Microarray analysis of trophoblast differentiation: gene expression reprogramming in key gene function categories
BRUCE J. ARONOW,
BRIAN D. RICHARDSON and
STUART HANDWERGER
Departments of Endocrinology and Molecular and Developmental Biology, Childrens Hospital Research Foundation and Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio 45229-2029
 |
ABSTRACT
|
---|
Placental development results from a highly dynamic differentiation program. We used DNA microarray analysis to characterize the process by which human cytotrophoblast cells differentiate into syncytiotrophoblast cells in a purified cell culture system. Of 6,918 genes analyzed, 141 genes were induced and 256 were downregulated by more than 2-fold. Dynamically regulated genes were divided by the K-means algorithm into 9 kinetic pattern groups, then by biologic classification into 6 overall functional categories: cell and tissue structural dynamics, cell cycle and apoptosis, intercellular communication, metabolism, regulation of gene expression, and expressed sequence tag (EST) and function unknown. Gene expression changes within key functional categories were tightly coupled to morphological changes. In several key gene function categories, such as cell and tissue structure, many gene members of the category were strongly activated while others were strongly repressed. These findings suggest that differentiation is augmented by "categorical reprogramming" in which the function of induced genes is enhanced by preventing the further synthesis of categorically related gene products.
placental development; gene regulation; pregnancy
 |
INTRODUCTION
|
---|
THE TROPHOBLAST LAYER of the human placental villous plays a critical role during pregnancy, regulating the exchange of substrates, gases, and other factors between the maternal and fetal circulations (2). Trophoblast cells also synthesize and secrete many protein and steroid hormones and growth factors important for the regulation of maternal and fetal metabolism and growth. The layer is composed of two cell types, syncytiotrophoblast and cytotrophoblast cells. Syncytiotrophoblast cells form the continuous, uninterrupted, multinucleated epithelium-like surface of the placental villous that separates maternal blood from the villous interior. Mononuclear cytotrophoblast stem cells (Langhans cells) are located between the syncytiotrophoblast layer and its basement membrane. During trophoblast differentiation, underlying cytotrophoblast cells proliferate and fuse to form a multinucleated syncytium (for review, see Ref. 2).
At present, the regulatory circuitry that controls trophoblast lineage determination and villous cytotrophoblast cell differentiation into syncytiotrophoblast cells is poorly understood. Many dynamic processes of trophoblast differentiation have been studied successfully using trophoblast cell cultures as a model system (14, 23, 26). Mononucleated cytotrophoblast cells isolated by enzymatic dispersion of placental tissue aggregate in culture and fuse to form a multinucleated syncytiotrophoblast that synthesizes and secretes human placental lactogen (hPL), human chorionic gonadotropin (hCG), and other protein and steroid hormones. This recapitulates important activities accomplished by normal cytotrophoblast cells during in vivo maturation. This in vitro model has implicated a critical relationship between the induction of hPL and hCG and several other placental hormones with the differentiation of cytotrophoblast cells into syncytiotrophoblast cells (4, 13). Epidermal growth factor (EGF; 15), hCG (29), leukemia inhibitory factor (LIF; 3), colony-stimulating factor-1 (CSF-1; 22), granulocyte-macrophage CSF (GM-CSF; 10), insulin-like growth factor-I (IGF-I; 16), and cAMP (32) have been shown to stimulate cytotrophoblast differentiation in vitro. Transforming growth factor-ß1 (TGF-ß1) alters the pathway of trophoblast differentiation from a villous syncytiotrophoblast phenotype to an anchoring phenotype (9, 17, 21). Tumor necrosis factor-
(TNF-
) and interferon-
have been shown to induce, and EGF to inhibit, trophoblast cell apoptosis in vitro (12, 17). Several transcription factors are implicated in the regulation of trophoblast differentiation in the mouse, including Hand1, Mash2, and other basic helix-loop-helix transcription factors (6, 28). However, in the human, little is known about the transcription factors involved in trophoblast differentiation, and the molecular mechanisms that underlie differentiation and apoptosis in the placenta are unclear.
Several studies have begun to define groups of genes that are induced during placental differentiation. Morrish and coworkers (19) used a subtraction cDNA library between undifferentiated and differentiating cytotrophoblast cells to identify six novel genes and four known syncytial products [hCG
, pregnancy-specific ß1-glycoprotein 1 (PSGP1), 3ß-hydroxysteroid dehydrogenase, and plasminogen activator inhibitor type I] whose mRNAs increased during differentiation. Ten other genes were also identified that increased during differentiation. Five of these (keratin 19, calreticulin, heat shock protein 27, serum and glucocorticoid-regulated kinase, and adrenomedullin) were not reported to be expressed in placenta. The other induced genes included keratin 8, fibronectin, mitochondrial ATP synthase, and superoxide dismutase-1. Dizon-Townson and coworkers (7) recently found 17 of 186 random clones of a cDNA library from first trimester placenta that represented potentially novel placental genes, which have not as yet been characterized.
