Abstract P4-07-06: Cyclin dependent kinase 7 (CDK7) inhibition with THZ1 induces mitotic failure and increases genomic instability in triple negative breast cancer

Triple negative breast cancer (TNBC) is highly proliferative and genomically unstable, making these tumors particularly sensitive to anti-proliferative chemotherapies. While efficacious, these drugs induce dose limiting toxicities. Agents that can selectively target proliferation within cancer cells without inducing systemic toxicity should greatly improve patient outcomes from TNBC. In this regard, we have found that TNBC cells are particularly vulnerable to suppression of the transcriptional regulator, LIN9, that controls cell cycle progression. LIN9 mRNA is overexpressed in 66% of TNBCs and is correlated with worse patient outcomes. Moreover, suppression of LIN9 expression induces multi-nucleation, micronucleation, mitotic catastrophe, and cell death and/or senescence. While transcription factors are generally considered “undruggable, LIN9 expression can be pharmacologically suppressed by blocking the activity of cyclin dependent kinase 7 (CDK7) with the selective, covalent CDK7 inhibitor, THZ1. CDK7 inhibitors inactivate RNA polymerase 2 and destabilize the super-enhancer mediated expression of oncogenes. Treatment of three TNBC cell lines (MDA-MB-231, MDA-MB-468, and HCC38) with THZ1 induces G2/M arrest and phenocopies genetic silencing of LIN9. Use of live cell imaging revealed that THZ1 increases the duration of mitosis and also leads to mitosis-associated cell death. Together, these data reveal that CDK7 inhibitors primarily inhibit TNBC growth by causing mitotic dysfunction and potentiating genomic instability by inducing micro- and multi-nucleation. In addition, they suggest that suppressing LIN9 expression by inhibiting CDK7 may lead to a selective approach for targeting proliferation of TNBC and improving patient outcomes from this disease.

Citation Format: Webb BM, Sahni JM, Keri RA. Cyclin dependent kinase 7 (CDK7) inhibition with THZ1 induces mitotic failure and increases genomic instability in triple negative breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P4-07-06.

Abstract P5-03-01: LIN9 regulation of NEK2 underlies taxol resistance in triple-negative breast cancer

Currently, there are no targeted strategies to combat triple negative breast cancer, resulting in poor patient survival. TNBC initially respond well to cytotoxic chemotherapies such as paclitaxel, yet resistance and metastatic recurrence are common. Paclitaxel causes defects in mitotic spindle formation and centrosome function, as well as improper chromosome segregation leading to cell death. While being some of the most effective drugs for this disease, taxanes are associated with high toxicity. Thus discovering new therapeutic targets that are selective for TNBC should yield novel approaches for improving patient outcomes. We discovered that LIN9, a transcriptional regulator of mitosis, is overexpressed in 66% of TNBC and associated with poor survival. We further found that both LIN9 mRNA and protein expression are upregulated in paclitaxel-resistant versus sensitive cells and directly correlates with paclitaxel IC50 values across nine breast cancer cell lines. In MDA-MB-231 and MDA-MB-468 cell lines, enforced overexpression of LIN9 increases multi- and micronucleation, indicators of chromosomal instability. Conversely, LIN9 silencing also results in multi- and micronucleation, and supernumerary centrosomes. Most importantly, LIN9 silencing increases sensitivity to paclitaxel in TNBC cells with intrinsic (BT549) or acquired (MDA-MB-231 and MDA-MB-468) resistance. We have previously reported that treating TNBC cells with Bromodomain and ExtraTerminal protein inhibitors (BETi) reduces LIN9 expression, thus we determined if BETi could reverse paclitaxel resistance. Treatment with the BETi, JQ1, and paclitaxel caused a greater induction of apoptosis compared to either drug alone. Dual treatment also resulted in a potentiation of abnormal centrosomes, multinucleation, and micronucleation compared to that caused by either BETi or JQ1 alone. To identify the mechanism(s) by which genetic or therapeutic suppression of LIN9 reverses paclitaxel resistance, we compared the transcriptomes of TNBC cells transiently transfected with non-targeting or LIN9-targeted siRNAs in MDA-MB-231 and HCC70 cell lines. We further narrowed the list of candidate LIN9 targets by identifying genes that were bound by LIN9 in a published ChIP-Seq dataset from HeLa cells whose expression is also correlated with LIN9 expression and associated with reduced breast cancer patient survival. Using this approach, we identified NIMA-related Kinase 2 (NEK2), a serine/threonine kinase required for centrosome separation during mitosis as a potential mediator of the effects of LIN9 suppression. NEK2 is overexpressed in 47% of basal breast cancers and is associated with poor survival. In addition, NEK2 is upregulated in paclitaxel-resistant cells and LIN9 silencing decreases expression of NEK2. Silencing NEK2 expression also restores sensitivity to paclitaxel in resistant cells. Together, these data indicate that increased LIN9 expression in TNBC promotes paclitaxel resistance by contributing to centrosome dysfunction through upregulation of NEK2. They also indicate that targeting LIN9 expression in addition to paclitaxel treatment may be a viable therapeutic approach for TNBC patients.

