This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jorgensen, J. S. Articles by Nilson, J. H. Search for Related Content PubMed PubMed Citation Articles by Jorgensen, J. S. Articles by Nilson, J. H.

Multiple and Overlapping Combinatorial Codes Orchestrate Hormonal Responsiveness and Dictate Cell-Specific Expression of the Genes Encoding Luteinizing Hormone

Department of Veterinary Biosciences (J.S.J.), University of Illinois, Urbana, Illinois 61802; Medical Sciences Program (C.C.Q.), Indiana University, Bloomington, Indiana 47405; and School of Molecular Biosciences (J.H.N.), Washington State University, Pullman, Washington 99164-4660

Correspondence: Address all correspondence and requests for reprints to: John H. Nilson, Ph.D., School of Molecular Biosciences, 639 Fulmer Hall, Washington State University, Pullman, Washington 99164-4660. E-mail: jhn{at}wsu.edu

	Abstract

Top Abstract I. Introduction II. Cell-Specific and Hormonal... III. Cell-Specific and Hormonal... IV. Conclusions and Future... References

 
Normal reproductive function in mammals requires precise control of LH synthesis and secretion by gonadotropes of the anterior pituitary. Synthesis of LH requires expression of two genes [-glycoprotein subunit (GSU) and LHß] located on different chromosomes. Hormones from the hypothalamus and gonads modulate transcription of both genes as well as secretion of the biologically active LH heterodimer. In males and females, the transcriptional tone of the genes encoding GSU and LHß reflects dynamic integration of a positive signal provided by GnRH from hypothalamic neurons and negative signals emanating from gonadal steroids. Although GSU and LHß genes respond transcriptionally in the same manner to changes in hormonal input, different combinations of regulatory elements orchestrate their response. These hormone-responsive regulatory elements are also integral members of much larger combinatorial codes responsible for targeting expression of GSU and LHß genes to gonadotropes. In this review, we will profile the genomic landscape of the promoter-regulatory region of both genes, depicting elements and factors that contribute to gonadotrope-specific expression and hormonal regulation. Within this context, we will highlight the different combinatorial codes that control transcriptional responses, particularly those that mediate the opposing effects of GnRH and one of the sex steroids, androgens. We will use this framework to suggest that GnRH and androgens attain the same transcriptional endpoint through combinatorial codes unique to GSU and LHß. This parallelism permits the dynamic and coordinate regulation of two genes that encode a single hormone.

I. Introduction

A. LH, one hormone encoded by two genes

B. Hypothalamic and gonadal hormones regulate synthesis and secretion of LH

C. Transcriptional regulation of LH: a critical foundation that underlies regulated LH secretion

II. Cell-Specific and Hormonal Regulation: The GSU Genomic Landscape

A. Regulatory elements that confer cell-specific expression

B. Regulatory elements that confer hormone responsiveness

C. Convergence of basal and hormonal regulation of GSU expression

III. Cell-Specific and Hormonal Regulation: The LHß Genomic Landscape

A. Regulatory elements that confer cell-specific expression

B. Regulatory elements that confer hormone responsiveness

C. Convergence of basal and hormonal regulation of LHß expression

IV. Conclusions and Future Directions

A. Conclusions

B. Future directions

	I. Introduction

Top Abstract I. Introduction II. Cell-Specific and Hormonal... III. Cell-Specific and Hormonal... IV. Conclusions and Future... References

 
REPRODUCTIVE VIABILITY IN males and females depends on coordinated synthesis and release of several hormones from the hypothalamic-pituitary-gonadal axis. Here we focus on LH, a pituitary-derived glycoprotein hormone that belongs to a family of hormones that includes FSH, TSH, and chorionic gonadotropin (CG) (1, 2). LH and FSH are produced in gonadotropes, whereas TSH is synthesized in thyrotropes of mammalian pituitaries. Production of CG in placental trophoblasts is well described only in equids and primates (1, 3).

A. LH, one hormone encoded by two genes
Like other glycoprotein hormones, LH is heterodimeric, consisting of an -glycoprotein subunit (GSU) common to all family members and a unique ß-subunit that associates noncovalently. Biological activity of LH occurs through selective binding of the heterodimer to the LH receptor, a member of the G protein-coupled family of receptors. Binding specificity of LH, as with all other members of the glycoprotein family, is conferred by the ß-subunit (1).

The chromosomal location, gene organization, and structural features of the - and LHß-subunits have been characterized and reviewed extensively (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12). With the exception of teleosts, the genome of vertebrates harbors a single copy of the -subunit gene. The LHß gene resides on a different chromosome and in all vertebrates is also present as a single copy. In primates, however, the closely related CGß gene cluster has arisen through a series of gene duplications of the single-copy LHß gene. The number of CGß genes ranges from 1 to 7, depending on the primate specie (13, 14, 15, 16, 17, 18, 19). After transcription and translation, the - and LHß-subunits are glycosylated and assembled noncovalently as they transit from the endoplasmic reticulum and Golgi apparatus to secretory granules where the heterodimers are stored (3). In short, synthesis of biologically active LH requires the coordinated transcription of two genes and subsequent posttranslational modification and noncovalent assembly of the two subunits.

B. Hypothalamic and gonadal hormones regulate synthesis and secretion of LH
Appropriate circulating levels of LH transpire as the result of three stages of production that are each intricately regulated: transcription, synthesis, and secretion. This is achieved through opposing positive and negative signals provided primarily by GnRH, synthesized and secreted from hypothalamic neurons, and sex steroids produced by the gonads. As illustrated below, changes in the balance of these hormonal contributions can result in production of too little or too much LH, either of which can have dire reproductive consequences.

Hypersecretion of LH in women has been implicated as a cause of infertility and miscarriages (20, 21, 22, 23). Whereas female reproductive disorders are part of a complex etiology, elevated LH is a feature of such conditions as polycystic ovarian syndrome (24, 25) or accelerated follicle exhaustion in perimenopausal women (26, 27, 28). Because it has been difficult to specifically implicate LH in female reproductive disorders, transgenic mice that chronically hypersecrete LH (LH-CTP) were developed (29). Whereas the males are overtly normal despite a 5-fold increase in LH (J. S. Jorgensen and J. H. Nilson, unpublished results), the females display a number of abnormal phenotypes. These include elevated serum LH, androgen, and estrogen concentrations, precocious puberty, polycystic ovaries, and anovulation leading to infertility (29, 30). Although a superovulation paradigm can induce ovulation in these mice, they fail to maintain pregnancy due to maternal defects leading to problems with embryo implantation (31). Finally, the chronically hyperstimulated ovary in LH-CTP mice leads to hormone-responsive cancer including granulosa cell tumors (29), pituitary adenomas (32, 33, 34), and mammary tumors (35, 36).