Xu and coworkers (33) have used the human choriocarcinoma cell line BeWo to study differentiation of human placental trophoblast cells in culture. Using differential display analysis, they identified seven genes that were induced during differentiation, four of which were novel compared with sequences in the GenBank database. The three other genes encoded human cytochrome P-450 IIC, inosine monophosphate dehydrogenase type II, and reducing agent and tunicamycin-responsive proteins (RTP). The function of many of these proteins in the placenta remains unclear.
We have previously shown that cytotrophoblast cells cultured in the presence of human maternal serum differentiate into a syncytiotrophoblast-like phenotype that expresses greater amounts of hPL and hCG than cells grown in the presence of other serums (23). This suggests that in vitro differentiation under these conditions parallels the in vivo differentiation of syncytiotrophoblast cells from cytotrophoblast cells. Using this model system, we have recently demonstrated a critical role for transcription factors AP-2 (25) and NF-IL6 (30) in the induction of hPL gene expression and villous cytotrophoblast differentiation. In the present study, we have used cDNA microarray technology to test the hypothesis that the previously observed activation of a small cadre of target genes under the control of specific factors represents a larger transcriptional program. The results have allowed us to observe multiple kinetic patterns of accumulation and decline in mRNA levels and implicate a series of genes not previously known to play roles in cytotrophoblast differentiation and placental development. In addition to providing groups of new genes to consider for their role in the health and differentiation of the placenta, our results point to a fundamental role for the selective loss of mRNAs during the differentiation process.
 |
MATERIALS AND METHODS
|
---|
Preparation and Culture of Cytotrophoblast Cells
Third trimester placentas were obtained from women with normal pregnancies and deliveries, and cytotrophoblast cells were isolated by enzymatic disaggregation and cultured essentially as described previously (23) using a modification of the pancreatin/protease method described by Bax et al. (1), except that the cytotrophoblast cells were purified by negative CD9 selection (18). The protocol for obtaining placentas was approved by the Human Investigation Committees of the University of Cincinnati and the Childrens Hospital Medical Center. The medium was changed at 12 h, when almost all cells were adherent to the bottom of the dish, and then at successive 24-h intervals until the experiment was stopped at 6.5 days after plating.
RNA Preparation
Poly(A)+ RNA from cultured trophoblast cells was obtained using the Oligotex mRNA isolation kit (Qiagen, Valencia, CA) according to the manufacturers instructions, following initial total RNA isolation (5). mRNA was subjected to two passes through the oligo d(T) columns.
Northern blot analysis.
Verification of microarray results was accomplished using RNA samples from independent trophoblast cultures via Northern blot analyses of fibronectin, prostate differentiation factor, cytochrome P-450 XIA, IGF binding protein-10 (IGFBP-10),
-actin (smooth muscle), and an expressed sequence tag (EST) (GenBank accession number AL048161) as previously described (24) except that 2 µg of total RNA was loaded. Equally loaded and membrane-transferred RNA was verified by methylene blue staining of ribosomal bands. Sizes of mRNA signals were determined by comparison with RNA standards (GIBCO-BRL).
RT-PCR analysis.
The relative expressions of hPL, hCG
, hCGß, corticotropin releasing hormone (CRH), carcinoembryonic antigen gene family member 6 (CEA6), PSGP1, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNAs were determined by semi-quantitative RT-PCR as described previously (31). The sense and antisense primers used in the PCR reactions are shown in Table 1.
DNA Microarray
DNA microarray analyses were performed using cyanin-3 (Cy3) and cyanin-5 (Cy5)-labeled probes prepared from poly(A)+ mRNA from human trophoblast cells after 1, 2, 3, 4, and 6 days of culture. Microarray hybridizations were performed by Incyte Genomics (Palo Alto, CA) using the human library GEM-V. The mRNAs were made from two enriched preparations of cytotrophoblast cells from two independent placentas as described above. One preparation was used to generate days 0 and 6 samples; another preparation was used to generate days 0, 1, 2, 3, and 4 samples. For each preparation, the reference "time 0" mRNA (Cy3) was prepared from cells 12 h after initial plating. This preincubation allows for the release and removal of nonadherent, damaged, and dying cells that might have confounded data interpretation.
Primary data were examined using Incyte Gemtools software and GeneSpring software (Silicon Genetics, Redwood City, CA). Defective cDNA spots (irregular geometry, scratched, or <40% area compared with average) were eliminated from the data set. Data sets were subjected to normalization within each microarray experiment such that the median of the Cy5 channel was balanced against the ratio of the Cy3 channel. Each microarray contained 192 control genes present as nonmammalian single gene "spikes" or "complex targets." The complex targets consisted of probe sets that contain a pool of cellular genes expressed in most cell types. In addition, each experimental mRNA sample was augmented with incremental amounts of nonmammalian gene RNA (2x, 4x, 16x, etc.) to permit assessment of the dynamic range attained within each microarray. Little variation was observed across the microarray series with respect to the 192 control genes (not shown), providing support for interarray comparisons of temporally regulated genes.