Citation Format: Shively MS, Gayle SS, Sahni JM, Keri RA. LIN9 regulation of NEK2 underlies taxol resistance in triple-negative breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P5-03-01.

Sustained trophism of the mammary gland is sufficient to accelerate and synchronize development of ErbB2/Neu-induced tumors

Epidemiological studies indicate that parity enhances HER2/ErbB2/Neu-induced breast tumorigenesis. Furthermore, recent studies using multiparous, ErbB2/Neu-overexpressing mouse mammary tumor virus (MMTV-Neu) mice have shown that parity induces a population of cells that are targeted for ErbB2/Neu-induced transformation. Although parity accelerates mammary tumorigenesis, the pattern of tumor development in multiparous MMTV-Neu mice remains stochastic, suggesting that additional events are required for ErbB2/Neu to cause mammary tumors. Whether such events are genetic in nature or reflective of the dynamic hormonal control of the gland that occurs with pregnancy remains unclear. We postulated that young age at pregnancy initiation or chronic trophic maintenance of mammary epithelial cells might provide a cellular environment that significantly increases susceptibility to ErbB2/Neu-induced tumorigenesis. MMTV-Neu mice that were maintained pregnant or lactating beginning at 3 weeks of age demonstrated accelerated tumorigenesis, but this process was still stochastic, indicating that early pregnancy does not provide the requisite events of tumorigenesis. However, bitransgenic mice that were generated by breeding MMTV-Neu mice with a luteinizing hormone-overexpressing mouse model of ovarian hyperstimulation developed multifocal mammary tumors in an accelerated, synchronous manner compared to virgin MMTV-Neu animals. This synchrony of tumor development in the bitransgenic mice suggests that trophic maintenance of the mammary gland provides the additional events required for tumor formation and maintains the population of cells that are targeted by ErbB2/Neu for transformation. Both the synchrony of tumor appearance and the ability to characterize a window of commitment by ovariectomy/palpation studies permitted microarray analysis to evaluate changes in gene expression over a defined timeline that spans the progression from normal to preneoplastic mammary tissue. These approaches led to identification of several candidate genes whose expression changes in the mammary gland with commitment to ErbB2/Neu-induced tumorigenesis, suggesting that they may either be regulated by ErbB2/Neu and/or contribute to tumor formation.

A single Pitx1 binding site is essential for activity of the LHbeta promoter in transgenic mice

Reproduction depends on regulated expression of the LHbeta gene. Tandem copies of regulatory elements that bind early growth response protein 1 (Egr-1) and steroidogenic factor 1 (SF-1) are located in the proximal region of the LHbeta promoter and make essential contributions to its activity as well as mediate responsiveness to GNRH: Located between these tandem elements is a single site capable of binding the homeodomain protein Pitx1. From studies that employ overexpression paradigms performed in heterologous cell lines, it appears that Egr-1, SF-1, and Pitx1 interact cooperatively through a mechanism that does not require the binding of Pitx1 to its site. Since the physiological ramifications of these overexpression studies remain unclear, we reassessed the requirement for a Pitx1 element in the promoter of the LHbeta gene using homologous cell lines and transgenic mice, both of which obviate the need for overexpression of transcription factors. Our analysis indicated a striking requirement for the Pitx1 regulatory element. When assayed by transient transfection using a gonadotrope-derived cell line (LbetaT2), an LHbeta promoter construct harboring a mutant Pitx1 element displayed attenuated transcriptional activity but retained responsiveness to GNRH: In contrast, analysis of wild-type and mutant expression vectors in transgenic mice indicated that LHbeta promoter activity is completely dependent on the presence of a functional Pitx1 binding site. Indeed, the dependence on an intact Pitx1 binding site in transgenic mice is so strict that responsiveness to GnRH is also lost, suggesting that the mutant promoter is inactive. Collectively, our data reinforce the concept that activity of the LHbeta promoter is determined, in part, through highly cooperative interactions between SF-1, Egr-1, and Pitx1. While Egr-1 can be regarded as a key downstream effector of GnRH, and Pitx1 as a critical partner that activates SF-1, our data firmly establish that the Pitx1 element plays a vital role in permitting these functions to occur in vivo.

