Aromatase expression of human osteoblast-like cells

Estrogen plays a major role in bone mineral homeostasis, maintaining a balance between bone formation and bone resorption not only in women but also in men. Extraglandular aromatization of circulating androgen is the major source of estrogen in post-menopausal women and men. In order to assess the capacity of bone cells as a local source of estrogen, osteoblast-like cells (OLCs) were obtained from human fetal bone in mid-trimester by the explant method and by mechanical disaggregation. The integrity of OLCs was confirmed by their ability to produce alkaline phosphatase and osteocalcin in response to vitamin D3 and also by their ability to deposit mineral. Aromatase activity was assessed by the formation of estrone from [1,2,6,7-3H]androstenedione and by the release of tritium from [1beta-3H]androstenedione into [3H]water. Formation of estrone was confirmed by thin layer chromatography (TLC) in OLCs stimulated with dexamethasone (DEX) + oncostatin M. The aromatase activity was 10 x higher in non-passaged OLCs than in passaged cells in the presence or absence of the stimulants (DEX + IL-1beta). The apparent Km and Vmax estimated by the release of [3H]water was 5.8+/-0.6 nM and 10.8+/-1.4 pmol/mg per 6 h in the presence of DEX + IL-1beta. The effects of several stimulants on aromatase activity in OLCs were examined: serum, IL-1beta, TNFalpha and type I cytokines stimulated activity in the presence of DEX, while PMA and PMA + dibutyryl cAMP did not. To confirm the expression of aromatase in OLCs, cells prepared from periosteal membranes were also examined: These cells in culture possessed aromatase activity corresponding to OLCs prepared from bone specimens. Moreover, the fresh periosteum expressed aromatase at higher levels than that of metaphyseal specimens. The aromatase gene employs several different promoters (I.1, 1.2, I.3, I.4, I.5, I.6, 2a, 1f and PII) and the usage of these promoters is known to be controlled in a tissue-specific fashion. Accordingly, promoter usage in OLCs and fetal long bone (tibia) tissue was examined using the 5' rapid amplification of cDNA ends (RACE) technique. The major promoter used was I.4, not only in stimulated and non-stimulated OLCs, but also in fetal tibia. Some minor transcripts were also found: 1f (brain-specific promoter), PII and I.6 in OLCs stimulated by DEX + IL-1beta, and PII and I.3 in OLCs stimulated by DEX + serum. Fetal tibia also expressed I.3 (15%) and I.6 (10%). Thus, regulation and promoter usage in OLCs was quite different from other tissues known as estrogen sources including adipose tissue, ovary and placenta. These results suggest that bone is an extraglandular source of local estrogen which plays an important role in bone mineral metabolism through autocrine and paracrine actions.

Multiple splicing events involved in regulation of human aromatase expression by a novel promoter, I.6

The expression of aromatase is regulated in a tissue-specific fashion through alternative use of multiple promoter-specific first exons. To date, eight different first exons have been reported in human aromatase, namely I.1., I.2, I.3. I.4, I.5, PII, 2a, and 1f. Recently, we have found a new putative exon I in a RACE-generated library of THP-1 cells and have conducted studies to characterize this new exon I. We confirmed that the constructs containing -1552/+17 or less flanking sequence of this exon function as a promoter in THP-1 cells, JEG-3 cells and osteoblast-like cells obtained from a human fetus. Results of transfection assays using a series of deletion constructs and mutation constructs indicate that a 1-bp mismatch of the consensus TATA-like box (TTTAAT) and the consensus sequence of the initiator site, which is located 45 bp downstream of the putative TATA box, were functioning cooperatively as a core promoter. The putative transcription site was confirmed by the results of RT-PCR southern blot analysis. We examined the regulation and the expression of this exon, I.6, in several human cells and tissues by RT-PCR Southern blot analysis. THP-1 cells (mononuclear leukemic origin) and JEG-3 cells (choriocarcinoma origin) expressed exon I.6 in serum-free media. The level of expression was increased by serum and phorbol myristyl acetate (PMA) in both cell lines. Adipose stromal cells also expressed exon I.6 in the presence of PMA. In fetal osteoblasts, the expression of exon I.6 was increased most effectively by serum and less so by dexamethasone (DEX) + IL-1beta and DEX + IL-11, whereas induction by serum was suppressed by the addition of DEX. The level of expression was low in granulosa cells in culture and did not change with forskolin. On the other hand, dibutyryl cAMP suppressed PMA-stimulated expression of exon I.6 in THP-1 cells and adipose stromal cells. This result supports the hypothesis that the expression of exon I.6 is regulated mainly via an AP-1 binding site that is found upstream of the initiator site of the promoter region. Expression of exon I.6-specific transcripts was examined in several human tissues. Testis and bone obtained from normal adults expressed exon I.6. Testicular tumor and hepatic carcinoma expressed high levels of exon I.6, whereas granulosa cell tumor did not. Fetal liver and bone also showed a significant level of exon I.6 expression, but not so much as testicular tumor and hepatic tumor. Several splicing variants of exon I.6 were detected especially in THP-1 and JEG-3 cells, and to a lesser extent in primary cultures and tissue samples. These variants were identified as an unspliced form, a form spliced at the end of exon I.4, a form spliced at the end of exon I.3 (truncated) and a form spliced 220 bp downstream of the 3' end of exon I.6. The last variant revealed a new splicing site. Because most of the splicing variants contain the sequence specific for exon I.3, RT-PCR specific for exon I.3 can coamplify these splicing variants of exon I.6 transcripts. These results suggests that it is necessary to examine the expression of I.6 in tissues that are known to express exon I.3 such as breast adipose tissue, in which promoter usage of exon I of the aromatase gene switches from exon I.4 to I.3 in the course of malignant transformation.

