Embryonic Expression of the Luteinizing Hormone β Gene Appears to Be Coupled to the Transient Appearance of p8, a High Mobility Group-related Transcription Factor*

A comparison between two pituitary-derived cell lines (αT3-1 and LβT2) that represent gonadotropes at early and late stages of development, respectively, was performed to further elucidate the genomic repertoire required for gonadotrope specification and luteinizing hormone β (LHβ) gene expression. One isolated clone that displayed higher expression levels in LβT2 cells encodes p8, a high mobility group-like protein with mitogenic potential that is up-regulated in response to proapoptotic stimuli and in some developing tissues. To test the functional significance of this factor in developing gonadotropes, a knockdown of p8 in LβT2 cells was generated. The loss of p8 mRNA correlated with loss of endogenous LHβ mRNA and the loss of activity of a transfected LHβ promoter-driven reporter, even upon treatment with gonadotropin-releasing hormone. In addition, expression of p8 mRNA in developing mouse pituitary glands mirrored its expression in the gonadotrope-derived cell lines and coincided with the first detectable appearance of LHβ mRNA. In contrast, p8 mRNA was undetectable in the pituitary glands of normal adults. Taken together, our data indicate that p8 is a stage-specific component of the gonadotrope transcriptome that may play a functional role in the initiation of LHβ gene expression during embryonic cellular differentiation.

With its ability to stimulate gonadal steroidogenesis and gametogenesis, luteinizing hormone (LH) 1 is essential for normal reproductive function in mammals (1,2). Luteinizing hormone is a heterodimeric protein composed of an ␣ glycoprotein hormone subunit (␣GSU) common to all members of the glycoprotein family of hormones that is non-covalently linked to a unique LH␤ subunit that confers its biological specificity (1,2). Biosynthesis of LH depends on the coordinated expression of both the ␣GSU and LH␤ subunit genes. In humans, the single copy ␣GSU gene resides on chromosome 6q12-q21 (Locus ID 1081) while the LH␤ gene resides amid a cluster of six chorionic gonadotropin-␤ genes on chromosome 19q13.32 (Locus ID 1082). In addition to their different locations within the human genome, the pattern of expression of the genes encoding ␣GSU and LH␤ are temporally and spatially distinct.
During development ␣GSU is seen throughout Rathke's pouch as early as embryonic day (e) 9.5 in the mouse (3). By later stages of pituitary development and in adult mammals, ␣GSU expression is limited to thyrotropes and gonadotropes (4). Gonadotrope-specific expression of the ␣GSU gene is controlled by an array of regulatory elements, including the pituitary glycoprotein hormone basal element (5,6), ␣ basal elements (6), gonadotrope-specific element (7), and tandemly repeated cAMP response elements (8 -10), as well as the intricate interplay between their cognate binding proteins (6).
As indicated above, many of the elements and proteins involved in regulation of LH␤ gene expression and terminal differentiation of gonadotropes have been characterized; however, the full scope of factors required for differentiation to a mature cell that expresses all phenotypic markers of a gonadotrope, including LH␤, have yet to be identified. To that end, we compared the gene expression of cell lines that represent two distinct stages of gonadotrope development by differential display. ␣T3-1 cells represent early gonadotrope progenitors that express ␣GSU and gonadotropin-releasing hormone (GnRH) receptor but not the unique ␤-subunits of the glycoprotein hormones (28). The L␤T2 cell line is characterized by expression of markers of fully differentiated gonadotropes, including the ␤-subunits of the gonadotropins (29,30). Because these two cell lines represent phenotypic gonadotropes at embryonic days before (␣T3-1) or after (L␤T2) the ability to express LH␤, characterization of differentially expressed factors may uncover components that are essential for expression of the LH␤ subunit gene and should refine our understanding of gonadotrope differentiation.