Chronic hypersecretion of luteinizing hormone in transgenic mice disrupts both ovarian and pituitary function, with some effects modified by the genetic background

When the pituitary or hypothalamus becomes resistant to steroid negative feedback, a vicious cycle ensues, resulting in chronic hypersecretion of luteinizing hormone (LH) from the pituitary and steroids from the ovaries. In women, LH hypersecretion is implicated in infertility, miscarriages, and development of granulosa cell tumors. Progress in defining the underlying mechanisms of LH toxicity, however, has been limited by the lack of well-defined animal models. To that end, we have developed a new transgenic mouse model (alpha-LHbetaCTP) wherein LH hypersecretion occurs chronically and results in several dire pathological outcomes. Chronic hypersecretion of LH was achieved by introducing a transgene containing a bovine alpha subunit promoter fused to the coding region of a chimeric LHbeta subunit. The alpha subunit promoter directs transgene expression only to gonadotropes. The LHbeta chimera contains the carboxyl-terminal peptide (CTP) of the human chorionic gonadotropin beta subunit linked to the carboxyl terminus of bovine LHbeta. This carboxyl extension extends the half-life of LH heterodimers that contain the chimeric beta subunit. In intact alpha-LHbetaCTP females, serum LH is elevated five- to ten-fold in comparison to nontransgenic littermates. Levels of testosterone (T) and estradiol (E2) also are elevated, with an overall increase in the T-to-E2 ratio. These transgenic females enter puberty precociously but are anovulatory and display a prolonged luteal phase. Anovulation reflects the absence of gonadotropin-releasing hormone (GnRH) and the inability to produce a pre-ovulatory surge of LH. The ovaries are enlarged, with reduced numbers of primordial follicles and numerous, giant, hemorrhagic follicles. Despite the pathological appearance of the ovary, females can be superovulated and mated. Although pregnancy occurs, implantation is compromised due to defects in uterine receptivity. In addition, pregnancy fails at midgestation, reflecting a maternal defect presumably due to estrogen toxicity. When the transgene is in a CF-1 background, all females develop granulosa cell tumors and pituitary hyperplasia by five months of age. They die shortly thereafter due to bladder atony and subsequent kidney failure. When the transgene is placed in other strains of mice, their ovaries develop a luteoma rather than a granulosa cell tumor and the pituitary develops pituitary hyperplasia followed by adenoma. In summary, alpha-LHbetaCTP mice provide a direct association between abnormal secretion of LH and development of a number of ovarian and pituitary pathological responses.

An NF-Y binding site is important for basal, but not gonadotropin-releasing hormone-stimulated, expression of the luteinizing hormone beta subunit gene

Regulated synthesis of luteinizing hormone (LH) requires coordinated transcriptional control of the alpha and LHbeta subunits in pituitary gonadotropes. Several cis-acting elements and trans-acting factors have been defined for control of the LHbeta promoter through heterologous cell culture models. In this report, we describe the identification of bipartite NF-Y (CBF/CP1) binding sites within the proximal bovine LHbeta promoter. When multimerized, one of these sites activates the heterologous, minimal HSV thymidine kinase promoter in the gonadotrope-derived cell line alphaT3-1. The functional role of the promoter-distal site in regulating the full-length bovine LHbeta promoter was assessed in vivo using transgenic mice harboring a mutant promoter linked to the chloramphenicol acetyltransferase reporter gene. While this element is important for conferring high level activity of the LHbeta promoter in pituitary, it does not appear to be essential for mediating gonadotropin-releasing hormone (GnRH) regulation. This is the first characterization of a cis-acting element within this GnRH-dependent promoter that is restricted to regulating basal expression and not GnRH-induced activity.