Stunted LH production or inactive LH culminates in infertility largely due to hypogonadism or defects in germ cell development. A lack of LH due to a primary pituitary defect is rare, and only one human with an inactivating mutation in the LHß gene has been reported (37). Alternatively, secondary causes of LH failure have been reported more frequently and can be caused by a hypothalamic defect that results in GnRH deficiency (reviewed in Ref.38), or an inactivating mutation in the LH receptor (reviewed in Ref.39). The cumulative effect of these studies underscores the need for precise control of all steps of LH production.

1. GnRH-mediated stimulation of LH.
Several animal models demonstrate that appropriate secretion of hypothalamic GnRH is an essential determinant of successful synthesis and secretion of LH. This has been unequivocally illustrated with the discovery of the hypogonadal (hpg) mouse, in which an autosomal recessive mutation in the GnRH gene resulted in hypogonadotropic hypogonadism (40, 41, 42, 43). Daily sc injections of GnRH or restoration of the GnRH gene rescued gonadotropin expression and ultimately reversed the phenotype (42, 43). Studies with hypothalamic-pituitary-disconnected (HPD) animals were among the earliest suggesting that GnRH regulates transcription of the genes encoding the subunits of LH. HPD ewes generated less than 10% of their gonadotropin mRNAs compared with intact animals. A simple treatment of these HPD animals with exogenous GnRH reversed the decline in gonadotropin subunit gene expression (44, 45, 46). In another study, castrated and testosterone-replaced rats treated with pulses of GnRH were used to determine that rapid pulse frequencies (every 8–30 min) stimulated an increase in pituitary content of both GSU and LHß mRNAs, whereas FSHß mRNA synthesis responded to slow (every 120–480 min) GnRH pulses (47, 48). In addition, the castrated and testosterone-replaced rats responded to GnRH with alterations in both GnRH-receptor (GnRH-R) and LHß-subunit mRNA that were subsequently used to show a correlation between GnRH hormone and LHß mRNA (49). The impact of GnRH on important regions of the regulatory sequences of GSU and LHß genes will be covered extensively in this review.

2. Sex steroid-mediated suppression of LH.
Scores of historical evidence have led to the acceptance that gonadal steroid hormones regulate serum LH concentrations and pituitary content of the heterodimer subunits, GSU and LHß (4, 50, 51, 52, 53, 54, 55, 56, 57). Methods such as cell-free translation assays (58, 59, 60, 61), Northern or RNA blot analysis (58, 59, 60, 61), and nuclear runoff assays from pituitary tissue (4, 62, 63, 64, 65) have illustrated that GSU and LHß-subunit mRNAs are increased after castration and subsequently decreased upon treatment with gonadal steroid hormones in various animal models. Since it was recognized that an increase in sex steroids will suppress LH production, several groups have attempted to determine how much of this repression is indirectly caused by altering GnRH production or is directly affecting LH subunit genes at the level of the pituitary.

Several studies reported a lack of sex steroid accumulation, or estrogen receptor (ER) or androgen receptor (AR) localization to GnRH-producing neurons in the hypothalamus (66, 67, 68, 69). Alternatively, both ER and ERß subtypes, along with AR, were localized to several neighboring hypothalamic neurons that also expressed regulatory neurotransmitters, -aminobutyric acid and somatostatin (67, 70). These findings led researchers to predict that sex steroid-negative feedback on GnRH production, and thus LH production, occurred indirectly through synaptic extensions from neighboring neurons (67).

Recently, however, exquisitely sensitive techniques, such as single-cell multiplex RT-PCR, have identified living GnRH-producing cells that also express mRNA for both ER and ERß (71). Further studies are required to similarly identify AR. In addition, GT1–7 cells, a model of GnRH-secreting neurons, were also found to express AR and both ER subtypes (72, 73, 74, 75). In fact, GT1–7 cells treated with estrogen resulted in suppression of GnRH mRNA presumably through binding ER (74). Thus, this recent work unmasks a mechanism for direct regulation of GnRH synthesis in GnRH-producing neurons, at least by estrogen.

In contrast, localization of steroid receptors to the gonadotrope has proved to be relatively straightforward. ER (76, 77, 78), ERß (79, 80), and AR (81, 82) have been localized specifically to gonadotropes and in gonadotrope-derived cell lines, T3–1 and LßT2 (83), suggesting that direct regulation by sex steroids can also occur in these cells of the pituitary.

3. Determining the estrogen site of action.
The precise location of estrogen-negative feedback appears to vary depending on the species studied and, even within species, is controversial. HPD ewes were studied to determine the extent of estrogen-negative feedback that is exclusive to the pituitary (84, 85). Serum LH decreased in HPD ewes, and then increased as a result of ovariectomy or the addition of hourly (85) or every other hour (84) pulses of GnRH. In contrast, studies in hpg mice, which are deficient in GnRH, suggest that estrogen may also work through the hypothalamus (40, 41, 42, 43). For example, although daily injections of estrogen blocked the stimulatory effect of daily doses of GnRH on pituitary gonadotropin content in female hpg mice, suggesting a pituitary effect, estrogen was unable to exert an effect caused by GnRH pulses given every other hour (86). In rats, one study also suggested that estrogen-negative feedback effects were mediated directly at the level of the pituitary (87). Assays were performed on ovariectomized rats that were also hypophysectomized. Harvested pituitaries were transplanted into the kidney capsule and stimulated by hourly infusions of exogenous GnRH. Whereas GnRH treatment stimulated an increase in serum LH concentrations, the addition of estradiol suppressed LH release by approximately 60% (87). In contrast, in vitro work performed with primary rat pituitary cell cultures determined that estrogen-negative effects on LH production were not likely mediated by the pituitary (65, 88). Both studies resulted in an increase in either GSU (88) or LHß (65) mRNAs upon the addition of estrogen, indicating a role for estrogen in the pituitary for stimulating LH production.

Finally, human and primate studies indicated a role for both the pituitary and hypothalamus in mediating estrogen’s suppressive effect (88, 89, 90). Monkey pituitary cell cultures perifused with GnRH and estrogen results first in an increase, and then a decrease of LH secretion into the media, supporting a role for the pituitary in mediating estrogen-negative feedback on LH production (88). To deduce the role of estrogen-negative feedback effects on the pituitary in men, hormonal profiles from normal men were compared with men with complete GnRH deficiency (89). Men with isolated GnRH deficiency exhibit normal pituitary gonadotropin and gonadal steroid secretion during long-term pulsatile GnRH replacement (89, 91, 92, 93, 94, 95, 96). Supraphysiological doses of estrogen suppressed LH secretion in GnRH-deficient and normal men to a similar degree, supporting previous work demonstrating estrogen-suppressive effects directly at the level of the pituitary in women and ovariectomized monkeys (97, 98). Whereas the preponderance of data supported a pituitary site of action for estrogen-negative feedback in men, later studies that used a highly selective aromatase inhibitor, anastrozole, in normal and GnRH-deficient men highlighted significant differences in the degree of estrogen-mediated suppression between these groups (90). These data demonstrated that the hypothalamus must be included when determining the mechanism of estrogen-negative feedback.

