Endocrinology, doi:10.1210/en.2006-1356
Endocrinology Vol. 148, No. 3 1246-1254
Copyright © 2007 by The Endocrine Society
Analysis of the Gene Regulatory Program Induced by the Homeobox Transcription Factor Distal-less 3 in Mouse Placenta
Li Han,
Monica Dias Figueiredo,
Kathie A. Berghorn,
Terri N. Iwata,
Patricia A. Clark-Campbell,
Ian C. Welsh,
Wei Wang,
Timothy P. O’Brien,
David M. Lin and
Mark S. Roberson
Department of Biomedical Sciences, Cornell University, Ithaca, New York 14853
Address all correspondence and requests for reprints to: Mark S. Roberson Ph.D., T3-004d Veterinary Research Tower, Department of Biomedical Sciences, Cornell University, Ithaca, New York 14853. E-mail: msr14{at}cornell.edu.
 |
Abstract
|
|---|
Dlx3, a homeodomain transcription factor, is essential for placental development in the mouse. The Dlx3–/– mouse embryo dies at embryonic d 9.5–10 putatively due to placental failure. To develop a more comprehensive understanding of the gene profile regulated by Dlx3, microarray analysis was used to determine differences in gene expression within the placenta of Dlx3+/+ and Dlx3–/– mice. Array analysis revealed differential expression of 401 genes, 33 genes in which signal to log ratio values of null/wild-type were lower than –0.5 or higher than 0.5. To corroborate these findings, quantitative real-time PCR was used to confirm differential expression for 11 genes, nine of which displayed reduced expression and two with enhanced expression in the Dlx3–/– mouse. Loss of Dlx3 resulted in a marked reduction (>60%) in mRNA expression of placental growth factor (Pgf), a member of the vascular endothelial growth factor family. Consistent with these results, Pgf secretion from placental explants tended to be reduced in the Dlx3–/– mice, compared with wild type. To investigate mechanisms of Dlx3 regulation of Pgf gene transcription, we cloned 5.2 kb of the Pgf 5' flanking sequence for use in reporter gene assays. Expression of the Pgf promoter luciferase reporter containing at least three Dlx3 binding sites was increased markedly by overexpression of Dlx3 supporting the conclusion that Dlx3 may have a direct effect on Pgf promoter activity. These studies provide a novel view of the transcriptome regulated by Dlx3 in mouse placenta. Dlx3 is specifically required for full expression and secretion of Pgf in vivo. Moreover, in vitro studies support the conclusion that Dlx3 is sufficient to directly modulate expression of the Pgf gene promoter in placental cells.
|
Introduction
|
|---|
THE COMPLEX PROCESS of determination of trophoblast cell lineages within the placenta is integral to the successful establishment of the maternal-fetal interface. Remarkably these tissues functionally give rise to cell populations that control fetal nutrient and gas exchange, participate in maternal-fetal immunological dialog and are critical sources of endocrine factors that communicate to the mother the presence of the conceptus. Failures associated with establishment of the maternal-fetal interface can result in fetal mortality or lead to intrauterine fetal growth abnormalities associated with fetal morbidity and the potential to contribute to the adult onset of diseases such as type 2 diabetes (recently reviewed in Refs. 1, 2, 3). Thus, a more comprehensive understanding of the genetic mechanisms that underlie the development and morphogenesis of the placenta will provide the foundation for understanding mechanisms of fetal morbidity and mortality during early pregnancy and potentially a better understanding of aspects of the origins of adult disease.
A great number of mouse knockout models have been instrumental in defining genetic determinants required for developmental decisions during placental morphogenesis. These have been recently reviewed in elegant detail (4). For example, knockouts of numerous receptors, signaling molecules and transcription factors have been shown to compromise syncytial trophoblast differentiation and morphogenesis of the placental labyrinth, a key site within the mouse placenta of interdigitations between fetal vascular beds and maternal blood spaces (5). Our laboratory has developed a keen interest in one of these transcription factors, Distal-less (Dlx) 3, based on two important observations. First, in human trophoblast cell lines Dlx3 plays a critical role in the basal regulation of the glycoprotein hormone 
-subunit gene, a key subunit of the heterodimeric placental borne hormone, chorionic gonadotropin (CG) (6). CG is expressed exclusively in primate and equine trophoblasts and is an important luteotropin during early human pregnancy. CG also appears to be an important ligand in the regulation of endometrial receptivity (7, 8, 9) through changes in MAPK signaling activity to affect morphology and function of epithelial and stromal compartments of the receptive endometrium (10). Second, in mouse models, disruption of the Dlx3 loci in knockout experiments resulted in fetal death at embryonic day (E) 9.5–10 due to failed placental labyrinth development (11). Dlx3 expression is largely restricted to trophoblasts within the labyrinth of the mouse placenta, and loss of Dlx3 appears to affect the expansion of the labyrinth during midgestation (12). Thus, Dlx3 appears to be playing a critical role in facilitating biosynthesis of a key placental hormone that serves as a luteotropin and a modulator of uterine receptivity in primates as well as being instrumental in facilitating organizational aspects of the mouse labyrinth morphogenesis and the establishment of the maternal-fetal interface.
In the present studies, we have furthered our understanding of Dlx3 expression within the mouse placenta. We have identified a rather large Dlx3-dependent transcriptome by microarray analyses comparing Dlx3+/+ and Dlx3–/– mice. Placental growth factor (Pgf) transcript levels and secretion appears to be markedly reduced by the loss of Dlx3 in these studies. The present studies provide important new evidence that Dlx3 may directly affect transcriptional regulation of Pgf through interactions of Dlx3 and the Pgf gene promoter in choriocarcinoma cells.
|
Materials and Methods
|
|---|
Histological staining of mouse placental sections
Implantation sites from Dlx3+/+ and Dlx3–/– placentas at E9.5 were fixed in 4% paraformaldehyde and embedded in paraffin. Sections were then cut and stained in hematoxylin and eosin to characterize relative differences in labyrinth expansion and morphogenesis.
