Pregnancy Zone Protein Research Articles

Synthesis and secretion of α2‐macroglobulin by human hepatocytes in culture

Hepatocytes were isolated by application of the two-step collagenase perfusion technique to pieces of human liver. The cells were incubated in serum-free medium or 10% FCS-medium supplemented with insulin, glucagon and dexamethasone, and kept in culture for more than 2 weeks. Seventy-five per cent of the medium was changed regularly and assayed for alpha 2-macroglobulin (alpha 2-M), pregnancy zone protein, alpha 1-antitrypsin and albumin by means of ELISA. Significant amounts of alpha 2-macroglobulin were present in all cultures. During incubation, alpha 2-M accumulated in the medium and the quantity of alpha 2-M released from the cells by far exceeded protein associated with hepatocytes prior to incubation. In 24 h 10(6) hepatocytes secreted 160.5 +/- 82.2 ng of alpha 2-M (mean +/- SD, n = 5). Cell-associated, as well as secreted alpha 2-M appeared to be on native form, as determined by immunoisolates from lysed cells and culture supernatants. Pregnancy zone protein was only detected in about 50% of the cultures and its rate of secretion was less than 2 ng 24 h-1 per 10(6) cells. In contrast, culture medium contained considerable quantities of alpha 1-antitrypsin and albumin. In 24 h, 10(6) hepatocytes released greater than 2 micrograms alpha 1-antitrypsin and greater than 5 micrograms albumin. The present study suggests the hepatocyte to be of major importance for the synthesis of intravascular alpha 2-M.

Receptors for α2-Macroglobulin- and pregnancy zone protein-proteinase complexes in the human placental syncytiotrophoblast

Receptors for complexes between alpha 2-macroglobulin (alpha 2M) and proteinases (e.g. trypsin) were identified and characterized in the human placenta. The receptors also bound the complex formed between pregnancy zone protein (PZP) and chymotrypsin, although with slightly lower affinity, whereas binding of alpha 2M or PZP in their native forms was negligible. Treatment with methylamine to cleave the internal thiol esters caused an increase in binding affinity of alpha 2M to the level of alpha 2M-trypsin but only a minor increase in the affinity of PZP. Chorionic villi prepared from normal full-term placentae were approximately half occupied by endogenous alpha 2M or PZP complexes. These ligands, as well as prebound 125I-labelled alpha 2M-trypsin, were rapidly removed by the addition of 5 mM EDTA. Binding was similar in villi from eight-week and full-term placentae. Autoradiography showed that labelled alpha 2M-trypsin was associated with the syncytiotrophoblast. The kinetics of 125I-labelled alpha 2M-trypsin binding at 4 degrees C was similar in isolated villi and microvillous membranes. Association was slow, with apparent equilibrium by about 16 h. Dissociation of prebound tracer was slow but markedly accelerated in the presence of unlabelled ligand at a saturating concentration. The concentration-dependence of binding at equilibrium yielded a non-linear Scatchard plot. Most of the binding of ligand at tracer concentration was accounted for by high-affinity receptors with a dissociation constant (Kd) of about 50 pM. The content of high-affinity receptors in one placenta was estimated as approximately 125 pmol, i.e., a significant fraction of the total receptor population in the pregnant woman.

Mapping of structure-function relationships in proteins with a panel of monoclonal antibodies: A study on human α2 macroglobulin

Monoclonal antibody (Mab) F2B2, directed to the receptor-recognition site of human α 2 macroglobulin ( α 2M), has been instrumental in the characterization of that site and in the isolation of the receptor-binding domain. We have now prepared a panel of Mab to study the structure-function relationships in α 2M, and in particular the expression of the receptor-recognition site. Reversed dot-blotting was very effective to screen hybridoma supernatant for specificities to either the native or complex form of α 2M. Reaction with the isolated 20 kDa receptor-binding domain of α 2M and cross-reaction with pregnancy zone protein was detected by the same technique. Eventually, a panel of 45 Mab was constructed consisting of essentially five types of specificities, although in fact no two Mab reacted with complete identity in all assays. In addition to the assays already mentioned, the Mab were tested for interference with binding of α 2M-trypsin to the cellular receptor, for competition with F2B2 for α 2M-trypsin and for inhibition of trypsin by α 2M. Finally, Western blotting was used as a first approximate mapping of the epitope relative to the internal thiolesters by exploiting the heat-induced fragmentation of α 2M at this site. The five categories of Mab thus detected were: (i) five Mab that react with native α 2M and not with α 2M trypsin; (ii) 18 Mab that react with both native α 2M and with α 2M-trypsin; (iii) 12 Mab, including F2B2 and F12A3, that react with the receptor-binding domain, neo-antigenically expressed on α 2M-trypsin, (iv) six Mab that are also specific for α 2M-trypsin but map outside the receptor-binding domain; (v) three Mab that define hidden determinants, not expressed on undenatured α 2M. For completeness, the panel includes the Mab obtained against pregnancy zone protein.

