Specific binding of 4-hydroxyestradiol to mouse uterine protein: evidence of a physiological role for 4-hydroxyestradiol.

There are several indications of a possible physiological role for 4-hydroxyestradiol (4-OHE(2)) in hormone-responsive tissues. To examine a hormonal activity of 4-OHE(2), we have studied the binding of (3)H-labeled 4-OHE(2) to mouse uterine cytosolic protein. In uteri of 3-week-old mice, total binding was 319.4 +/- 13.9 fmol/mg protein. Binding in the presence of excess unlabeled 4-OHE(2) dropped to 82.1 +/- 1.7 fmol/mg protein, whereas 214.6 +/- 9.4 fmol/mg protein bound while incubating in an excess of unlabeled 17beta-estradiol (E(2)). The difference between the two binding values in the presence of excess steroid (132.5 +/- 11.1 fmol/mg protein) is taken as selective binding of 4-OHE(2) to a specific protein. In mice older than 4 weeks, the specific 4-OHE(2) binding declined: 32.0 +/- 4.0 fmol/mg protein at 8 weeks, 54.8 +/- 6.3 fmol/mg protein at 12 weeks and 54.6 +/- 5.2 fmol/mg protein at 9 months. Of other organs tested (liver, kidney, lung and whole brain) only lung showed significant selective binding of 4-OHE(2). When E(2)-binding sites are blocked, binding follows first-order kinetics, yielding a dissociation constant (K(d)) value of 11.8 +/- 2.1 nM. The specific binding of 4-OHE(2) was not inhibited by any other steroids or estrogen metabolites that were tested, except for 2-hydroxyestradiol (2-OHE(2)), which displayed competitive inhibition of 4-OHE(2) binding with an inhibition constant (K(i)) value of 98.2 +/- 12.6 nM. These results lead us to conclude that 4-OHE(2) binds to a specific binding protein, distinct and different from binding to estrogen receptors (ERalpha and ERbeta). The physiological role of this binding remains to be elucidated.

Genotoxicity of the steroidal oestrogens oestrone and oestradiol: possible mechanism of uterine and mammary cancer development.

Oestrogens, including the natural hormones oestrone and oestradiol, induce various tumours in laboratory animals and have been recognized to be carcinogens in humans, raising the risk for breast and uterine cancer. As part of the search for the mechanism of hormone-induced carcinogenesis, various types of DNA damage have been detected which have been induced by oestrogens in cell-free systems, in cells in culture, or in vivo. Nevertheless, oestrogens have been postulated to act only as promoters of mammary carcinogenesis by receptor-mediated growth stimulation without consideration of their genotoxicity because these hormones failed to induce mutations in commonly used assays. More recently, oestradiol-induced numerical chromosomal changes (aneuploidy) and structural chromosomal aberrations have been detected in cells in culture and in hamster kidney, a target of oestrogen-induced cancer. In this animal model, oestradiol generates c-myc gene amplification and microsatellite instability. Mutations of the hprt gene have been induced by oestradiol in V79 cells and by catecholoestrogen metabolites in Syrian hamster embryo cells. Sequencing of this gene isolated from V79 mutant clones revealed point mutations and deletions. It is concluded that oestradiol plays a dual role as mutagen/carcinogen and as growth-stimulating hormone in the induction of tumours.

Role of estrogen in alcohol promotion of breast cancer and prolactinomas.

This article represents the proceedings of a symposium at the 2000 ISBRA Meeting in Yokohama, Japan. The chair was Dipak K. Sarkar. The presentations were (1) Dual role of estrogen as hormone and carcinogen in mammary carcinogenesis, by Joachim G. Liehr; (2) Alcohol and breast cancer: Studies using animals, by Keith W. Singletary; and (3) Evaluation of the role of estrogen in mediation of ethanol effect on prolactinoma: Studies using animals, by Dipak K. Sarkar.

Role of iron in estrogen-induced cancer.