Microarray
All experimental paradigms using animals in this research were approved by the Cornell University Institutional Animal Care and Use Committee. Male and female Dlx3+/– (heterozygous) mice were time mated and killed 9.5 d after conception. Gravid uteri were dissected free and fetal tissues were harvested for genotyping as previously described (12). The placental disks (with some maternal decidua) were carefully removed and immediately frozen in liquid nitrogen pending completion of genotyping. Total RNA was extracted by homogenizing dissected, thawed tissues in Trizol reagent (Invitrogen, Carlsbad, CA) using a TissueLyser (QIAGEN, Valencia, CA). The RNA was examined for quality and degradation through an Agilent BioAnalyzer by the microarray core facility (Cornell University, Ithaca, NY). Double-stranded cDNA was synthesized from mRNA and in vitro transcription reactions were then used to produce biotin-labeled cRNA from the cDNA using GeneChip IVT labeling kit (Affymetrix, Santa Clara, CA). The labeled cRNA was hybridized with GeneChip mouse genome 430 2.0 array (Affymetrix) and scanned by GeneArray 3000 scanner (Affymetrix). The raw array data of three pairs of Dlx3–/– vs. Dlx3+/+ samples were analyzed using Affymetrix GCOS software to obtain signal, signal to log ratio values, change call, and associated P values, using the wild-type sample in each pair as baseline. Paired t test was performed on signal to log ratio values vs. mean of zero for each gene. Multiple test correction was not used because of the heterogeneous nature of the tissue samples, the relatively small differences in gene expression levels observed between samples, and the limited number of replicates that were performed. We therefore chose an arbitrary P value of 0.05 to generate an initial set of genes for further validation. A set of 401 differentially expressed genes were selected, which had t test P value of less than 0.05 and maximum signal value of at least 512. Gene networks were identified by database analyses (Ingenuity Pathway Analysis; Ingenuity Systems, Redwood City, CA) using all 401 identified genes in the array analysis.
Quantitative real-time PCR
Quantitative (q) real-time PCR using probe based chemistry was used to confirm differential expression of genes identified by array experiments. Briefly, a total of 9.0 µg total mRNA from individual Dlx3+/+ and Dlx3–/– mouse placentas (n = 3/genotype) was reversed transcribed using a high-capacity cDNA archive kit (Applied Biosystems, Foster City, CA). TaqMan universal PCR master mix (Applied Biosystems, Foster City, CA) with uracil-N-glycosylase was used to set up PCRs containing forward and reverse primers, amplicon-specific probes, and 20.0 ng cDNA template in a total volume of 20.0 µl. All PCRs were performed in triplicate. Forward and reverse primers flanking an exon junction were designed by ProbeFinder version 2.10 (Roche Applied Science, Indianapolis, IN) and synthesized by Integrated DNA Technologies Inc. (Coralville, IA). The amplicon-specific fluorogenic probes used were identified from the Roche universal probe library that were labeled with 6-carboxy fluorescein and 6-carboxy-tetra-methyl rhodamine for the 5' reporter and 3' quencher, respectively. The oligonucleotide sequences of primers and identification number for the corresponding Roche mouse universal probe library are listed in Table 1
. PCRs were performed on an ABI 7500 real-time PCR system and analyzed with sequence detection software version 1.3 (Applied Biosystems). Negative controls were run with each sample set, in which the template was replaced by PCR-grade water. ß-Actin was used as the endogenous control for normalization of cDNA input. Relative quantification of gene expression was determined by the 2-(
Ct) method after first determining that the amplification efficiency of all targets and ß-actin was identical (13). We also used commercially available ABI TaqMan gene expression assays to quantify synuclein- (Snca) expression levels as well as confirm the specificity of the Roche probe sets to the hemoglobin gene family members Hba-a1, Hba-x, and Hbb-y (listed in Table 2).
Northern blotting
The Pgf cDNA was amplified from E9.5 mouse placenta using PCR, the PCR product was cloned into pGEM T-Easy vector (Promega Corp., Madison, WI), and nucleotide sequence analysis was used to confirm the Pgf cDNA. Male and female Dlx3+/– (heterozygous) mice were time mated and killed 9.5 d after conception. Gravid uteri were dissected free and embryos were harvested for genotyping as described above. The placental disk (with some maternal decidua) was removed and immediately frozen in liquid nitrogen, pending completion of genotyping. Total RNA was extracted from E9.5 placenta using Trizol reagent (Invitrogen). RNA samples (15 µg) from different genotypes were separated in conventional denaturing formaldehyde-containing agarose gel electrophoresis, ribosomal bands documented, and then RNA transferred to nylon membranes using capillary action. The membranes were prehybridized in Church’s hybridization buffer [7% (wt/vol) sodium dodecyl sulfate, 0.5 M sodium phosphate buffer (pH 7.2), and 10 mM EDTA] overnight. Radiolabeled (32P) Pgf cDNA probe was prepared by random primer labeling to high specific activity and then added to the hybridization buffer. Hybridizations were carried out at 65 C overnight. Membranes were then washed under high stringency and bands were visualized by autoradiography. Densitometry analyses of band intensity were performed using ImageJ (National Institutes of Health, Bethesda, MD).
Placental explant studies and Pgf ELISA
Male and female Dlx3+/– (heterozygous) mice were time mated and pregnant females were killed 9.5 d after conception. Gravid uteri were dissected free and embryos were harvested for genotyping as described above. The placental disks with some maternal decidua [Dlx3+/+ (n = 5) and Dlx3–/– (n = 7)] were removed and immediately placed into explant culture as described previously (12). Media samples were collected at time 0 and after 3 h in this culture system. Media samples were then analyzed for secreted Pgf protein levels using a mouse Pgf ELISA-based immunoassay commercially available (R & D Systems, Minneapolis, MN) as per the manufacturer’s instructions. Media samples were assayed in duplicate in two separate assays. In media samples collected at time 0, Pgf concentrations were below detectable limits of the assay (data not shown). According to the manufacturer, interassay coefficient of variation for this assay was 5.6%. Within-assay variation was reported as 9.5%. The sensitivity of this assay (minimum detectable dose) was 1.49 pg Pgf/ml.