We assembled a pool of genes that exhibited significant change in their expression during the cell culture period using a multistep procedure. Genes were selected for consideration based on whether their apparent change in their expression relative to the time 0 reference sample was greater than two standard deviations at one time point or more than one standard deviation change at two time points. Because of the highly correlated behavior and the low average deviation of the microarray series, genes exhibiting a change in expression of as little as 60% could be evaluated for their behavior over the entire experimental time course. However, we chose to limit our subsequent analyses to the 397 genes that were in the upper 75th percentile in their level signal intensity in either Cy3 or Cy5 and whose expression changed by more than 2-fold at one or more times. These dynamically regulated genes were clustered according to their expression pattern dynamics by subjecting the log2-transformed data set [R = log2(xt=i/xt=0), where R is the expression ratio for each gene] to the K-means (11) and hierarchical tree clustering algorithms (8) as implemented in the GeneSpring program (Silicon Genetics). The hierarchical tree analysis was performed using a minimum distance value of 0.001, a separation ratio of 0.5, and the standard correlation distance definition. The hierarchical tree structure was used to suggest group numbers for K-means clusters, and this was empirically confirmed by observation of similarity patterns among the resulting sets. Too few groups resulted in sets composed of heterogeneously behaving genes, and too many groups led to adjacent sets that exhibited indistinguishable behavior. Genes were classified into biological function groups by adaptation of the gene ontology classification of Drosophila gene products (20) and iterative PubMed exercise. All gene expression data, cluster groups, and the functional categories of the dynamically regulated genes have been placed onto our microarray database web server (http://genet.chmcc.org).
The reliability of microarray quantitative data was independently corroborated by the use of RT-PCR or Northern blot analysis of the mRNAs used in the microarray experiments, as well as by replicate analyses using additional cell and mRNA preparations. As shown in Fig. 1, the expression of CRH, cytochrome P-450, PSGP1, and CEA6 mRNAs based on RT-PCR or Northern analysis followed the same induction profiles as determined by the microarray analyses. Similar results were obtained for fibronectin, collagen type I receptor, and several ESTs (data not shown). All patterns shown by RT-PCR and Northern blots agreed with the patterns indicated by microarray analyses.

View larger version (21K):
[in this window]
[in a new window]
|
Fig. 1. Comparison of microarray, RT-PCR, and Northern analyses for the determination of gene expression profiles of representative genes during trophoblast differentiation. The patterns of corticotropin releasing hormone (CRH), carcinoembryonic antigen (CEA), and pregnancy-specific ß1-glycoprotein 1 (PSGP1) mRNA levels determined by microarray analysis are compared with the patterns of expression determined by RT-PCR from the same mRNA samples. Independent placental isolations and cultures of trophoblast cells also showed consistent pattern similarity between microarray analysis and Northern analysis for cytochrome P-450 subfamily XIA.
|
|
 |
RESULTS
|
---|
Trophoblast cells grown in the presence of second trimester maternal serum underwent a progressive change in morphology and the induction of hPL, hCG
, and hCGß gene expression as previously reported by our laboratory (23). Twelve hours following plating, greater than 90% of the cells had attached (time 0). By day 1, the cytotrophoblast cells had formed large aggregates, 20% of which had fused to form multinucleated cells that contained two or three nuclei. By day 2,
50% were multinucleated containing five or more nuclei. By day 4, greater than 95% of the cells were multinucleated, and large syncytiotrophoblast cells appeared to form a network with numerous nuclei arranged in linear arrays. Some clusters consisted of cells with only a few nuclei; most were in larger clusters of 10 or more nuclei. This pronounced and relatively homogeneous behavior recapitulates the dynamics of cytotrophoblast differentiation in vivo.
The levels of hPL, hCG
, and hCGß mRNA, measured by semi-quantitative RT-PCR, increased markedly during the 6-day culture period, with net induction similar to that previously reported by our laboratory (23) (Fig. 2). The hPL, hCGß, and hCG
mRNAs were barely detectable at 0 and 1 days and then increased significantly over the next 3 days, reaching peak levels at 4, 4, and 5 days, respectively. The peak amounts of hPL, hCG
, and hCGß mRNAs were 36.8-, 10.1-, and 60.4-fold greater, respectively, than on day 2.