Three different conformational states of pregnancy zone protein identified by monoclonal antibodies.

Human pregnancy zone protein (PZP), related to human alpha 2-macroglobulin, forms dimeric/tetrameric (360/720 kDa) species. PZP binds proteinases which cause the cleavage of internal thiol esters in the molecule. Both the binding of proteinases, i.e. chymotrypsin, (CT) to PZP, forming PZP.CT complexes, or reaction with methylamine (MA) forming PZP.MA complexes, cause transition to a new similar conformational state. Reactivity of selected monoclonal antibodies against PZP towards the three PZP derivatives demonstrated differences in the reactivity pattern. PZP and PZP.MA share one determinant, which is missing on the PZP.CT complex. PZP after transition to PZP.CT, but not to PZP.MA, presents a neodeterminant detected by one of six monoclonal antibodies. The findings demonstrate that at least three different conformational states exist for PZP and its derivatives. Access to discriminating immunochemical tools makes possible an evaluation of the relative abundance of the different complexes in vivo.

Interaction between pregnancy zone protein and plasmin.

Pregnancy zone protein (PZP, alpha 2-PAG, SP3) was found to bind to plasmin in crossed affino-immunoelectrophoresis using sodium caseinate in the first dimension gel. The plasmin presence in the PZP-plasmin complex was confirmed by addition of antiserum against plasminogen to the gel. In crossed affino-immunoelectrophoresis using plasmin in the first dimension gel a non migrative PZP immunoreactive peak appeared, similar to the peak obtained with casein in the first dimension gel. Incubation of mixtures of PZP and plasmin also demonstrated complex formation between PZP and plasmin. The complex between PZP and plasmin could be precipitated not only by anti-PZP antibodies, but also by anti-plasminogen antibodies, confirming the interaction between the two molecules. The significance of the binding between plasmin and PZP remains to be elucidated, but it is tempting to speculate that PZP, present on the trophoblastic surface, immobilizes plasmin, rendering this molecule able to perform a local fibrinolytic activity.

Cell association and degradation of pregnancy zone protein-chymotrypsin complex in cultured human monocytes

125I-labelled pregnancy zone protein complexed with chymotrypsin (PZP-ct) bound to freshly isolated and to cultured human monocytes. Binding of uncomplexed 125I-pregnancy zone protein (PZP) was less than 20% of complex binding. At 4 degrees C, labelled PZP-ct complex bound to cultured monocytes with a half-time of about 4 h. Dissociation at 4 degrees C was slow at low receptor occupancies (t 1/2 greater than 24 h). At high receptor occupancies, dissociation was biphasic (initial k-1 = 8 X 10(-3) min-1) and 85% of the cell-associated label had dissociated within 24 h. At near equilibrium, binding of 20 pmol/l 125I-PZP-ct was half maximally inhibited at a ligand concentration of about 450 pmol/l. The Scatchard plot was linear giving an estimated concentration of 2 X 10(4) receptors/cell. Ligand bound at 4 degrees C was rapidly internalized at 37 degrees C (half-time = 0.5 min) and after a 5-10 min lag time, radioactivity comprising monoiodotyrosine was released into the medium following a sigmoidal curve. At 37 degrees C the uptake of 125I-PZP-ct complex was initially linear and within 15 min internalized label accumulated, reaching a steady state at about 85% of the cell-associated radioactivity. Following a lag time of 15 min, radioactivity soluble in trichloroacetic acid appeared in the medium. Similar results were obtained with 125I-alpha 2-macroglobulin-trypsin complex (alpha 2M-T). It is concluded that high affinity receptors mediate binding, uptake and degradation of pregnancy zone protein--and alpha 2-macroglobulin-proteinase complex in human monocyte-macrophages.