Cloning the Pgf promoter sequences
Mouse Pgf (5.2 kb) promoter was obtained by PCR using mouse genomic DNA and the following primers: 5' primer, 5'-GGTACCGATGAGAGAGGGCGGAAGAGAA-3' and 3' primer, 5'-GGTACCTTCAAGGCACAATCACCATGC-3'. To facilitate cloning, a KpnI restriction enzyme site was incorporated at the end of both 5' and 3' primers. The PCR product was cloned into pGEM T-Easy vector (Promega). The nucleotide sequence was verified by sequence analysis. After KpnI digestion, the promoter fragment was subcloned into a luciferase reporter vector and orientation confirmed using restriction analysis. The Pgf promoter-luciferase reporter gene plasmid was prepared by two cycles through cesium chloride using standard protocols.
Cell culture and transient transfection studies
JEG3 cells were cultured in DMEM supplemented with 10% fetal bovine serum. Before transfection studies, cells were split to 35-mm dishes and subconfluent cultures were used in all experiments. The Pgf promoter-luciferase reporter containing 5.2 kb 5' flanking sequences as described above was used in these studies. The Pgf-Luc reporter was transiently cotransfected into JEG3 cells with increasing doses of Dlx3 expression vector (pKH3-Dlx3) using FuGENE 6 transfection reagent (Roche Diagnostics Inc., Indianapolis, IN) as per the manufacturer’s instructions. All transfections were carried out with an equivalent amount of total DNA by supplementing reactions with the parent vector (pKH3). Twenty-four hours after transfection, luciferase activity was determined using reagents from the dual-luciferase reporter assay system (Promega Corp., Madison, WI). Luciferase activity was standardized by total cell protein content as determined by Bradford assay. All transfection studies were performed in triplicate on at least three separate occasions with similar results. Data are shown as means (n = 3) ± SEM of representative experiments. To determine ectopic Dlx3 expression level in transfected cells, Western blot analysis was performed on cell lysates obtained for luciferase assay. Equal protein amounts of cell lysate were resolved by SDS-PAGE and transferred to nitrocellulose membranes using electroblotting. Membranes were blocked using Tris-buffered saline [10 mM Tris (pH 7.6) and 150 mM NaCl] containing 0.1% Tween 20 and 5% nonfat dried milk. After 1 h of blocking, membranes were incubated with a rabbit polyclonal Dlx3 antiserum or ß-actin as previously described (13), and protein bands representing ectopic expression were visualized using enhanced chemiluminescence (PerkinElmer, Boston, MA).
Identification of putative Dlx3 binding sites within the Pgf 5' flanking sequences and DNA binding assays
Inspection of the Pgf 5' flanking sequences revealed three putative Dlx3 binding sites based on comparison with the consensus binding site sequence A/C/G TAATT G/A C/G. These sites were located at positions –4297, –3061, and –1716 relative to the start site of transcription. The site at position –4297 was a consensus site, whereas the sites at positions –3061 and –1761 were near consensus (a single nucleotide mismatch in the 3' termini of the site). Oligonucleotides (top and bottom strands) were synthesized for each of these sites, annealed, and labeled for use in EMSAs. Recombinant Dlx3 was prepared by in vitro transcription and translation using a wheat germ lysate kit commercially available (Promega) for use in EMSA studies. Nuclear extracts were prepared from JEG3 choriocarcinoma cells and used in EMSA as described previously (13, 14).
Statistical analyses
Statistical methods for the array analyses are described above. Where appropriate for Pgf secretion data and all luciferase data, paired t tests were used to examine statistical differences between means. Statistical significance is reported for each experiment.
|
Results
|
|---|
Genetic perturbation of Dlx3 in knockout mice results in a failure to expand the placental labyrinth
In the mouse, Dlx3 has been shown to be required for normal placental development and morphogenesis (11). Dlx3 is expressed within the trophoblasts of the labyrinth beginning sometime after E8.5 and is clearly detectable by immunohistochemistry on E9.5 (12). Dlx3 was not detectable in giant cell trophoblasts. Targeted deletion of Dlx3 resulted in embryonic lethality at E9.5–10 characterized by a failure in the development of the murine placenta. Consistent with previous observations by Morasso et al. (11), histological examination of the murine placenta on E9.5 revealed a reduction in the thickness of the labyrinth in Dlx3–/– implantation sites, compared with wild-type (Fig. 1A).
View larger version (53K):
[in this window]
[in a new window]
|
FIG. 1. Effect of Dlx3 on labyrinth morphology and gene profile of 33 differentially expressed genes in Dlx3–/– mouse placenta. A, Hematoxylin and eosin staining of E9.5 mouse Dlx3–/– and Dlx3+/+ implantation sites identifying labyrinth (lab; demarcated by dashed lines). Bar, 200 µm. B, Thirty-three genes from microarray analysis whose magnitude of log ratio was lower than –0.5 or higher than 0.5 are depicted. Genes were identified by Affymetrix ID (Gene ID), symbol, and gene name. Each column represents a pair wise sample comparison. The signal to log ratio values of Dlx3–/–/Dlx3+/+ from Affymetrix GCOS software were used in the cluster analysis. The shades of green indicate repressed genes and the shades of red are induced genes. Colored pixels represent the magnitude of the gene response (log fold change).
|
|
Gene profiling of the Dlx3–/– mouse placenta
The Dlx3–/– mouse placenta is characterized by a putative failure in the expansion of the labyrinth compartment, a key developmental step effectively enlarging the fetal vascular compartment and thus the maternal-fetal interface (Fig. 1A
). To better understand the transcriptome regulated by Dlx3 within the labyrinth, we used microarray analysis of placental tissues from the Dlx3–/– mouse model. Our studies focused on the gene profile at E9.5, a time when the placenta appears to be causal in the demise of the developing embryo (11). Mouse Genome 430 2.0 arrays containing 45,000 probe sets analyzing the expression level of approximately 39,000 transcripts were used. From this analysis, we identified 401 differentially expressed genes. Of these 401 genes, RNA levels from 194 genes were shown to be reduced, whereas RNA levels from the remaining 207 genes were increased. Using the criteria of a significant paired t test and a signal to log ratio value of lower than –0.5 and higher than 0.5 when comparing Dlx3–/– to Dlx3+/+ mice, analyses of array data demonstrated differential expression of 33 genes in three pair-wise comparisons (Fig. 1B). RNA levels from 25 of the 33 genes were reduced, whereas the remaining eight genes were expressed at elevated levels in the placenta of Dlx3–/– mice, compared with Dlx3+/+ mice.