4. Determining the androgen site of action.
Androgens are also a source of negative feedback regulation for LH. Similar to estrogen, the precise location that mediates androgen-negative feedback varies between species but has not been studied to the same extent. The majority of data from rodent models point to the pituitary for direct androgen regulation of LH (88, 99, 100, 101, 102, 103). Androgens suppress levels of GSU and LHß mRNA in rat pituitary cells devoid of hypothalamic input (88, 99, 100, 101, 102). In addition, rat pituitary cultures perifused with GnRH and testosterone secreted a significantly decreased amount of LH into media compared with GnRH treatment alone (88).

Studies in nonhuman primates suggest that testosterone has little effect on the pituitary (88, 104). Plant (104) found that testosterone administration in primates retards the frequency of the GnRH pulse generator. Comparative studies between normal men and men deficient in GnRH determined that testosterone decreased mean LH concentrations as well as LH pulse amplitude (105). However, concomitant administration of an aromatase inhibitor and testosterone prevented LH suppression in both normal and GnRH-deficient men, indicating that a conversion of testosterone to estrogen is necessary (105, 106). Testosterone itself had a minimal direct effect at the level of the pituitary (106).

In summary, research in several animal models, including humans, documents the difficulty in isolating suppressive effects of sex steroids on LH production to either the pituitary or the hypothalamus exclusively. Problems appear to stem from the use of supraphysiological dosages of sex steroids, administering pharmacological agents that are later found to exhibit unwanted secondary effects, and the challenge of mimicking a true GnRH pulse generator. Obviously, it is extremely difficult to single out specific components that are a part of a highly complex regulatory system. Thus, we predict that these controversial findings demonstrate that both the pituitary and the hypothalamus participate in coordinating the negative feedback effect on LH by both sex steroids. For this review, we will focus on the suppressive effects of androgen on GSU and LHß promoter sequences because there is direct evidence for its action in gonadotropes of the pituitary.

5. Hormonal regulation of synthesis and secretion.
Posttranslational modification and secretion of LH are also controlled by several neural and hormonal mediators. For example, gonadotropin glycosylation is a highly complex process that includes sialylation and sulfation of the oligosaccharide attachments. Manipulation of these N- and O-linked oligosaccharide chains is essential for intracellular folding, assembly, secretion, metabolic clearance, and biological potency of LH (107). Release of LH from the pituitary occurs as a pulse or a surge, depending on the stage of the estrous cycle and the hormonal milieu. Regulation of this differential secretion depends on several factors including hormonal stimulation or inhibition, intrapituitary paracrine controls, and packaging into LH-specific secretory granules (for reviews see Refs.4, 11 , and 108, 109, 110, 111, 112, 113). Indeed, LH secretion is the final critical step in the process of controlling appropriate serum hormone levels. Here, we focus on the transcription of the two subunits that make up the hormone itself, although details of LH packaging and release are worthy of a review in their own right.

C. Transcriptional regulation of LH: a critical foundation that underlies regulated LH secretion
The ability to maintain pulsatile secretion of LH over the lifetime of a mammal requires constant replacement of gonadotropin liberated from secretory granules. Understanding the transcriptional regulation of LH is complicated because its synthesis requires the expression of two genes located on different chromosomes. Moreover, synthesis of the complete spectrum of glycoprotein hormones requires targeted expression of GSU in three cell types: gonadotropes and thyrotropes in the pituitary and trophoblasts in the placenta. As might be expected, achieving this spatial pattern of gene expression requires an extensive array of regulatory elements with subsets acting as cell-specific combinatorial codes. Although expression of LHß occurs only in gonadotropes, arrays of regulatory elements are still used to form a cell-specific combinatorial code. Because hormones generally modulate expression of transcriptionally active genes, it follows that some of the regulatory elements that form cell-specific codes also act as endocrine-specific targets. The goal of this review is to develop a genomic picture of the critical transcriptional components that both target expression of GSU and LHß to gonadotropes and render them responsive to regulation by GnRH and androgens. Because both genes respond transcriptionally in the same manner to either hormone, it is important to determine the nature of the codes and the underlying mechanism that leads to this coordinated response. Recent progress now permits us to place transcriptional control in a more specific genomic context and allows us to propose global mechanisms of regulation that are common to both genes.

	II. Cell-Specific and Hormonal Regulation: The GSU Genomic Landscape

Top Abstract I. Introduction II. Cell-Specific and Hormonal... III. Cell-Specific and Hormonal... IV. Conclusions and Future... References

 
Steadily emerging data suggest that inputs for gonadotrope-specific expression come from promoter-regulatory elements that contribute to basal activity, with some also mediating hormone responsiveness. We will begin by describing the elements and factors necessary for gonadotrope-specific expression, highlighting those that play critical roles in mediating hormone responsiveness, and then predict the mechanisms whereby GnRH and steroids regulate transcriptional homeostasis for both LH subunit genes.

A. Regulatory elements that confer cell-specific expression
As the common subunit of all glycoprotein hormones, GSU expression is required in gonadotropes, thyrotropes, and trophoblasts. As summarized in Fig. 1A and Table 1, there are at least 15 regulatory elements that bind an even greater number of transcription factors in the 5'-flanking region of the human and mouse GSU (HGSU, mGSU) promoters. Distinct sets of these regulatory sequences comprise codes that direct GSU expression to specific cell types. Targeting sequences mediate responses to transactivators that may or may not be cell type specific.

View larger version (30K): [in this window] [in a new window]   FIG. 1. Several elements define GSU expression in a cell-specific manner. A, Specific promoter elements along with their cognate binding factors are responsible for GSU gene expression. Human and rodent sequences are similar but have unique properties that define both basal and GnRH-responsive expression. Elements that bind shaded factors contribute to gonadotrope-specific expression and include the E boxes, Pitx1 binding element, CRE, ACT, GSE, BE1 and -2 (human specific), PGBE, GnRH-RE, and distal enhancer sequences (both murine specific). B, The HGSU promoter depends on synergistic participation from factors that bind the tandem CREs, BE1 and -2, and PGBE. These studies imply that the factors that bind these elements interact with each other, potentially through an adapter complex.

View this table: [in this window] [in a new window]   TABLE 1. Elements and transcription factors in the GSU promoter region

Thyrotrope-specific GSU activity is conferred by a non-steroidogenic factor 1 (SF-1)-binding gonadotrope-specific element (GSE) (124, 125), Msx-1 binding element (126), and thyrotrope-specific element (122, 127, 128). In addition, nonspecific elements, such as the proximal E boxes (129, 130), pituitary homeobox 1 (Pitx1)-responsive element (131), CRE (125), ACT, pituitary glycoprotein hormone basal element (PGBE) (122, 124, 132, 133), GnRH-response element (GnRH-RE) (128, 134), and mouse distal enhancer (135, 136), also contribute to GSU activity in the thyrotrope.

