MECHANISMS IN THE REGULATION OF AROMATASE IN DEVELOPING OVARY AND PLACENTA

1. Introduction

In most vertebrates, expression of the aromatase P450 (P450arom/CYP19) gene is restricted to the gonads and brain; however, in humans, aromatase is expressed in specific cell populations of a variety of estrogen-producing tissues, including the syncytiotrophoblast layer of the placenta, granulosa and luteal cells of the ovary, Leydig, Sertoli and germ cells of the testis, stromal cells of adipose tissue, bone, discrete nuclei within the brain and in fetal liver [1,2]. hCYP19 gene expression in various estrogen-producing tissues appears to be driven by tissue-specific promoters upstream of alternative first exons, which encode the tissue-specific 5′-untranslated regions of P450arom mRNA transcripts. These unique promoters not only control tissue-specific expression of these P450arom mRNA transcripts, but also mediate their differential regulation by hormones and factors. The alternative first exons, which are located from ∼110 to ∼100,000 bp upstream of the hCYP19 translation initiation site in exon II, are alternatively spliced onto a common site just upstream of the translation start site in exon II, so that the protein encoded in each of these tissues is identical (Fig. 1). In placenta, the majority of the P450arom mRNA transcripts contain sequences encoded by exon I.1, which lies ∼100,000 bp upstream of the start site of translation in exon II, whereas in ovary, P450arom mRNA transcripts contain 5′-untranslated sequences encoded by exon IIa which lies 110 bp upstream of the translation start site [2] (Fig. 1).

2. Use of transgenic mice and transfected cells to define genomic sequences that mediate placenta-specific hCYP19 expression

In previous studies using primary cultures of human placental cells, we observed that differentiation of cytotrophoblasts to syncytiotrophoblast is oxygen-dependent and associated with a marked induction of aromatase activity and hCYP19 gene expression [3,4]. Transfection of placental and non-placental cells with reporter gene constructs revealed that placenta-specific exon I.1 5′-flanking sequences between -501 and -42 bp mediates trophoblast-specific hCYP19 gene expression [3]. Studies using transgenic mice also suggested that as little as 501 bp of exon I.1 5′-flanking DNA directed reporter gene expression exclusively to the placenta and specifically to the labyrinthine trophoblast layer, which is region of mouse placenta most analogous to the human syncytiotrophoblast [5]. Collectively, these findings suggest that the 5′-flanking DNA within 501 bp of exon I.1 of the hCYP19 gene contains cis-acting elements that bind placenta-specific transcription factors. Since mouse placenta does not express aromatase, it is likely that placental transcription factors that mediate hCYP19 gene expression are conserved between mouse and human, while the genetic response elements that bind these factors are not.

More recently, we created transgenic mice carrying fusion genes containing 246, 201 and 125 bp of exon I.1 5′-flanking sequence fused either to a mutated (hGX) or wild-type (hGH) human growth hormone reporter gene (Fig. 1). We found that little as 246 bp of hCYP19 exon I.1 5′-flanking sequence was sufficient to direct placenta-specific expression of hGX or hGH in transgenic mice (Fig. 2). By contrast, transgenes containing 201 bp or 125 bp of exon I.1 5′-flanking DNA were not expressed in mouse placenta [6] (Fig. 2). Furthermore, hCYP19I.1_-246:hGX transgene expression was developmentally regulated; expression was observed as early as embryonic day (E) 7.5 in several cells of the trophoblast ectoderm, at E8.5 in some trophoblast giant cells and by E9.5 throughout the giant cell layer (Fig. 3). Very low levels of hCYP19I.1_-246:hGX transgene expression were detected at E9.5 in the primitive labyrinthine layer. However, by E10.5, relatively high levels of hCYP19I.1_-246:hGX transgene expression were observed both in the well-vascularized labyrinthine and in the trophoblast giant cell layers (Fig. 3). By contrast, expression of the hCYP19I.1_-501:hGH transgene was first observed at E10.5 and restricted to the labyrinthine layer (Fig. 3). This suggests the presence of regulatory elements between -501 and -246 bp that may bind inhibitory transcription factors expressed in giant cells.