One factor that is more highly expressed in L␤T2 cells than ␣T3-1 cells is the HMG-like nuclear phosphoprotein known as p8 or candidate of metastasis 1 (com1). The HMG class of proteins to which p8 is related (HMG-I/Y or HMGA) function as architectural transcription factors that promote gene activa-tion by relieving histone H1-mediated repression of transcription (31) and facilitating the formation of enhanceosomes as a consequence of both protein/DNA and protein/protein interactions (32). Like p8, these proteins, which have the capacity to bend, straighten, unwind, and induce loop or supercoil formation in linear DNA molecules in vitro, are at maximal levels of expression during embryonic development and in rapidly proliferating cells (32). While p8 lacks characteristic "A-T hook" DNA binding domains found in the HMGA class of non-histone chromatin-binding proteins (32,33), it does appear to bind DNA, especially upon phosphorylation (34), and thus, p8 may perform a comparable role within cells. In the gonadotrope, p8 appears to be a stage-specific component of the cellular transcriptome that may play a functional role in the initiation of LH␤ gene expression. This potential is explored herein.

EXPERIMENTAL PROCEDURES
PCR Differential Display-Total RNA was extracted from confluent ␣T3-1 and L␤T2 cells using the method of Chomczynski and Sacchi (35). Removal of chromosomal DNA contamination from samples, PCR differential display, and reamplification of DNA were performed as described in the Current Protocols in Molecular Biology (36). Each assay was performed using one of four degenerate anchored oligo(dT) primer sets (T 12 MN; M can be G, A, or C and N is G, A, T, and C) where each primer set is dictated by the 3Ј base (N) with degeneracy in the penultimate (M) position by itself or each in combination with one of 26 decameric primers that were originally designed to allow for coverage of the expressed genome (37). Clones of interest were subcloned into the pCR®II-TOPO® vector (Invitrogen, Carlsbad, CA) using the standard protocol described by the manufacturer. The DNA sequence of each clone was found by dideoxynucleotide sequencing using Sp6 and/or T7 primers. Identification of clones was determined by comparing each sequence against the BLAST database (38).
Plasmid Vectors-All plasmid DNAs were prepared from overnight bacterial cultures using Qiagen DNA plasmid columns according to the protocol of the supplier (Qiagen, Chatsworth, CA). Murine p8-pcDNA3 was a gift from Juan Iovanna and colleagues (39). To produce the p8-antisense vector (p8AS-pcDNA3), p8-pcDNA3 was digested with BamHI and ApaI restriction endonucleases. The p8 insert was then blunt-ended and ligated into the EcoRV site of pcDNA3. Clones were dideoxynucleotide sequenced to verify selection of a reverse orientation clone. The LH␤ promoter-driven luciferase reporter vector (23) as well as the ␣GSU promoter-driven luciferase reporter vector (6) have been described previously.
Cell Lines-The gonadotrope-derived cell lines, ␣T3-1 and L␤T2, were maintained in high glucose Dulbecco's modified Eagle's medium containing 2 mM L-glutamine and supplemented with 10% fetal bovine serum and antibiotics (complete medium). To produce the control (C)-L␤T2, p8-knockdown (KD)-L␤T2, and p8-overexpressing (OE)-L␤T2 cell lines, 24 ϫ 10 6 L␤T2 cells on a 100-mm culture dish were transfected with 16 g of pcDNA3, p8AS-pcDNA3, and p8-pcDNA3, respectively, using 80 l of LipofectAMINE reagent (Invitrogen) and medium lacking serum and antibiotics. Stably transfected cells were selected by introducing 500 g/ml G418 (Invitrogen) into the complete medium 3 days following the transfection. The cells utilized in these studies, which were maintained in the above-described complete medium supplemented with G418, represent pools of clones for each cell line derived from the parent L␤T2 line.