Characterization of a region upstream of exon I.1 of the human CYP19 (aromatase) gene that mediates regulation by retinoids in human choriocarcinoma cells

The biosynthesis of estrogens is catalyzed by aromatase P450 (P450arom), the product of the CYP19 gene. The tissue-specific expression of the CYP19 gene is regulated by means of tissue-specific promoters through the use of alternative splicing mechanisms. Thus, transcripts containing various 5'-untranslated termini are present in ovary, brain, adipose stromal cells, and placenta. Sequence corresponding to untranslated exon I.1 is present uniquely in 5'-termini of transcripts expressed in human placenta and choriocarcinoma cells, as a consequence of expression driven by a distal promoter, I.1. The goal of the present study was the identification of regulatory elements in this promoter region. Various deletion mutations of the upstream flanking region of exon I.1 were constructed using the PCR or restriction enzyme digestion. The genomic fragments were fused upstream of the luciferase reporter gene. These constructs were transfected into human choriocarcinoma (JEG3) cells. The longest construct employed, -924/+10 bp, expressed the highest luciferase reporter gene activity. The -64/+10 bp and -125/+10 bp constructs showed no reporter gene expression. Transfection of the -201/+10 bp construct resulted in reporter gene expression, but at a lower level than that of the -924/+10 bp construct, and this expression was induced by serum as well as by LG69 and TTNPB, ligands specific for RXR and RAR respectively, as well as by vitamin D. These results parallel the actions of the ligands on aromatase activity. Mutation or deletion of an imperfect palindromic sequence (AGGTCATGCCCC) located at -183 to -172 bp upstream of the transcriptional start site of exon I.1 resulted in loss of basal- and retinoid-induced reporter gene expression. Gel retardation analysis using nuclear extracts of JEG3 cells treated with retinoids and the imperfect palindromic sequence as probe, showed that proteins present in the nuclear extracts bound to this sequence in a specific fashion. The binding activities were elevated by incubation of the cells with LG69 and TTNPB, ligands specific for RXR and RAR respectively. Binding of nuclear proteins to the palindromic sequence was displaced either by anti-RXR alpha serum or by anti-VDR serum, suggesting the formation of a heterodimer of RXR alpha and VDR. These results suggest that the imperfect palindromic sequence upstream of exon I.1 plays an important but novel role in the regulated expression of the CYP19 gene in choriocarcinoma cells.

Estrogen biosynthesis in THP1 cells is regulated by promoter switching of the aromatase (CYP19) gene

The expression of aromatase, the enzyme responsible for estrogen biosynthesis, has been studied in THP-1 cells of human mononuclear leukemic origin, which exhibit high rates of aromatase activity. These cells have the capacity to differentiate in the presence of vitamin D into cells with osteoclast-like properties. Differentiated cells displayed higher rates of aromatase than undifferentiated cells, and, in both cases, activity was stimulated 10- to 20-fold by dexamethasone. Phorbol esters also increased aromatase activity, but the effect was the same in differentiated as in undifferentiated cells. In a similar fashion to adipose stromal cells, serum potentiated the response to dexamethasone but had no effect on phorbol ester-stimulated activity. By contrast to its action in adipose stromal cells, (Bu)2cAMP markedly inhibited aromatase activity of THP-1 cells, as did factors whose actions are mediated by cAMP, such as PTH and PTH-related peptide. This was true of control cells, as well as of dexamethasone- and phorbol ester-stimulated cells. Previously we have shown that type 1 cytokines as well as tumor necrosis factor-alpha stimulate aromatase activity of adipose stromal cells in the presence of dexamethasone. By contrast, interleukin-6, interleukin-11, and leukemia-inhibitory factor had no effect on aromatase activity of THP-1 cells, whereas tumor necrosis factor-alpha, oncostatin M, and platelet-derived growth factor were slightly inhibitory of aromatase activity. Exon-specific Southern analysis of rapid amplification of cDNA ends-amplified transcripts was employed to examine the distribution of the various 5'-termini of aromatase transcripts. In the control group, most of the clones contained transcripts specific for the proximal promoter II, whereas in dexamethasone-treated cells, most transcripts contained exon I.4. In the phorbol ester-treated cells, a broader spectrum of transcripts was present, with equal proportions of I.4, II, and I.3-containing clones. Additionally, one clone containing a new sequence, exon I.6, was found. This was shown to be located about 1 kb upstream of exon II. By contrast, all clones from cells treated with (Bu)2cAMP contained promoter II-specific sequences. In addition to these transcripts, two clones in the library from the dexamethasone-treated cells contained the sequence previously defined as the brain-specific sequence, 1f. In one of these, the 1f sequence was fused downstream of exon I.4, indicative that its expression likely employed promoter I.4. These results point to similarities and important differences between aromatase expression in THP-1 cells and other cells such as adipose stromal cells, indicative of unique regulatory pathways governing aromatase expression in these cells.