To confirm these array findings, 12 selected genes were examined by qPCR methods (Table 3). Selection of these genes was made from the entire set of 401 differentially expressed genes; however, the majority was selected from the list of 33 genes with the greatest fold change in transcript level (Fig. 1B). qPCR results confirmed the fidelity of the array analysis for 11 of 12 genes with consistent fold changes in RNA levels (Table 3). For example, four hemoglobin genes (Hba-a1, Hba-x, Hbb-y, and Hbb-bh1) examined were all reduced in the Dlx3–/– placenta to approximately 30–40% of wild-type RNA levels, whereas lumican and thrombospondin 2 were both elevated 1.8- to 2.2-fold using these two assays. Remarkably we did not obtain consistent results for only 1 (IGF binding protein 5) of the 12 genes examined. We attribute this inconsistency to a lack of an effective mouse universal probe library assay in the qPCR approach rather than true differences between the assays used. Therefore, the qPCR results confirm the accuracy of the array analysis in determining differential gene expression in the placenta in the absence of Dlx3.
To gain further insight into how these differentially regulated genes may be organized into specific gene networks, the entire array data set (401 differentially expressed genes) was subjected to gene network analysis using Ingenuity pathway analysis (Ingenuity.com). This analysis revealed two interrelated gene networks strongly impacted by the loss of Dlx3 (Fig. 2). Loss of Dlx3 resulted in the altered expression of 11 genes within a cell-to-cell signaling and interaction gene network. This network is composed of a number of important signaling ligands/receptors [placental growth factor (Pgf) and epidermal growth factor receptor (EGFR)], signaling enzymes (the low molecular weight GTP binding protein ras, phospholipase D, protein kinase C inhibitor), and extracellular matrix-related molecules [matrix metalloproteinase 9 (MMP9), thrombospondin 2 (THBS2)]. This cell-to-cell signaling and interaction gene network was linked to a second network composed of genes related to hematological disease. These two networks were linked via the putative regulation of hemoglobin genes through EGFR signaling. Within the hematological disease network, loss of Dlx3 resulted in the down-regulation of five specific hemoglobin genes, some of which were embryonic/fetal hemoglobin forms (i.e. Hbb-bh1). Regulation of embryonic/fetal forms of hemoglobin in fetal, nucleated erythrocytes may be consistent with the role of Dlx3 in supporting an emerging fetal vascular compartment associated with the placental labyrinth (11, 12). Thus, changes in the Dlx3-modulated cell-to-cell signaling and interaction network affects another gene network related to hematological disease. These types of gene network relationships provide an important opportunity to examine the specific impact of related genes more broadly on the functional development and morphogenesis of the mouse placenta.
View larger version (28K):
[in this window]
[in a new window]
|
FIG. 2. Dlx3 impact on gene networks. The microarray data (401 differentially expressed genes) were subjected to Ingenuity pathway analysis. Two networks were identified; one was related to hematological disease and the other was related to cell-to-cell signaling and interaction. The genes identified by circles are Dlx3 dependent.
|
|
Pgf gene appears to be an important target of Dlx3 action in placental cells
Examination of our array data immediately drew our attention to changes in expression of the Pgf gene, a member of the vascular endothelial growth factor (VEGF) family. Consistent with qPCR results, Northern blot analysis revealed a decrease in a single 1.7-kb mRNA in the placental tissue when using mouse Pgf cDNA as the probe (Fig. 3A). Densitometric analysis indicated that the expression level of Pgf mRNA in placenta of Dlx3–/– mice was 43% of that in wild-type mice (Fig. 3B). To determine whether changes in Pgf transcript level were functionally coupled to reduced Pgf secretion, we performed placental explant studies (12) using implantation sites from Dlx3+/+ and Dlx3–/– mice. After 3 h in explant culture, concentrations of Pgf detected in explant culture medium tended to be reduced by more than 50% (P = 0.063) in Dlx3–/– placentas, compared with Dlx3+/+ (Fig. 3C) providing an indication of coupled Pgf gene expression and secretion.
View larger version (9K):
[in this window]
[in a new window]
|
FIG. 3. Pgf expression and secretion in placenta of Dlx3+/+ and Dlx3–/– mice. A, Total RNA was isolated from placenta of different genotypes. RNA (15 µg) was hybridized with Pgf cDNA probe. Ethidium bromide staining of 28s and 18s rRNAs was used as the lane loading control. B, Band intensity from the Pgf Northern blots was quantified using ImageJ software and depicted as a histogram. (a and b reflect statistical differences, P < 0.05.) C, Dlx3–/– and Dlx3+/+ implantations were prepared and placed in explant cultures for 3 h. Media samples were assayed for Pgf using a commercially available ELISA-based assay. (a and b reflect statistical tendency, P = 0.063.)
|
|
The Pgf gene 5' flanking region is responsive to Dlx3 overexpression in choriocarcinoma cells
To examine the hypothesis that the Pgf gene promoter may be regulated by Dlx3 at the transcriptional level, we cloned a portion of the Pgf promoter into a luciferase reporter gene construct. The promoter fragment was based on studies previously reported (15) and additional sequences obtained from Ensembl genome browser. Initial studies compared basal expression of the Pgf promoter-luciferase reporter gene activity with other promoter fragments known to be expressed in placental trophoblasts [Esx-1 and Dlx3 (13, 14) or constitutively expressed (thymidine kinase)]. The luciferase reporter containing 5.2 kb of the Pgf gene 5' flanking region displayed transcriptional activity approximately 10-fold higher that the parent luciferase reporter gene. Basal Pgf promoter activity was greater than Esx-1 promoter activity (13) but less than the Dlx3 or thymidine kinase promoters (14) when transfected as luciferase reporters into the JEG3 choriocarcinoma cell line (a trophoblast cell model; data not shown).
