinexpression precedes histological changes in epitheliaof vitamin A-deficient rats. Cell Tiss. Res., 268,197-203.

Hickey, M., Fraser, I., Dwarte, D. and Graham, S.(1996) Endometrial vasculature in Norplant users:Preliminary results from hysteroscopic study. Hum.Reprod., 11, 35^4 .

Howard, G., Blair, M., Chen, J.K., Fortherby, K. et al.(1982) Clinical trial of norethysterone-enanthate(Norigest®) injected every two months.Contraception, 225, 333-340.

Howard, G., Blair, M., Fotherby, D. et al. (1986)Seven years clinical experience of the injectablecontraceptive, norethysterone-enanthate. Br. J. Fam.Plan., 11, 9-16.

Johannisson, E., Brossens, I., Cornillie, F. et al. (1991)Morphometric study of the human endometriumfollowing continuous exposure to levonorgestrelreleased from vaginal rings during 90 days.Contraception, 43, 61-74.

Joshi, J.V., Hazari, K.T., Shah, R.S. et al. (1988) Serumprogesterone and northisterone levels followinginjection of northisterone-enanthate in different sitesand doses. Steroid, 35, 751-761.

195



Lessey, B.A., Killam, A.P., Metzger, D.A. et al. (1998)Immunohstochemical analysis of human uterineestrogen and progesterone receptors throughout themenstrual cycle. J. Clin. Endocrinol. Metab., 67,334-40.

Lodish, H., Baltimore, D., Berk, A. et al. (1995) Multi-adhesive matrix proteins and their cell surfacereceptors. Mol. Cell Biol., 3rd edn. Scientific AmericanBook, New York, pp. 11431-1145.

Marbaix, E., Kokorine, I., Donnez, J. et al. (1996)Regulation and restricted expression of interstitialcollagenase suggest pivotal role in the initiation ofmenstruation. Hum. Reprod., 11, 134-143.

Nagle, R.B. (1989) Intermediate filaments. Am. J. Clin.Pathol., 94, S14-S18.

Palmer, J.A., Lau, T.M., Hickey, M. and Rogers, P.A.W.(1996) Immunohistochemical study of endometrialmicrovascular basement membrane components inwomen using Norplant. Hum. Reprod., 11, 2142-2150.

Quinlan, R.A., Schiller, D.L., Hatzfeld, M. et al. (1985)Pattern of expression and organization of cytokeratinintermediate filaments. Ann. NY Acad. Sci., 455,282-306.

Rogers, P.A.W. (1996) Endometrial vasculature inNorplant® users. Hum. Reprod., 11, 45-50.

Rogers, P.A.W., Au, C.L. and Affandi, B. (1993)Endometrial microvascular density during the normalmenstrual cycle and following exposure to long-termlevonorgestrel. Hum. Reprod., 8, 1396-1404.

Salamonsen, L. and Wooley, D.E. (1996) Matrixmetalloproteinases in normal menstruation. Hum.Reprod., 11, 124-33

Skalli, O., Jones, J.C., Gagescu, R. and Goldman, R.D.(1994) IFAP 300 is common to desmosomes andhemidesmosomes and is possible linker ofintermediate filaments to these junctions. J. Cell Biol.,125, 159-70.

Stoval, D.W., Anners, J.A. and Halme, J. (1992)Immunohistochemical detection of type I, III, and IVcollagen in endometrial implants. Fertil. Steril., 57,984-989.

World Health Organization, Special Program ofResearch, Development and Training in HumanReproduction, Task Force on Long-acting SystemicAgents for Fertility Regulation (1996) Clinicalevaluation of the therapeutic effectiveness of ethinylestradiol and oestrone sulphate on prolonged bleedingin women using depot medroxyprogesterone acetatefor contraception. Hum. Reprod., 11, 1-13.

Witjaksono, J., Lau, T.M., Affandi, B. and Rogers,P.A.W. (1996) Estrogen treatment for increasedbleeding in Norplant users. Preliminary results. Hum.Reprod., 11, 109-114.

Wonodirekso, S., Au, C.L., Hadisaputra, W. et al. (1993)Cytokeratins 8, 18 and 19 in endometrial epithelialcells during the normal menstrual cycle and in womenreceiving Norplant. Contraception, 48, 481^-93.

Wonodirekso, S., Hadisaputra, W, Affandi, B. et al.(1996) Cytokeratin 8, 18 and 19 in endometrialepithelium of Norplant and norethisterone enanthate(NET-EN) injectable progestogen contraceptive users.Hum. Reprod., 11, 144-149.

196


Documents

Endometrial epithelial integrity and subepithelial reticular fibre