Redox cycling of catecholestrogen metabolites between quinone and catechol forms is a mechanism of generating potentially mutagenic oxygen radicals in estrogen-induced carcinogenesis. Consistent with this concept, multiple forms of oxygen radical-generated DNA damage are induced by estrogen in cell-free systems, in cells in culture and in rodents prone to estrogen-induced cancer. Metal ions, specifically iron, are necessary for the production of hydroxy radicals. Iron has not received much attention in discussions of estrogen-induced carcinogenesis and human hormone-associated cancer, and is the focus of this review. An elevated dietary iron intake enhances the incidence of carcinogen-induced mammary cancer in rats and estrogen-induced kidney tumors in Syrian hamsters. Estrogen administration increases iron accumulation in hamsters and facilitates iron uptake by cells in culture. In humans, elevated body iron storage has been shown to increase the risk of several cancers including breast cancer. A role of iron in hormone-associated cancer in humans offers attractive routes for cancer prevention by regulating metal ion metabolism and interfering with iron accumulation in tissues.

Frequency and molecular analysis of hprt mutations induced by estradiol in Chinese hamster V79 cells.

The natural hormone estradiol (E2) induces tumors in rodents and various types of DNA damage in vitro and in vivo, but has not been mutagenic in bacterial or mammalian assays. Recent reports of chromosomal and genetic lesions induced by E2 has led us to re-examine the mutation frequency and molecular alterations of the hypoxanthine-guanine phosphoribosyltransferase (hprt) gene in Chinese hamster V79 cells. E2 at both physiological and pharmacological concentrations (10-11, 10-10, and 10-7, 10-6 M) significantly increased the mutation frequency of the hprt gene by 2. 57-, 3.45-, 2.63-, and 8.78-fold, respectively, compared to the controls, while 10-13, 10-12, 10-9, or 10-8 M E2 induced little change (< or =0.93-fold). PCR and a molecular analysis of the hprt coding sequence identified genetic lesions in the cDNA and/or genomic DNA in 15 of the 21 picked E2-induced mutants (71%). Simple base substitutions, such as Tright curved arrow G or Tright curved arrow A transversions, were the most common mutations (8/21 or 38%) and frequently occurred at 122 bp or 407 bp of the hprt coding sequence. Deletion mutations were detected in 6 of the 21 clones (29%). An Aright curved arrow G and a Cright curved arrow T transition and a four-base insertion (TATT) were identified each in one mutant clone. A RT-PCR analysis demonstrated an abundant expression of the estrogen receptor-alpha (ERalpha). However, ICI 182,780, an antagonist of ERalpha, acted in an additive manner with E2 and increased the hprt mutation frequency. In conclusion, E2 induces a low frequency of mutations (deletions and point mutations) in V79 cells, which is consistent with the weak carcinogenic activity of this hormone. The mutagenic effects of E2 in V79 cells are not mediated by the ERalpha.

Estrogens as endogenous genotoxic agents--DNA adducts and mutations.

Estrogens induce tumors in laboratory animals and have been associated with breast and uterine cancers in humans. In relation to the role of estrogens in the induction of cancer, we examine formation of DNA adducts by reactive electrophilic estrogen metabolites, formation of reactive oxygen species by estrogens and the resulting indirect DNA damage by these oxidants, and, finally, genomic and gene mutations induced by estrogens. Quinone intermediates derived by oxidation of the catechol estrogens 4-hydroxyestradiol or 4-hydroxyestrone may react with purine bases of DNA to form depurinating adducts that generate highly mutagenic apurinic sites. In contrast, quinones of 2-hydroxylated estrogens produce less harmful, stable DNA adducts. The catechol estrogen metabolites may also generate potentially mutagenic oxygen radicals by metabolic redox cycling or other mechanisms. Several types of indirect DNA damage are caused by estrogen-induced oxidants, such as oxidized DNA bases, DNA strand breakage, and adduct formation by reactive aldehydes derived from lipid hydroperoxides. Estradiol and the synthetic estrogen diethylstilbestrol also induce numerical and structural chromosomal aberrations and several types of gene mutations in cells in culture and in vivo. In conclusion, estrogens, including the natural hormones estradiol and estrone, must be considered genotoxic carcinogens on the basis of the evidence outlined in this chapter.

Tissue-specific synthesis and oxidative metabolism of estrogens.