Examination of the nucleotide sequences of this Pgf promoter fragment revealed three consensus or near consensus Dlx3 binding sites (using A/C/G TAATT G/A C/G consensus sequence for a Dlx3 binding site). For identification of near-consensus sites, a single nucleotide mismatch outside the central TAATT core of the binding site was considered. These sites were located on the sense strand at positions –4297 (consensus site), –3061, and –1761 (both near consensus sites). We synthesized oligonucleotide probes for the putative Dlx3 binding sites for use in EMSA using recombinant Dlx3 and nuclear extracts purified from JEG3 choriocarcinoma cells to determine whether Dlx3 could bind these elements in vitro (Fig. 4, A and B). EMSA revealed that all of these sites were sufficient to bind recombinant Dlx3. The junctional regulatory element (JRE; a consensus Dlx3 binding site) from the glycoprotein hormone -subunit promoter was used as a positive control in these studies (Fig. 4A). Binding sites at –4297 and –1761 bound recombinant Dlx3 in a manner consistent with the JRE, whereas the site located at –3061 appeared to be a higher relative affinity binding site, compared with the others despite being a near consensus site.
View larger version (15K):
[in this window]
[in a new window]
|
FIG. 4. Dlx3 binds to sites within the Pgf gene promoter. A, EMSAs were used to determine Dlx3 DNA binding to three putative Dlx3 binding sites found within the Pgf 5' flanking region. The positions of these sites (designated as probes) were at –4297 (a consensus site), –3061 (a near consensus site), and –1761 (a near consensus site) relative to the start site of transcription for the Pgf promoter. The JRE from the glycoprotein hormone -subunit was used as a positive control. Recombinant (r) Dlx3 was added in increasing doses (gray triangles) and Dlx3 complexes are designated by the arrow. Free probe at the bottom of the gel is also designated. B, EMSAs were performed using JEG3 nuclear extracts (NE) and the –3061 probe as described above. All binding reactions received the same amount of JEG3 NE. In some binding reactions, NRS, Dlx3, or C/EBP antisera were added in an effort to perturb the Dlx3 binding complex. NRS and C/EBP antisera served as negative controls. To identify the electrophoretic mobility of the Dlx3-containing complex, rDlx3 was added as a standard. Free probe, the Dlx3 complex, and the Dlx3 complex bound to the Dlx3 antibody (Dlx3 supershift) are designated with arrows.
|
|
Using the –3061 binding site, we examined DNA binding complex formation using nuclear extracts from JEG3 choriocarcinoma cells (Fig. 4B). These binding reactions resulted in the formation of discrete complexes with an electrophoretic mobility consistent with recombinant Dlx3. Further using Dlx3 antiserum, these studies demonstrated a Dlx3-like immunoreactivity within the DNA binding complex as measured by a more retarded electrophoretic mobility often referred to as a supershift. In these studies, normal rabbit serum (NRS) and the CCAAT/enhancer-binding protein (C/EBP) antisera were used as negative controls. Similar results were observed using the –4297 and –1761 binding sites and JEG3 nuclear extracts (data not shown). These studies support the conclusion that Dlx3 or a highly similar epitope is present in the binding complexes formed using these sites and JEG3 nuclear extracts. Moreover, these studies support speculation that Dlx3 may have a direct effect on Pgf promoter activity in cells of the trophoblast lineage.
To determine the effect of Dlx3 on the expression of Pgf promoter-luciferase reporter, cotransfection studies in JEG3 choriocarcinoma cells were used. Overexpression of Dlx3 resulted in a dose-dependent increase in expression of epitope tagged Dlx3 (Fig. 5A). Addition of three HA epitopes in frame with the Dlx3 coding sequence resulted in expression of HA-Dlx3, which was slightly larger in molecular size relative to endogenous Dlx3. Recombinant Dlx3 was used as a positive control in these studies. Importantly, Dlx3 overexpression increased Pgf promoter luciferase activity in a dose-dependent manner (Fig. 5B). At the highest dose of Dlx3 expression vector, Pgf promoter-luciferase reporter expression was increased greater than 7-fold. These studies provide clear evidence that Dlx3 is sufficient to modulate expression of the Pgf promoter in cells of trophoblast origin and support the conclusion that the loss of Dlx3 in vivo likely impacts Pgf production at the transcriptional level.
View larger version (12K):
[in this window]
[in a new window]
|
FIG. 5. Dlx3 overexpression is sufficient to activate the Pgf gene promoter in JEG3 cells. A, Transient overexpression studies were carried out using the vector control (2 µg) or increasing doses of an expression vector for hemagglutinin (HA) epitope-tagged Dlx3 (HA-Dlx3). Whole-cell lysates were prepared and subjected to Western blot analysis using the Dlx3 antibody to determine the extent of ectopic expression of HA-Dlx3. HA-Dlx3 migrates slightly slower than endogenous Dlx3 due to the addition of three HA epitopes. Recombinant (r)Dlx3 was used as a positive control for the Western blot. To control for lane loading, the samples were reprobed with an antibody to actin. B, Transient transfection studies in JEG3 cells were used to determine the effect of Dlx3 on Pgf promoter activity. Cotransfection of the Pgf gene promoter-luciferase reporter with increasing doses of Dlx3 expression vector (0, 0.5, 1.0, 1.5, and 2.0 µg) resulted in a dose-dependent increase in expression of the Pgf luciferase reporter. Data are reported as relative luciferase activity standardized by total cellular protein content from representative studies (n = 3/treatment). (a and b reflect statistical differences P < 0.05.)