Transfection Assays-The day prior to transient transfection, C-L␤T2 and p8-KD-L␤T2 cells were plated at a density of ϳ2 ϫ 10 6 cells/35-mm well. Transfections were carried out using medium lacking serum, antibiotics, and G418 selection medium with 10 l of Lipo-fectAMINE reagent, 2 g each test vector, and 100 ng of pRL-CMV (Promega, Madison, WI), which was used to normalize data for transfection efficiency. Cell cultures were incubated with the transfection mixtures for ϳ18 h at 37°C in a humidified atmosphere with 5% CO 2 . Complete medium was then added to the cells, which where indicated, were also supplemented with 100 nM GnRH. Twenty-four hours following the addition of fresh medium and hormonal treatments, cells were lysed in passive lysis buffer (Promega) and a dual-luciferase assay was performed on each cellular lysate as per standard procedures (Promega). Transient transfections were performed a minimum of three times with at least two separate plasmid preparations for each construct that was tested.
Northern Blot Analyses-For each Northern blot, 10 or 20 g of total RNA were separated by electrophoresis in a 1% denaturing agarose gel and transferred to nylon membrane (Hybond-N ϩ ; Amersham Biosciences) by gravity and capillary action. After UV cross-linking to fix RNA to the nylon, the membrane was prehybridized for approximately 3 h and hybridized with the appropriate radiolabeled probe overnight at 45°C in a roller-bottle hybridization oven (Techne, Inc., Princeton, NJ). The hybridization solution consisted of 40% deionized formamide, 20 mM PIPES, 800 mM NaCl, 2 mM EDTA, 4% SDS, 80 g/ml salmon sperm DNA. All probes were made by radiolabeling of cDNAs with [ 32 P]deoxy-CTP or ATP (3000 Ci/mmol, PerkinElmer Life Sciences) using DECAprime II kit as per the suggestions of the manufacturer (Ambion, Austin, TX). The final washes following hybridization were in 0.5ϫ SSC, 0.5% SDS at 65°C. The membranes were exposed to Biomax MR film (Eastman Kodak Co., Rochester, NY) for one or more days at Ϫ80°C. In addition, some blots were exposed to storage phosphor screens (Amersham Biosciences) for 2-18 h followed by densitometric scanning and analysis using a Storm 820 PhosphorImager (Amersham Biosciences). Between hybridizations, each blot was stripped of radioactivity using the protocol enclosed with the nylon membrane. Northern blots were replicated at least three times, and densitometric scanning was performed on representative blots. Random-prime labeled probes used in the Northern blot analyses consisted of the full murine p8 cDNA (39), murine LH␤ cDNA representing bases 35 through 213 (40), murine cytoskeletal ␤-actin cDNA representing bases 1210 through 1657 (41), a murine GnRH receptor cDNA representing bases 247 through 537 (42), and a cDNA encompassing the murine 18 S rRNA contained within primers included in the QuantumRNA 18 S Internal Standards (Ambion). In all cases a very small amount of flanking DNA from the multiple cloning sites of each vector was included when making probes.
In Situ Hybridization-Timed pregnancies were obtained by mating CF1 or FVB/N male ϫ CF1 female. Noon on the day of copulatory plug detection was considered e0.5. Appropriately aged embryos were frozen in 2-methylbutane at approximately Ϫ30°C and sectioned at 15 m on a Hacker-Bright cryostat (Hacker Instruments and Industries, Fairfield, NJ). In situ hybridization was performed as described by Cushman et al. (43). Briefly, sections were warmed to room temperature for 30 min and fixed in 4% paraformaldehyde/phosphate-buffered saline (PBS) at 37°C. Following proteinase K treatment (0.1 g/ml), sections were fixed again in 4% paraformaldehyde/PBS and washed in PBS. Sections were acetylated using a 0.1 M triethanolamine, 0.25% acetic anhydride mixture and incubated with hybridization solution minus the probe (50% formamide, 5ϫ SSC, 2% Boehringer blocking powder, 0.1% Triton X-100, 0.5% CHAPS, 1 mg/ml yeast tRNA, 5 mM EDTA, 50 mg/ml heparin in diethyl pyrocarbonate-treated water) at 55°C. Riboprobe was diluted in hybridization solution and allowed to hybridize overnight at 55°C in a chamber humidified with 5ϫ SSC. The next morning, sections were washed in a 50% formamide, 0.5ϫ SSC mixture at 55°C followed by a wash in 0.5ϫ SSC at room temperature and blocked with the following solution in a chamber humidified with water: 10% heat-inactivated sheep serum; 2% BSA; 0.02% sodium azide in 50 mM Tris-Cl (pH 7.5), 100 mM NaCl, 0.1% Triton X-100. Anti-digoxigenin Fab fragments (Roche Molecular Biochemicals) were diluted 1:1000 in blocking buffer, incubated on sections for 1 h at room temperature in a humidified chamber and washed in PBS. After equilibration in several washes of chromagen buffer (100 mM Tris-Cl (pH 9.5), 100 mM NaCl, 50 mM MgCl 2 ), sections were developed overnight in the chromagen buffer with 4.5 l/ml 4-nitro blue tetrazolium chloride and 3.5 l/ml 5-bromo-4-chloro-3-indolyl-phosphate added as substrate for alkaline phosphatase activity. Sections were rinsed with PBS, fixed in 4% paraformaldehyde/PBS, washed, and mounted. In situ hybridization was repeated at least two times for each embryonic age.