Elevated luteinizing hormone induces expression of its receptor and promotes steroidogenesis in the adrenal cortex

Transgenic (TG) female mice expressing bLHbeta-CTP (a chimeric protein derived from the beta-subunit of bovine luteinizing hormone [LH] and a fragment of the beta-subunit of human chorionic gonadotropin [hCG]) exhibit elevated serum LH, infertility, polycystic ovaries, and ovarian tumors. In humans, increased LH secretion also occurs in infertility and polycystic ovarian syndrome, often concomitant with adrenocortical dysfunction. We therefore investigated adrenal function in LH overexpressing bLHbeta-CTP female mice. The size of their adrenals was increased by 80% with histological signs of cortical stimulation. Furthermore, adrenal steroid production was increased, with up to 14-fold elevated serum corticosterone. Primary adrenal cells from TG and control females responded similarly to ACTH stimulation, but, surprisingly, the TG adrenals responded to hCG with significantly increased cAMP, progesterone, and corticosterone production. LH receptor (LHR) expression and activity were also elevated in adrenals from female TG mice, but gonadectomized TG females showed no increase in corticosterone, suggesting that the dysfunctional ovaries of the intact TG females promote adrenocortical hyperfunction. We suggest that, in intact TG females, enhanced ovarian estrogen synthesis causes increased secretion of prolactin (PRL), which elevates LHR expression. Chronically elevated serum LH, augmented by enhanced PRL production, induces functional LHR expression in mouse adrenal cortex, leading to elevated, LH-dependent, corticosterone production. Thus, besides polycystic ovaries, the bLHbeta-CTP mice provide a useful model for studying human disorders related to elevated LH secretion and adrenocortical hyperfunction.

Luteinizing hormone induction of ovarian tumors: oligogenic differences between mouse strains dictates tumor disposition

The use of fertility drugs has continued to grow since their introduction in the 1960s. Accompanying this increase has been the speculation that repetitive use of these drugs can cause ovarian tumors or cancer. We recently reported that transgenic mice with chronically elevated luteinizing hormone (LH), an analog of which is commonly used in fertility regimens, develop granulosa cell (GC) tumors. In this report we show that LH induction of these tumors is highly dependent on genetic background. In CF-1 mice, chronically elevated LH invariably causes GC tumors by 5 months of age. However, in hybrid mice generated by crossing CF-1 males with C57BL/6, SJL, or CD-1 females, elevated levels of this same hormone cause a completely different phenotype resembling a luteoma of pregnancy. We also show that three genes likely control these alternative hormonal responses. This clinical correlate of elevated LH reveals remarkably distinct, strain-dependent, ovarian phenotypes. In addition, these results support the rare incidence of GC tumors in the human population, and suggest that the ability of certain fertility drugs to cause ovarian tumors may depend on an individual's genetic predisposition.

Transgenic mice with chronically elevated luteinizing hormone are infertile due to anovulation, defects in uterine receptivity, and midgestation pregnancy failure

Elevated levels of LH have been associated with infertility and miscarriage in women. Previously, we have reported generating a transgenic mouse model that hypersecretes LH. Female transgenics exhibit extensive pathology including enlarged, cystic, and hemorrhagic ovaries; elevated testosterone:estradiol ratios; and infertility primarily due to anovulation. Here we show that anovulation can be reversed in transgenics and that, despite development within a pathological ovary, oocytes from transgenics are remarkably healthy. Fertilized ova from transgenics are capable of normal development to term when transferred into nontransgenic pseudopregnant recipients. However, reciprocal transfers of nontransgenic embryos into transgenic recipients failed due to lack of uterine receptivity. In addition, while superovulated and mated transgenics appear to have normal early pregnancy, embryos are resorbed at midgestation due to maternal hormonal defects. Transgenic infertility can be rescued by ovariectomy with progesterone and estradiol replacement. These studies are particularly intriguing in light of data indicating an increased rate of miscarriage among women undergoing infertility treatments who are diagnosed with polycystic ovarian syndrome.

Characterization of the equine glycoprotein hormone alpha-subunit gene reveals divergence in the mechanism of pituitary and placental expression