|
|
|
Discussion
|
|---|
Dlx3 is a member of a larger group of homeobox containing transcription factors within the Dlx family (16). Pairs of Dlx family members are grouped in clusters aligned with Hox genes on separate chromosomes in mouse and human. For example, Dlx3 and Dlx4 (also referred to as Dlx7) are tandemly arrayed on human chromosome 17 separated by approximately 17.6 kb of intervening sequence and aligned with the Hox B cluster (17). Dlx3 and 4 are both expressed within placental trophoblasts (18). The Dlx3–/– mouse dies in utero at E9.5–10, presumably due to failed placental development (11). Morphologically, the labyrinth of the Dlx3–/– placenta appears compact and does not appear to enlarge at a similar rate, compared with wild-type mice potentially reflecting a defect in the expansion of the fetal vascular bed (Fig. 1A). Morasso et al. (11) found that defects in both yolk sac and fetal vasculature within the labyrinth were evident in the Dlx3–/– placenta; however, it remains unclear whether this is causal in the demise of the developing embryo. Placental trophoblast expression of Dlx4 (18) is likely not simply redundant because the presence of Dlx4 in the Dlx3–/– mouse embryo cannot rescue the placental phenotype providing evidence for the distinct role of Dlx3 in placental morphogenesis. In primate species, Dlx3 expression is conspicuous throughout gestation within syncytial and cytotrophoblasts in microvilli in human placenta (14) and baboon placental samples obtained throughout gestation (Fazleabas, A., and M. S. Roberson, unpublished observations). This is consistent with the endocrine role of Dlx3 in directing transcriptional regulation of the glycoprotein hormone -subunit of CG, whose expression is restricted to trophoblasts of primates and horses. The mouse placenta does not produce a similar placental borne gonadotropin; as a consequence questions regarding a potential role of Dlx3 in regulating placental endocrine function have not been fully explored.
The present studies are the first to provide insight into the underlying transcriptome modulated by Dlx3 in the developing placenta. The loss of Dlx3 impacts the expression of a functionally diverse group of genes. For example, genes affected by Dlx3 can be categorized as ligands (such as reduced expression of Pgf), cell signaling molecules (such as down-regulation of Rab6B, Rac/Cdc42 guanine nucleotide exchange factor 6, and up-regulation prostaglandin E receptor 3), regulators of cytoarchitecture (such as down-regulation of keratin complex 2 and phosphatase and actin regulator 1), and extracellular matrix-related molecules (such as down-regulation of MMP9 and up-regulation of lumican and thrombospondin 2). These changes in gene expression point to a spectrum of distinct cellular processes dependent on transcriptional activity and regulation by Dlx3. Focusing on more global changes, we identified a potential impact of loss of Dlx3 on gene networks related to cell-to-cell signaling and interactions and a cohort of hemoglobin genes related to hematological disease. This gene network analysis lends support to the conclusion that Dlx3 may be involved in development of the fetal vascular bed consistent with a reduction in labyrinth depth. During mouse development, initial hematopoiesis begins in the yolk sac around E7.0 within extraembryonic blood islands (19). Later (E11.0/12.0), fetal liver becomes the primary site of hematopoiesis. Because it is unlikely that the fetal hemoglobin genes observed in our array analysis are specifically expressed in labyrinth trophoblasts (the specific site of Dlx3 expression), it seems more likely that hemoglobin mRNA expression occurs in fetal nucleated erythrocytes gradually populating the labyrinth vascular bed primarily from yolk sac at E9.5. Consistent with this notion, fetal erythrocytes were greatly reduced within the labyrinth vasculature in the Dlx3–/– mouse at E12.5 (11). Thus, it is reasonable to suspect that loss of the hemoglobin mRNAs in the Dlx3–/– placenta in our array analysis is consistent with reduced fetal vasculature and potential for reduced fetal erythrocytes within the labyrinth of the midgestation placenta.
One of the more notable gene transcripts identified by our array analysis was Pgf. Array analysis, qPCR, and Northern analysis all consistently demonstrated a reduced expression of the Pgf transcript in the Dlx3–/– placenta. Consistent with coupled biosynthesis and secretion, the Dlx3–/– implantation sites tended to secrete less Pgf than wild-type placentas in explant cultures, suggesting a functional consequence of reduced Pgf gene expression in vitro. Our additional studies based on the array results implicate Dlx3 as a putative direct activator of the Pgf gene promoter in studies using a choriocarcinoma cell model. Examining the initial 5.2 kb of the mouse Pgf 5' flanking sequence revealed several consensus or near consensus Dlx3 binding sites (Fig. 4). All of these sites could bind Dlx3 in vitro, albeit with variable relative affinity and overexpression of Dlx3, was sufficient to increase transactivation of the Pgf luciferase reporter gene (Fig. 5). Interestingly, the extent of Dlx3 overexpression necessary to induce Pgf promoter-luciferase reporter transcriptional activation was higher than expected. This may be related to our recent observations that the inhibitory Smad, Smad6, can functionally interact with Dlx3 and prevent Dlx3 DNA binding and thus transactivation of target genes in choriocarcinoma cells (13). In our Dlx3 overexpression studies, high doses of Dlx3 expression vector leading to relatively high levels of Dlx3 may have been necessary to overcome the negative effects of Smad6 interaction. As the dose of Dlx3 increased, the stoichiometry between Dlx3 and Smad6 would be shifted in favor of unbound Dlx3 and a positive transcriptional signal. Thus, our in vivo Dlx3–/– mouse implicates Dlx3 as a necessary factor for the expression and secretion of Pgf. Using in vitro and cell culture models, our studies provide direct evidence that Dlx3 is sufficient to bind to and transactivate the Pgf gene promoter.
