Chaperoning of estrogen receptor and induction of mammary gland proliferation by neuronal protein synuclein gamma.

Synucleins are emerging as central players in the formation of pathologically insoluble deposits characteristic of neurodegenerative diseases. However, synuclein gamma (SNCG), previously identified as a breast cancer specific gene (BCSG1), is also highly associated with breast cancer progression. Using transgenic mouse model, we demonstrated a role of SNCG in induction of highly proliferative pregnancy-like phenotype of mammary epithelial cells and branching morphology. SNCG participated in the heat shock protein-based multiprotein chaperone complex for steroid receptor signaling. Expression of SNCG in mammary epithelium resulted in a significant stimulation of ERalpha transcriptional activity. SNCG-induced mammary gland proliferation can be effectively blocked by antiestrogen and ovariectomy, indicating that the induced proliferation is mediated by ERalpha signaling and requires estrogen stimulation. These data indicate the chaperone activity of SNCG on stimulation of steroid receptor signaling in mammary gland and, thus induces extensive mammary gland proliferation and contributes to the hormonal impact on mammary tumorigenesis.

Effects of the liver tumor promoter ethinyl estradiol on epidermal growth factor-induced DNA synthesis and epidermal growth factor receptor levels in cultured rat hepatocytes.

The objective of this study was to determine whether DNA synthesis induced in the livers of female rats treated with ethinyl estradiol (EE) was due to direct effects of this synthetic estrogen on hepatocytes. Hepatocytes, obtained by collagenase perfusion from female Lewis rats, were cultured in serum-free medium containing low or no phenol red and supplemented with insulin, transferrin, and selenium. When present at 10-15 microM for the initial 30 h of culture, EE caused a subsequent 2-2.7-fold increase in hepatocyte DNA synthesis. Pretreatment of the hepatocytes with EE during the first 30 h of culture caused an EE concentration-dependent enhancement of their subsequent DNA synthetic response to epidermal growth factor (EGF). Pretreatment with EE shifted the EGF dose-response curve, causing a dramatic enhancement of the response to EGF beginning at 2 ng EGF/ml. The response to a saturating (25 ng/ml) dose of EGF was also greatly enhanced. Determination of the effect of EE on hepatocyte surface EGF receptors revealed that the increased responsiveness of DNA synthesis to EGF was accompanied by a twofold increase in EGF receptor number per cell. These results indicate that EE has direct, growth-related effects on hepatocytes which may contribute to liver growth induced in vivo by this tumor promoter.

Stimulation of breast cancer invasion and metastasis by synuclein gamma.

We recently identified and cloned novel breast cancer-specific gene BCSG1 by direct differential cDNA sequencing. BCSG1 has a great sequence homology with the Alzheimer's disease related neural protein synuclein (SNC); thus, it was also named SNC-gamma. Overexpression of SNC-gamma in breast cancer cells leads to a significant increase in motility and invasiveness in vitro and a profound augmentation of metastasis in vivo. Our data suggest that this member of the neural protein SNCs might have important functions outside the central nervous system and may play a role in breast cancer progression.

Identification and characterization of a novel matrix-degrading protease from hormone-dependent human breast cancer cells.

A novel matrix-degrading enzyme was identified from human breast cancer cells. This enzyme appears as major gelatinase in hormone-dependent breast cancer cell lines and has as an apparent molecular mass of 80 kDa on gelatin zymography. The 80-kDa enzyme has a unique metal ion specificity. In addition to calcium ions, the gelatinolytic activity can be supported by manganese and/or magnesium. Unlike 92- and 72-kDa gelatinases and other known members of the metalloproteinase family, the 80-kDa protease is not activated by p-aminophenylmercuric acetate and its gelatinolytic activity is not inhibited by tissue inhibitor of metalloproteinase 2. It is active over the pH range 7.5-9.5 with an optimum at pH 8.5. The enzyme degrades gelatin and type IV collagen. The proteolytic activity of the enzyme is inhibited by EDTA and leupeptin. These unique features clearly distinguish the 80-kDa protease from the known 92-and 72-kDa gelatinases. The expression of 80-kDa enzyme can be detected in hormone-dependent human breast cancer cell lines in vitro and in tumors grown from these cells in athymic nude mice.

Induction of mammary differentiation by mammary-derived growth inhibitor-related gene that interacts with an omega-3 fatty acid on growth inhibition of breast cancer cells.