Expression analysis using the Incyte Human GemV microarray allowed for identification of other genes exhibiting strongly regulated behavior during in vitro differentiation. Figure 3 shows scattergrams comparing intensities of Cy5 signals from cytotrophoblast cells cultured for sequential days to Cy3 signals derived from day 0 cell mRNAs. Each point thus represents the ratio of an mRNAs signal intensity at the two time points. Based on this data, we selected 397 genes of the 6,918 genes analyzed that exhibited robust changes during differentiation using the criterion of more than 2-fold induced or repressed. Of these, 141 genes were induced and 256 were repressed. Over the culture period, the total numbers of regulated genes increased progressively, with 73% and 93% of the selected genes exhibiting changes at two or more time points, respectively. Only three genes were repressed and subsequently induced, and only two genes were initially activated and subsequently repressed. Taken together, these observations strongly suggest a high degree of experimental consistency and minimal experimental noise. Moreover, the relative numbers of induced and repressed genes highlight the prominence of selective mRNA repression or decay during cytotrophoblast cell differentiation.

View larger version (29K):
[in this window]
[in a new window]
|
Fig. 3. Scattergram depiction of primary microarray data used to generate expression profiles. Each data point represents the relative intensity of Cy3 and Cy5 hybridization to one of the 6,918 spots on the microarray. The y-axes indicate Cy5 spot intensities from mRNAs purified from days 1, 2, 3, 4, and 6 of culture, respectively, each in relation to time 0 mRNA labeled with Cy3. For each microarray, a balance coefficient was derived based on the overall average of median intensity values for Cy3 vs. Cy5. Thus Cy5 signal intensity values were multiplied by the coefficient causing the average gene on the microarray to be unchanged in its expression between the sample time and time 0. For each time point, the vast majority of genes (shown in gray) were unchanged in their expression compared with the time 0 reference control. The balance coefficients for the individual arrays were 1.63, 1.65, 1.57, 2.04, and 0.41. The overall correlation coefficients for each of the microarrays, excluding their control elements, were 0.988, 0.982, 0.979, 0.972, and 0.962 for days 1, 2, 3, 4, and 6, respectively. One standard deviation changes in gene expression were 2 (2±0.29, 2±0.31, 2±0.37, 2±0.44, and 2±0.44, respectively).
|
|
All genes on the microarray previously known to be regulated during trophoblast regulation were in fact confirmed by our analyses. However, most of the dynamically regulated genes identified by microarray had not been previously identified in earlier studies. Moreover, the occurrence of varying kinetics with respect to early and late induction and selective mRNA loss had also not been previously appreciated. The most regulated genes detected by microarray analysis include PSGP1; prostate differentiation factor; protease, serine, 11 (IGF binding); CEA6; fibronectin 1; syndecan 1; ADAM12; cytochrome P-450 XIA; the placental growth factor known as vascular endothelial growth factor (VEGF)-related protein; and LIF receptor 1. The genes whose mRNAs exhibited the greatest loss include superoxide dismutase 2, IGFBP-10, integrin B6, integrin-
6, integrin-
2, filamin A, vinculin, and actin-
2. A complete list of the 102 most regulated genes, along with the accession numbers and fold change from day 0 level of expression, is shown in Table 2.
To categorize better the patterns of gene expression changes that occurred in the cytotrophoblast cells, we subjected the log2-transformed Cy5/Cy3 expression ratios of the data set to two forms of mathematical clustering using hierarchical tree and K-means algorithms. Application of this approach to the 397 strongly regulated genes allowed the detection of coordinately regulated gene groups within the pool of dynamically regulated genes without a dilution effect generated by the inclusion of genes whose expression did not change significantly during the differentiation process. As expected, the hierarchical tree structure (Fig. 4A) revealed a major division between induced and repressed genes, with the principal variations within each major division attributable to the delay period prior to induction or repression. Figure 4B shows the distribution of genes grouped by the K-means clustering algorithm based on their position within the hierarchical tree structure. Genes that belonged to the same K-means groups were frequently distributed over separate branches of the hierarchical tree. Consequently, clusters of coregulated genes formed by one algorithm were further subdivided by the other algorithm into additional subclusters. These results demonstrate that the hierarchical tree and K-means algorithms weigh different aspects of the regulatory kinetics and that the two clustering methods discern somewhat different gene expression behavior relationships with neither method providing full pattern discrimination. Increasing the numbers of groups used in the K-means analysis did not improve the concordance of the two algorithms and led to the occurrence of poorly differentiated subclusters.