As discussed earlier, Pgf is a member of the VEGF gene family and is thought to play a role in placental and endometrial angiogenic potential (20, 21). The Pgf receptor is expressed within the endothelial compartment of the labyrinth and in placental trophoblasts, suggesting a complex role of Pgf in placental cell biology. In the human, alternative splicing of the Pgf transcript generates four isoforms: Pgf-1, Pgf-2, Pgf-3, and Pgf-4 (22, 23, 24, 25). In mouse, only one Pgf mRNA encoding the equivalent of human Pgf-2 has been identified (24). Structurally related to VEGF, Pgf binds to VEGF receptor (VEGFR) 1 and appears to be involved in modulation of vascular permeability during pathological angiogenesis (26). The mechanistic role of Pgf and VEGFR1 during placental development and angiogenesis has not been fully elucidated. Pgf–/– mice appear to be fertile without evidence of placental vascular pathologies (26); however, Pgf has been found to be an important marker for preeclampsia in women, a syndrome of late pregnancy characterized by maternal hypertension, proteinuria, insufficient trophoblast invasiveness, a reduction of Pgf levels, and increased soluble antiangiogenic fms-like tyrosine kinase activity during mid- to late gestation. All of these symptoms are resolved with expulsion of the placenta (27, 28, 29, 30, 31, 32). An important mouse model recapitulating many of the cardiovascular symptoms of preeclampsia has recently been described (33, 34). Pgf expression is markedly reduced in this mouse model of preeclampsia as well. The possibility exists that Dlx3-dependent changes in Pgf may be related to the molecular determinants driving the progression of events leading to preeclampsia. Additional studies will be required to fully appreciate this possibility.
Dlx3-dependent biosynthesis and secretion of Pgf also affords a potential mechanism for the indirect actions of Dlx3 on target gene expression. Because Pgf is a secreted angiogenic/growth factor and the VEGFR1 is expressed in cell types that do not necessarily express Dlx3, it is conceivable that Dlx3-dependent loss of Pgf would have a marked impact on other genes within the Dlx3-dependent transcriptome through activation of VEGFR1. We would predict that some proportion of the genes identified by our array analysis are indirectly regulated by Dlx3 through Pgf-induced signaling and transcriptional activation. This may be of particular importance in situations like preeclampsia in which Pgf appears to be a significant component of the progression of this disease in women.
|
Acknowledgments
|
|---|
The authors thank Dr. Maria I. Morasso for generously providing the Dlx3–/– mice and Dr. Asgi Fazleabas for critical reading of this manuscript. We are indebted to Ms. Gillian Angliss for help with sectioning and staining mouse placental tissues.
|
Footnotes
|
|---|
This work was supported by a grant (HD39354) from the National Institutes of Child Health and Human Development (to M.S.R.), a Fellowship from Cornell University Center for Vertebrate Genomics (to L.H.), and a summer Fellowship from the Cornell University Veterinary Investigator Program (to T.N.I.).
Disclosure Statement: The authors have nothing to disclose.
First Published Online November 16, 2006
Abbreviations: C/EBP, CCAAT/enhancer-binding protein; CG, chorionic gonadotropin; Dlx, Distal-less; E, embryonic day; EGFR, epidermal growth factor receptor; JRE, junctional regulatory element; MMP9, matrix metalloproteinase 9; NRS, normal rabbit serum; Pgf, placental growth factor; q, quantitative; Snca, synuclein-; THBS2, thrombospondin 2; VEGF, vascular endothelial growth factor; VEGFR, VEGF receptor.
Received October 4, 2006.
Accepted for publication November 9, 2006.
|
References
|
|---|
- Simmons R 2006 Developmental origins of adult metabolic disease. Endocrinol Metab Clin North Am 35:193–204[CrossRef][Medline]
- Simmons DG, Cross JC 2005 Determinants of trophoblast lineage and cell subtype specification in the mouse placenta. Dev Biol 284:12–24[CrossRef][Medline]
- Jones RL, Stoikos C, Findlay JK, Salamonsen LA 2006 TGF-ß superfamily expression and actions in the endometrium and placenta. Reproduction 132:217–232[Abstract/Free Full Text]
- Cross JC 2005 How to make a placenta: mechanisms of trophoblast cell differentiation in mice—a review. Placenta 26(Suppl A):S3–S9
- Adamson SL, Lu Y, Whiteley KJ, Holmyard D, Hemberger M, Pfarrer C, Cross JC 2002 Interactions between trophoblast cells and the maternal and fetal circulation in the mouse placenta. Dev Biol 250:358–373[CrossRef][Medline]
- Roberson MS, Meermann S, Morasso MI, Mulvaney-Musa JM, Zhang T 2001 A role for the homeobox protein Distal-less 3 in the activation of the glycoprotein hormone subunit gene in choriocarcinoma cells. J Biol Chem 276:10016–10024[Abstract/Free Full Text]
- Fazleabas AT, Donnelly KM, Srinivasan S, Fortman JD, Miller JB 1999 Modulation of the baboon (Papio anubis) uterine endometrium by chorionic gonadotrophin during the period of uterine receptivity. Proc Natl Acad Sci USA 96:2543–2548[Abstract/Free Full Text]
- Lobo SC, Srisuparp S, Peng X, Fazleabas AT 2001 Uterine receptivity in the baboon: modulation by chorionic gonadotropin. Semin Reprod Med 19:69–74[CrossRef][Medline]
- Filicori M, Fazleabas AT, Huhtaniemi I, Licht P, Rao C, Tesarik J, Zygmunt M 2005 Novel concepts of human chorionic gonadotropin: reproductive system interactions and potential in the management of infertility. Fertil Steril 84:275–284[CrossRef][Medline]