The equine glycoprotein hormone alpha-subunit gene is expressed in both pituitary and placenta, unlike that of all other nonprimate mammals studied, in which expression is limited to pituitary. Previous studies of the 5'-flanking region of the equine alpha-subunit promoter have revealed unique characteristics as well as similarities with the human alpha-subunit promoter, which demonstrates a similar pattern of tissue-specific expression. We have cloned and sequenced the equine alpha-subunit gene and have used tissue culture systems and transgenic mice to characterize its expression. Unlike the human promoter, the cloned equine alpha-subunit promoter failed to direct trophoblast-specific expression in either tissue culture or transgenic mouse models, suggesting an entirely different mechanism for expression. In contrast, the equine alpha-subunit promoter was able to direct gonadotroph expression in both tissue culture and transgenic mouse models. In alphaT3-1 cells, 550 base pair (bp) was sufficient for expression. This expression involves promoter elements identified in other species as playing a role in gonadotroph expression, but mutation of these elements reveals differences in their relative contributions to promoter activity. In mice, 2800 bp of 5'-flanking sequence allowed specific expression in gonadotrophs but not in thyrotrophs or placenta. The pattern of estrogen regulation observed in transgenic mice matched neither the repression that has been observed with human and bovine alpha-subunit promoters in transgenic mice nor the stimulation in mRNA levels reported in mares, suggesting a unique mechanism that is not recapitulated in the transgenic model. Thus the equine alpha-subunit promoter uses a combination of conserved and unique features of gene regulation to direct its pattern of tissue-specific expression.

A steroidogenic factor-1 binding site is required for activity of the luteinizing hormone beta subunit promoter in gonadotropes of transgenic mice

Analysis of luteinizing hormone (LH) beta subunit promoters from a broad range of species including teleosts and humans revealed strict conservation of a sequence homologous to the steroidogenic factor-1 (SF-1) regulatory element of cytochrome P-450 steroid hydroxylase genes. Interaction between SF-1 and this putative response element in the bovine LH beta promoter was confirmed by electrophoretic mobility shift assays. Furthermore, cotransfection of alpha T3-1 cells with an expression vector encoding SF-1 induced binding site-dependent transcription from the bovine LH beta promoter. Physiological significance of the LH beta SF-1 consensus sequence was established using transgenic mice containing either the wild type bovine promoter or a promoter with a site-specific mutation of this site. Mutation of the SF-1 binding site nearly eliminated promoter activity, and the mutant transgene remained inactive following induction of gonadotropin-releasing hormone accomplished by castrating male and female mice. Thus, increases of gonadotropin-releasing hormone within a physiological range did not compensate for the loss of the SF-1 binding site. Together, these findings indicate that the SF-1 binding site is a key regulator of LH beta promoter activity in vivo and implicate SF-1 as at least one of the transcription factors that acts through this site.

Implementing transgenic and embryonic stem cell technology to study gene expression, cell-cell interactions and gene function

This review highlights the use of transgenic mice and gene targeting in the study of reproduction, pituitary gene expression, and cell lineage. Since 1980 numerous applications of transgenic animal technology have been reported. Altered phenotypes resulting from transgene expression demonstrated that introduced genes can exert profound effects on animal physiology. Transgenic mice have been important for the study of hormonal and developmental control of gene expression because gene expression in whole animals often requires more DNA sequence information than is necessary for expression in cell cultures. This point is illustrated by studies of pituitary glycoprotein hormone alpha- and beta-subunit gene expression (Kendall et al., Mol Endocrinol 1994; in press [1]. Transgenic mice have also been invaluable for producing animal models of cancer and other diseases and testing the efficacy of gene therapy. In addition, cell-cell interactions and cell lineage relationships have been explored by cell-specific expression of toxin genes in transgenic mice. Recent studies suggest that attenuated and inducible toxins hold promise for future transgene ablation experiments. Since 1987, embryonic stem (ES) cell technology has been used to create numerous mouse strains with targeted gene alterations, contributing enormously to our understanding of the functional importance of individual genes. For example, the unexpected development of gonadal tumors in mice with a targeted disruption of the inhibin gene revealed a potential role for inhibin as a tumor suppressor (Matzuk et al., Nature 1992:360: 313-319 [2]. The transgenic and ES cell technologies will undoubtedly continue to expand our understanding and challenge our paradigms in reproductive biology.

The proximal promoter of the bovine luteinizing hormone beta-subunit gene confers gonadotrope-specific expression and regulation by gonadotropin-releasing hormone, testosterone, and 17 beta-estradiol in transgenic mice