Pregnancy zone protein binds plasmin and counteracts the inhibitory activity of α2-antiplasmin on plasmin

Pregnancy zone protein binds plasmin and counteracts the inhibitory activity of α₂-antiplasmin on plasmin

Binding and receptor-mediated endocytosis of pregnancy zone protein-proteinase complex in rat macrophages

125I-labelled pregnancy zone protein complex was injected intravenously in rats and after 6 min uptake into cells of the liver and spleen was determined by electron microscopic autoradiography. The liver took up 68% of the injected radioactivity; 61% was in the hepatocytes and 7% was in the liver macrophages (Kupffer cells). The spleen took up 3–4% and nearly all the radioactivity was in the macrophages of the red pulp. The uptake per cell volume was several times higher in the macrophage than in the hepatocyte. The radioactivity associated with macrophages was largely in endocytotic vacuoles and lysosomes. Binding of labelled pregnancy zone protein complex to peritoneal macrophages at 4°C was 2–3-times higher than binding of the homologous α 2-macroglobulin complex. The two ligands competed for binding to the same receptors and the difference was due to a higher affinity of the pregnancy zone protein complex ( K d approx. 60 pM). After binding to the receptor, this ligand was internalised within 2–3 min at 37°C and radioactivity inside the cells largely represented intact pregnancy zone protein complex. Radioactivity was released from the cell as iodotyrosine after a lag time of about 10 min. It is concluded that pregnancy zone protein complex is bound with a high affinity to the α 2-macroglobulin receptors in rat macrophages followed by receptor-mediated endocytosis and degradation of the ligand in the lysosomes.

Application of an enzyme-linked immunoassay for the measurement of pregnancy zone protein (PZP) in cell culture supernatants and sera

A simple and sensitive enzyme-linked immunosorbent assay (ELISA) measuring specifically the pregnancy zone protein (PZP) was constructed. The assay range was 2.0-500 micrograms/l. The intra-assay coefficient of variation (CV%) was 5.9% at the level of 100 micrograms/l and 3.5% at 10 micrograms/l. The imprecision between runs was 4.5% at 100 micrograms/l and 7.6% at 10 micrograms/l. Recovery of the native PZP standard added to serum-free cell culture medium was 98.1 +/- 3.7% (mean +/- SD), and recovery from serum of women in late pregnancy was 96.0 +/- 9.3%. Recovery from PZP-chymotrypsin (PZP-CT) complexes added to serum-free medium was 141 +/- 4.3%. There was no detectable cross-reactivity between the anti-human PZP antibody and human alpha 2-macroglobulin (alpha 2-M). The dose-response of two PZP standards and the PZP serum concentrations of 100 blood donors were determined. Furthermore, the serum level of PZP from 11 patients suffering from IgA myeloma was quantitated and found within the normal range when compared to serum levels of healthy blood donors of the same age and sex. Finally, supernatants from serum-free cultures of different human peripheral blood mononuclear cell (PBM) subpopulations were assayed. Neither of them were found to exhibit any detectable increase in PZP concentration during culture, but cultures of monocytes were found to produce alpha 2-M.

Application of an enzyme-linked immunoassay for the measurement of pregnancy zone protein (PZP) in cell culture supernatants and sera

A simple and sensitive enzyme-linked immunosorbent assay (ELISA) measuring specifically the pregnancy zone protein (PZP) was constructed. The assay range was 2.0-500 micrograms/l. The intra-assay coefficient of variation (CV%) was 5.9% at the level of 100 micrograms/l and 3.5% at 10 micrograms/l. The imprecision between runs was 4.5% at 100 micrograms/l and 7.6% at 10 micrograms/l. Recovery of the native PZP standard added to serum-free cell culture medium was 98.1 +/- 3.7% (mean +/- SD), and recovery from serum of women in late pregnancy was 96.0 +/- 9.3%. Recovery from PZP-chymotrypsin (PZP-CT) complexes added to serum-free medium was 141 +/- 4.3%. There was no detectable cross-reactivity between the anti-human PZP antibody and human alpha 2-macroglobulin (alpha 2-M). The dose-response of two PZP standards and the PZP serum concentrations of 100 blood donors were determined. Furthermore, the serum level of PZP from 11 patients suffering from IgA myeloma was quantitated and found within the normal range when compared to serum levels of healthy blood donors of the same age and sex. Finally, supernatants from serum-free cultures of different human peripheral blood mononuclear cell (PBM) subpopulations were assayed. Neither of them were found to exhibit any detectable increase in PZP concentration during culture, but cultures of monocytes were found to produce alpha 2-M.

Human pregnancy zone protein and alpha 2-macroglobulin. High-affinity binding of complexes to the same receptor on fibroblasts and characterization by monoclonal antibodies.