Probes Used for in Situ Hybridization-Plasmids were linearized as follows to generate sense and antisense probes: 435-bp ␣GSU cDNA (GenBank TM accession number NM0099889, bp numbers 181 through 616 of ␣GSU) in pGEM3ZF ϩ was linearized with BamHI for antisense and HindIII for sense probes; 233-bp LH␤ cDNA (GenBank TM accession number NM_008497, bp numbers 12 through 245) in BSK Ϫ was linearized with HindIII for antisense and BamHI for sense probes; and full murine p8 cDNA in pcDNA3 was linearized with EcoRI for antisense and ApaI for sense probes. Digoxigenin-labeled riboprobe was generated with in vitro transcription using 10ϫ DIG RNA labeling mix (Roche Molecular Biochemicals) and purified with the RNeasy minikit (Qiagen).
Experimental Animals-All animal studies were conducted in accordance with the principles and procedures approved by the Institutional Animal Care and Use Committee of Case Western Reserve University.

p8 mRNA Is Up-regulated in Cells That Represent Mature
Gonadotropes-To identify factors that may be essential for maturation of gonadotropes, differential display was employed to compare the mRNA populations from gonadotrope-derived cells that represent distinct developmental stages, namely ␣T3-1 and L␤T2. Because over 300 differentially expressed cDNAs were isolated using this method, reverse Northern blots were performed to verify the expression seen by differential display (data not shown). After reamplification and nucleotide sequencing of several clones, we focused our attention on clone 100 (Fig. 1A), which was identified as the 3Ј-untranslated region of p8 using the BLAST database available through the National Center for Biotechnology Information (38). p8 shows both sequence and structural similarities with factors that have proven vital for embryonic development (34,44). To confirm the differential expression of p8 in ␣T3-1 and L␤T2 cells, we examined total RNA from each cell line by Northern blot analysis and found that mRNA encoding p8 was significantly up-regulated in L␤T2 compared with ␣T3-1 cells (Fig.  1, B and C).
Silencing of Endogenous p8 in L␤T2 Cells-To determine the functional significance of p8 mRNA, a plasmid vector driving expression of p8-antisense was used to stably transfect L␤T2 cells (p8-KD-L␤T2). We anticipated that this aberrant message would form double-stranded RNA with the endogenous p8 transcript and either inhibit its translation or target it for degradation, thus allowing for sequence-specific, post-transcriptional silencing of the p8 gene. Our data support the latter possibility as no detectable levels of p8 mRNA were observed in p8-KD-L␤T2 cells (Fig. 2). In contrast, in L␤T2 cells that were stably transfected with a plasmid encoding the sense strand of p8 (p8-OE-L␤T2), a 579-nucleotide p8 mRNA representing the transcription product from the transfected vector was identified. Importantly, endogenous p8 transcript (639 nucleotides) was detected in total RNA from untransfected L␤T2 cells as well as cells that have been stably transfected with the empty, negative control vector (C-L␤T2). Thus, stable transfection of the p8 antisense vector effectively removed p8 mRNA from L␤T2 cells.