View larger version (32K):
[in this window]
[in a new window]
|
Fig. 4. Cluster analysis of the 397 genes dynamically regulated during trophoblast differentiation using hierarchical-tree and K-means algorithms applied to the log2 values for the ratio of each genes expression. Top: the distribution of gene behaviors across a hierarchical-tree structure. The color code for the signal strength in the classification scheme is shown in the box below the top panel, in which induced genes are indicated by shades of red and repressed genes are indicated by shades of green. Bottom: gene classification into clusters as determined by K-means algorithm using 9 estimated sets. Middle: the distribution of K-means-categorized genes in relation to their position within the hierarchical tree. Middle and bottom: each K-means group is both colored and numbered. Note that the hierarchical tree divides each K-means group into separate subclusters.
|
|
Figure 4C shows the individual traces of the 397 dynamically regulated genes separated into the different K-means cluster groups. The individual gene traces followed simple kinetic patterns without marked fluctuations between adjacent time points, suggesting a relatively low noise component within the expression ratio values. This was particularly true for genes that were expressed above the 50th percentile of signal intensity in both Cy3 and Cy5 channels (data not shown). Pattern 1 genes of the K-means clustering algorithm were induced strongly at day 1, then either slowed in their rate of accumulation or declined. Pattern 2 genes reached peak induction at day 2, then leveled or declined at later days. Genes in patterns 3 and 4 exhibited further time delays in their induction, increasing after the induction of the hPL and hCG genes. Patterns 59 were composed of repressed genes that exhibited varying delays prior to their decline. Only patterns 4 and 6 genes exhibited a significant delay prior to initiation of induction or repression. Thus most gene pattern groups exhibited rapid initiation of their transcriptional pattern. This suggests that cytotrophoblast cells are fully poised to enter their differentiation program. The distribution of these regulated genes into the nine patterns of gene expression is shown in Table 3 and Fig. 5.

View larger version (59K):
[in this window]
[in a new window]
|
Fig. 5. Classification of the 102 most dynamically regulated genes during trophoblast differentiation by hierarchical analyses. The 102 genes induced by more than 4-fold or repressed to less than 25% of their expression level at time 0 were grouped according to functional classification and subjected to hierarchical tree clustering. Colors indicate expression values as in Fig. 4.
|
|
Following extensive manual inspection and literature study of the 342 genes dynamically regulated during differentiation that are annotated or partially annotated, we divided the genes into 6 functional categories: cell and tissue structural dynamics (110 genes), cell cycle and apoptosis (21 genes), intercellular communication (45 genes), metabolism (79 genes), regulation of gene expression (85 genes), and unknown function (2 genes). To correlate gene functions with different patterns of gene expression, individual gene groups within the functional categories were further divided into groups that exhibited similar expression pattern behaviors. Table 3 divides the functional categories into the different K-means groups, and Fig. 5 divides the hierarchical tree into constituent functional categories. The results revealed that single functional categories were frequently divided into groups of strongly induced and strongly repressed genes. This was particularly true within the category of cell and tissue structural dynamics. Six cell adhesion genes within this category (fibronectin I, integrin ß4, lymphocyte function-associated antigen III, collagen type I receptor) and two members of the carcinoembryonic antigen gene family (CEA6 and PSGP1) were induced prior to the aggregation and fusion of the cytotrophoblast cells. The two CEA genes were among the most highly induced genes during the differentiation process, increasing by 17- and 70-fold. Four adhesion genes were induced at later times. In contrast, 24 of the 34 adhesion genes were repressed, with 10 of these belonging to pattern 6. Among the genes involved in cytoskeletal organization, 17 of 21 were repressed, with 7 of these exhibiting pattern 6.
To illustrate better the behavioral pattern distribution of the most regulated genes within each functional category, we selected genes that changed by fourfold or more (102 genes), divided them into major functional categories, and subjected each category to hierarchical tree cluster analysis (Fig. 6). Figure 6 shows that strongly divergent behaviors occurred within tightly related categories, suggesting that functional reprogramming is necessary to accomplish differentiation. These findings suggest the hypothesis that efficient execution of some biologic processes, such as adhesion and tissue remodeling, is best accomplished by both the induction and downregulation of individual genes within the specific functional category. Other gene categories illustrated the same theme of coordinate gene induction and repression, including metabolism and regulation of gene expression.

View larger version (28K):
[in this window]
[in a new window]
|
Fig. 6. The relationship between expression behavior and gene function of the 397 dynamically regulated genes during trophoblast differentiation. Individual genes within each functional classification group are depicted in relation to their position within the expression pattern-based hierarchical tree as shown in Fig. 4. The colors of each gene function group are arbitrary.
|
|
Some categories were composed of genes that were more uniformly induced or repressed. For example, most of the genes relating to intercellular communication were induced, with 9 of the polypeptide hormone genes following a pattern identical to hPL, hCG
, and hCGß (pattern 2). All 13 of the regulated inflammatory response genes were repressed, as were 16 of the 20 inflammatory mediator genes, with most following pattern 6. Many of the genes related to gene expression were also repressed, including 27 of the 28 translation-related genes, 11 of which followed pattern 7. However, genes for transcription and signal transduction molecules were more evenly split, with 9 of 31 and 8 of 26 exhibiting upregulated behavior, respectively.