A CRE-like sequence that binds CREB and contributes to cAMP-dependent regulation of the proximal promoter of the human aromatase P450 (CYP19) gene

The major physiological regulator of human aromatase P450 gene expression in the ovary is follicle stimulating hormone (FSH), which acts by increasing intracellular cAMP levels. This study describes the identification of an element in the aromatase proximal promoter that is critical for the full transcriptional response of this promoter to cAMP. The cAMP-responsive element (CRE)-like sequence (CLS) was originally identified by its sequence similarity to a palindromic CRE, from which it differs by the insertion of a single cytosine. Mutation of the CLS in the context of 278 bp of 5'-flanking DNA resulted in the loss of cAMP-induced reporter gene expression in transfected ovarian luteal cells. A cell line survey EMSA revealed that CLS binding factors are ubiquitously distributed, although the migration pattern of CLS-nuclear protein complexes varied among different nuclear extracts. An extended half-site for binding members of the basic-leucine zipper class of transcription factors was found to be responsible for ovarian luteal cell nuclear protein binding and cAMP-dependent transcriptional transactivation. Competition and supershift EMSAs revealed that the CLS-nuclear protein complexes that regulate cAMP-induced transcription were indistinguishable from homodimeric CREB bound to the CRE oligonucleotide, yet the interaction with the CLS was of lower affinity.

The 5'-flanking region of the ovarian promoter of the bovine CYP19 gene contains a deletion in a cyclic adenosine 3',5'-monophosphate-like responsive sequence

Conversion of C19 steroids to estrogens is catalyzed by aromatase P450 (P450arom; the product of the CYP19 gene). In the ovary, P450arom is expressed in granulosa cells of both human (h) and bovine (b) follicles. After the ovulatory surge of gonadotropins, however, P450arom expression is maintained only in the luteinized granulosa cells of the human ovary and is absent from the bovine corpus luteum. We compared the regulation of expression of the ovary-specific human CYP19 (hCYP19ov) and the bovine CYP19 (bCYP19ov) gene by cAMP (forskolin) and sought to determine whether the divergence in the expression of P450arom with the onset of luteinization could be explained by specific cis-acting elements present uniquely in the 5'-flanking DNA of the hCYP19ov or bCYP19ov gene. We, therefore, subcloned DNA encompassing the promoters and 5'-flanking regions of the hCYP19ov or bCYP19ov gene into a promoterless luciferase vector. These constructs were transfected into luteinized bovine granulosa cells or bovine luteal cells in primary culture. Neither cell type exhibits endogenous expression of bovine P450arom. After transfection, cells were treated with either vehicle or 25 microM forskolin. There was little or no increase in luciferase activity after forskolin treatment in cells transfected with any of the bCYP19ov constructs, whereas all of the corresponding hCYP19ov constructs (-693/-16 to -214/-16 bp) expressed reporter activity in the presence of forskolin. This dramatic difference between the activities of the constructs of the two species occurred despite the fact that there is an 88% sequence identity between the bovine and human promoters in the region between -214 to -16 bp. One possible explanation for this variability may be that the bCYP19ov gene has a 1-bp deletion in a cAMP-response element-like sequence (CLS) present at -208 to -201 bp in the hCYP19ov gene that we have shown to be critical for cAMP-stimulated transcription of hCYP19ov in the ovary. When this region of the bCYP19ov promoter was mutated to the hCLS, a partial restoration in luciferase activity was observed after forskolin treatment. Therefore, these results suggest that another sequence in this -214 bp region of the bCYP19ov gene is also contributing to the lack of expression of P450arom after luteinization in the bovine ovary. This lack of expression of the bCYP19ov gene may be due to the presence of a repressive trans-acting factor expressed with the onset of luteinization of the bovine granulosa cell. These results further suggest that in the cow, elements upstream of those employed by the hCYP19ov gene may have been recruited to facilitate regulated expression of the bCYP19ov gene in the absence of a functional CLS.

A putative binding site for Sp1 is involved in transcriptional regulation of CYP17 gene expression in bovine ovary

In the bovine ovary, thecal cells are the only cell type capable of expressing the CYP17 gene in response to LH. With the onset of ovulation and luteinization in the cow, there is complete loss of P450c17alpha expression. To characterize the molecular mechanisms involved in tissue-specific regulation of the CYP17 gene in the bovine ovary, deletion mutations of the bovine CYP17 promoter were ligated into a promoterless luciferase expression vector, and reporter constructs were transiently transfected into primary cultures of bovine thecal and luteal cells. Deletion of the promoter sequences between -191 and 101 bp dramatically decreased the levels of reporter gene activity in both thecal and luteal cells. Computer-assisted analysis revealed the presence of a putative inverted Sp1-like binding site at -188/-180 bp. Deletion or mutation of this sequence caused a decrease in both basal and forskolin-stimulated reporter gene activity. In addition, mutation or deletion of this sequence also decreased reporter gene expression induced by overexpression of the protein kinase A catalytic subunit. Electrophoretic mobility shift assays showed that this sequence binds to a nuclear protein(s) from both thecal and luteal cells that is related to Sp1, as suggested by the results of gel mobility supershift assay employing an antibody raised against Sp1. DNA-binding activity was not increased by the addition of forskolin to thecal or luteal cells. We conclude that this inverted Sp1-like binding sequence is involved in constitutive as well as cAMP-dependent expression of the CYP17 gene in the bovine ovary.

Ultrasound biomicroscopic imaging of the anterior aspect of peripheral choroidal melanomas

PURPOSE

To correlate ultrasound biomicroscopic features of the anterior aspect of peripheral choroidal melanoma with respect to histopathology.