It is known that double-stranded RNA in the cytoplasm of mammalian cells can trigger profound effects, including a global suppression of translation and mRNA degradation in a sequence-nonspecific manner as part of an interferon response (45), as well as apoptosis (46). To provide evidence that the loss of endogenous p8 was due to sequence-specific rather than nonspecific gene silencing, Northern blot hybridization with a radiolabeled ␤-actin probe was performed (Fig. 2). Levels of detectable ␤-actin mRNA were comparable across RNA samples from p8-KD-L␤T2 cells, p8-OE-L␤T2 cells, as well as C-L␤T2 and normal, untransfected L␤T2s, indicating that the product of this architectural gene, and likely other transcripts, was not affected by the presence of double-stranded RNA in the p8 knockdown cell line.
The LH␤ Promoter Is Non-functional When p8 Is Knocked Down in L␤T2 Cells-Expression of the p8 gene appears to follow a pattern that corresponds to LH␤ expression in gonadotrope-derived cell lines. To determine the potential impact of p8 on regulation of the LH␤ gene, p8-KD-L␤T2 and C-L␤T2 cells were tested for their ability to support activity of a LH␤ promoter-driven luciferase reporter after transient transfection. As expected, the C-L␤T2 cell line displayed levels of luciferase activity ϳ4-fold higher than that from cell lines transfected with a promoterless reporter vector (Fig. 3A), levels that are comparable with normal L␤T2 cells (data not shown). In cells lacking p8 (p8-KD-L␤T2), the LH␤ promoter was nonfunctional, even in the presence of 100 nM GnRH (Fig. 3B). To confirm that the effect of p8 knockdown on the LH␤ promoter was not due to an overall hindrance of cellular processes, an ␣GSU promoter-driven construct was also tested. This vector was shown to be functional in p8-KD-L␤T2 cells as well as C-L␤T2 cells. Thus, p8 appears to be crucial for expression of LH␤ in L␤T2 cells. However, overexpression of p8 does not allow for activation of LH␤-driven luciferase expression in heterologous cells (COS7 and ␣T3-1 cells) or enhanced activation in homologous cells (L␤T2 cells) (data not shown). While there are many potential explanations for the lack of an increase in LH␤ gene expression upon overexpression of p8, the simplest may be that endogenous p8 in L␤T2 cells may already be exerting a maximal effect on the LH␤ gene. In the heterologous cells it is possible that other factors vital to the impact of p8 on FIG. 1. p8 mRNA is higher in cells that represent fully differentiated gonadotropes. A, PCR differential display comparison between ␣T3-1 cells and L␤T2 cells illustrates the band (arrow) found to be up-regulated in cells representing the mature gonadotrope. This band (clone 100) was found to be a partial murine p8 cDNA. B, Northern blot analysis of ␣T3-1 and L␤T2 cell RNA. Ten micrograms of total RNA each extracted from cells indicated across the top were hybridized to radiolabeled murine p8 cDNA. C, ethidium-stained gel used for Northern blot above. Approximately equal RNA loading was observed by the intensity of ethidium staining in each 28 S and 18 S rRNA subunit band. LH␤ promoter activity may be absent. In other words, p8 may be necessary but not sufficient for expression of the LH␤ gene, and the mechanism by which p8 regulates LH␤ gene expression may be very complex and/or quite indirect.
LH␤ mRNA Is Down-regulated in p8-KD-L␤T2 Cells Even in the Presence of GnRH-To determine the importance of p8 on activation of endogenous genes in gonadotrope-derived cells, we extended our investigation to include Northern blot analysis of the LH␤ gene. As shown in Fig. 4, there is a lack of LH␤ gene expression in cells that stably express the p8 antisense strand (p8-KD-L␤T2). In contrast, low levels of LH␤ gene expression were detected in normal, untreated L␤T2 cells, C-L␤T2 cells and p8-OE-L␤T2 cells. To further characterize the impact of p8 removal on LH␤ activity, the various L␤T2 cell lines were treated with 100 nM GnRH for 6 h prior to harvest and collection of RNA for analysis by Northern blot. While this treatment paradigm increased LH␤ gene expression in normal L␤T2, C-L␤T2, and p8-OE-L␤T2 cells, it could not rescue expression of LH␤ in the p8-KD-L␤T2 cells. This provides further functional evidence for the importance of the p8 transcription factor in LH␤ gene expression in L␤T2 cells.