Estrogen exposure represents the major known risk factor for development of breast cancer in women and is implicated in the development of prostate cancer in men. Human breast tissue has been shown to be a site of oxidative metabolism of estrogen due to the presence of specific cytochrome P450 enzymes. The oxidative metabolism of 17beta-estradiol (E2) to E2-3,4-quinone metabolites by an E2-4-hydroxylase in breast tissue provides a rational hypothesis to explain the mammary carcinogenic effects of estrogen in women because this metabolite is directly genotoxic and can undergo redox cycling to form genotoxic reactive oxygen species. In this chapter, evidence in support of this hypothesis and of the role of P4501B1 as the 4-hydroxylase expressed in human breast tissue is reviewed. However, the plausibility of this hypothesis has been questioned on the grounds that insufficient E2 is present in breast tissue to be converted to biologically significant amounts of metabolite. This critique is based on the assumption that plasma and tissue E2 levels are concordant. However, breast cancer tissue E2 levels are 10-fold to 50-fold higher in postmenopausal women than predicted from plasma levels. Consequently, factors must be present to alter breast tissue E2 levels independently of plasma concentrations. One such factor may be the local production of E2 in breast tissue through the enzyme aromatase, and the evidence supporting the expression of aromatase in breast tissue is also reviewed in this chapter. If correct, mutations or environmental factors enhancing aromatase activity might result in high tissue concentrations of E2 that would likely be sufficient to serve as substrates for CYP1B1, given its high affinity for E2. This concept, if verified experimentally, would provide plausibility to the hypothesis that sufficient E2 may be present in tissue for formation of catechol metabolites that are estrogenic and which, upon further oxidative metabolism, form genotoxic species at levels that may contribute to estrogen carcinogenesis.

Nuclear localization of catechol-O-methyltransferase in neoplastic and nonneoplastic mammary epithelial cells.

Catechol-O-methyltransferase (COMT) plays both a regulatory and protective role in catechol homeostasis. It contributes to the regulation of tissue levels of catecholamines and catecholestrogens (CEs) and, by blocking oxidative metabolism of catechols, prevents endogenous and exogenous catechols from becoming a source of potentially mutagenic electrophiles. Evidence implicating CEs in carcinogenesis, in particular in the hamster kidney model of estrogen-induced cancer, has focused attention on the protective role of COMT in estrogen target tissues. We have previously reported that treating hamsters with estrogens causes translocation of COMT to nuclei of epithelial cells in the renal cortex, the site of CE biosynthesis and where the cancers arise. This finding suggested that nuclear COMT may be a marker of a threat to the genome by catechols, including CEs. It is postulated that CEs play a role in the genesis of breast cancer by contributing to a state of chronic oxidative stress that is presumed to underlie the high incidence of this disease in the United States. Therefore, here we used immunocytochemistry to re-examine human breast parenchyma for nuclear COMT. In addition to confirming previous reports of cytoplasmic COMT in mammary epithelial cells, we identified nuclear COMT in foci of mammary epithelial cells in histologically normal breast tissue of virtually all control (macromastia) and cancer patients and in breast cancer cells. There was no correlation between tissue histology and the numbers of cells with nuclear COMT, the size of foci containing such cells, or intensity of nuclear COMT immunostaining. The focal nature of the phenomenon suggests that nuclear COMT does not serve a housekeeping function but that it reflects a protective response to an increased local catechol load, presumably of CEs and, as such, that it may be a characteristic of the population of women studied who share the same major risk factor for developing breast cancer, that of living in the industrialized West.

Is estradiol a genotoxic mutagenic carcinogen?

The natural hormone 17 beta-estradiol (E2) induces tumors in various organs of rats, mice, and hamsters. In humans, slightly elevated circulating estrogen levels caused either by increased endogenous hormone production or by therapeutic doses of estrogen medications increase breast or uterine cancer risk. Several epigenetic mechanisms of tumor induction by this hormone have been proposed based on its lack of mutagenic activity in bacterial and mammalian cell test systems. More recent evidence supports a dual role of estrogen in carcinogenesis as a hormone stimulating cell proliferation and as a procarcinogen inducing genetic damage. Tumors may be initiated by metabolic conversion of E2 to 4-hydroxyestradiol catalyzed by a specific 4-hydroxylase (CYP1B1) and by further activation of this catechol to reactive semiquinone/quinone intermediates. Several types of direct and indirect free radical-mediated DNA damage are induced by E2, 4-hydroxyestradiol, or its corresponding quinone in cell-free systems, in cells in culture, and/or in vivo. E2 also induces various chromosomal and genetic lesions including aneuploidy, chromosomal aberrations, gene amplification, and microsatellite instability in cells in culture and/or in vivo and gene mutations in several cell test systems. These data suggest that E2 is a weak carcinogen and weak mutagen capable of inducing genetic lesions with low frequency. Tumors may develop by hormone receptor-mediated proliferation of such damaged cells.