The occurrence of distinct temporal behaviors for gene activation and repression within different functional categories correlated well with the progressive morphological changes that underlie trophoblast differentiation (Fig. 7). Thirty annotated genes were induced on day 1 (pattern 1) during the time the cells initiate aggregation and fusion, but prior to the greatest induction of hPL and hCG. Thirty-two annotated genes were induced with a delayed pattern similar to that of hPL, hCG
, and hCGß (pattern 2). On the other hand, 75 genes were repressed prior to the greatest induction of hPL and hCG gene expression. During the early stages of differentiation, when the cells are aggregating and fusing to form a syncytium, many specialized adhesion genes were induced (PSGP1, CEA6, fibronectin 1, integrin B4, CD58, and CD36) as well as genes for the long form of the prolactin receptor and hydroxy
5-steroid dehydrogenase. Endoglin, a component of the TGF-ß receptor complex that binds ß1 and ß3 isoforms and is expressed at high levels on syncytiotrophoblast cells throughout pregnancy, was induced in the villous cytotrophoblast cells early in differentiation (pattern 1).

View larger version (26K):
[in this window]
[in a new window]
|
Fig. 7. Categorical reprogramming-based scheme for cytotrophoblastic differentiation. Major genes and functional groups that exhibit dynamic expression changes are depicted in relationship to the morphological changes that accompany syncytiotrophoblast cell formation during the 6-day culture period. The genes and functional groups are taken from Table 3. Integrins refers to integrins- 2, - 6, and -ß6. Adhesion proteins refers to cadherin 1, 3 and 5, annexins A3 and A8, CD24, and transgelin. TIMP3, tissue inhibitor of metalloproteinase 3; CEA6, carcinoembryonic antigen gene family member 6; and MAO, monoamine oxidase A. We hypothesize that there is categorical reprogramming of gene expression, particularly within the category of cell and tissue structure genes, which is necessary to accomplish the marked cell morphology changes that occur during trophoblast differentiation. Categorical reprogramming represents the simultaneous activation, repression, or degradation of mRNAs from within a given functional group. The diagram of cellular morphology changes is adapted from Kliman et al. (14).
|
|
Following initial syncytial formation, there was marked induction of genes involved in intercellular communication. These included genes for nine polypeptide hormone genes: hPL, hCG
, hCGß, luteinizing hormone-ß, granulin, and VEGF, and three members of the TGF-ß superfamily, prostate differentiation factor, bone morphogenetic protein 1 (BMP-1), and BMP-7 (osteogenic protein 1). BMP-7 (osteogenic protein-1) and BMP-1 play strong roles in development and differentiative transformation of many different organ systems. CRH, which is known to be a specific marker for terminally differentiated syncytiotrophoblast cells, was not significantly induced until day 4, following the inductions of hPL and hCG. Expression of genes for IGFBP-3 and IGFBP-10 were repressed. However, CSF-1, which is known to induce trophoblast differentiation in vitro, was also repressed. The role of CSF-1 in normal placental differentiation is unclear. Other interesting genes include ADAM12, a member of "a disintegrin and metalloprotease" family of cell-surface proteins, which was induced with a pattern similar to that of hPL and hCG. AIM 1 ("absent in melanoma 1"), a novel non-lens member of the ß
-crystallin superfamily that is associated with the control of tumorigenicity in human malignant melanoma, was also induced prior to cell aggregation and fusion (pattern 1).
 |
DISCUSSION
|
---|
A major finding of these results is that there are several different inductive and repressive kinetic patterns associated with the differentiation of cytotrophoblast cells into syncytiotrophoblast cells. These patterns consisted of variable delays prior to the induction or repression of different groups of genes. In contrast, we did not observe genes that exhibited induction followed by repression or vice versa. We interpret this behavior as an indication that the cells are poised to enter directly the differentiation process and that the activation of the cytotrophoblast-to-syncytiotrophoblast gene program results from the release of one or more sequential regulatory triggers. In this model, genes that undergo activation could be in a transcriptionally poised configuration. Similarly, genes undergoing repression may do so transcriptionally, or posttranscriptionally (for review, see Ref. 27), via the action of machinery already in place within the cytotrophoblast cells. However, to our knowledge, specific machinery for selective mRNA decay in the cytotrophoblast has not been described.
It will now be of considerable interest to determine roles of the dynamically regulated genes in cytotrophoblast differentiation. Moreover, it will also be critical to understand the molecular basis by which cytotrophoblast cells normally conform to an extremely dynamic terminal differentiation process and, on untoward occasion, undergo malignant transformation to choriocarcinoma cells, one of the most aggressive of all cancers. One clue for this may lie within the cytotrophoblast cells extremely dynamic gene program for controlling cell and tissue structure and angiogenesis.