We previously identified and characterized a novel tumor growth inhibitor and a fatty acid-binding protein in human mammary gland and named it the mammary-derived growth inhibitor-related gene (MRG). Here, the effects of MRG on mammary gland differentiation and its interaction with omega-3 polyunsaturated fatty acids (omega-3 PUFAs) on growth inhibition were investigated. MRG protein expression was associated with human mammary gland differentiation, with the highest expression observed in the differentiated alveolar mammary epithelial cells from the lactating gland. Overexpression of MRG in human breast cancer cells induced differentiation with changes in cellular morphology and a significant increase in the production of lipid droplets. Treatment of mouse mammary gland in organ culture with MRG protein resulted in a differentiated morphology and stimulation of beta-casein expression. Treatment of human breast cancer cells with the omega-3 PUFA docosahexaenoic acid resulted in a differential growth inhibition proportional to their MRG expression. MRG-transfected cells or MRG protein treated cells were much more sensitive to docosahexaenoic acid-induced growth inhibition than MRG-negative or untreated control cells. Our results suggest that MRG is a candidate mediator of the differentiating effect of pregnancy on breast epithelial cells and may play a major role in omega-3 PUFA-mediated tumor suppression.

Stimulation of mammary tumorigenesis by systemic tissue inhibitor of matrix metalloproteinase 4 gene delivery.

Tissue inhibitors of matrix metalloproteinase (TIMPs) are multifunctional proteins with both matrix metalloproteinase (MMP) inhibitory effects and growth-regulatory activity. TIMPs inhibit MMP activity, suggesting a use for cancer gene therapy. However, here we report that systemic administration of human TIMP-4 by electroporation-mediated i.m. injection of naked TIMP-4 DNA stimulates tumorigenesis of human breast cancer cells in nude mice. Consistent with tumor stimulation, TIMP-4 up-regulates Bcl-2 and Bcl-X(L) protein. TIMP-4 also inhibits apoptosis in human breast cancer cells in vitro and mammary tumors in vivo. A synthetic MMP inhibitor BB-94 did not have such antiapoptotic effect. Analysis of TIMP-4 expression in human mammary specimens indicates that TIMP-4 protein is increased in mammary carcinoma cells compared with normal mammary epithelial cells. These data indicate an antiapoptotic activity in breast cancer cells and a tumor-stimulating effect of TIMP-4 when administrated systemically.

Identification of a breast cancer-specific gene, BCSG1, by direct differential cDNA sequencing.

A high-throughput direct-differential cDNA sequencing approach was employed to identify genes differentially expressed in normal breast as compared with breast cancer. Approximately 6000 expressed sequence tags (ESTs) from cDNA libraries of normal breast and breast carcinoma were selected randomly and subjected to EST-sequencing analysis. The relative expression levels of more than 2000 unique EST groups were quantitatively compared in normal versus cancerous breast. Of many putative differentially expressed genes, a breast cancer-specific gene, BCSGC1, which was expressed in high abundance in a breast cancer cDNA library but scarcely in a normal breast cDNA library, was identified as a putative breast cancer marker. In situ hybridization analysis demonstrated stage-specific BCSG1 expression as follows: BCSG1 was undetectable in normal or benign breast lesions, showed partial expression in ductal carcinoma in situ, but was expressed at an extremely high level in advanced infiltrating breast cancer. The predicted amino acid sequence of BCSG1 gene has a significant sequence homology to the non-amyloid beta protein fragment of the Alzheimer's disease amyloid protein. BCSG1 overexpression may indicate breast cancer malignant progression from benign breast or in situ carcinoma to the highly infiltrating carcinoma.

RIZ1, but not the alternative RIZ2 product of the same gene, is underexpressed in breast cancer, and forced RIZ1 expression causes G2-M cell cycle arrest and/or apoptosis.

The retinoblastoma protein-interacting zinc finger gene RIZ maps to the distal short arm of human chromosome 1 (1p36), a region thought to harbor tumor suppressor genes for a variety of human cancers including breast cancer. The RIZ gene normally produces two protein products of different length, RIZ1 and RIZ2. RIZ2 is generated by an internal promoter and lacks an NH2-terminal motif of RIZ1, the PR domain conserved in a subfamily of zinc finger genes that function as negative regulators of tumorigenesis. We have here studied whether the RIZ gene may play a role in human neoplasia. We found that expression of RIZ1 is commonly decreased or at undetectable levels in breast cancer tissues and cell lines. Decreased RIZ1 expression was also found in other tumor types including neuroblastoma and lung cancer. Remarkably, RIZ2 is normally expressed in all cases examined, suggesting that the abnormality observed for RIZ1 is specific. Forced RIZ1 expression in breast cancer cells caused cell cycle arrest in G2-M and/or programmed cell death. These observations suggest an exclusive negative selection for RIZ1 but not RIZ2 in breast cancer and a role for RIZ1 in tumor suppression.