Induction of uterine adenocarcinoma in CD-1 mice by catechol estrogens.

Catechol estrogens may mediate estrogen-induced carcinogenesis because 4-hydroxyestradiol induces DNA damage and renal tumors in hamsters, and this metabolite is formed in the kidney and estrogen target tissues by a specific estrogen 4-hydroxylase. We examined the carcinogenic potential of catechol estrogen in an experimental model previously reported to result in a high incidence of uterine adenocarcinoma after neonatal exposure to diethylstilbestrol. Outbred female CD-1 mice were treated with 2- or 4-hydroxyestradiol, 17beta-estradiol, or 17alpha-ethinyl estradiol on days 1-5 of neonatal life (2 microg/pup/day) and sacrificed at 12 or 18 months of age. Mice treated with 17beta-estradiol or 17a-ethinyl estradiol had a total uterine tumor incidence of 7% or 43%, respectively. 2-Hydroxyestradiol induced tumors in 12% of the mice, but 4-hydroxyestradiol was the most carcinogenic estrogen, with a 66% incidence of uterine adenocarcinoma. Both 2- and 4-hydroxylated catechols were estrogenic and increased uterine wet weights in these neonates. These data demonstrate that both 2- and 4-hydroxyestradiol are carcinogenic metabolites. The high tumor incidence induced by 4-hydroxyestradiol supports the postulated role of this metabolite in hormone-associated cancers.

Role of DNA adducts in hormonal carcinogenesis.

In this review, a mechanism of estrogen-induced cancer has been examined which features a dual role of estrogen as hormone and carcinogen. Evidence exists that estrogens are metabolically activated to 4-hydroxylated metabolites by a specific cytochrome P450 in tissues prone to estrogen-induced cancer. These metabolites and their semiquinone/quinone oxidation products may cause various types of DNA damage. Preliminary data also exist that estrogens induce various genetic mutations. Tumors may develop from cells carrying such mutations and responding to receptor-mediated proliferation signals.

Pharmacology of camptothecin esters.

An intact lactone ring of camptothecins is a structural requirement for their anticancer activity. Propionate esters of camptothecin (CPT) and 9-nitrocamptothecin (9NC), CZ48 and CZ112, respectively, have been synthesized as derivatives resistant to lactone hydrolysis and are chemotherapeutically active. In this study, we have examined the mechanism of action of CZ48 and CZ112 and their distribution, metabolism, and toxicity. CZ112 incubated in human plasma retained its lactone structure longer than 9NC (t1/2: 10.5 and < 1 hr for CZ112 and 9NC, respectively). This resistance to lactone hydrolysis was also observed in mouse plasma or albumin solutions. Neither CZ48 nor CZ112 inhibit topoisomerase I and thus are prodrugs dependent on hydrolysis to CPT or 9NC, respectively. Rates of hydrolysis of CZ48 to CPT are higher by homogenates of mouse liver, spleen, lung, and kidney than by plasma. Rates of hydrolysis by tumor cells in culture vary and were higher by breast cancer and melanoma cells than by colon cancer cells. On the basis of these and other data, it is proposed that CZ48 and CZ112 may act as anticancer agents by resisting hydrolysis to camptothecins while in circulation. Hydrolysis in tissues may release intact lactone in target tissues.

Lack of mutations in DNA polymerase beta of estradiol-induced hamster kidney tumors: sequence of hamster DNA polymerase beta cDNA.

We examined the effects of estradiol (E2), the natural estrogenic hormone, on the structure and expression of DNA polymerase beta (DNA pol beta), a DNA repair gene, from E2-induced primary kidney tumors of twelve Syrian hamsters, their metastases, and from kidney tissues surrounding the tumors. We sequenced the coding region of the hamster DNA pol beta and found it to differ from that of the human by 11%. No mutations were detected in the entire coding region including the catalytic domain of the DNA pol beta from E2-induced primary kidney tumors, their metastases, or from kidney tissues surrounding the tumors. The expression of the DNA pol beta mRNA was also not significantly altered in E2-induced kidney tumors or in kidney tissues surrounding the tumors compared to that of control kidney tissues. These results suggest that mutations in the DNA pol beta gene may not be involved in the induction or malignant progression of hamster kidney tumors induced by E2. The nucleotide sequence of the hamster DNA pol beta described here will be useful for the study of the structure and expression of this gene.