2. Targeting to gonadotropes.
The fundamental elements and factors in the GSU promoter that are necessary for gonadotrope-specific activity are strikingly similar in human and mouse. The most important contributors to HGSU promoter activity in gonadotropes are the CREs (137). In fact, a number of proteins can bind the palindromic pair of CREs present in the HGSU promoter; these include CRE binding protein (CREB), CRE modulator, c-Jun, activating transcription factor (ATF)1, and ATF2 (138). The GSU gene promoter regions from all other mammalian species harbor a single CRE with a sequence that differs by one base from the human sequence (TGATGTCA vs. TGACGTCA) and binds heterodimers of c-Jun/ATF2 with much higher affinity than CREB (138, 139). Interestingly, the mouse CRE is unique in that the alleles are heterozygous; one allele contains the sequence corresponding to the human element whereas the second allele is made up of the variant sequence (Fig. 1A; only the variant allele shown in the mouse sequence) (122). Whereas an GSU promoter harboring the human pair of palindromic CREs targeted gene expression to both gonadotrope- and trophoblast-derived cell lines, a construct engineered with the single variant CRE sequence was expressed only in gonadotrope-derived cells (120). In addition, HGSU promoter activity was suppressed by c-Jun in placenta-derived cells (140), but was slightly activated in pituitary-derived cells; in contrast, pituitary activity of GSU was suppressed upon overexpression of CREB (141). Together, these data support the notion that regulatory sequences and their cognate binding factors can determine cell specificity and suggest that c-Jun/ATF2 heterodimers are the preferred binding partners for CRE sequences in the gonadotrope.

Three neighboring elements in the distal 5'-regulatory region, PGBE and -basal element 1 and 2 (BE1 and -2), are also essential for basal expression of the GSU promoter in the pituitary (132, 133, 137, 142). Whereas LH2 was originally presented as the trans-acting partner for PGBE in the mouse, and therefore postulated to extend to the human counterpart (122, 132, 134, 137), additional studies suggest that another member of the LIM homeodomain family, Lhx3, may be the functional binding partner (143, 144). Lhx3 knockout mice fail to develop the anterior and intermediate lobes of the pituitary gland (143). In addition, cells from a gonadotrope-derived cell line, T3–1, that were stably transfected with an antisense Pitx1 sequence, and were thus deficient in Pitx1, lost both Lhx3 and GSU gene expression, whereas transcription of LH2 remained unchanged (144). Furthermore, recent studies determined that a point mutation in the LIM domain of Lhx3 reduced activation of a modified GSU promoter (145). Gel mobility-shift experiments indicated that two factors, BP1 and -2, bind the BE region that neighbors PGBE (137). Southwestern blot analysis identified BP1 as 54- and 56-kDa bands that may represent two proteins, or two forms, of the same protein. Attempts to identify BP2 by Southwestern blot analysis were unsuccessful, suggesting that this protein may be heterodimeric. These studies also suggested that there may be interactions between proteins binding PGBE and BE, resulting in synergistic effects on GSU transcription (137).

SF-1, an orphan nuclear receptor that binds the GSE, was originally described as a defining factor for pituitary-specific expression of GSU (142, 146, 147, 148, 149). Mutational analysis of the HGSU promoter indicated a requirement for the GSE that does not depend on any other sequences but contributes independently to transcription (137). Yet, as will be discussed later, the contribution of SF-1 is undoubtedly much more significant to LHß expression than to GSU promoter activity.

Other sequences are involved in basal transcription of the GSU gene in the gonadotrope but appear to provide only supportive or minor contributions to its expression (137, 142). Two elements, trophoblast-specific element (TSE) and ACT, together make up the upstream regulatory element and have been extensively characterized for their importance in placental expression (120, 150, 151). Whereas TSE binding proteins such as AP2 have been isolated from placental cells (152), there is evidence that proteins that bind ACT, such as GATA2 or GATA3, play a role in gonadotrope GSU expression (153). Finally, loss of Pitx1 (144), or basic helix-loop-helix protein expression (129), resulted in suppressed GSU gene activity in T3–1 cell lines, indicating they may play a role in binding these sequences to promote gene expression.

3. Synergistic interactions determining GSU promoter activity.
The above data indicate that GSU gene expression in the gonadotrope is not controlled exclusively by a single dominant element but rather by an array of weaker elements. This suggests an intricate interplay between these regulatory elements and their cognate binding proteins. Studies of the HGSU promoter that used mutations in multiple elements revealed that the transcription factors binding PGBE, BE1 and -2, and the CREs influence each other’s activities. In contrast, mutations in the GSE had the same relative impact on promoter activity in the presence or absence of these factors, suggesting its autonomy (137). In summary, these data suggest that the activation pathways of PGBE, BEs, and CREs converge to stimulate transcription and illuminate the possibility that the proteins binding these elements share a common coactivator to synergistically elevate GSU gene expression (Fig. 1B) (137).

Similar results were obtained when multiple block mutations of the mouse GSU promoter were tested for activity (133). Both PGBE and GSE sequences, in addition to a previously identified distal enhancer region (–4600 to –3700 bp), were necessary for gonadotrope-specific gene expression (133, 135, 136). Sequences homologous to BE are absent in the mouse promoter, and the CRE was not tested. The final outcome, however, mimics that of its human counterpart: GSU promoter expression depends on active participation of a specific combination of elements and their binding partners that link the distal and proximal regions of the promoter, most likely via adaptor complexes. Integral elements include the distal enhancer region, PGBE, BE, and GSE sequences.

As will become apparent, many of the sequences that have been described as necessary for gonadotrope-specific expression of the GSU gene are also critical for transcriptional regulation by GnRH and sex steroids. Subsequent sections will focus on specific mechanisms and pathways that utilize gonadotrope-defining sequences of the GSU promoter to provide moment-to-moment regulation by these hormones.

B. Regulatory elements that confer hormone responsiveness
1. GnRH stimulation of transcription
a. GnRH-responsive sequences.
GnRH responsiveness has been extensively investigated for both mouse and human GSU genes. In the systems employed, the mouse GSU promoter confers very robust GnRH-induced activity compared with other species. A composite element made up of two binding sites mediates the GnRH response in rodent promoters. An upstream region, GnRH-RE, was isolated in the mouse promoter corresponding to –416 to –385 bp (134) and has been duplicated only in the rat GSU promoter (154). The protein that binds this site in vivo has not been established; however, a likely candidate is a member of the Ets family because overexpression of a dominant-negative form blocks the GnRH response (155). A second GnRH-responsive region located at –344 to –300 bp of the murine GSU gene corresponds to PGBE sequences (132, 134) and likely binds a LIM-homeodomain factor such as LH2 (132) or Lhx3 (144). Both the GnRH-RE and PGBE regions are required for GnRH-responsive expression of the mouse GSU gene in vivo (Fig. 1A) (134).