Antitumor activity of the novel human breast cancer growth inhibitor, mammary-derived growth inhibitor-related gene, MRG.

A novel human tumor growth inhibitor was identified by differential cDNA sequencing. The predicted amino acid sequence of this tumor-suppressing factor has a significant sequence homology to mouse mammary-derived growth inhibitor and thus was named mammary-derived growth inhibitor-related gene (MRG). MRG was found to be expressed in normal and benign human breast tissues but not in breast carcinomas. In situ hybridization analysis demonstrated a stage-specific MRG expression as follows. MRG was barely detectable in breast carcinomas, showed partial and weak expression in benign hyperplasia, but was expressed at a high level in normal breast epithelial cells. To determine if MRG can modulate in vivo growth of human breast cancers, we transfected a full-length MRG cDNA into MDA-MB-231 human breast cancer cells and studied the orthotopic growth of MRG transfectants versus control transfectants in the mammary fat pad of athymic nude mice. Overexpression of MRG in human breast cancer cells significantly suppressed cell proliferation in vitro and tumor growth in an orthotopic nude mouse model. These results suggest that MRG has tumor-suppressing activity, and the loss of MRG expression may be involved in the development and progression of breast cancer.

Inhibition of tumor growth and metastasis of human breast cancer cells transfected with tissue inhibitor of metalloproteinase 4.

We recently identified, cloned, and characterized a novel human tissue inhibitor of metalloproteinases-4, TIMP-4 (Greene et al., 1996). To determine if TIMP-4 can modulate the in vivo growth of human breast cancers, we transfected a full-length TIMP-4 cDNA into MDA-MB-435 human breast cancer cells and studied the orthotopic growth of TIMP-4-transfected (TIMP4-435) versus control (neo-435) clones in the mammary fat pad of athymic nude mice. TIMP4-435 clones expressed TIMP-4 mRNA and produced anti-metalloproteinase (MMP) activity, while neo-435 clones did not express TIMP-4 mRNA or produce detectable anti-MMP activity. Overexpression of TIMP-4 inhibited the invasion potential of the cells in the in vitro invasion assay. When injected orthotopically into nude mice, TIMP-4 transfectants were significantly inhibited in tumor growth by 4-10-fold in primary tumor volumes; and in an axillary lymph node and lung metastasis as compared with controls. These results suggest the therapeutic potential of TIMP-4 in treating cancer malignant progression.

Inhibition of Wilms' tumor growth by intramuscular administration of tissue inhibitor of metalloproteinases-4 plasmid DNA.

Extracellular matrix (ECM) degrading matrix metalloproteinases (MMPs) lead to ECM turnover, a key event in cancer growth and progression. The tissue inhibitors of matrix metalloproteinases (TIMPs) limit the activity of MMPs, which suggests their use for cancer gene therapy. Here we report that systemic administration of naked TIMP-4 DNA significantly inhibited Wilms' tumor growth in nude mice. TIMP-4, whose expression was lost in Wilms' tumor, inhibited the growth of G401 Wilms' tumor cells at a concentration lower than those required for MMP inhibition. This inhibition was associated with internalization of exogenous recombinant TIMP-4. Electroporation-mediated intramuscular injection of TIMP-4 expression plasmid resulted in sustained plasma TIMP-4 levels and significant tumor suppression. Our data demonstrate a tumor suppressive effect of TIMP-4 against Wilms' tumor and the potential utility of intramuscular delivery of TIMP gene for treatment of kidney derived cancers.

Synuclein γ protects Akt and mTOR and renders tumor resistance to Hsp90 disruption.