METHODS

We examined 17 eyes of 17 patients who had clinically diagnosed peripheral choroidal melanomas that approached the ora serrata or extended into the ciliary body and who had been assessed with ultrasound biomicroscopy before enucleation. Comparisons were made between anterior tumor margins imaged by ultrasound biomicroscopy and histopathologic specimens. Anatomic features noted on ultrasound biomicroscopy before enucleation were correlated with enucleation specimens, including supraciliary effusion, rotation of the ciliary body, angle involvement, and internal reflectivity patterns. Anterior tumor margin position was determined with reference to the scleral spur.

RESULTS

Mean distances from the anterior tumor margin to the scleral spur were 1.47 mm on ultrasound biomicroscopy and 1.65 mm on pathologic examination. This difference was not statistically significant (P = .325). Tumor features evident on ultrasound biomicroscopy were also seen on pathologic examination: supraciliary choroidal effusions in seven of seven, ciliary body rotation in seven of eight, and angle involvement in seven of eight. All tumors were mixed-cell melanomas, and 12 of 17 (70%) demonstrated homogeneous ultrasound biomicroscopic internal reflectivity. Irregular internal reflectivity was seen in five of 17 tumors (29%) and was related to prominent internal vascularity on pathology in three of five.

CONCLUSIONS

Ultrasound biomicroscopy is an accurate imaging technique for the in vivo assessment of anterior tumor margins of peripheral choroidal melanomas and can provide detailed imaging of the tumor's interface with the ciliary body.

Endocrine disorders associated with inappropriately high aromatase expression

Aromatase P450 (P450arom) is responsible for conversion of C19 steroids to estrogens in a number of human tissues, such as the placenta, gonads, adipose tissue, skin and the brain. Aromatase expression in human tissues is regulated by use of alternative promoters in the placenta (promoter I.1), adipose tissue (promoters I.4, I.3 and II) and gonads (promoter II). Aromatase expression is absent in the disease-free adult liver, adrenal and uterine tissues. Excessive or inappropriate aromatase expression in adipose fibroblasts and endometriosis-derived stromal cells, as well as in testicular, hepatic, adrenal and uterine tumors, is associated with abnormally high circulating estrogen levels and/or with increased local estrogen concentrations in these tissues. Whether systemically delivered or locally produced, elevated estrogen levels will in turn promote the growth of hormone-responsive tissues. We recently studied aromatase expression in testicular tumor and adipose tissue samples from prepubertal boys with gynecomastia, in hepatocellular cancer and adrenocortical tumor samples from adult men with gynecomastia, in breast adipose tissue samples proximal to breast tumors, and in endometrial cancer, leiomyoma and endometriosis tissues. Excessive aromatase activity and P450arom transcript levels were found in these tissue samples or in cultured cells derived from these tissues. In these neoplastic or non-neoplastic tissues or cells, the regulation of aromatase expression was studied in terms of alternative promoter use, both in vivo and in response to various hormonal stimuli. Our results were suggestive of a common metabolic abnormality associated with activation of a cyclic AMP-dependent signalling pathway that gives rise to transcriptional transactivation of aromatase expression via promoters I.3 and II in all of the above tissues. This article describes the common pathophysiological and molecular features of excessive aromatase expression in these disease states.

Aromatase in axonal processes of early postnatal hypothalamic and limbic areas including the cingulate cortex

It has been shown that sexual dimorphic morphology of certain hypothalamic and limbic areas underlie gender-specific sexual behavior and neuroendocrine mechanisms. The key role played by locally formed estrogen in these developmental events has been revealed during a critical perinatal period. In this study, we aimed to document the presence of estrogen-synthetase (aromatase)-immunoreactive elements in the involved limbic system and hypothalamus of the developing rat brain. On postnatal day 5, animals of both sexes were perfusion-fixed, and sections from the forebrain and hypothalamus were immunolabelled for aromatase using an antiserum that was generated against a 20 amino acid sequence of placental aromatase. Aromatase-immunoreactivity was present in neuronal perikarya and axonal processes in the following limbic structures: the central and medial nuclei of the amygdala, stria terminalis, bed nucleus of the stria terminalis (BNST), lateral septum, medial septum, diagonal band of Broca, lateral habenula and all areas of the limbic (cingulate) cortex. In the hypothalamus, the most robust labelling was observed in the medial preoptic area, periventricular regions, ventromedial and arcuate nuclei. The most striking feature of the immunostaining with this antiserum was its intracellular distribution. In contrast to the heavy perikaryal labelling that can be observed with most of the currently available aromatase antisera, in the present experiments, immunoperoxidase was predominantly localized to axons and axon terminals. All the regions with fiber staining corresponded to the projection fields of neuron populations that have previously been found to express perikaryal aromatase. Our results confirm the presence of aromatase-immunoreactivity in developing limbic and hypothalamic areas. The massive expression of aromatase in axonal processes raises the possibility that estrogen formed locally by aromatase may not only regulate the growth, pathfinding and target recognition of its host neuronal processes, but may also exert paracrine actions on structures in close proximity, including the target cells.