Unlike LH␤, the GnRH receptor gene is expressed in the L␤T2 cells lacking p8 (Fig. 4). In the absence of GnRH stimu-lation, this level of gene expression is higher in p8-KD-L␤T2s than p8-OE-L␤T2, C-L␤T2, or normal L␤T2 cells. However, upon treatment of each cell line with 100 nM GnRH for 6 h, GnRH receptor gene expression is lowest in the p8-KD-L␤T2 cells. While a modest increase in expression was observed in the other cell lines, the GnRH receptor gene does not appear to be GnRH responsive in p8-KD-L␤T2 cells. In addition, other genes (␣GSU, ␤-actin, and Egr-1) that are GnRH-responsive in gonadotropes, normal L␤T2 cells, and C-L␤T2 cells do not show an increase in gene expression upon GnRH stimulation in p8-KD-L␤T2 cells (data not shown), opening the intriguing possibility that p8 impacts the GnRH signaling cascade in L␤T2 cells.
Temporal Expression of p8 Gene in Vivo Mirrors That of the Gonadotrope-derived Cells-Expression of the p8 gene is very low in gonadotrope-derived precursor cells (␣T3-1), while higher levels of expression are seen in cells that represent a later stage of gonadotrope differentiation (L␤T2). To determine whether a similar pattern of temporal expression of p8 mRNA occurs in vivo, in situ hybridization was performed in murine tissues at distinct developmental time points (Fig. 5). Whereas p8 was found to be undetectable in the developing pituitary   FIG. 3. The LH␤ promoter is non-functional when p8 is knocked down in L␤T2 cells. A, C-L␤T2 and p8-KD-L␤T2 stable cell lines were transiently transfected with promoter-driven luciferase constructs consisting of the bovine LH␤ promoter (LH␤-luc), the human ␣GSU promoter (␣GSU-luc), or a promoterless luciferase plasmid (luc). In both the untreated C-L␤T2 and p8-KD-L␤T2 cell lines, the -luc construct showed only very low levels of luciferase activity, as expected. In contrast, the LH␤-luc construct displayed normal luciferase activity in C-L␤T2 cells, ϳ4-fold higher than the promoterless control, but was non-functional in the p8-KD-L␤T2 cells. On the other hand, the ␣GSUluc construct displayed its normal, high levels of luciferase in both untreated cell lines. Values are means Ϯ S.D. of firefly luciferase activity normalized with renilla luciferase activity. B, C-L␤T2 and p8-KD-L␤T2 stable cell lines were transiently transfected with LH␤-luc and treated for 24 h with 100 nM GnRH. The LH␤ promoter remains non-functional in the p8-KD-L␤T2 cells, even in the presence of GnRH. Values are means Ϯ S.D. of firefly luciferase activity normalized with renilla luciferase activity.

FIG. 4. Even in the presence of GnRH, mRNA encoding LH␤ is undetectable in p8-KD-L␤T2 cells.