- Srisuparp S, Strakova Z, Brudney A, Mukherjee S, Reierstad S, Hunzicker-Dunn M, Fazleabas AT 2003 Signal transduction pathways activated by chorionic gonadotropin in the primate endometrial epithelial cells. Biol Reprod 68:457–464[Abstract/Free Full Text]
- Morasso MI, Grinberg A, Robinson G, Sargent TD, Mahon KA 1999 Placental failure in mice lacking the homeobox gene Dlx3. Proc Natl Acad Sci USA 96:162–167[Abstract/Free Full Text]
- Berghorn KA, Clark PA, Encarnacion B, Deregis CJ, Folger JK, Morasso MI, Soares MJ, Wolfe MW, Roberson MS 2005 Developmental expression of the homeobox protein Distal-less 3 and its relationship to progesterone production in mouse placenta. J Endocrinol 186:315–323[Abstract/Free Full Text]
- Berghorn KA, Clark-Campbell PA, Han L, McGrattan M, Weiss RS, Roberson MS 2006 Smad6 represses Dlx3 transcriptional activity through inhibition of DNA binding. J Biol Chem 281:20357–20367[Abstract/Free Full Text]
- Holland MP, Bliss SP, Berghorn KA, Roberson MS 2004 A role for CCAAT/enhancer-binding protein ß in the basal regulation of the distal-less 3 gene promoter in placental cells. Endocrinology 145:1096–1105[Abstract/Free Full Text]
- Zhang H, Palmer R, Gao X, Kreidberg J, Gerald W, Hsiao L, Jensen RV, Gullans SR, Haber DA 2003 Transcriptional activation of placental growth factor by the forkhead/winged helix transcription factor FoxD1. Curr Biol 13:1625–1629[CrossRef][Medline]
- Beanan MJ, Sargent TD 2000 Regulation and function of Dlx3 in vertebrate development. Dev Dyn 218:545–553[CrossRef][Medline]
- Sumiyama K, Irvine SQ, Ruddle FH 2003 The role of gene duplication in the evolution and function of the vertebrate Dlx/distal-less bigene clusters. J Struct Funct Genomics 3:151–159[CrossRef][Medline]
- Quinn LM, Johnson BV, Nicholl J, Sutherland GR, Kalionis B 1997 Isolation and identification of homeobox genes from the human placenta including a novel member of the Distal-less family, DLX4. Gene 187:55–61[CrossRef][Medline]
- Car BD, Eng VM 2001 Special considerations in the evaluation of the hematology and hemostasis of mutant mice. Vet Pathol 38:20–30[Abstract/Free Full Text]
- Desai J, Holt-Shore V, Torry RJ, Caudle MR, Torry DS 1999 Signal transduction and biological function of placenta growth factor in primary human trophoblast. Biol Reprod 60:887–892[Abstract/Free Full Text]
- Arroyo J, Torry RJ, Torry DS 2004 Deferential regulation of placenta growth factor (PlGF)-mediated signal transduction in human primary term trophoblast and endothelial cells. Placenta 25:379–386[CrossRef][Medline]
- Cao Y, Ji WR, Qi P, Rosin A, Cao Y 1997 Placenta growth factor: identification and characterization of a novel isoform generated by RNA alternative splicing. Biochem Biophys Res Commun 235:493–498[CrossRef][Medline]
- Maglione D, Guerriero V, Viglietto G, Ferraro MG, Aprelikova O, Alitalo K, Del VS, Lei KJ, Chou JY, Persico MG 1993 Two alternative mRNAs coding for the angiogenic factor, placenta growth factor (PlGF), are transcribed from a single gene of chromosome 14. Oncogene 8:925–931[Medline]
- DiPalma T, Tucci M, Russo G, Maglione D, Lago CT, Romano A, Saccone S, Della VG, De GL, Dragani TA, Viglietto G, Persico MG 1996 The placenta growth factor gene of the mouse. Mamm Genome 7:6–12[CrossRef][Medline]
- Yang W, Ahn H, Hinrichs M, Torry RJ, Torry DS 2003 Evidence of a novel isoform of placenta growth factor (PlGF-4) expressed in human trophoblast and endothelial cells. J Reprod Immunol 60:53–60[CrossRef][Medline]
- Luttun A, Brusselmans K, Fukao H, Tjwa M, Ueshima S, Herbert JM, Matsuo O, Collen D, Carmeliet P, Moons L 2002 Loss of placental growth factor protects mice against vascular permeability in pathological conditions. Biochem Biophys Res Commun 295:428–434[CrossRef][Medline]
- Torry DS, Wang HS, Wang TH, Caudle MR, Torry RJ 1998 Preeclampsia is associated with reduced serum levels of placenta growth factor. Am J Obstet Gynecol 179:1539–1544[CrossRef][Medline]
- Tidwell SC, Ho HN, Chiu WH, Torry RJ, Torry DS 2001 Low maternal serum levels of placenta growth factor as an antecedent of clinical preeclampsia. Am J Obstet Gynecol 184:1267–1272[CrossRef][Medline]
- Thadhani R, Mutter WP, Wolf M, Levine RJ, Taylor RN, Sukhatme VP, Ecker J, Karumanchi SA 2004 First trimester placental growth factor and soluble fms-like tyrosine kinase 1 and risk for preeclampsia. J Clin Endocrinol Metab 89:770–775[Abstract/Free Full Text]
- Shibata E, Rajakumar A, Powers RW, Larkin RW, Gilmour C, Bodnar LM, Crombleholme WR, Ness RB, Roberts JM, Hubel CA 2005 Soluble fms-like tyrosine kinase 1 is increased in preeclampsia but not in normotensive pregnancies with small-for-gestational-age neonates: relationship to circulating placental growth factor. J Clin Endocrinol Metab 90:4895–4903[Abstract/Free Full Text]
- Sankaralingam S, Arenas IA, Lalu MM, Davidge ST 2006 Preeclampsia: current understanding of the molecular basis of vascular dysfunction. Expert Rev Mol Med 8:1–20[Medline]
- Robinson CJ, Johnson DD, Chang EY, Armstrong DM, Wang W 2006 Evaluation of placenta growth factor and soluble Fms-like tyrosine kinase 1 receptor levels in mild and severe preeclampsia. Am J Obstet Gynecol 195:255–259[CrossRef][Medline]
- Davisson RL, Hoffmann DS, Butz GM, Aldape G, Schlager G, Merrill DC, Sethi S, Weiss RM, Bates JN 2002 Discovery of a spontaneous genetic mouse model of preeclampsia. Hypertension 39:337–342[Abstract/Free Full Text]
- Dokras A, Hoffmann DS, Eastvold JS, Kienzle MF, Gruman LM, Kirby PA, Weiss RM, Davisson RL 2006 Severe feto-placental abnormalities precede the onset of hypertension and proteinuria in a mouse model of preeclampsia. Biol Reprod 75:899–907[Abstract/Free Full Text]
Top
Abstract
Introduction