Fifteen hundred base pairs of the proximal HGSU promoter were sufficient to target chloramphenicol acetyl transferase (CAT) reporter activity to gonadotropes and was responsive to GnRH stimulation in transgenic mice (156). Transfection assays were used to localize the GnRH-responsive region of the HGSU promoter to –346 to –244 bp, corresponding to the PGBE and BE elements (142, 157). Whereas LIM homeodomain proteins have been implicated in PGBE binding, specific BE-binding proteins have not yet been fully characterized. Furthermore, a potential Ets binding sequence similar to the GnRH-RE in the mouse promoter has been identified between the PGBE and BE sequences of the human promoter (137). In this regard, recent data indicate that the BE, including the putative Ets sequence, have autonomous GnRH responsiveness when added in multiple copies upstream of a minimal promoter (H. P. Mohammad and J. H. Nilson, unpublished data). Recently, GSE and upstream regulatory element sequences (–180 to –156 bp) were also implicated in a minor role for GnRH responsiveness (142, 158). Based on data obtained from the mouse GSU promoter, it is proposed that the human sequences that include PGBE, BE, and the putative Ets binding site together form a composite GnRH-responsive region (Fig. 1A).

Mutational analyses suggested synergistic interplay between proximal and distal elements of the HGSU promoter, namely the GnRH-responsive region that includes PGBE and BE and the tandem CREs (137). Thus, whereas the tandem CREs themselves do not appear to mediate GnRH responsiveness, they may, together with PGBE and BE, participate in a combinatorial code that provides a communication link between the proximal and distal GSU promoter to mediate GnRH signaling, as depicted in the model in Fig. 2A2. It is plausible that the GnRH-RE sequences may contribute to this code in the mouse promoter; additional studies are necessary to verify this hypothesis. Finally, participants of the combinatorial code likely include additional coactivators and adapter proteins, such as the CREB-binding protein (CBP), that link the proteins that bind promoter sequences and ultimately the RNA polymerase holoenzyme to initiate GSU transcription (Fig. 2, A2 and B2).

View larger version (44K): [in this window] [in a new window]   FIG. 2. AR and GnRH use the same regulatory sequences to provide moment-to-moment regulation of the GSU gene. A, HGSU promoter: 1) AR-mediated suppression of HGSU gene activity requires the tandem CREs and BE sequences. Functional studies indicate that AR interacts with bZip binding proteins, c-Jun and ATF2, to block transcriptional activity. Because AR suppression requires both CRE and BE1 and -2 sequences, it is proposed that AR may also interact with BE binding proteins and disrupt the link between proximal and distal sequences that mediate synergistic transactivational properties. 2) GnRH stimulates the assembly of a transcriptional complex composed of PGBE, Ets binding sequences, and BE1 and -2 linked to the proximal pair of CREs through an adaptor complex. The pulsatile release of GnRH stimulates the production of this synergistic complex to disrupt chronic repression of AR. B, Murine GSU promoter: 1) It is proposed that AR binds the c-Jun/ATF2 heterodimer to block transactivation of the murine GSU promoter using a similar mechanism as predicted with the human sequences. 2) Factors that bind GnRH-responsive sequences including the distal enhancer region, GnRH-RE (Ets binding sequences), and PGBE are linked to the proximal CRE via an adaptor complex to stimulate gene activity and disrupt AR-mediated suppression.

Although reports indicated that MAPK-related factors were mainly involved in basal activity of GSU transcription (160), studies by Roberson et al. (155) suggested that MAPK may be a downstream signaling effector for GSU gene transcription. There is evidence supporting the ability of PKC to activate MAPK kinase in both Ras-dependent and Ras-independent pathways (161). In addition to stimulating MAPK kinase and MAPK activity in T3–1 cells, a constitutively active Raf mutant also increased mouse GSU promoter activity through two regions of the promoter shown to elicit the GnRH response, GnRH-RE and PGBE (155). In addition, GnRH promoted MAPK activity, indicated by phosphorylation of tyrosine residues on ERK1 and -2, and activation of the MAPK-responsive transcription factor, Elk1. GnRH stimulation of mouse GSU promoter activity was attenuated upon cotransfection of kinase-defective ERK or MAPK phosphatases, providing further support for MAPK-mediated signaling. MAPK activity was linked to a specific Ets binding site within the GnRH-RE on the mouse GSU promoter that is likely mediated by an Ets factor (155).

Similarly, the rat GSU promoter was stimulated only by PMA in both rat pituitary cell culture and T3–1 cells (162), whereas addition of a mitogen-activated ERK-activating kinase (MEK) inhibitor eliminated GSU gene expression. In contrast, elimination of intra- or extracellular calcium activity had no effect on the rat GSU promoter (162). PKC also stimulated GSU gene activity in a somatolactotrope cell line that was stably transfected with the GnRH-R, GGH3–1' (163). However, in contrast to Weck et al. (162), this study found that an increase in intracellular calcium stimulated GSU promoter activity to a greater extent than PKC (163). Dihydropyridine, a calcium channel blocker, inhibited GnRH-induced stimulation of GSU.

Examination of GnRH-mediated stimulation of the HGSU promoter indicates that regulation is calcium dependent and augmented by PKC. ERK, but not c-Jun N-terminal kinase (JNK), activates the HGSU promoter ostensibly through sequences corresponding to PGBE and BE1 and -2. Furthermore, an additional MAPK effecter, c-Src (164, 165), was found to increase GSU transcription by affecting sequences that include the GSE (–280/–180 bp) (142). An additional study suggested that GATA mediates ERK activation of HGSU transcription in LßT2 cells (158). Thus, the GnRH-mediated increase in GSU gene expression appears to rely on different downstream signaling mechanisms depending on the test system that is used (Fig. 3).

View larger version (32K): [in this window] [in a new window]   FIG. 3. Regulation of GSU transcription depends on AR and GnRH-R activation. AR, but not ER, directly suppresses GSU promoter expression, whereas GnRH-R activation stimulates multiple signaling cascades to ultimately increase transcription of the GSU subunit gene.

2. Sex steroid negative feedback on GSU.

AR-mediated suppression of the GSU gene occurs in the absence of direct binding of the activated transcription factor to DNA. Instead, the AR interaction maps to the BE/BP and CRE/bZip complexes (166). Several assays were employed to determine the functional interactions between AR and specific bZip proteins that may bind the CREs. Of the proteins tested, it was found that only cotransfected c-Jun and ATF2 could prevent AR-mediated suppression of GSU promoter activity. Cotransfected CREB failed to prevent the suppressive effect of the androgens. In fact, transfection of CREB diminished activity of the GSU promoter on its own (141). In addition, binding assays carried out with a glutathione-S-transferase-AR fusion protein indicated specific interactions of AR with both c-Jun and ATF2, as individual proteins and as a heterodimer; however, no specific interaction with CREB was detected. These selective responses with c-Jun and ATF2 imply that only a limited number of CRE-binding proteins can mediate the suppressive effects of AR when bound to the tandem CREs of the GSU promoter in gonadotropes. In addition, these data suggest that AR suppresses activity of the GSU promoter through direct protein-protein interactions with at least c-Jun and ATF2 (141) (Fig. 2A1). Although this hypothesis has not been tested directly on mouse GSU promoter activity, it is suggested that exposure to androgens causes a similar suppressive effect via protein-protein interactions with the c-Jun/ATF2 heterodimer and additional unknown sequences in the distal 5'-regulatory region (Fig. 2B1). Further analysis of the BE/BP complex in the HGSU promoter is underway to ascertain its involvement in AR-mediated suppression.