Pregnancy zone protein (PZP) was isolated from late pregnancy serum and examined for binding to normal skin fibroblasts in culture. A high-affinity binding site on these cells is demonstrated for PZP reacted with methylamine. Experiments with alpha 2-macroglobulin (alpha 2M) and PZP, both modified by methylamine, showed this receptor to be identical to the previously characterized receptor for alpha 2M-proteinase complexes (Van Leuven, F., Cassiman, J.J., and Van den Berghe, H. (1979) J. Biol. Chem. 254, 5155-5160). With available monoclonal antibodies directed toward alpha 2M and prepared toward PZP, only a limited cross-reaction was observed. We obtained a monoclonal antibody which defines a neo-antigenic site on PZP-methylamine, completely analogous to the monoclonal antibody F2B2, which was previously shown to define a neo-antigenic site on alpha 2M complexes (Marynen, P., Van Leuven, F., Cassiman, J.J., and Van den Berghe, H. (1981) J. Immunol. 127, 1782-1786). These results provide evidence for the homologous function of alpha 2M and PZP as proteinase scavengers. The need for an extra proteinase inhibitor of the alpha 2M-type in pregnancy is discussed. The monoclonal antibodies now available will prove helpful in quantitation and eventually isolation of proteinase complexes of alpha 2M and PZP.

Evidence for binding of human pregnancy zone protein-proteinase complex to α2-macroglobulin receptors

125-I-labelled human pregnancy zone protein complexed with chymotrypsin was removed from the circulation with a half-time of 2.3 min after intravenous injection in rats. After 6 min about 67% of the label was present in the liver and about 3% was in the spleen, both in male and in female pregnant rats. The half-time of removal was more than 30 min for native pregnancy zone protein. Uptake into other organs, including placentae and feti, was negligible. 30 pM labelled pregnancy zone protein · chymotrypsin was specifically bound to rat hepatocytes and adipocytes and to human fibroblasts and monocyte-derived macrophages at 4°C. Binding was almost completely abolished by a saturating concentrating of unlabelled α 2-macroglobulin · trypsin. Binding of 15 pM labelled macroglobulin complex was completely abolished by a saturating concentration of pregnancy zone protein · chymotrypsin. In rat hepatocytes, binding of pregnancy zone protein complex was lower than that of α 2-macroglobulin complex at low ligand concentrations. Half-maximal receptor occupancy was obtained with about 300 pM pregnancy zone protein complex. Unlabelled α 2-macroglobulin or pregnancy zone protein complex failed to accelerate dissociation of the labelled pregnancy zone protein complex under conditions where dissociation of α 2-macroglobulin was markedly enhanced. It is concluded that pregnancy zone protein and α 2-macroglobulin complexes bind to the same receptors. The quantitative differences may be related to the fact that α 2-macroglobulin is a tetramer whereas the functional unit of pregnancy zone protein is probably a dimer.

The Molecular Genetics of Components of Complement

Rapid progress has been made in establishing linkages and in chromosome allocation of the genes of some 9 complement components. In the MHC, C2, Factor B, and two C4 or C4 related genes have been placed in some detail in both man and mouse. The gene coding for the cytochrome P-450 21-hydroxylase has been shown to be duplicated and immediately 3' to the two C4 genes, though it appears to be functionally and structurally unrelated to the complement components. Thus six genes have been mapped to this region where particular haplotypes are associated with increased susceptibility to a number of diseases, some of which are autoimmune in character. The complete gene structure of Factor B has been solved in man and rapid progress is being made with the C2 and C4 genes. The structural basis of the polymorphisms of these genes is being established. In C4, the polymorphism is exceptionally complex with varying numbers of loci and probably more than 50 allotypes occurring in man. A structural basis has also been found for the big differences in the biological activity of some of the C4 allotypes in man. Apart from the genes in the MHC, linkage has been found between the genes coding for C4bp, CR1, and Factor H. Remarkably there are sequence homologies between these proteins and C2 and Factor B, probably related to the ability to bind to one or other of the structurally similar proteins C3b and C4b. The complete cDNA sequences of C3 and C4 in mouse and man have given much information on the many posttranslational modifications of these proteins. A partial structure has been obtained for the C3 gene and the homology shown between C3, C4, C5, alpha 2-macroglobulin, and pregnancy zone protein. Although the amount of detailed information in the molecular genetics of complement components is accumulating rapidly, there appears to be a reasonable prospect that linkages and homologies will classify the data into a comprehensible form.

Characterization of human pregnancy zone protein. Comparison with human alpha 2-macroglobulin.