Both the -501 bp- and -246 bp-containing transgenes were highly expressed in the labyrinthine trophoblast at E10.5. At this stage, the embryonic vasculature has invaded and branched extensively into the labyrinth to facilitate efficient transport of nutrients and oxygen to the embryo [7,8]. This indicates that the transcription factors required for activation of these transgenes are expressed in the labyrinth and raises the possibility that O₂ may play a permissive role in their expression. In studies using human trophoblast cells in culture, we observed that syncytiotrophoblast differentiation and induction of hCYP19 gene expression were prevented when the cells were cultured under hypoxic (2% O₂) conditions [4]. This suggests the presence of response elements within the 501 bp region that bind hypoxia/O₂-regulated transcription factors, which in turn control hCYP19I.1 promoter activity. We observed that increased expression of the basic-helix-loop-helix transcription factor mammalian achaete scute homologous protein-2 (Mash-2) under hypoxic conditions prevented induction of hCYP19 gene expression in cultured human trophoblast cells [4]. The inhibitory effect of Mash-2 appears to be mediated directly by increased binding of upstream stimulatory factors (USFs) 1 and 2 as heterodimers to E-boxes within the 5′-flanking region (-58 bp) and first exon (+26 bp) of the hCYP19 gene [9]. In mouse placenta, Mash-2 is expressed in the spongiotrophoblast and labyrinthine layers [7,10]. Interestingly, Mash-2 expression decreases at E10.5 [11], a time that coincides with increased vascularization of the labyrinth and induced expression of the 501 bp- and 246 bp-containing transgenes. Therefore, elevated Mash-2 under the hypoxic conditions existing in mouse placenta prior to E10.5 could possibly inhibit hCYP19I.1_-501:hGH transgene expression by preventing binding of stimulatory transcription factor(s).

To further define potential stimulatory and inhibitory sequences within the 501-bp exon I.1 5′-flanking region, we carried out transfection experiments using human trophoblast cells in primary culture. In previous cell transfection studies, we observed that hCYP19I.1_-246:hGH fusion genes were highly expressed in human syncytiotrophoblast but also were expressed in human lung and kidney cell lines. By contrast, expression of -501 bp-containing fusion genes was syncytiotrophoblast-specific [3]. To further define the region between -246 bp and -501 bp upstream of exon I.1 that may be involved in labyrinth/syncytiotrophoblast-specific hCYP19 gene expression, hCYP19I.1:hGH fusion genes comprised of -246, -300, -350, -400 or -501 bp of exon I.1 5′-flanking DNA were incorporated into recombinant replication defective adenoviral viral particles and introduced into freshly isolated human trophoblast cells and into A549 lung adenocarcinoma cells by infection (multiplicity of infection = 0.5). Fusion gene expression was analyzed three days after infection (when most of the primary trophoblast cells had fused to form syncytia) by assay of levels of hGH secreted into the medium over a 24 h period of culture. As observed previously [3], fusion genes containing 246 bp of 5′-flanking sequence were highly expressed in the human syncytiotrophoblast cells and in lung A549 cells (Fig. 4). On the other hand, expression of the hCYP19I.1_-300:hGH fusion gene was essentially undetectable, suggesting the presence of transcriptional repressors in both cell types that bind to elements between -246 bp and -300 bp. In the human placental cells, expression levels of fusion genes containing 350-, 400- and 501-bp of exon I.1 5′-flanking DNA were increased as compared to the 300 bp construct, while in the lung A549 cells, expression of these longer fusion genes was barely detectable (Fig. 4). This suggests that the repression in the trophoblast cells was overcome, in part, by syncytiotrophoblast-specific enhancers binding to the region between -300 and -501 bp. It is possible that analogous repressors also prevent expression of the 501 bp transgene in the trophoblast giant cell layer in transgenic mice.