In summary, we have identified a temporal event sequence that underlies cytotrophoblast differentiation based on the induction and repression of a series of genes not recognized previously to play a role in placental development. We have shown that cytotrophoblast-to-syncytiotrophoblast cell differentiation comprises a highly dynamic gene program that significantly affects the mRNA levels of 400 of 7,000 individual genes queried. Several distinct kinetic patterns of gene induction and repression were observed. As shown in Fig. 7, the early events in this process involve strong repression, possibly consisting of accelerated mRNA degradation of many genes as well as the brisk induction of other genes within the same overall function category. We therefore hypothesize that trophoblast differentiation requires both activation and repression of a substantial number of genes. We further postulate the existence of two classes of gene regulatory processes that are necessary to accomplish cellular differentiation. The first is the induction of gene products that are responsible for cell functions that were not necessary prior to differentiation but are subsequently required for differentiated cell functions. Examples would include hormone production, unique metabolic processes, or as mediators of differentiation per se. The other class represents induction of genes that replace existing gene products with those that cause the cell to switch structure and function. To accomplish this, we envision that the cell must eliminate mRNAs of gene products that could compete or interfere with the induced gene set. Thus our hypothesis is that the successful accomplishment of cellular differentiation requires both the induction of the effectors of the differentiated cell as well as dynamic reprogramming of genes within functional pathways that are critical for both precursor and product cell lineages.
 |
ACKNOWLEDGMENTS
|
---|
We thank Michael Hubert for technical assistance and Michael Eisen, Paul Spellman, and Andrew Conway for valuable discussions.
This research was supported by National Institutes of Health Grant HD-07447
 |
FOOTNOTES
|
---|
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
Address for reprint requests and other correspondence: S. Handwerger, Dept. of Endocrinology, Childrens Hospital Medical Center, 3333 Burnet Ave., Cincinnati, OH 45229-2029 (E-mail: Stuart.Handwerger{at}CHMCC.org).
 |
REFERENCES
|
---|
-
Bax CM, Ryder TA, Mobberley MA, Tyms AS, Taylor DL, and Bloxam DL. Ultrastructural changes and immunocytochemical analysis of human placental trophoblast during short-term culture. Placenta 10: 179194, 1989.[ISI][Medline]
-
Benirschke K and Kaufmann P. Early development of the human placenta. In: Pathology of the Human Placenta, edited by Benirschke K and Kaufmann P. New York: Springer-Verlag, 1995, p. 4956.
-
Bischof P, Haenggeli L, and Campana A. Effect of leukemia inhibitory factor on human cytotrophoblast differentiation along the invasive pathway. Am J Reprod Immunol 34: 225230, 1995.[ISI][Medline]
-
Boime I. Human placental lactogen production is linked to the stage of trophoblast differentiation. In: Trophoblast Research, edited by Miller RK and Thiede HA. Rochester: Plenum Medical, 1991, p. 5760.
-
Chomczynski P and Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162: 156159, 1987.[ISI][Medline]
-
Cross JC. Genetic insights into trophoblast differentiation and placental morphogenesis. Semin Cell Dev Biol 11: 105113, 2000.[ISI][Medline]
-
Dizon-Townson DS, Lu J, Morgan TK, and Ward KJ. Genetic expression by fetal chorionic villi during the first trimester of human gestation. Am J Obstet Gynecol 183: 706711, 2000.[ISI][Medline]
-
Eisen MB, Spellman PT, Brown PO, and Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95: 1486314868, 1998.[Abstract/Free Full Text]
-
Feinberg RF, Kliman HJ, and Wang CL. Transforming growth factor-beta stimulates trophoblast oncofetal fibronectin synthesis in vitro: implications for trophoblast implantation in vivo. J Clin Endocrinol Metab 78: 12411248, 1994.[Abstract]
-
Garcia-Lloret MI, Morrish DW, Wegmann TG, Honore L, Turner AR, and Guilbert LJ. Demonstration of functional cytokine-placental interactions: CSF-1 and GM-CSF stimulate human cytotrophoblast differentiation and peptide hormone secretion. Exp Cell Res 214: 4654, 1994.[ISI][Medline]
-
Gordon AD. Classification. Boca Raton, FL: Chapman and Hall/CRC, 1999, p. 256.