Native human pregnancy zone protein (PZP), a close homolog of alpha 2-macroglobulin (alpha 2M), can be obtained in approximately 20% yield from pooled late pregnancy plasma or serum by a combination of polyethylene glycol precipitation, euglobulin precipitation, DEAE-Sephacel chromatography, zinc-chelate affinity chromatography, and negative affinity chromatography on insolubilized antibodies against human serum proteins. Both proteins are similarly organized as disulfide-bridged dimers of 360 kDa containing 180-kDa subunits. These dimers constitute the proteinase-binding units of PZP, and in contrast to alpha 2M, they appear to be only loosely associated, indicating a subtle difference in the quaternary structure of these alpha-macroglobulins. The preparations contain functionally intact beta-cysteinyl-gamma-glutamyl thiol esters, located in the same nonapeptide sequence as found in alpha 2M, and form complexes with a variety of proteinases in which a large fraction of the proteinase is bound covalently. Proteinases bound to PZP are still active and poorly accessible to reaction with large inhibitors like alpha 1-proteinase inhibitor. The structural and functional features of PZP indicate that PZP and alpha 2M, although extremely similar, may have different yet overlapping sets of proteinases as targets. It is possible that PZP mainly controls the activity of cellular proteinases released under conditions of increased cellular turnover and that PZP could be the human equivalent to the acute phase alpha-macroglobulins known in other species.

Does ultrasound examination render biochemical tests obsolete in the prediction of early pregnancy failure?

Serum levels of fetal, placental and maternal hormones and proteins [alpha-fetoprotein (AFP), human chorionic gonadotrophin, human placental lactogen, schwangerschaftsprotein 1, pregnancy associated plasma protein-A (PAPP-A), oestradiol-17 beta, progesterone, pregnancy zone protein] were measured in 108 women with bleeding during the first half of pregnancy. Ultrasound examination at the time of each blood sampling revealed a fetal heart action on at least one occasion in 77 women. Spontaneous abortion occurred in 42 pregnancies, 31 of these showed no ultrasound sign of fetal life, whilst the fetal heart action was observed repeatedly until abortion in the remaining 11 women. Abnormally low levels of PAPP-A were most likely to indicate pregnancy failure, in particular if the fetal heart action was seen at the time of blood sampling. The predictive value, sensitivity and relative risk of a single depressed PAPP-A level were respectively 49, 89 and 41%, the predictive value of a normal result being 99%. With the exception of AFP, all other biochemical indices examined were consistently in the normal range in this group of women. If ultrasound findings were not considered, the biochemical indices were of comparable value in the prediction of spontaneous abortion. PAPP-A levels were uniformly depressed in all patients who spontaneously aborted, frequently weeks before this event, in the presence of a live fetus.

Oral progesterone and estrogen/progestogen therapy. Effects of natural and synthetic hormones on subfractions of HDL cholesterol and liver proteins.

The pharmacokinetics of progesterone after oral administration were investigated. After a single dose of 100 mg, serum concentrations of progesterone around 50-60 nM and of 20 alpha-hydroxy-4-pregnene-3-one around 5-6 nM were recorded within 1-4 hours and elevated levels persisted for 8-12 hours. There were only minor changes in 17 alpha-hydroxyprogesterone concentrations and serum levels of androstenedione and cortisol were unaffected. A significant conversion of circulating progesterone into deoxy-corticosterone was demonstrated. During clinical treatment with oral progesterone the individual serum concentrations of this potent mineralocorticoid were closely correlated to the progesterone concentrations. This observation could be important as regards the etiology of certain clinical disorders characterized by fluid retention. According to competitive binding analyses there was no evidence that progesterone or medroxyprogesterone acetate could cause significant displacement of cortisol from corticosteroid-binding globulin during clinical treatment. The effects of natural and synthetic hormones upon subfractions of high density lipoprotein (HDL) cholesterol and liver proteins were followed in postmenopausal women during replacement therapy with various estrogen-progestogen combinations. The total serum cholesterol level was significantly reduced during treatment with all estrogen regimens. The concentrations of HDL cholesterol, HDL2 cholesterol, apolipoprotein A I and A II were increased in a dose-dependent pattern during unopposed estrogen therapy. The potency of 10 micrograms of ethinyl estradiol was estimated to be equivalent to 3-4 mg of estradiol valerate. Estrogen effects were significantly reversed by levonorgestrel 250 micrograms and also by medroxyprogesterone acetate 10 mg. During treatment with natural progesterone, no changes were recorded in HDL cholesterol or its subfractions. As compared with the lipoproteins, pregnancy zone protein and sex hormone binding globulin (SHBG) were found to be more sensitive to the alkylated than to the non-alkylated estrogen. Both levonorgestrel and medroxyprogesterone acetate clearly reduced SHBG levels after 3 months, whereas micronized progesterone had no such effect. While tamoxifen counteracted the therapeutic and metabolic effects of estrogen the sequential addition of estriol had no apparent influence. Unopposed estrogen treatment enhanced liver lecithin synthesis along pathway I, i.e. reduced the amount of stearic acid and increased the amount of palmitic acid. The effects were dose-dependent and no qualitative differences between ethinyl estradiol and estradiol valerate were recorded.