Transcriptional regulation of CYP19 gene (aromatase) expression in adipose stromal cells in primary culture

Estrogen biosynthesis in adipose tissue increases with age and obesity, and has been implicated in the development of endometrial cancer and breast cancer. In normal human adipose tissue, expression of the CYP19 gene which encodes aromatase P450, the enzyme responsible for estrogen biosynthesis, is regulated by a distal promoter, namely promoter I.4. Stimulation of expression in adipose stromal cells by members of the type 1 cytokine family, i.e. interleukin (IL)-6, IL-11, leukemia inhibitory factor (LIF) and oncostatin M (OSM), is mediated via a Jak-STAT3 signaling pathway and a GAS element upstream of promoter I.4. In contrast, aromatase expression in breast adipose tissue proximal to tumor is increased three- to four-fold to the utilization of another promoter, namely promoter II, proximal to the translation initiation site. In the present report, we show that prostaglandin (PG) E2 is the most potent factor which stimulates aromatase expression via cyclic AMP and promoter II. PGE2 acts via EP1 and EP2 receptor subtypes to stimulate both the PKC and PKA pathways. The combined stimulation of both of these pathways results in the maximal expression of promoter II-specific CYP19 transcripts. Because PGE2 is a major secretory product both of breast tumor epithelial cells and fibroblasts, as well as of macrophages infiltrating the tumor site, then this could be the mechanism whereby estrogen biosynthesis is stimulated in breast sites adjacent to a tumor, leading in turn to increased growth and development of the tumor itself.

Regulation of aromatase expression in the ovary and placenta: a comparison between two species

The conversion of C19 steroids to estrogens is catalysed by aromatase P450 (P450arom; product of the CYP19 gene). Tissue sites of expression include the gonads and brain; however, in a small subset of mammals, P450arom is also expressed in the placenta. In humans, gonadal expression employs a promoter proximal to the start site of translation, whereas expression in the placenta relies on a promoter which is distal to this site. We characterized the bovine CYP19 gene in both the ovary (OV) and placenta (PL), to determine if this method of regulation of tissue-specific expression is employed in other species. Both bovine and human species express P450arom in these tissues, however, the pattern of expression is very different. In both humans and cattle, P450arom is expressed in the granulosa cells of the ovary, however, after ovulation only the luteinized granulosa cells of the human continue to express P450arom. There is a high degree of sequence identity (>70%) shared between humans and cattle in the OV-specific 5'-flanking DNA, whereas little identity (<40%) was found between humans and cattle in the PL-specific DNA. Promoters and 5'-flanking regions of human and bovine CYP19 genes were subcloned into a luciferase (LUC) vector. Bovine PL-specific constructs transfected into JEG-3 human choriocarcinoma cells failed to express LUC activity. When OV-specific constructs were transfected into luteinized bovine granulosa cells, there was no change in LUC activity in cells transfected with the bovine constructs after treatment with forskolin, whereas all of the corresponding human constructs expressed LUC activity. The bovine 5'-flanking DNA lacks a cAMP-responsive element-like sequence (CLS) critical for cAMP-stimulated transcription of P450arom in the human ovary. With the absence of this CLS sequence in the bovine gene, there appears to be little enhancement of transcription by cAMP in the portion of the 5'-flanking region studied so far. When the analogous region of the bovine promoter was mutated to the human CLS, only a partial restoration of LUC activity was observed. An additional element therefore appears to be important in preventing the full expression of the bovine CYP19 gene in luteinized granulosa cells.

Steroidogenic factor 1 (SF-1) and SP1 are required for regulation of bovine CYP11A gene expression in bovine luteal cells and adrenal Y1 cells

Cholesterol side-chain cleavage cytochrome P450 (CYP11A; P450scc) gene expression is regulated by gonadotropins via cAMP in the ovary and by ACTH via cAMP in adrenal cortical cells. Previously, we have characterized a response element located at -118 to -101 bp in the 5'-flanking region of the bovine P450scc gene required for cAMP-stimulated transcription in both mouse adrenocortical Y1 cells and bovine ovarian cells in primary culture. It was shown that this region contains a binding site for the transcription factor Sp1. Deletion of this sequence abolished cAMP-stimulated transcription in both Y1 cells and bovine ovarian luteal cells. Another sequence element located at -57 to -32 bp upstream from the transcription initiation site, which is highly conserved in CYP11A of other species, contains the motif TAGCCTTG, similar to the consensus binding site of steroidogenic factor-1, SF-1 (or Ad4-BP), but in the inverted orientation. In the present study, gel shift analysis using nuclear extracts of either Y1 cells or bovine luteal cells demonstrated that the sequence between -57 and -32 bp bound SF-1. A mutation of the SF-1-binding site that abolished binding of the nuclear protein to DNA reduced markedly the basal transcription of the reporter gene as well as the responsiveness to cAMP, when the mutated fragments containing the region from -186 to +12 bp were cloned into a luciferase construct and transfected into mouse adrenal Y1 cells and bovine luteal cells. The role of SF-1 in P450scc transcription was further confirmed by transactivation of the -186/+12Luc construct employing an SF-1 expression vector after transfection into nonsteroidogenic COS-1 cells. In addition, results obtained employing a double mutation of the Sp1- and SF-1-binding sites, and from a construct containing both Sp1 and SF-1 elements upstream of the CYP11A TATA box, indicated that Sp1 and SF-1 function cooperatively in the transactivation of the bovine CYP11A promoter in both bovine luteal cells and Y1 cells. Finally, a mammalian two-hybrid system was employed to demonstrate that Sp1 and SF-1 can associate in vivo. These results establish that basal and cAMP-stimulated activity of the bovine P450scc promoter in both Y1 cells and bovine luteal cells requires the combined action of at least two transcription factors, Sp1 and SF-1.