3. Use of transgenic mice and transfected cells to define the response elements and transcription factors that mediate ovary-specific CYP19 gene expression

In studies using transgenic mice to define the regions of the hCYP19 gene involved in ovary-specific expression, we observed that as little as 278 bp of DNA flanking the 5′-end of ovary-specific hCYP19 exon IIa was sufficient to _target ovary-specific expression [12]. This region is highly conserved in the CYP19 genes of rodents and humans and contains cis-acting elements crucial for cAMP induction of hCYP19 promoter activity of human and rat CYP19 genes, including a cAMP response element (CRE)-like sequence (CLS), which binds the CRE-binding protein (CREB) transcription factor [13,14], a binding site for GATA-4 [15,16], and a putative nuclear receptor response element that was suggested to bind the orphan nuclear receptor steroidogenic factor 1 (SF-1/NR5A1) [17,18], which is essential for development of the ovaries, testes, adrenals and ventromedial nucleus of the hypothalamus in mice [19] and plays an important role in the regulation of various members of the cytochrome P450 family of steroid hydroxylases. Within the ovary, SF-1 mRNA and protein are expressed in the granulosa and theca cells of the follicle, the interstitial region and in cells of the corpus luteum. The highest levels of SF-1 expression, however, are found in the theca and interstitium of the ovary [20,21]. Based on these findings, SF-1 has been thought to serve a role in regulation of estrogen biosynthesis in the follicle.

It has been observed that the orphan nuclear receptor, closely related to SF-1, LRH-1/NR5A2, which is highly expressed in liver, pancreas, colon and intestine [22], also is expressed at elevated levels in the ovary [23-25]. In the digestive tract, LRH-1 plays an important role in the control of bile acid synthesis, through regulation of expression of a number of rate-limiting enzymes in this pathway, such as cholesterol-7α-hydroxylase [26-28] and sterol 12α-hydroxylase [29] (products of the CYP7A1 and CYP8B1 genes, respectively). LRH-1 is most closely related to SF-1 in its DNA-binding domain; thus, both transcription factors can potentially bind to the same hexameric response element in DNA [30]. We also have identified a second putative nuclear receptor binding site within the 278-bp 5′-flanking region of the hCYP19 gene [31]. In studies using electrophoretic mobility shift assays and granulosa cell transfection, we observed that both response elements were functionally required for cAMP induction of hCYP19 promoter IIa activity [31]. It also was found that expressed SF-1 and LRH-1 had an equivalent capacity to bind to these response elements and to markedly stimulate hCYP19 promoter IIa activity [31].

4. Use of in situ hybridization to evaluate the potential roles of SF-1 and LRH-1 in gonadal development and in the regulation of ovarian steroidogenesis

To evaluate their potential roles in gonadal development and in ovarian steroidogenesis during early postnatal development and in the post-pubertal and pregnant ovary, temporal and spatial regulation of LRH-1 and SF-1 mRNA in the rodent ovary during pre- and postnatal development [32] and during the estrous cycle [31,33,34] and pregnancy [31,34] have been compared.

5. Conclusions

The hCYP19 gene is controlled by tissue-specific promoters that lie upstream of tissue-specific first exons. Using transgenic mice, we observed that 501 bp of hCYP19 exon I.1 5′-flanking DNA, which lies ∼100,000 bp upstream of exon II containing the hCYP19 translation initiation site, is sufficient to mediate placenta-specific expression. More recent deletion mapping studies in transgenic mice revealed that response elements critical for directing placenta-specific expression are contained within the genomic region between 201 and 246 bp upstream of hCYP19 exon I.1. Whereas, the 501-bp-containing transgene was expressed specifically in the labyrinthine trophoblast, which is structurally and functionally analogous to the human syncytiotrophoblast, the 246-bp-containing transgene was expressed both in the labyrinth and in giant cells. The differences in temporal and spatial expression of the -246 bp- and -501 bp-containing transgenes suggest the presence of silencer elements between -246 and -501 bp that prevent activation of hCYP19 promoter I.1 in giant cells. These findings from transgenic experiments, together with deletion mapping studies using transfected human placental cells, indicate that the concerted interaction of strong placenta-specific enhancers and silencers within the 501 bp region mediate labyrinthine and syncytiotrophoblast-specific CYP19 gene expression.

By use of in situ hybridization studies to compare and contrast expression of the related orphan nuclear receptors, LRH and SF-1, with two cell-specific steroidogenic enzymes, cytochromes P450 aromatase and P450 17α-hydroxylase/17,20 lyase, we provide evidence for a potential role for LRH-1 in gonadal development, the initiation of folliculogenesis and regulation of estrogen biosynthesis within the ovary.