Effects of an antiestrogen on subfractions of HDL cholesterol during estrogen replacement therapy.

The antiestrogenic effects of tamoxifen and estriol were compared in 39 postmenopausal women during estrogen replacement therapy. Subfractions of HDL cholesterol and its apolipoproteins and the serum levels of two estrogen sensitive liver proteins were followed during three cycles of unopposed estrogen therapy with 10 micrograms ethinyl estradiol daily. During the last 10 days of the following three cycles the women received sequential addition of either 10 mg tamoxifen twice daily or 2 mg estriol twice daily. Tamoxifen clearly reduced the estrogen-induced increase of apolipoprotein AI, HDL2 cholesterol and total HDL cholesterol. In comparison the pregnancy zone protein and sex hormone-binding globulin were more sensitive to the estrogenic as well as to the antiestrogenic effect than the lipoproteins. Tamoxifen also counteracted the therapeutic effect on climacteric symptoms and it seems unlikely that this compound may be clinically useful as an alternative to progestogens during estrogen replacement therapy. The sequential addition of estriol had no apparent effects as compared to unopposed estrogen treatment.

Percutaneous oestrogen therapy opposed by lynestrenol or natural progesterone - effects on circulating oestrogens, FSH, sex hormone binding globulin and pregnancy zone protein

Three groups of post-menopausal women were treated with percutaneous 17β-oestradiol 3 mg daily opposed by lynestrenol 5 mg or natural micronized progesterone 200 mg and 300 mg, respectively. During 6 mth of cyclic replacement oestrogen serum levels were increased corresponding to follicular phase values. The pre-treatment ratio E 1/E 2 was unchanged which is in contrast to oral therapy. Progesterone had a much weaker effect on FSH levels than lynestrenol. Women treated with progestogen had regular withdrawal bleedings, while the micronized progesterone was insufficient in this respect. No increase of the oestrogen-sensitive liver proteins, sex hormone binding globulin (SHBG) and pregnancy zone protein (PZP), was recorded in spite of a marked increase in circulating oestrogens.

An animal model for the study of the function of pregnancy-specific beta 1-glycoprotein and pregnancy zone protein: a review.

Two pregnancy-associated murine proteins, PAMP-1 and PAMP-2, have been found to be immunologically cross-reacting analogues to the human proteins pregnancy zone protein (PZP) and pregnancy-specific beta 1-glycoprotein (SP-1), respectively. In this study, the physiochemical and biological properties of the proteins are compared and the effect on pregnancy of administration of monospecific antibodies against the proteins is reported.

Abstract

Abstract. A hydroalcoholic gel containing estradiol‐17β (0.6 mg/g) for percutaneous application was used for estrogen replacement therapy in postmenopausal women. During treatment the serum concentrations of estrogens were significantly increased and stable. The conversion of estradiol into estrone was far less pronounced after percutaneous vis‐à‐vis oral therapy. In terms of FSH suppression the estrogenic potency of percutaneous estradiol‐17β 3 mg daily was similar to that of oral estradiol‐ 17β 2 mg. The serum concentrations of three estrogen‐sensitive liver proteins—sex hormone binding globulin (SHBG), pregnancy zone protein (PZP) and ceruloplasmin—were unchanged during percutaneous treatment, whereas during oral therapy all three proteins showed a marked increase. The endometrial response was quantified by morphometry. The mean relative endometrial gland volume increased from 10% to 40% and was correlated to serum levels of estradiol‐17β. Single‐cell growth did not reveal the same correlation and the sequential addition of lynestrenol, 5 mg daily for ten days, did not reduce the estrogen‐induced endometrial growth. No obviously harmful local effects were observed in skin samples. Topical application of estradiol‐17P may develop into an alternative treatment for climacteric complaints and may even have some metabolic advantages over oral treatment.