Transient transfection studies have proven useful in unraveling the molecular mechanisms underlying gonadotrope-specific expression and hormonal regulation of the gene encoding the alpha-subunit of the glycoprotein hormones. In contrast, similar studies performed with the LH beta gene have been less informative. When assayed by transient transfection into alpha T3-1 cells, activity of a 776-basepair bovine LH beta promoter-chloramphenicol acetyltransferase fusion gene (bLH beta CAT) was no greater than that of a promoterless control. To determine whether limited activity in vitro reflected the absence of critical regulatory elements, we examined activity of bovine LH beta fusion genes after stable integration in transgenic mice. In contrast to transient transfection studies, the LH beta promoter targeted high levels of CAT expression specifically to the pituitary. In addition, a bLH beta TK fusion gene was active only in gonadotropes. The bLH beta CAT transgene was also evaluated for responsiveness to gonadal steroids and GnRH. Testosterone and 17 beta-estradiol were capable of suppressing activity 70-80% in castrated males, despite the absence of high affinity binding sites for androgen or estrogen receptors. This suggests that feedback inhibition of LH beta CAT transgene expression by gonadal steroids may occur through an indirect mechanism, possibly at the level of the hypothalamus. To address whether the bLH beta CAT transgene could be regulated by GnRH, we treated ovariectomized females with antide, a GnRH antagonist. Antide suppressed transgene activity by 60%. Thus, the proximal promoter of the bovine LH beta subunit gene directs appropriate patterns of cell-specific expression and retains responsiveness to gonadal steroids and GnRH. In light of the robust activity of the LH beta promoter in transgenic mice, we suggest that this animal model can be exploited further to dissect the complex mechanisms that underlie gonadotrope-specific expression and hormonal regulation of the LH beta gene.

Transcriptional repression of the glycoprotein hormone alpha subunit gene by androgen may involve direct binding of androgen receptor to the proximal promoter

Testicular androgens suppress the synthesis and secretion of the pituitary gonadotropins, in particular, luteinizing hormone. This suppressive effect includes transcription of both the common alpha subunit gene and the unique beta subunit genes. Herein, we demonstrate that 1500 base pairs (bp) of proximal 5'-flanking region derived from the human alpha subunit gene and a shorter 315-bp segment of the bovine alpha subunit gene confer negative regulation by androgen to the gene encoding bacterial chloramphenicol acetyltransferase in transgenic mice. Cotransfection assays with human androgen receptor indicated that the 1500-bp promoter region of the human alpha subunit gene also confers androgen regulation (transcriptional suppression) to reporter genes in both pituitary and placental cell lines. This raises the possibility of a role for DNA binding in suppression of alpha subunit transcription by activated androgen receptor. Consistent with this possibility, we have used a gel-mobility shift assay to detect several high affinity binding sites for androgen receptor located in the proximal promoter of the human alpha subunit gene. The strongest androgen receptor binding site is located at approximately -101 in the proximal 5'-flanking region. This steroid receptor binding site overlaps another binding site that defines one of several contiguous cis-acting regulatory elements required for basal transcriptional activity. Thus, binding of activated androgen receptor to this region may block the binding of a requisite trans-acting factor and lead to an attenuation in transcription. We conclude that this interaction, which occurs directly at the level of the pituitary, represents one of several physiological avenues through which androgens regulate gonadotropin gene expression.

Gonadotrope- and thyrotrope-specific expression of the human and bovine glycoprotein hormone alpha-subunit genes is regulated by distinct cis-acting elements

The proximal 5'-flanking region of the alpha-subunit gene from humans and cattle confers pituitary-specific expression to heterologous reporter genes in transgenic mice. To investigate whether these promoter regions also contain the necessary regulatory elements for cell-specific expression and hormonal regulation, we used three independent lines of transgenic mice. Two lines of transgenic mice contained chimeric genes consisting of either 1.6 kilobasepairs (kbp) of human or 3 15 basepairs of bovine alpha-subunit proximal 5'-flanking sequence linked to the bacterial gene encoding chloramphenicol acetyltransferase (CAT). A third line of transgenic mice contained the proximal 1.6 kbp of 5'-flanking sequence of the human alpha-subunit gene linked to the bacterial lacZ gene encoding beta-galactosidase (beta gal; H alpha beta gal transgenic mice). Hormonal replacement paradigms indicate that both human and bovine alpha CAT transgenes are regulated by GnRH, suggesting that their expression occurs in gonadotropes. Thus, the proximal 5'-flanking regions of both the human and bovine alpha-subunit genes must contain regulatory elements that confer both gonadotrope-specific expression and responsiveness to GnRH. In contrast to the human alpha-subunit promoter, the bovine alpha-subunit promoter lacks a functional cAMP response element, suggesting that transduction of both cell-specific and GnRH transcriptional signals occurs through cAMP response element-independent pathways. Thyrotropes also express the glycoprotein hormone alpha-subunit gene. Yet, hormone replacement paradigms with propylthiouracil and T3 were ineffective in altering CAT activity in the pituitary of human or bovine alpha CAT transgenic mice. Because a thyroid hormone response element has been localized to the proximal 5'-flanking region of the human alpha-subunit gene, these data suggest that the alpha CAT transgenes lack sufficient information to direct expression to thyrotropes. Direct evidence for this possibility was obtained through immunocytochemical studies performed on pituitaries from H alpha beta gal transgenic mice. beta-Galactosidase activity appeared in gonadotropes, but not thyrotropes. We conclude, therefore, that distinct and separable regulatory elements mediate the expression of the alpha-subunit gene in gonadotropes and thyrotropes.