Differential regulation of c-fos expression in estrogen-induced hamster renal tumors compared with kidney not due to creation of an estrogen-response element by point mutation in the gene's flanking sequence.

The conversion of a palindromic sequence, GGTCTnnnAGACC, in the 5'-flanking region of the murine c-fos proto-oncogene into a functional estrogen-response element by a single base change into GGTC(A/G)nnnAGACC has previously been postulated [Nawaz et al., 1993] as a possible mechanism of the induction of tumors by estrogens. This attractive hypothesis has been investigated in estradiol-induced Syrian hamster kidneytumors, in H-301 kidney tumor cells (a cell line derived from the Syrian hamster tumor), and in normal kidney tissue. The c-fos gene is differentially regulated by a classical estrogen receptor-mediated process in tumors, whereas in the acutely treated kidney, estradiol induces c-fos expression independent of estrogen-receptor function. In this study, we identified in the 5'-flanking region of the hamster kidney c-fos gene the sequence AGTCCnnnAGACC, which closely resembled but did not appear to function as an estrogen-response element. No mutations were detected in this sequence or in the 5'-flanking region of c-fos genes from three different primary tumors and from H-301 tumor cells. To rule out the possibility of a low copy number of mutant alleles in a tumor sample, polymerase chain reaction-based single-strand conformation polymorphism analysis was performed on 372 base pairs of the 5'-flank of the c-fos gene (-367 to +5 base pairs relative to the transcription start point). Nine different kidney tumor DNA samples and five normal kidney tissue samples (controls) produced an identical pattern of DNA bands, suggesting a lack of natural polymorphisms and mutations in this region of the c-fos gene. Acute treatment of hamsters with 17beta-estradiol for 6 h significantly induced renal c-fos mRNA expression, whereas control levels of c-fos were restored by co-treatment with estradiol and either N-acetyl-L-cysteine or alpha-naphthoflavone. We concluded that the previously observed change in regulatory control of c-fos expression in kidney versus estradiol-induced tumors does not involve the creation of a functional estrogen-response element by single point mutation in the 5'-flanking region of the gene. Additionally, c-fos expression in estradiol-treated hamster kidneys appears to be mediated by free radicals generated by the catechol metabolites of estradiol and not by the activation of any estrogen receptor.

Estrogen, DNA damage and mutations.

Estrogen administration to rodents results in various types of DNA damage and ultimately leads to tumors in estrogen-responsive tissues. Yet these hormones have been classified as nonmutagenic, because they did not induce mutations in classical bacterial and mammalian mutation assays. In this review, we have discussed the induction by estrogens of DNA and chromosomal damage and of gene mutations, because the classical assays were designed to uncover mutations only at one specific locus and could not have detected other types of mutations or changes in other genes. Various types of estrogen-induced DNA damage include: (a) direct covalent binding of estrogen quinone metabolites to DNA; (b) enhancement of endogenous DNA adducts by chronic estrogen exposure of rodents; (c) free radical generation by metabolic redox cycling between quinone and hydroquinone forms of estrogens and free radical damage to DNA such as strand breakage, 8-hydroxylation of purine bases of DNA and lipid hydroperoxide-mediated DNA modification. Two different types of chromosomal damage have also been induced by estrogen in vivo and in cells in culture such as numerical chromosomal changes and also structural chromosomal aberrations. Gene mutations have been induced in several cell types in culture either by the parent estrogen or by reactive estrogen quinone metabolites. Furthermore, in estrogen-induced kidney tumors in hamsters, several mutations have been observed in the DNA polymerase beta gene mRNA. Estradiol also induces microsatellite instability in these kidney tumors and in premalignant kidney exposed to estradiol. Although this work is still ongoing, it can be concluded that estrogens are complete carcinogens capable of tumor initiation by mutation potentially in critical genes. The hormonal effects of estrogens may complete the development of tumors.

Concentration dependence of prooxidant and antioxidant properties of catecholestrogens.