Cytochromes P450 11: expression of the CYP19 (aromatase) gene: an unusual case of alternative promoter usage

Family 19 of the P450 super family is responsible for the conversion of C19 androgenic steroids to the corresponding estrogens, a reaction known as aromatization because it involves conversion of the delta4-3-one A-ring of the androgens to the corresponding phenolic A-ring characteristic of estrogens. The gene encoding human aromatase has been cloned and characterized and shown to be unusual compared to genes encoding other P450 enzymes, because there are numerous untranslated first exons that occur in aromatase transcripts in a tissue-specific fashion due to differential splicing as a consequence of the use of tissue-specific promoters. Thus, expression in the ovary uses a proximal promoter that is regulated primarily by cAMP. On the other hand, expression in the placenta uses a distal promoter located at least 40 kb upstream of the start of transcription that is regulated by retinoids. Other promoters are used in brain and adipose tissue. In the latter case, class I cytokines such as IL-6 and IL-11, as well as TNF-alpha, are important regulatory factors. A common 3'-splice junction located upstream of the start of translation is used in all of the splicing events involved in the use of these various promoters. Thus, the coding region of the transcripts, and hence the protein, are identical regardless of the tissue site of expression; what differs in a tissue-specific fashion is the 5'-end of the transcripts. This pattern of expression has great significance both from a phylogenetic and ontogenetic standpoint, as well as for the physiology and pathophysiology of estrogen formation, as will be discussed in this review.

Alternatively spliced transcripts of the aromatase cytochrome P450 (CYP19) gene in adipose tissue of women

Estrogen biosynthesis in adipose tissue has assumed great significance in terms of a number of estrogen-related diseases. The biosynthesis of estrogens from C19 steroids is catalyzed by a specific form of cytochrome P450, namely aromatase cytochrome P450 (P450arom; the product of the CYP19 gene). The human CYP19 gene comprises nine coding exons, II-X, and its transcripts are expressed in the ovary, placenta, testes, adipose tissue, and brain. Tissue-specific expression of the CYP19 gene is determined at least in part by the use of tissue-specific promoters, which give rise to transcripts with unique 5'-noncoding termini. Thus, the distal promoter I.1 is responsible for expression uniquely in placenta. On the other hand, the proximal promoter II, which regulates expression via a cAMP-dependent signaling pathway, is responsible for expression in the gonads. Transcripts in breast adipose tissue contain 5'-termini corresponding to expression derived from promoters I.4, II, and I.3, with I.4-specific termini predominating. The latter are derived from promoter I.4, which contains a glucocorticoid response element and an interferon-gamma activation site element and is responsible for expression in the presence of glucocorticoids and members of the class I cytokine family. The object of the present study was to determine the distribution of these various transcripts in adipose tissue from abdomen, buttocks, and thighs of women, as this would provide important clues to the factors regulating aromatase expression in these sites. To achieve this, we employed competitive reverse transcription-PCR to amplify unique 5'-ends of each of the transcripts of the CYP19 gene that are expressed in adipose tissue as well as for the coding region to evaluate total CYP19 gene (P450arom) transcript levels. We observed that exon I.4-specific transcripts were predominantly present in adipose tissue samples obtained from women regardless of the tissue site or the age of the individual. In these tissues, promoter II- and exon I.3-specific transcripts were present in lower copy numbers. We also demonstrated that in these sites total or exon-specific P450arom transcripts levels increased in direct proportion to advancing age and that transcript levels were the highest in buttocks, followed by thighs, and lowest in the abdomen. These results suggest that in normal human adipose tissue, aromatase expression is mainly under local control by a number of cytokines via paracrine and autocrine mechanisms in the presence of systemic glucocorticoids.

Aromatase expression in health and disease

Family 19 of the P450 superfamily is responsible for the conversion of C19 androgenic steroids to the corresponding estrogens, a reaction known as aromatization, since it involves conversion of the delta 4-3-one A-ring of the androgens to the corresponding phenolic A-ring characteristic of estrogens. Its members occur throughout the entire vertebrate phylum. The reaction mechanism of aromatase is very interesting from a chemical point of view and has been studied extensively; however, a detailed examination of structure-function relationships has not been possible due to lack of a crystal structure. Recent attempts to model the three-dimensional structure of aromatase have permitted a model that accounts for the reaction mechanism and predicts the location of aromatase inhibitors. The gene encoding human aromatase has been cloned and characterized and shown to be unusual compared to genes encoding other P450 enzymes, since there are a number of untranslated first exons that occur in aromatase transcripts in a tissue-specific fashion, due to differential splicing as a consequence of the use of tissue-specific promoters. Thus, expression in ovary utilizes a proximal promoter that is regulated primarily by cAMP. On the other hand, expression in placenta utilizes a distal promoter that is located at least 40 kb upstream of the start of transcription and that is regulated by retinoids. Other promoters are employed in brain and adipose tissue. In the latter case, class I cytokines such as IL-6 and IL-11 as well as TNF alpha are important regulatory factors. PGE2 is also an important regulator of aromatase expression in adipose mesenchymal cells via cAMP and PGE2 appears to be a major factor produced by breast tumors that stimulates estrogen biosynthesis in local mesenchymal sites. In all of the splicing events involved in the use of these various promoters, a common 3'-splice junction is employed that is located upstream of the start of translation; thus, the coding regions of the transcripts- and hence the protein-are identical regardless of the tissue site of expression; what differ in a tissue-specific fashion are the 5'-ends of the transcripts. This pattern of expression has great significance both from a phylogenetic and ontogenetic standpoint as well as for the physiology and pathophysiology of estrogen formation. Recently, a number of mutations of the aromatase gene have been described, which give rise to complete estrogen deficiency. In females this results in virilization in utero and primary amenorrhea with hypergonadotropic hypogonadism at the time of puberty. In men the most striking feature is continued linear bone growth beyond the time of puberty, delayed bone age, and failure of epiphyseal closure, thus indicating an important role of estrogens in bone metabolism in men. In both sexes the symptoms can be alleviated by estrogen administration.