C. Convergence of basal and hormonal regulation of GSU expression
At least one element in the 5'-flanking region, BE, is common to both AR-mediated repression and GnRH-mediated stimulation of the HGSU gene. However, close examination of the sequences involved in both hormonal regulatory processes and those important for basal activity suggest a common theme. PGBE, BE, and the tandem CREs link the distal and proximal regions of the HGSU promoter to synergistically increase basal activity, and BE and the tandem CREs also associate with each other during androgen-mediated suppression. In addition, known GnRH-responsive sequences include PGBE, putative Ets binding sequences, and BE1 and -2. Thus, it appears that moment-to-moment regulation of the HGSU gene may be simplified by the reliance on a core comprised of five elements, PGBE, Ets binding sequences, BE1 and -2, and the CREs, as their interactions may link distal and proximal regulatory regions to the RNA polymerase holoenzyme. We suggest that a complex is built on the HGSU promoter based on these regulatory elements. This complex can be stimulated by GnRH signaling to communicate with the transcriptional machinery to initiate transcription or can be repressed by the presence of ligand-bound AR and its protein-protein interactions (Fig. 2A). Similarly, common elements including the CRE and PGBE, in addition to the Ets-containing GnRH-RE, in the mouse GSU promoter provide a combinatorial code remarkably like that of the human sequences. Thus, it is likely that GnRH and androgen signaling converge through these elements to regulate the mouse GSU gene as well (Fig. 2B).

We will now consider the requirements for regulated expression of the gene encoding the unique LHß-subunit. Whereas it is remarkable that similar layers of complexity, including regulatory elements, transcription factors, and adapter proteins, arise in hormonal regulation of the LHß gene, the players in this combinatorial code are entirely distinct.

	III. Cell-Specific and Hormonal Regulation: The LHß Genomic Landscape

Top Abstract I. Introduction II. Cell-Specific and Hormonal... III. Cell-Specific and Hormonal... IV. Conclusions and Future... References

 
Characterization of the LHß-subunit gene has progressed due to the development of several lines of transgenic mice and a gonadotrope-derived cell line, LßT2, that expresses the LHß gene (167, 168, 169, 170, 171). In each species studied thus far, all of the regulatory elements identified as important for LHß promoter activity reside in the proximal 500 bp (Fig. 4 and Table 2). The proximal 140 bp of the LHß promoter are remarkably conserved across species (172), suggesting their functional significance.

View larger version (20K): [in this window] [in a new window]   FIG. 4. Sequences in the proximal 150 bp of the LHß promoter are highly conserved, whereas distal regions diverge depending on species. A, An array of GSE, Egr-1, and Pitx1 sequences is highly conserved across species. Distal elements on the LHß promoter are less characterized; bovine regulatory sequences include a pair of NF-Y binding sequences, whereas a pair of Sp1 sequences and an overlapping CArG element characterizes the rat promoter. B, Synergistic transactivation between the proximal binding elements, GSE, Egr-1, and Pitx1.

Studies that focus on DNA elements of the LHß promoter have illustrated that NF-Y (169), SF-1 (168), and Pitx1 (170) binding sequences are all required for LHß activity in transgenic mice. Each of these elements was mutated in the context of the –779/+10 bovine (b) LHß promoter, and linked to the CAT reporter gene. Compared with wild-type LHß promoter-driven CAT activity, there was little to no expression of CAT in µ5'NF-Y (169), µ5'SF-1 (168), or µPitx1 (170) LHß transgenic mice.

Whereas transgenic studies illuminated the importance of several DNA elements for LHß expression, mice with targeted deletions of the proteins that bind these elements provide evidence that there may be a functional redundancy in these transcription factors. Due to an absence of LHß gene expression, Egr-1 null females are anovulatory and therefore infertile; however, Egr-1 null males have no defects in fertility (173, 178). Egr-4 null males, however, had normal levels of serum LH but were infertile due to defects in maturation of germ cells; the females displayed no fertility defects (179). Whereas the female Egr-1/Egr-4 double-mutant mice exhibited a phenocopy of the Egr-1 null mice, the males had deficient LH production that caused low testosterone and atrophy of androgen-dependent organs. These studies suggest a male-specific redundancy between Egr-1 and Egr-4 in gonadotrope-derived regulation of LH (180). Importantly, consensus sequences for Egr-1 and Egr-4 are not identical, and gel-shift assays indicate specificity for each binding protein (181, 182). Thus, additional Egr-4-binding sequences that have not yet been characterized may be present in the LHß promoter.

SF-1 knockout mice that have been maintained on corticosteroid therapy lacked immunoreactivity for LH, and expression of GSU and LHß mRNA was markedly decreased (149, 183). However, treatment of these mice with GnRH restored pituitary expression of LH (184). Together, these data suggest that SF-1 expression may be imperative for basal, but not GnRH-induced, expression of the LH subunit genes. Alternatively, another protein may bind the GSEs to account for LHß gene expression after rescue with GnRH treatment.

By postpartum d 0, Pitx1 null mice express both GSU and LHß, albeit from a diminished number of cells (185). In contrast, Pitx1 antisense studies performed in T3–1 cells resulted in the absence of GSU and Lim3/Lhx3 expression (144). It thus appears that Pitx1 protein expression may be critical in the early stages of gonadotrope development but later may be compensated for by related proteins.

Taken together, these transfection assays and transgenic or knockout mouse studies illustrated the requirement for specific DNA sequences for successful expression of the LHß gene but suggest that multiple proteins may have the capacity to bind these elements to ensure production of the LHß-subunit.

2. Synergistic interactions determining LHß promoter activity.
Coexpression of Egr-1 and SF-1 results in synergistic activation of the LHß promoter that is completely ablated upon mutation of any of the paired Egr-1 or GSE sequences (172, 176, 186, 187, 188), indicating that this synergistic relationship requires the cognate DNA binding elements for SF-1 and Egr-1. Additionally, binding assays with various Egr-1 mutant constructs determined that the zinc finger domain is necessary for the direct interaction with SF-1 (176). However, addition of another orphan nuclear receptor, Dax1, a repressor of SF-1 activity (189, 190), altered the Egr-1/SF-1 interaction, causing a reduction in their synergistic activation of the LHß promoter (188). Together, these data illustrate the dependence of LHß activity on the presence of both Egr-1 and SF-1.