Estradiol is an established antioxidant in vitro and in vivo. In contrast, prooxidant effects such as 8-hydroxylation of guanine bases of DNA have been induced by various estrogens in hamsters and by 4-hydroxyestradiol or -estrone and a microsomal activating system in vitro. As part of an examination of these conflicting reports, we studied the enhancement or inhibition of lipid peroxidation (conjugated diene formation monitored at 240 nm) by catecholestrogens in human low-density lipoprotein (LDL) incubated with cupric sulfate in phosphate buffer. Addition of 2- or 4-hydroxyestradiol, 2- or 4-methoxyestradiol, or estradiol or estriol (0.5-50 microM) increased lag times for diene formation by 30 to <300% over control values in the absence of estrogens (lag time, 1.6 h). In contrast, low concentrations (5 pM-100 nM) of catecholestrogens decreased lag times by about 40-50%, demonstrating their prooxidant activities. The prooxidant capabilities of catecholestrogens were examined by assaying the reduction by estrogens of Cu(II) to Cu(I) and of Fe(III) to Fe(II). Both 2- and 4-hydroxyestradiol and 2- and 4-methoxyestradiol reduced Cu(II) and Fe(III) ions to their lower oxidation state. In conclusion, the reduction of Cu(II) to Cu(I) by catecholestrogens is proposed to initiate lipid peroxidation and thus oxidation of LDL. In contrast, at high concentrations of catecholestrogens, the scavenging of oxygen radicals may predominate over lipid peroxidation and free radical generation by analogy to the action of similar phenolic antioxidants. With estradiol, estriol, and the methoxyestrogen metabolites, only antioxidant effects were observed.

Accumulation of albumin in renal cytosol of hamsters treated with estradiol and in estrogen-induced hamster kidney tumors.

The intracellular accumulation of albumin has been observed in cytosols of benign and malignant human breast tumors and in mammary tumors of rodents induced by carcinogens. Additionally, cellular uptake of albumin has been detected in MCF-7 human breast cancer cells in culture. The clinical relevance of the albumin accumulation in human and rodent mammary tumors is not clear. In this study, we investigated the accumulation of albumin in an estrogen-induced and -dependent hamster kidney tumor model to understand the mechanisms and the role of hormones in this process. Protein accumulation patterns were examined by Western blot analyses in kidney homogenates of hamsters treated with 17beta-estradiol for various lengths of time and in kidney tumors which are induced with 100% incidence by this treatment for at least six months. Such analyses were also carried out in tissues of hamsters treated with the weakly carcinogenic estrogen 17alpha-ethinylestradiol (10% tumor incidence after nine months of treatment). Our data demonstrate the accumulation of albumin in kidney of hamsters treated with 17beta-estradiol but not with 17alpha-ethinylestradiol. Albumin accumulates specifically in the target organ of carcinogenesis, the kidney, however, with no increase in the serum concentrations or in the liver. Tumors do not develop in the livers of hamsters under these conditions of 17beta-estradiol treatment. This accumulation of albumin in hamster kidney may be the result of damage to the glomerulum which may be compromised by estradiol-induced toxicity and therefore unable to filter out excess albumin.

Enhancement of estrogen-induced renal tumorigenesis in hamsters by dietary iron.

Iron participates in the generation of hydroxyl radicals by the iron-catalyzed Fenton reaction. Its role in estrogen-induced carcinogenesis has been examined in this study by investigating the effects of iron content of hamster diets on tumor induction by estradiol. The renal tumor incidence and number of tumor nodules in hamsters treated with estradiol plus a diet enriched with iron (384 p.p.m. Fe as ferric citrate) for 5 months were 2- and 4-fold higher, respectively, than those observed in animals on an iron-poor diet plus estradiol (3.9 p.p.m. Fe, as ferric citrate). Tumor incidence and number of tumor nodules in estradiol-treated hamsters on the iron-deficient diet were not different from those of animals on a normal rodent chow. No tumors were detected in hamsters treated only with the low or high iron diets. Total serum iron was significantly increased in animals treated with the high iron diet plus estradiol compared with the low iron diet plus estradiol group and the high and low iron controls. Estrogen treatment increased non-heme iron in liver of both high and low iron treatment groups and in kidney of the hamsters on the low iron diet. It is concluded that dietary iron enrichment enhances the incidence and severity of estrogen-induced tumor induction.