Heat shock protein (Hsp)90 regulates many key pathways in oncogenesis, including Akt and mammalian target of rapamycin (mTOR). The strengths of disruption of Hsp90 in cancer therapy include their versatility in inhibiting a wide range of oncogenic pathways. The present study demonstrated that synuclein γ (SNCG) protects the functions of Akt and mTOR in the condition when the function of Hsp90 is blocked. Disruption of Hsp90 abolished Akt activity and mTOR signaling. However, expression of SNCG restored Akt activity and mTOR signaling. SNCG bound to Akt and mTOR in the presence and absence of Hsp90. Specifically, the C-terminal (Gln106-Asp127) of SNCG bound to the loop connecting αC helix and ß4 sheet of the kinase domain of Akt. SNCG renders resistance to 17-AAG-induced apoptosis both in vitro and in tumor xenograft. A clinical follow-up study indicates that patients with an SNCG-positive breast cancer have a significantly shorter disease-free survival and overall survival than patients with SNCG-negative tumors. The present study indicates that SNCG protects Hsp90 client proteins of Akt and mTOR, and renders drug resistance to Hsp90 disruption.

Expression of 67 kDa laminin receptor in human breast cancer cells: regulation by progestins.

The level of 67 kDa laminin receptor (67LR) expression on breast and colon tumor cell surfaces was previously shown to be correlated with the capacity of tumor cells to metastasize. In the present work we investigate the effects of progestins and estrogen on the expression of 67LR in two sublines of the T47D human breast cancer cells: weakly tumorigenic, poorly invasive parental T47D cells and a highly tumorigenic, more invasive T47Dco subclone. Immunoblotting with an affinity purified antibody directed against a synthetic peptide recognizes the 67LR in these cells. 67LR expression in the T47Dco subclone is 5.5-fold higher than in their parental T47D cells. Treatment of T47D cells with 1 nM of the synthetic progestin R5020 results in a 4-fold increase in 67LR protein expression. Estrogen also induced 67LR expression, but only by 1.5-fold. The progestin-stimulated expression of the 67LR correlates with a 4.3-fold increase in attachment of T47D cells to laminin. A monoclonal antibody, mAb 13, directed against beta 1 integrin, completely blocks the attachment of T47D cells to fibronectin, only partially inhibits the attachment of T47D cells to laminin, and appears not to affect the progestin-stimulated laminin attachment of T47D cells. A new antiprogestin, ZK 112.993, significantly inhibits both progestin-stimulated 67LR expression and the increased attachment to laminin. These results suggest a possible role for progestin in mediating one of the multiple events thought to be important in metastasis of steroid receptor positive human breast cancer cells.

Pharmacokinetics of norethisterone in humans.

The pharmacokinetics of a dose of 1mg norethisterone administered with 50 micrograms ethynyloestradiol was studied in 83 subjects. The dose was rapidly absorbed and there were wide variations in the serum NET concentrations at any particular time after dosing; the concentrations at 24 h varied from 100 to 1700 pg/ml. There was a significant negative correlation between the serum NET concentration and the time after dosing in all women. There were large inter-subject variations in the pharmacokinetic parameters, the elimination half-life and bioavailability showing 3- and 5- fold variability, respectively. Mean values for the parameters were t1/2, 7.6 h; bioavailability, 53.6 ng/ml/h; C max, 4.63 ng/ml; clearance, 22.6 l/h; and Vd 2361. There were a number of statistically significant correlations between the pharmacokinetic parameters and analysis of the correlations suggested that clearance was an important determinant of the bioavailability and of C max whereas the elimination half-life was the determinant of the NET concentration at 24 h. The pharmacokinetics of NET are compared with those of ethynyloestradiol. The wide variation in pharmacokinetics is likely to be important in determining inter-subject variations in efficacy and, particularly, side-effects of oral contraceptives especially now that low-dose formulations are widely used.

PIP: The pharmacokinetics of a dose of 1 mg norethisterone administered with 50 mcg ethinyl estradiol was studied in 83 subjects. There were marked variations in age, height, and weight between subjects in the 14 centers, but few of the differences were statistically significant. Table 1 shows the mean values for the serum norethisterone (NET) concentration in samples from each center. Variations between the values observed at any particular time both between the subjects recruited in any 1 center and between subjects in different centers were considerable. Considering only mean values, there was a 2-3-fold variation between the centers at any 1 time. Absorption was quick in most of the 83 subjects. The peak serum concentration occurred within 1 hour in 10 women, between 1-2 hours in 47 women, and between 2-4 hours in 25 women. Elimination was slow, and in only 4 of the 83 subjects had the serum NET concentration declined to 100 pg/ml by 24 hours. Thus, in the majority of women, significant concentrations of NET still were present 24 hours after administration of the dose. For the elimination phase (2-24 hours), there was a significant negative correlation between the logarithm of the serum NET concentration and time after administration of the dose, the correlation coefficient being greater than 0.9 in all but 2 of the 83 women. Due to the large inter-subject variation and the small number of subjects studied in each center, few of the correlations between the pharmacokinetic parameters or of the parameters with the ponderal index in each center were significant although in 10 of the 14 centers there was a significant correlation of the serum NET concentration at 24 hours with the elimination half-life. There was a 3-fold variation between the lowest and highest values for the half-life and a 5-fold variation for area under the curve. In no subject were undetectable NET levels reached in less then 20 hours, and only in 18.5% had the levels decrease below 400 pg/ml. In 63% of the women, a time of 30-50 hours was necessary for serum NET levels to become undetectable, but in 22.2% of the subjects detectable levels were still present at more than 50 hours after administration of the dose. The wide variation in pharmacokinetics is likely to be important in determining inter-subject variations in efficacy and, particularly, side-effects of oral contraceptives.