Targeted ablation of pituitary gonadotropes in transgenic mice

LH, FSH, and TSH are heterodimeric glycoprotein hormones composed of a common alpha-subunit and unique beta-subunits. The alpha-subunit is produced in two distinct specialized cell types of the pituitary gland: gonadotropes, which synthesize LH and FSH, and thyrotropes, which synthesize TSH. We have demonstrated that 313 base pairs of the bovine-alpha subunit promoter direct expression of diphtheria toxin A chain specifically to the gonadotropes in transgenic mice. Animals carrying this transgene generally exhibit reproductive failure and lack of gonadal differentiation, consistent with gonadotrope ablation. Lack of gonadotrope activity was verified by RIA and immunohistochemical staining for LH. The phenotype of these transgenic mice is nearly identical to mice homozygous for the spontaneous mutation, hpg, which is due to a deletion in the gene encoding GnRH. Thyrotrope function was judged normal based on overall growth of the animals, appearance of their thyroids, T4 levels measured by RIA, and immunohistochemical staining for TSH. The ablation of gonadotropes but not thyrotropes suggests that separate cis-acting elements are necessary for expression of the alpha-subunit gene in these two cell types. Pituitary content of ACTH and GH was apparently normal, while PRL synthesis and storage were reduced. Thus, in a pituitary almost completely devoid of gonadotropes, most other pituitary functions were normal. This suggests that most pituitary cells are able to differentiate independently of terminal gonadotrope differentiation and can function in the absence of paracrine signaling provided by gonadotropes.

Estradiol inhibits transcription of the human glycoprotein hormone alpha-subunit gene despite the absence of a high affinity binding site for estrogen receptor

Chronic administration of estradiol inhibits transcription of the gene encoding the alpha-subunit of pituitary glycoprotein hormones. Here, we show, using transfection analyses and a filter binding assay, that 1500 basepairs of proximal 5' flanking sequence of the human alpha-subunit gene lack a functional estrogen response element when transfected into heterologous cell lines, and fail to bind estrogen receptor purified from calf uterus. Yet, this same region of the alpha-subunit gene confers estradiol responsiveness (transcriptional suppression) to the bacterial chloramphenicol acetyltransferase gene in transgenic mice. A smaller promoter fragment of the bovine alpha-subunit gene also confers responsiveness to estradiol in transgenic mice, suggesting that the same element may mediate the steroid responsiveness of both promoters. Furthermore, regulation by estradiol of the chimeric human or bovine alpha-chloramphenicol acetyltransferase genes is pituitary specific, underscoring the physiological significance of these studies. Based on these results, we conclude that estradiol regulates expression of the alpha-subunit gene in vivo through a mechanism that does not involve high affinity binding of estrogen receptor to the alpha-subunit gene. Whether this mechanism is manifest at the level of the pituitary or hypothalamus remains to be determined.

Different combinations of regulatory elements may explain why placenta-specific expression of the glycoprotein hormone alpha-subunit gene occurs only in primates and horses

Expression of the glycoprotein hormone alpha-subunit gene occurs in the pituitary of all mammals but in placenta of only primates and horses. In humans, two different elements, termed upstream regulatory element (URE) and cAMP response element (CRE), are required for placenta-specific expression of the alpha-subunit gene. The URE binds a protein unique to placenta whereas the CRE binds a ubiquitous protein. Comparative analysis of the promoter-regulatory region of the alpha-subunit gene from a number of mammals indicates that a functional URE has been retained and suggests the potential for placenta-specific expression. Indirect evidence also indicates that the URE-binding protein has been conserved, even in placenta from mammals that fail to express the alpha-subunit gene. Lack of expression of the alpha-subunit gene in placenta of rodents and cattle can be traced to a single nucleotide change that renders the CRE-like sequence of these genes incapable of binding the protein that confers responsiveness to cAMP. In contrast, although expression of the alpha-subunit gene occurs in horse placenta, the promoter-regulatory region lacks a functional CRE but appears to retain a functional URE. This suggests that either a different accessory element and cognate protein interacts with the horse URE to provide placenta-specific expression or that a completely different set of regulatory elements is required for placenta-specific expression in horses.