Northern blot analysis of RNA from several untreated and GnRH-treated (for 6 h) gonadotrope-derived cell lines. Twenty and 10 g of total RNA extracted from several treated and untreated gonadotrope-derived cell lines indicated across the top and bottom were hybridized to radiolabeled murine LH␤ and GnRH receptor (R), respectively. After stripping each blot, they were hybridized with radiolabeled murine 18 S rRNA sequence. Message encoding LH␤ and GnRHR was found to be up-regulated upon GnRH treatment in all cells containing p8 (p8-OE-L␤T2, C-L␤T2, and L␤T2). However, LH␤ mRNA was undetectable in both untreated and GnRH treated p8-KD-L␤T2 cells, while GnRHR mRNA, which was expressed in p8-KD-L␤T2 cells, was not responsive to GnRH. The graph shows relative signal intensity of LH␤ or GnRHR message normalized to 18S rRNA as measured by exposure to a storage phosphor screen and scanning on a densitometer. gland at e13.5 (when gonadotropes are phenotypically similar to ␣T3-1 cells) and e15.5, abundant expression was observed throughout the pituitary gland at e16.5 (prior to detection of LH␤), e17.5 (gonadotropes at this stage have a phenotype similar to that of L␤T2 cells), and e18.5, indicating that expression of the p8 gene in the developing pituitary gland is stage-specific and transient, as it was again undetectable in the pituitary glands of normal adults. Thus, p8 follows a transient pattern of expression, similar to other factors vital to pituitary development (47,48).
As expected, ␣GSU mRNA was detectable throughout the pituitary gland at each embryonic day tested as well as in presumptive gonadotropes and thyrotropes of the adult. In addition, while LH␤ mRNA was observed in the pituitary glands of adult mice and at later stages of embryonic development (e17.5 and e18.5), we were unable to detect the transcript at earlier time points, such as e16.5. While others have been able to detect very low levels of LH␤ gene expression in pituitary glands at e16.5 by use of 35 S-labeled oligonucleotides (49), we were unable to do so using digoxygenin-labeled cRNA. However, we were able to detect p8 expression using the same type of probe at this developmental age. Thus, it appears that expression of the p8 gene precedes LH␤ gene expression in the developing pituitary gland.

DISCUSSION
Like Egr1, SF-1, and Pitx1, which have all been described as playing a role in terminal differentiation of gonadotropes (11-14;18 -20), the p8 nuclear phosphoprotein may play a functional role in LH␤ gene expression as well as a potential role in cellular differentiation. However, unlike Egr1, SF-1, and Pitx1, p8 expression in the developing pituitary gland is transient. High levels of expression can be detected during later embryogenesis, while no p8 is detectable in the pituitary glands of normal adult mice. Thus, p8 may represent a true embryonic factor vital to gonadotrope development and LH␤ gene expression.
Activation of LH␤ gene expression likely involves transcription factors that bind directly to DNA regulatory elements as well as additional transcriptional regulators that influence the binding or activity of these proteins, some of which are also likely to affect the chromatin mechanics of the gene. Clearly, p8, a new member of the HMG family of chromatin-binding proteins, plays a role in expression of the LH␤ gene as its absence in L␤T2 cells corresponds to an absence of LH␤ gene expression. What is not known, however, is how p8 functions to direct this expression.
In addition to their ability to relieve histone H1-mediated repression of transcription, the HMGA class of proteins are also known for their ability to specifically interact with numerous transcription factors to form stereospecific multiprotein enhanceosome complexes (50 -52). This allows the architectural transcription factors considerable flexibility in regulating the expression of a large number of genes (33,53). p8 has been shown to enhance the activity of other transcription factors (54) as well as interact directly with both factors and coactivators of transcription (55). Thus, p8 may function in a similar fashion to activate the LH␤ promoter, forming an enhanceosome with factors known to be important for basal expression of LH␤. In fact, one of the numerous factors that HMGA has been shown to interact with is NF-Y (NF-YA subunit) (56), a factor that is known to be important for activation of LH␤ (27) and FSH␤ (57) gene expression. In this regard, determining the protein partners of p8 in L␤T2 cells may provide insight toward its function in LH␤ gene expression.