The effect of levonorgestrel administered in large doses at different stages of the cycle on ovarian function and endometrial morphology.

In a pharmacokinetic study, levonorgestrel (L-NOG) 0.75 mg was administered orally to 10 swedish women in the early follicular phase of the menstrual cycle. L-NOG levels were measured after L-NOG administration. A peak level of 16 nmol/l was reached after 2 hours, T 1/2 was estimated to be 14.5 hours (8.5-18.5) in the 24-48-hour interval after dosing. Seventy-two women (in Stockholm, Bombay and Shanghai) were assigned to 4 treatment groups and studied during a control cycle, a treatment cycle and a posttreatment cycle when 0.75 mg L-NOG was administered orally for 4 days in the follicular phase, periovulatory period or luteal phase. Peripheral blood was drawn 3 times weekly during the entire study for the assay of estradiol and progesterone. In 22 women in Stockholm, an endometrial biopsy was obtained on cycle day 20-22 in all 3 cycles studied. When L-NOG was administered on periovulatory days 9, 11, 13, and 15, 3 women showed follicular activity only, 7 exhibited follicular activity followed by insufficient luteal function and 7 women ovulated normally. When L-NOG was administered on periovulatory days 11, 12, 16 and 19, 7 women ovulated during treatment, 6 women exhibited follicular activity followed by insufficient luteal function and 5 exhibited follicular activity only. When L-NOG was administered in the follicular or luteal phase, no effect on ovarian function was seen. No significant prolongation of the cycle lengths was seen when L-NOG was taken during the follicular phase. Only minor effects in the endometrium were observed during treatment.

Pharmacokinetic study of levonorgestrel used as a postcoital agent.

The pharmacokinetics and pharmacodynamics of levonorgestrel (LNG) were studied in six women given 0.75 mg LNG orally for seven days during the periovulatory phase of the menstrual cycle. Steady-state concentrations of LNG were reached within three days and serum LNG concentrations at various times on day 7 were generally lower than on day 1, presumably due to a reduced serum level of SHBG. On day 7 the volume of distribution was significantly increased and Co significantly decreased and both the clearance and elimination half-life were higher on day 7 than on day 1. Half-lives varied from 5.6 to 25.1 hours. The day-to-day intra-subject variations in serum LNG concentrations ranged from 23% to 80%. Serum concentrations of pituitary and ovarian hormones suggested that ovulation was not inhibited in four of the six subjects and was delayed in the remaining two. No significant changes in serum prolactin levels were observed.

Pharmacokinetic study of orally administered RU 486 in non-pregnant women.

A method based on HPLC was devised for the estimation of RU 486 in blood and utilised to study the pharmacokinetics of a single dose of 50 mg RU 486 administered orally to 12 women on day 7 of the cycle. The dose was rapidly absorbed with peak plasma concentration between 1 and 2 hours. Distribution was also rapid (mean t1/2 alpha: 1.4h), whereas elimination was slow (mean t1/2 beta: 28.3 h). RU 486 was still detectable in some women at 72 h after administration. The plasma concentrations fitted the equation for a two-compartment open model from which the pharmacokinetic parameters were calculated. The mean total plasma clearance was 3.0 l/h, and the comparison of our data with those published studies suggests that the pharmacokinetics of RU 486 in Chinese women are similar to those of other populations.