In addition to interacting with each other, Egr-1 and SF-1 also interact with Pitx1 to form a tripartite protein complex that synergistically activates the LHß promoter (191). This interaction is also completely ablated upon mutation of the Pitx1 binding site in the proximal LHß promoter (170, 172). Many genes encoding hormones require Pitx1 expression and its interaction with a cell type-specific factor for promoter activation. For example, Pitx1 interacts with NeuroD1/Pan1 to induce POMC gene activity (192), or with Pit1 for prolactin and GH gene expression (193). Furthermore, an interaction between Pitx1, SF-1, and Pit1 was found to be essential for regulation of the Lim3/Lhx3 gene (144). Binding assays with a series of truncated Pitx1 mutants isolated a requirement of the C-terminal domain in binding both Egr-1 and SF-1 (172, 191). In addition, the binding of Pitx1 to SF-1 stimulated LHß promoter activity to a similar extent as the constitutively active SF-1 mutant, SF-1LBD, alone (191). These studies suggested that the protein-protein interaction between Pitx1 and SF-1 force a conformational change in SF-1, thereby inducing its activity (191). Together, these data imply that Egr-1, SF-1, and Pitx1 interact directly and cooperatively to stimulate LHß promoter activity (Fig. 4B).

B. Regulatory elements that confer hormone responsiveness
1. GnRH stimulation of transcription
a. GnRH-responsive sequences.
The LHß promoter is exquisitely sensitive to GnRH stimulation. Work performed by several groups has identified GnRH-responsive regions within the 5'-flanking region of the LHß gene, in the proteins that are expressed and interact as a result of GnRH activity, and in the potential signaling pathways that mediate the primary GnRH signal. Like the GSU promoter, there appears to be some species variability in many of these regions.

Two GnRH-responsive regions were identified on the rat LHß promoter, –490 to –353 and –207 to –82 (194). Of these two potential GnRH-regulatory sequences, only deletion of the more distal caused a decrease in GnRH-stimulated activity. Although the rat LHß promoter minus the more proximal sequences between –207 and –82 bp was still responsive to GnRH, it was clear that this region was necessary for GnRH stimulation, as linkage of the distal and proximal sequences increased GnRH-mediated activity from 4- to 10-fold (194). Two Sp1 elements, as well as an overlapping CArG box (154, 177), define the distal domain of the rat LHß promoter that is sensitive to GnRH stimulation. No other species promoter tested thus far, including human, equine, or bovine, duplicates this distal region of GnRH responsiveness (Fig. 4A).

The SF-1, Pitx1, and Egr-1 binding sequences that reside within the proximal 140 bp of the LHß promoter are faithfully conserved among many species (Fig. 4A) (172). This region has been defined in LHß promoters from bovine (168, 170, 172), equine (186, 187), human (195), mouse (195), and rat (175, 176, 188, 194, 196) to be important for mediating the GnRH signal. Mutation of either the 5'-GSE or Pitx1 elements in the bLHß promoter resulted in loss of reporter activity in transgenic mice (170). Furthermore, an ovariectomy-induced elevation in GnRH failed to stimulate these mutant promoters (168). In addition, mutated GSE sequences, in the context of the rat LHß promoter, markedly diminished a 56-fold enhancement in promoter activity resulting from the presence of an SF-1 expression vector in CV-1 cells (175). These data suggest that the GSE, the Pitx1 element, or both, are required for GnRH-induced LHß gene activity (168, 170, 175).

Egr-1 mRNA and protein expression is stimulated by GnRH (172, 187, 188, 196); however, no change in either SF-1 (172, 187, 188) or Pitx1 (172) expression is elicited by this hormone. PMA treatment can mimic GnRH induction of Egr-1 expression (172, 187, 196), but protein kinase A, JNK, or calmodulin kinase are incapable of this activity (172). PMA or GnRH treatment of GGH3–1', T3–1, or LßT2 cells resulted in an increase in Egr-1 mRNA and an increase in protein binding to Egr-1 sequences as measured by gel mobility-shift assays (187, 188, 197). Mutations in either the proximal or distal Egr-1 binding sequences caused a loss of protein binding (188). Furthermore, transfection of an Egr-1 expression vector resulted in an increase in LHß message in LßT2 cells (188). Thus, GnRH induction of Egr-1 appears to be tightly linked to an increase in LHß-subunit synthesis.

GnRH also stimulated the expression of nuclear growth factor 1-A (NGF1A, Egr1)-binding protein (NAB) 1 in T3–1 cells (187). NAB proteins suppress Egr-1 activity by binding to an internal inhibitory domain of 34 amino acids within the Egr-1 protein (198). Wolfe and Call (187) found that Egr-1 induction of the equine LHß promoter in T3–1 cells was diminished upon the addition of a NAB1 expression vector. In contrast, NAB1 and -2 unexpectedly enhanced Egr-1, -2, and -3 stimulation of the rat LHß promoter in CV-1 cells (199). Differences in these studies may be due to differences in the species origin of the promoter or to the origin of cell lines used. To test this hypothesis, a study was conducted within the context of two gonadotrope-derived cell lines, T3–1 and LßT2, to determine whether a commonality existed among the different species of LHß promoters in their use of Egr-1 (195). Nonrodent promoters, including human, bovine, and equine, bind only Egr-1 in the distal Egr-1 sequences, whereas the same sequence in mouse or rat promoters may bind Egr-1, Sp1, or Sp3. In all species tested, Egr-1, Sp1, Sp3, or a zinc finger protein may bind to proximal Egr-1 sequences (195). Together, these results imply that a large number of genes are potential targets of GnRH stimulation in the gonadotrope and must be further investigated.

The expression of Egr-1 in response to GnRH stimulation sparks a cascade of events that culminate in transcription of the LHß-subunit gene. Once Egr-1 protein synthesis is induced, a massive stoichiometric change occurs in the proximal LHß promoter that results in synergistic activation of gene activity. A tripartite GnRH-responsive unit, different from bovine, has been reported for the rat LHß promoter. Using the GGH3–1' cell line, Kaiser et al. (197) found that the GnRH-induced activation of the rat LHß promoter was muted when an Sp1 element in the distal promoter was deleted. Furthermore, combined mutations in Sp1, SF-1, and Egr-1 binding sequences determined that these sites are required to alter GnRH responsiveness. Together, these data suggest that communication exists between the distal and proximal domains of the rat LHß promoter to mediate GnRH responsiveness (172, 176, 187, 188, 197).

With the determination of a communicative link between the proximal and distal LHß promoter, a parallel reference can be made to the combinatorial code that we proposed to mediate the GnRH signal on the GSU promoter. The GnRH signal for LHß requires the distal binding proteins (Sp1 and CArG in the rat and NF-Y in the bovine) to communicate with the tripartite complex made up of Egr-1, SF-1, and Pitx1 in the proximal promoter. Thus, GnRH signaling initiates a series of cascading events causing the appearance of Egr-1; this results in significant restructuring of the proximal and distal sequences of the LHß promoter and brings about synergistic interactions between key factors to promote gene expression (Fig. 5, A2 and B2).