Estrogen biosynthesis proximal to a breast tumor is stimulated by PGE2 via cyclic AMP, leading to activation of promoter II of the CYP19 (aromatase) gene

In the present report, we show that prostaglandin (PG) E2 is the most potent factor which stimulates aromatase expression via cyclic AMP and promoter II. PGE2 acts via EP1 and EP2 receptor subtypes to stimulate both the PKC and PKA pathways. The combined stimulation of both of these pathways results in maximal expression of promoter II-specific CYP19 transcripts. Since PGE2 is a major secretory product both of breast tumor epithelial cells and fibroblasts, as well as of macrophages infiltrating the tumor site, then this could be the mechanism whereby estrogen biosynthesis is stimulated in breast sites adjacent to a tumor, leading in turn to increased growth and development of the tumor itself.

Use of alternative promoters to express the aromatase cytochrome P450 (CYP19) gene in breast adipose tissues of cancer-free and breast cancer patients

Estrogen biosynthesis in adipose tissue has assumed great significance in terms of a number of estrogen-related diseases. Recent evidence suggests that estrogen synthesized locally in the breast is of singular significance in the development of breast cancer in elderly women. The biosynthesis of estrogen from C19 steroids is catalyzed by a specific form of cytochrome P450, namely aromatase cytochrome P450 (P450arom; the product of the CYP19 gene). The human CYP19 gene comprises nine coding exons, II-X, and its transcripts are expressed in the ovary, placenta, testes, adipose tissue, and brain. Tissue-specific expression of the CYP19 gene is determined, at least in part, by the use of tissue-specific promoters, which give rise to transcripts with unique 5'-noncoding termini. Transcripts in adipose tissue contain 5'-termini derived from specific untranslated exons, corresponding to expression derived from the proximal promoter II and its splice variant I.3, as well as a distal promoter, I.4. The object of the present study was to determine the distribution of these various exon-specific transcripts in breast adipose tissues from cancer-free women undergoing reduction mammoplasty and from patients with breast cancer, because this would provide important clues as to the nature of the factors regulating aromatase expression in these sites. To achieve this, we employed competitive RT-PCR, utilizing an internal standard for each exon-specific transcript of the CYP19 gene, as well as for the coding region, to evaluate total CYP19 gene transcripts. In cancer patients (n = 18), total CYP19 gene transcript levels were significantly higher in adipose tissue proximal to a tumor in comparison with adipose tissue distal to a tumor, in agreement with previous findings. Moreover, total transcript levels were higher in breast adipose tissue of cancer patients in comparison with those of cancer-free individuals (n = 9), even when the adipose tissue from the cancer patient was taken from a quadrant with no detectable tumor. We observed that exon I.4-specific transcripts were predominant in breast adipose obtained from cancer-free women. In this tissue, promoter-II-specific and exon I.3-specific transcripts were present in low copy number. On the other hand, in breast cancer patients, CYP19 gene transcripts from breast adipose tissue had primarily promoter-II-specific and exon I.3-specific sequence, whereas comparatively few transcripts had exon I.4-specific sequence at the 5'-terminus. We conclude that CYP19 gene transcription in breast adipose tissue of cancer-free individuals uses preferably promoter I.4, implicating a role of glucocorticoids and members of the IL-6 cytokine family in the regulation of this expression. On the other hand, the increased expression in breast adipose tissue bearing a carcinoma results from expression from promoters II and I.3, which are regulated by unknown factors acting via increased cAMP formation, which are presumably secreted by the tumor or associated cells.

Tumor necrosis factor-alpha stimulates aromatase gene expression in human adipose stromal cells through use of an activating protein-1 binding site upstream of promoter 1.4