-
Ho S, Winkler-Lowen B, Morrish DW, Dakour J, Li H, and Guilbert LJ. The role of Bcl-2 expression in EGF inhibition of TNF-alpha/IFN-gamma-induced villous trophoblast apoptosis. Placenta 20: 423430, 1999.[ISI][Medline]
-
Hoshina M, Hussa R, Pattillo R, Camel HM, and Boime I. The role of trophoblast differentiation in the control of the hCG and hPL genes. Adv Exp Med Biol 176: 299312, 1984.[ISI][Medline]
-
Kliman HJ, Nestler JE, Sermasi E, Sanger JM, and Strauss JF. Purification, characterization, and in vitro differentiation of cytotrophoblasts from human term placentae. Endocrinology 118: 15671582, 1986.[Abstract]
-
Maruo T, Matsuo H, Otani T, and Mochizuki M. Role of epidermal growth factor (EGF) and its receptor in the development of the human placenta. Reprod Fertil Dev 7: 14651470, 1995.[ISI][Medline]
-
Maruo T, Murata K, Matsuo H, Samoto T, and Mochizuki M. Insulin-like growth factor-I as a local regulator of proliferation and differentiated function of the human trophoblast in early pregnancy. Early Pregnancy 1: 5461, 1995.[Medline]
-
Morrish DW, Bhardwaj D, and Paras MT. Transforming growth factor beta 1 inhibits placental differentiation and human chorionic gonadotropin and human placental lactogen secretion. Endocrinology 129: 2226, 1991.[Abstract]
-
Morrish DW, Dakour J, Li H, Xiao J, Miller R, Sherburne R, Berdan RC, and Guilbert LJ. In vitro cultured human term cytotrophoblast: a model for normal primary epithelial cells demonstrating a spontaneous differentiation programme that requires EGF for extensive development of syncytium. Placenta 18: 577585, 1997.[ISI][Medline]
-
Morrish DW, Linetsky E, Bhardwaj D, Li H, Dakour J, Marsh RG, Paterson MC, and Godbout R. Identification by subtractive hybridization of a spectrum of novel and unexpected genes associated with in vitro differentiation of human cytotrophoblast cells. Placenta 17: 431441, 1996.[ISI][Medline]
-
Myers EW, Sutton GG, Delcher AL, Dew IM, Fasulo DP, Flanigan MJ, Kravitz SA, Mobarry CM, Reinert KH, Remington KA, Anson EL, Bolanos RA, Chou HH, Jordan CM, Halpern AL, Lonardi S, Beasley EM, Brandon RC, Chen L, Dunn PJ, Lai Z, Liang Y, Nusskern DR, Zhan M, Zhang Q, Zheng X, Rubin GM, Adams MD, and Venter JC. A whole-genome assembly of Drosophila. Science 287: 21962204, 2000.[Abstract/Free Full Text]
-
Nachtigall MJ, Kliman HJ, Feinberg RF, Olive DL, Engin O, and Arici A. The effect of leukemia inhibitory factor (LIF) on trophoblast differentiation: a potential role in human implantation. J Clin Endocrinol Metab 81: 801806, 1996.[Abstract]
-
Pollard JW, Bartocci A, Arceci R, Orlofsky A, Ladner MB, and Stanley ER. Apparent role of the macrophage growth factor, CSF-1, in placental development. Nature 330: 484486, 1987.[ISI][Medline]
-
Richards RG, Hartman SM, and Handwerger S. Human cytotrophoblast cells cultured in maternal serum progress to a differentiated syncytial phenotype expressing both human chorionic gonadotropin and human placental lactogen. Endocrinology 135: 321329, 1994.[Abstract]
-
Richardson B, Palgunachari MN, Anantharamaiah GM, Richards RG, Azrolan N, Wiginton D, and Handwerger S. Human placental tissue expresses a novel 22.7 kDa apolipoprotein A-I-like protein. Biochemistry 35: 75807585, 1996.[ISI][Medline]
-
Richardson BD, Cheng YH, Langland RA, and Handwerger S. Differential expression of AP-2
and AP-2
during human trophoblast differentiation Life Sci In press.
-
Ringler GE and Strauss JF III. In vitro systems for the study of human placental endocrine function. Endocr Rev 11: 105123, 1990.[ISI][Medline]
-
Ross J. mRNA stability in mammalian cells. Microbiol Rev 59: 42350, 1995.[Abstract]
-
Scott IC, Anson-Cartwright L, Riley P, Reda D, and Cross JC. The HAND1 basic helix-loop-helix transcription factor regulates trophoblast differentiation via multiple mechanisms. Mol Cell Biol 20: 53041, 2000.[Abstract/Free Full Text]
-
Shi QJ, Lei ZM, Rao CV, and Lin J. Novel role of human chorionic gonadotropin in differentiation of human cytotrophoblasts. Endocrinology 132: 138795, 1993.[Abstract]
-
Stephanou A and Handwerger S. The nuclear factor NF-IL6 activates human placental lactogen gene expression. Biochem Biophys Res Commun 206: 21522, 1995.[ISI][Medline]
-
Stephanou A, Myatt L, Eis AL, Sarlis N, Jikihara H, and Handwerger S. Ontogeny of the expression and regulation of interleukin-6 (IL-6) and IL-1 mRNAs by human trophoblast cells during differentiation in vitro. J Endocrinol 147: 48796, 1995.[Abstract]
-
Wice B, Menton D, Geuze H, and Schwartz AL. Modulators of cyclic AMP metabolism induce syncytiotrophoblast formation in vitro. Exp Cell Res 186: 30616, 1990.[ISI][Medline]
-
Xu B, Lin L, and Rote NS. Identification of a stress-induced protein during human trophoblast differentiation by differential display analysis. Biol Reprod 61: 6816, 1999.[Abstract/Free Full Text]