Expression of the glycoprotein hormone alpha-subunit gene in the placenta requires a functional cyclic AMP response element, whereas a different cis-acting element mediates pituitary-specific expression

The single-copy gene encoding the alpha subunit of glycoprotein hormones is expressed in the pituitaries of all mammals and in the placentas of only primates and horses. We have systematically analyzed the promoter-regulatory elements of the human and bovine alpha-subunit genes to elucidate the molecular mechanisms underlying their divergent patterns of tissue-specific expression. This analysis entailed the use of transient expression assays in a chorionic gonadotropin-secreting human choriocarcinoma cell line, protein-DNA binding assays, and expression of chimeric forms of human or bovine alpha subunit genes in transgenic mice. From the results, we conclude that placental expression of the human alpha-subunit gene requires a functional cyclic AMP response element (CRE) that is present as a tandem repeat in the promoter-regulatory region. In contrast, the promoter-regulatory region of the bovine alpha-subunit gene, as well as of the rat and mouse genes, was found to contain a single CRE homolog that differed from its human counterpart by a single nucleotide. This difference substantially reduced the binding affinity of the bovine CRE homolog for the nuclear protein that bound to the human alpha CRE and thereby rendered the bovine alpha-subunit promoter inactive in human choriocarcinoma cells. However, conversion of the bovine alpha CRE homolog to an authentic alpha CRE restored activity to the bovine alpha-subunit promoter in choriocarcinoma cells. Similarly, a human but not a bovine alpha transgene was expressed in placenta in transgenic mice. Thus, placenta-specific expression of the human alpha-subunit gene may be the consequence of the recent evolution of a functional CRE. Expression of the human alpha transgene in mouse placenta further suggests that evolution of placenta-specific trans-acting factors preceded the appearance of this element. Finally, in contrast to their divergent patterns of placental expression, both the human and bovine alpha-subunit transgenes were expressed in mouse pituitary, indicating differences in the composition of the enhancers required for pituitary- and placenta-specific expression.

CRE-binding proteins interact cooperatively to enhance placental-specific expression of the glycoprotein hormone alpha-subunit gene

The 18-bp direct repeat occurring between positions -146 and -111 in the 5'-flanking region of the human alpha-subunit gene serves two functions: it mediates the transcriptional effect of cAMP and it acts in conjunction with an adjacent cis-acting element (URE) to confer properties of placental-specific expression to the alpha-subunit promoter. Functional activity of the URE and CRE requires binding of a trans-acting factor; each element binds a different factor. Analysis of saturation isotherms provides good evidence that cooperativity is involved in binding of CREB to the 18-bp direct repeat. This cooperativity could account for the synergistic effect of two CRE on both basal and cAMP-stimulated transcription. It remains to be determined whether heterotropic cooperativity is involved in binding of trans-acting factors to the URE and CRE. A major difference between the 5'-flanking region of the human alpha-subunit gene and comparable regions from bovine, rat, and mouse alpha-subunit genes is that the latter contain a single CRE homolog which appears incapable of binding the trans-acting factor that binds to the human alpha CRE. Lack of a functional CRE provides at least one explanation for inactivity of the bovine alpha-subunit promoter in choriocarcinoma cells and probably in bovine placenta as well. Yet, the same bovine promoter-regulatory region that lacks a functional CRE is capable of conferring pituitary-specific expression to the CAT gene in transgenic mice (data not shown). This suggests that the CRE is not required for pituitary-specific expression of the bovine alpha-subunit gene. Instead, another cis-acting element(s) must confer this property to the alpha-subunit promoter. While it is tempting to suggest that bovine, rat, and mouse alpha-subunit genes are not regulated by cAMP because of their inactive CRE homolog, it is also quite possible that other CRE are located further upstream. Accordingly, it will be of interest to obtain additional 5'-flanking sequence and determine whether functional homologs of the human alpha CRE are present in the bovine, rat, and mouse alpha-subunit genes, or whether another class of cis-acting elements provide cAMP-responsiveness.