PIP: A method based on high performance liquid chromatography (HPLC) was developed for the estimation of RU-486 in blood and used to analyze the pharmacokinetics of a single dose of 50 mg of RU-486 administered orally to 12 Chinese women on day 7 of the menstrual cycle. In the 1st 6 subjects, blood samples were taken immediately before and 1, 2, 4, 8, 24, and 48 hours after RU-486 administration; in the 2nd group of 6 women, additional samples were taken at 0.33, 0.5, 0.75, and 72 hours. Since there were no significant differences between the 2 groups in values, the data were combined. RU-486 was found to be rapidly absorbed, with significant amounts present in the 20 minute samples and peak concentrations at 1-2 hours. In contrast, elimination was slow and significant levels of RU-486 were still detectable in most subjects at 72 hours. The mean value for the half-life of absorption was less than 1 hour, while the mean half-life of distribution was 1.4 hours. Mean total plasma clearance was 3.0 liters/hour. The plasma concentrations fitted the equation for a 2-compartment open model from which the pharmacokinetic parameters were calculated. Comparison of these findings with those from other studies suggest that the pharmacokinetics of RU-486 in Chinese women are similar to those in non-Asian populations.

Pharmacokinetic study of RU 486 and its metabolites after oral administration of single doses to pregnant and non-pregnant women.

RU 486 and three of its metabolites (RU 42633-monodemethyl, RU 42848-didemethyl, and RU 42698-hydroxymetabolite) were determined by HPLC in plasma from nine non-pregnant and 36 pregnant women. Each non-pregnant subject took an oral dose of RU 486 (25, 100, 400 and 600 mg consecutively) once per menstrual cycle. Six of the nine women also received a dose of 200 mg. The 36 pregnant women were randomized into four groups which were given a single dose of 25, 100, 400 or 600 mg RU 486. Blood samples were taken up to 120 h after dosing. Peak concentrations of RU 486 occurred on most occasions within 2 h. Plasma concentrations at 1 h and at 24 h increased in proportion to log dose. There was a wide variability (up to ten-fold) in the pharmacokinetic parameters within each dose group. Plasma concentrations of RU 42633 were similar to those of RU 486 but concentrations of RU 42848 and RU 42698 were much lower. As with RU 486, the plasma concentrations of the metabolites were maintained at high levels for up to 48-72 h after dosing. The findings were consistent with a rapid metabolism of RU 486 to RU 42633; removal of the second methyl group leading to RU 42698 occurred much more slowly and to a much less extent than removal of the first. There appeared to be no significant differences between the non-pregnant and pregnant women in either the plasma concentrations or pharmacokinetic parameters of RU 486 and its metabolites.

Comparative cross-over pharmacokinetic study on two types of postcoital contraceptive tablets containing levonorgestrel.

A pharmaceutical and pharmacokinetic study was carried out on levonorgestrel tablets from two different sources (Hungarian- and Chinese-made). Both preparations contained 0.75 mg levonorgestrel and had been shown to have similar contraceptive efficacy and side effects when used for postcoital contraception. Absorption and bioavailability of the Hungarian-made tablets were greater as evidenced by higher serum concentrations of levonorgestrel, a greater area under the concentration-time curve during the first 24 hours, and a more marked suppressive effect on SHBG levels. These differences most probably reflect differences in their pharmaceutical formulation, in particular the extent of tablet dissolution and the degree of micronisation of levonorgestrel.

Postcoital contraception with levonorgestrel during the peri-ovulatory phase of the menstrual cycle. Task Force on Post-ovulatory Methods for Fertility Regulation.

The contraceptive efficacy and side effects of postcoital levonorgestrel used repeatedly during the peri-ovulatory period of one cycle was examined in 259 women. All subjects were of proven fertility in their present union and had ovulatory cycles as assessed from pre-treatment BBT charts. The mean number of coital acts during the treatment cycle was 7.5 (SD:2.6) and the mean number of 0.75 mg levonorgestrel tablets taken during the peri-ovulatory period was 4.0 (SD:1.2). Two pregnancies, both considered to be method failures, occurred, giving a failure rate of 0.8% per treated cycle. Although the overall effect of levonorgestrel on menstrual cycle length was small and insignificant, menstrual cycle disturbances were not uncommon. Intermenstrual bleeding or spotting occurred in 8.5% of the treated cycles and 12.5% of the cycles were less than 20 or more than 35 days. Other side effects, mainly nausea, headache and dizziness, were reported by about 20% of the subjects but the apparent incidence of these complaints varied markedly between the nine participating centres from 0% to just over 50%. The data suggest that repeated postcoital use of levonorgestrel is probably not a viable approach to fertility regulation for the majority of women who have regular intercourse and wish to limit the number of their pregnancies.