View larger version (55K): [in this window] [in a new window]   FIG. 5. AR and GnRH regulation of the LHß-subunit gene is mediated by the same distal and proximal sequences. A, bLHß promoter: 1) AR binds SF-1 to block communication between Pitx1, SF-1, and the initiation complex and thus suppresses LHß promoter activity. AR can interact with SF-1 in the presence of Pitx1. Without GnRH stimulus, Egr1 is virtually absent. Evidence suggests that additional sequences in the distal promoter may also be necessary for AR suppression. 2) A pulse of GnRH stimulates Egr-1 production to synergistically impact the microenvironment of protein interactions between Egr-1, Pitx1, SF-1, and distal regulatory elements of the LHß promoter, and consequently disrupts the AR/SF-1 binary complex. B, Racine LHß promoter: 1) AR binds to Sp1 and Egr1 to prevent LHß promoter activity. Thus, AR prevents interactions between these factors that are necessary to provide the link between the proximal and distal promoter sequences needed for hormonal regulation. 2) It is proposed that pulsatile release of GnRH disrupts chronic AR-mediated suppression by stimulating synergistic transactivation mediated by the distal Sp1/CArG and GSE/Pitx1/Egr-1 elements.

View larger version (34K): [in this window] [in a new window]   FIG. 6. AR directly suppresses the LHß promoter, whereas GnRH potentially initiates several signaling cascades to activate the gene.

In contrast, Saunders et al. (163) used the somatolactotrope cell line, GGH3–1', in transient transfection assays to test GnRH stimulation of the rat LHß promoter. In this case, PKC was found to activate both GSU and LHß promoter activity, with a larger effect on LHß. In addition, chronic PMA treatment that results in specific depletion of PKC-mediated signaling (204, 205) inhibited LHß gene activity. Finally, treatment with GF 109203X, a specific inhibitor of PKC, caused a larger degree of inhibition of the maximal GnRH response for the LHß promoter than for the GSU gene (163). Treatment with nimodipine had little effect on the LHß gene. Thus, these studies suggest a role for PKC signaling in promoting LHß synthesis rather than calcium.

Using another model system, different results were reported by Yokoi et al. (203). In these studies, PKC signaling was found to mediate ERK, but not JNK, activation by GnRH in LßT2 cells. In addition, a GnRH-induced elevation of intracellular calcium resulted in the activation of ERK, but not JNK, activity. Neither the MEK inhibitor, PD 998059, nor the catalytically inactive MAPK, iMAPK, had any bearing on the ability of GnRH to stimulate the rat LHß promoter in LßT2 cells. Additionally, neither pretreatment with calcium channel blockers nor calcium-chelating agents had an effect on GnRH-induced LHß activity (203). Instead, cotransfection with either a dominant-negative SAPK/JNK or a dominant-negative c-Jun resulted in an attenuated GnRH-induced activation of LHß gene expression. Furthermore, LßT2 cells stably transfected with the dominant-negative c-Jun construct diminished GnRH induction of JNK and LHß but had no effect on GnRH induction of ERK activity (203). Together, these studies indicate that JNK or SAPK mediates GnRH signaling rather than MAPK or ERK.

Recently, several studies have supported previous work and implicated other signaling pathways in promoting a rise in LHß transcripts. GnRH activated ERK, JNK, and p38MAPK in LßT2 cells and induced LHß protein expression through calcium and MEK-dependent mechanisms (201). Studies by Harris et al. (202) supported GnRH-induced rat LHß promoter activity via ERK and JNK mechanisms. Furthermore, experiments with constitutively active or dominant-negative cSrc constructs suggested that this factor participates in LHß stimulation (202). cSrc participates in JNK and, to a lesser extent, in ERK activation by GnRH (164, 165). Additional studies in LßT4 cells suggest that GnRH uses PKC for acute induction of LHß transcripts, whereas calcium influx is responsible for long-term repression of the same gene via different regulatory sequences (200).

Although these studies implicate different signaling pathways in the role of GnRH-mediated activation of the LHß-subunit gene, they also used distinct modeling systems. Together, these data suggest that GnRH stimulates a multitude of complex signaling pathways that participate in an even larger cross-talk network. This network includes transcription factors such as Sp1, SF-1, Egr-1, Pitx1, and probably other as yet unidentified proteins.

2. Sex steroid-negative feedback on LHß.
The proximal promoter (–779/+10 bp) of the bLHß-subunit gene has been shown to be responsive to both estrogen and testosterone in a transgenic mouse model (206). However, it lacks high-affinity binding sites for either ligand-occupied ER or AR (206). It follows that if steroids act at the pituitary to directly suppress activity of the LHß promoter, the mechanism must involve protein-protein interactions between steroid nuclear receptors and specific DNA-binding proteins regulating transcriptional activity of the LHß gene. In contrast, the rat LHß promoter does harbor a putative ERE (–1173 to –1159 bp); however, this site has been associated with stimulation of the gene rather than suppression (207). Studies using transient transfection assays in gonadotrope-derived LßT2 cells determined that ligand-bound AR, but not ER, directly suppressed activity of the bLHß promoter (208) (Fig. 5). Thus, these findings instigated a series of in vitro studies that focused on AR-mediated suppression of the LHß-subunit promoter.

When tested by cotransfection with AR, overexpression of Egr-1, Pitx1, and constitutively active SF-1 missing the ligand-binding domain (SF-1LBD), each individually rescued androgen-mediated suppression of the bLHß promoter. In contrast, overexpression of full-length SF-1 was incapable of relieving the bLHß promoter from the suppressive effect of AR, indicating that the ligand-binding domain plays an important role in functional interactions between SF-1 and AR. Furthermore, these findings suggest a functional interaction between Egr-1, Pitx1, SF-1, and AR (208) and imply that the synergistic interaction of these factors plays a role in androgen-mediated suppression. Binding assays performed with a glutathione-S-transferase-AR-DBD construct further supported this notion; these studies identified SF-1 as a key interactive partner with AR and localized the interaction to the ligand-binding domain of the protein (208). Additional binding studies indicated that whereas AR-DBD did not directly interact with Pitx1 or Egr-1, the addition of both these proteins interfered with formation of a binary complex that contains AR and SF-1.

Androgen, but not estrogen, also suppressed the rat (r) LHß promoter (–617/+44 bp) (209). The major difference between this work and the bovine promoter studies outlined above is the requirement of the distal Sp1 elements for AR-mediated suppression in the rat studies (209). Although the bLHß promoter lacks comparable elements (197), individual deletion of regulatory elements within the proximal region of the bLHß promoter did not result in a loss of AR-mediated suppression. Thus, it was considered that elements in the distal region of the bovine promoter were responsible for the compensation and perhaps even necessary, but not sufficient, for mediating the negative response to androgens (208).

As overexpression of Sp1 cDNA prevented AR-mediated suppression of the rat LHß promoter, binding assays also determined a direct interaction between AR and Sp1. Additionally, an interaction between AR and Egr-1 suggested a contribution of proximal promoter sequences in mediating androgen suppression (interactions with SF-1 or Pitx1 were not tested) (209). Taken together, the bovine and rat studies suggest that AR suppresses LHß transcription via protein-protein interactions with specific DNA binding factors, namely Sp1, Egr-1, and SF-1, that li