Alternatively, or perhaps concurrently, p8 may be playing a role in the remodeling of chromatin during development, moving the gene from transcriptionally inactive heterochromatic regions to euchromatin, allowing for the induction of LH␤ gene expression. This may explain why p8 is not necessary for expression of the LH␤ gene in adult mice and why overexpression of p8 in L␤T2 cells does not enhance LH␤ promoter activity or endogenous LH␤ gene expression. In addition, with its mitogenic potential (44,58) and possible anti-apoptotic role (59), p8 may be involved in gonadotrope cellular proliferation during development. Its absence of expression in the adult may be critical to maintaining/limiting the gonadotrope population within the anterior pituitary gland as p8 has been shown to be up-regulated in proliferating and tumorigenic tissues (44, 59 -62). A similar temporal pattern of expression has been observed with other factors vital to developmental processes (59 -62), including other members of the HMGA family (53).
Determination of pituitary cell types during organogenesis occurs in response to a series of extrinsic and intrinsic signaling molecules. In Rathke's pouch, multiple molecular gradients help to establish distinct patterns of transcription factors that allow for commitment, positional determination, and differentiation of pituitary cell types. For example, Rosenfeld and colleagues have shown that a ventral to dorsal gradient of BMP2 expression in Rathke's pouch induces a ventral to dorsal gradient of GATA2 expression (63,64). The high levels of GATA2 in the most ventral aspect of the developing gland restrict Pit-1 gene expression in presumptive gonadotropes, allowing for induction of the transcription factors that are critical for differentiation of the cell type, including SF-1 (64). In fact, in vivo expression of a dominant-negative GATA2 inhibited terminal differentiation of gonadotropes while extended expression of GATA2 dorsally expanded the gonadotrope population. Perhaps like SF-1, expression of the p8 gene is regulated by GATA2. In this regard, several potential GATA-binding sites are located in the 5Ј flanking region of the murine p8 gene (39) FIG. 5. Temporal expression of p8 gene in vivo mirrors that of the gonadotrope-derived cells. In situ hybridization for detection of mRNA encoding ␣GSU, LH␤, and p8 in developing and adult pituitary glands at e13.5, e15.5, e16.5, e17.5, e18.5, and approximately post-natal day 56. While ␣GSU is detectable in pituitary tissue across all time points tested, LH␤ is detectable beginning at e17.5 through adulthood. On the other hand, detection of p8 mRNA is temporal and transient. p8 appears to arise from the ventral aspect of the developing pituitary gland at e16.5. However, p8 is undetectable in the pituitary glands of normal adult mice (bar ϭ 300 m).
that have yet to be characterized.
While p8 is important for LH␤ gene expression in vitro, a definitive answer to the significance of p8 for gonadotrope maturation and initiation of LH␤ gene expression in vivo will lie in studies involving temporally disregulated expression or targeted disruption of the p8 gene in mice. In regards to the latter, Iovanna and colleagues have recently developed knockout mice to study the role of p8 on cell growth, apoptosis, and tumor development (65,66). No abnormalities were found in organs of these mice by histological analysis, including liver, lung, intestine, pancreas, testis, brain, kidney, and heart (65). To this point, the p8 knock-out mice have yet to be assessed for defects in fertility or gonadotrope development. To determine whether p8 is a dominant member of a signal transduction pathway or serves as a component of a complex mechanism (perhaps even redundant or compensatory) for initiation of LH␤ gene expression, a close examination of the pituitary glands, reproductive fitness, and reproductive system organ development in these mice will be necessary. In addition, the identification and characterization of p8-responsive genes may better elucidate the role of p8 in LH␤ gene expression and the development of gonadotropes from committed precursors to differentiated cells.
Of the factors that appear to be essential for terminal differentiation of gonadotropes and expression of LH␤, p8 is unique among them because of the transient component to its expression. p8 gene expression is detectable at e16.5 in pituitary cells, one developmental day prior to detection of LH␤ mRNA in our system. By its temporal pattern of expression, p8 appears to be a marker of differentiating gonadotropes that acquire the ability to express LH␤. However, p8 appears to be more than simply a marker of the developing gonadotrope as the knockdown of p8 in L␤T2 cells corresponds with an inability of these cells to express the LH␤ gene. Thus, our data indicate that p8 is a stage-specific component of the gonadotrope transcriptome that may play a functional role in LH␤ gene expression during embryonic cellular differentiation.