Expression of aromatase P450 (P450arom; the product of the CYP19 gene) in human adipose stromal cells in primary culture is markedly stimulated by serum in the presence of dexamethasone (DEX). Under these conditions, the majority of P450arom transcripts contain untranslated exon 1.4 at their 5'-ends. Previously, we observed that the region of the CYP19 gene upstream of exon 1.4 contains a TATA-less promoter, a glucocorticoid response element, and an interferon-gamma-activating sequence. These act to mediate the action of interleukin-6 and related cytokines to stimulate aromatase expression in the presence of DEX. In the present study, we found that tumor necrosis factor-alpha (TNF alpha) also acts synergistically with DEX to stimulate aromatase expression in adipose stromal cells in serum-free medium. We observed that the action of TNF alpha can be mimicked by ceramide. Maximal aromatase activity was obtained when cells were incubated with 5 ng/ml TNF alpha or 100 nM ceramide in the presence of 250 nM DEX. Levels of c-fos and c-jun proteins also were increased by TNF alpha or ceramide in the presence of DEX. Upstream of the interferon-gamma-activating sequence site there is an imperfect activating protein-1 (AP-1) binding site (2-bp mismatch). Gel retardation analysis using nucleotide probes containing the putative AP-1-binding sequence and nuclear extracts of human adipose stromal cells cultured in the presence of TNF alpha or ceramide plus DEX revealed that adipose stromal cells nuclear proteins bind to this site and that binding was competed by a 100-fold excess of a consensus AP-1 sequence. In addition, binding activity was competed by both anti-c-fos and anti-c-jun sera. Mutation or deletion of the putative AP-1 element resulted in the loss of TNF alpha- plus DEX-induced activity of reporter constructs comprised of 515 bp of the exon 1.4 flanking sequence linked to the luciferase gene. These results suggest that TNF alpha, probably acting through ceramide formation, stimulates the binding of both c-fos and c-jun to the AP-1 element upstream of exon 1.4. These act cooperatively with the ligand-activated glucocorticoid receptor to induce aromatase expression in adipose stromal cells in primary culture. We conclude that this TNF alpha signal transduction pathway may play an important role in the regulation of estrogen biosynthesis in adipose tissue.

17alpha-Hydroxylase gene expression in the bovine ovary: mechanisms regulating expression differ from those in adrenal cells

17alpha-Hydroxylase cytochrome P450 (P450(17alpha)) is the enzyme which synthesizes C19 steroids in a two-step reaction in which 17alpha-OH pregnenolone is an intermediate. In the bovine and human adult female, 17alpha-hydroxylase is expressed in adrenocortical cells where 17alpha-OH pregnenolone and 17alpha-OH progesterone are precursors of cortisol, and in theca cells of the ovary where these intermediates are precursors of C19 steroids. In both adrenal cortex and theca, 17alpha-hydroxylase gene expression is stimulated by cyclic AMP (cAMP). The aim of this study was to determine the mechanism regulating 17alpha-hydroxylase gene expression in the bovine ovary. Our results indicate that the bovine 17alpha-hydroxylase gene is regulated in a tissue-specific fashion. Primer extension and S1 nuclease protection assays reveal that the start site of transcription in the theca is identical to that in the adrenal. Transfection studies employing beta-globin reporter gene constructs fused to successive deletions of the 5' regulatory region of the bovine 17alpha-hydroxylase gene indicate that sequences between -80 and -37 basepairs (bp) (CRS2) confer cAMP-regulated transcription in bovine theca cells in culture. These results are in contrast to similar studies conducted in bovine adrenocortical cells, which indicate that the major cAMP response element (referred to as CRS1) is located at -243 to -225 bp. The Ad4 element (AGGTCA, -42 to -37 bp) within CRS2, which has been shown to be involved in cAMP responsiveness in other steroidogenic P450 genes, cannot by itself confer cAMP-regulated reporter gene expression in bovine cells. These results indicate that in the cow, 17alpha-hydroxylase gene expression is regulated in a tissue-specific fashion, and that this regulation may be conferred, at least in part, by the use of tissue-specific cis-acting elements in the bovine 17alpha-hydroxylase gene.

Cellular characterization of adipose tissue from various body sites of women

Adipose tissue is the primary site of estrogen biosynthesis in postmenopausal women. The two main histologic components of adipose tissue are mature adipocytes and fibroblasts. Aromatase P450 expressed in the fibroblast component of adipose tissue is responsible for catalyzing conversion of C19 steroids to estrogens. We previously have demonstrated that in women, aromatase expression in adipose tissue of various body sites increases with age and that aromatase expression in the hip is markedly higher than in the abdomen. To determine whether this age- and regional-dependent variation in aromatase expression is caused by an alteration in the ratio of fibroblasts to mature adipocytes, we collected sc adipose tissue samples from 19 women (age range: 21-93 yr) at the time of autopsy. Using a computerized image analysis system, we determined by morphometry the proportions of adipocytes, fibroblasts, and vascular endothelial cells within histologic sections of adipose tissue from midabdomen, both breasts, and both hips. The percentage of each cell component at each body site was expressed as the mean of triplicate replicates. Statistical analysis of our results did not indicate any correlation between advancing age and fibroblast to adipocyte ratios in the breast, abdomen, or hip. Fibroblast to adipocyte ratios were found to be significantly higher in the breast and abdomen compared with the hip ( P < 0.05). No statistical differences were found between the breast and abdomen. These findings suggest that the increase in aromatase expression with advancing age and the higher aromatase expression in the hip compared with the abdomen in women may be caused by alterations in specific signal transduction mechanisms rather than a simple increase in local adipose fibroblast numbers.

Detection of aromatase cytochrome P450, 17 alpha-hydroxylase cytochrome P450 and NADPH:P450 reductase on the surface of cells in which they are expressed

Using the membrane impermeant probe NHS-LC-biotin, we show in this report that a fraction of aromatase P450 (P450arom), the enzyme that catalyzes estrogen biosynthesis, is present at the surface of cells in which it is expressed, either endogenously or as a consequence of transfection. The same findings were obtained for a truncated form of P450arom lacking the putative membrane-spanning region, thus suggesting the presence of other membrane-spanning region(s) within its structure. P450arom is not unique in this regard as we find that a fraction of 17 alpha-hydroxylase P450 as well as NADPH:P450 reductase also are present at the cell surface. It is therefore possible that a number of microsomal P450s are expressed at the cell surface in this fashion.