Large-scale reprogramming of cranial neural crest gene expression by retinoic acid exposure

Sarah S. Williams1, John P. Mear2, Hung-Chi Liang3, S. Steven Potter3, Bruce J. Aronow1 and Melissa C. Colbert2

1 Divisions of Biomedical Informatics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio 45229
2 Molecular Cardiovascular Biology, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio 45229
3 Molecular Developmental Biology, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio 45229


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Although retinoic acid (RA), the active form of vitamin A, is required for normal embryonic growth and development, it is also a powerful teratogen. Infants born to mothers exposed to retinoids during pregnancy have a 25-fold increased risk for malformations, nearly exclusively of cranial neural crest-derived tissues. To characterize neural crest cell responses to RA, we exposed murine crest cultures to teratogenic levels of RA and subjected their RNA to microarray-based gene expression profile analysis using Affymetrix MG-U74Av2 GeneChips. RNAs were isolated from independent cultures treated with 10–6 M RA for 6, 12, 24, or 48 h. Statistical analyses of gene expression profile data facilitated identification of the 205 top-ranked differentially regulated genes whose expression was reproducibly changed by RA over time. Cluster analyses of these genes across the independently treated sample series revealed distinctive kinetic patterns of altered gene expression. The largest group was transiently affected within the first 6 h of exposure, representing early responding genes. Group 2 showed sustained induction by RA over all times, whereas group 3 was characterized by the suppression of a time-dependent expression increase normally seen in untreated cells. Additional patterns demonstrated time-dependent increased or decreased expression among genes not normally regulated to a significant extent. Gene function analysis revealed that more than one-third of all RA-regulated genes were associated with developmental regulation, including both canonical and noncanonical Wnt signaling pathways. Multiple genes associated with cell adhesion and cell cycle regulation, recognized targets for the biological effects of RA, were also affected. Taken together, these results support the hypothesis that the teratogenic effects of RA derive from reprogramming gene expression of a host of genes, which play critical roles during embryonic development regulating pathways that determine subsequent differentiation of cranial neural crest cells.

microarray; retinoic acid; cell signaling; cell proliferation


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
THE USE OF RETINOIDS OR RETINOID derivatives to treat a variety of clinical conditions expands yearly. Despite continued warnings and programs for pregnancy prevention, the risk of fetal exposure to retinoic acid (RA), primarily via Accutane, remains high (29, 67). Infants born to mothers exposed to retinoids during pregnancy have a 25-fold increased risk for congenital malformations. In a landmark paper, Lammer et al. (44) described a distinctive constellation of defects associated with RA exposure and characterized this new retinoid embryopathy. They include microtia/anotia, micrognathia, cleft palate, abnormalities of the aortic arch arteries, conotruncal defects such as common arterial trunk, thymic abnormalities, and malformation of the central nervous system and optic nerve. In animal studies, embryonic exposure to RA produces many of the same defects (23, 76, 91). Furthermore, this retinoid embryopathy recapitulates many of the malformations seen with ablation of a specific portion of the cranial neural crest in chick embryos and has established the neural crest as a major target for the teratogenic actions of RA (35).

The neural crest, a transient, multipotential population of cells, originates from the dorsal neural folds, and cells migrate to a variety of sites within the developing embryo (47). Crest cells move in coherent streams and follow highly defined migratory patterns, a hallmark behavior of these cells (38, 54). Cranial crest arises anterior to the fifth somite and contributes much of the mesenchyme that forms the bones of the face and skull, facial musculature, stroma for the thymus and thyroid, the smooth muscle that surrounds the great vessels, the septum that separates the outflow tract of the heart into pulmonary and systemic circuits, in addition to components of the peripheral and enteric nervous system (26). Kirby and coworkers (see Refs. 14 and 36 for reviews) demonstrated that removal of a subset of the cranial neural crest, termed the cardiac crest, early in development produced aortic arch and conotruncal defects as well as loss of the thymus, thyroid, and parathyroids. In unablated embryos, these streams of crest cells pause briefly to proliferate, populating the pharyngeal arches to provide mesenchyme for the developing arch arteries and other structures (41, 42, 60, 90).

One common, consistent observation in developing systems exposed to high RA levels suggests that too few cells are available to populate the target tissue, resulting in organ malformations (63). There is ample evidence on the biological consequences of RA exposure to support this hypothesis. Retinoids decrease neural crest cell adhesion to substrate and their ability to migrate. Lamellipodial extensions and translocations are inhibited in chick cultures treated with RA (80, 85). We have shown RA inhibits migration of neural crest exposed either in vivo or in vitro (53). Inhibition was dose dependent and associated with reduced activation of c-jun N-terminal kinase (JNK). Following in vivo exposure, RA also disrupts migratory pathways such that crest cells end up at the wrong targets (51, 62). Retinoids inhibit cell proliferation and promote or suppress cell death or cell differentiation in a cell type-specific context. In crest cells, RA reduces bromodeoxyuridine (BrdU) incorporation and significantly prolongs the duration of the cell cycle (53, 70). In some tumor cells, notably MCF7 breast cancer and several neuroblastomas (17, 69, 86, 95), RA inhibits cell growth and promotes differentiation. However, in developing embryos, decreased proliferation, enhanced cell death, and inhibition or induction of differentiation at an inappropriate time would have dire consequences: failed embryonic development of target organs.

The effects of RA are mediated through specific nuclear receptors, members of the steroid hormone superfamily of receptors (57). Canonical retinoid signaling involves binding of the ligand (RA) to a heterodimer complex, containing one member of the RA receptor family (RARs) and one member of the retinoid X receptor family (RXRs). Receptor dimers recognize and bind to DNA response elements embedded in the regulatory region of target genes (RAREs). Ligand binding relieves repression by coregulators and recruits coactivator complexes and general transcription factors to activate target gene transcription (see Refs. 4 and 93 for recent reviews and references therein). However, the expression level of many genes can be altered in response to RA through less direct mechanisms such as mRNA stability (10) or protein turnover (6, 44) and are real though indirect targets. Thus genes may be classified as either direct or indirect targets for the action of RA (3).

An abundance of literature has provided examples of RA-responsive genes. Balmer and Blomhoff (3) have compiled a catalog of more than 532 genes from over 1,100 citations that are regulated in some fashion by RA. With the advent of techniques for large-scale gene discovery including subtractive hybridization, differential display, and microarray interrogation, the number of reports on cell type-specific or developmental changes in gene expression in response to RA has steadily increased (8, 13, 56, 96). These studies add to the growing number of potential targets and, in some cases, correlate the changes in gene expression with the biology of what RA may be doing in the specific cell type under examination. Most of these reports have focused on how RA induces differentiation of embryonic stem cell or tumor cell lines (21, 34, 52, 86, 95, 96).

To begin to examine mechanisms that mediate the pleiotropic effects of RA, we have examined the changes in gene expression patterns in microdissected neural crest exposed to teratogenic levels of RA. These cells are a critical embryonic target cell population of in utero RA exposure. We hypothesize that multiple changes in gene expression contribute collectively to dysregulation of normal biological functions in response to RA. We have previously shown that this exposure drastically alters the migration and proliferation potential of these cells in vitro and in vivo (53). We have now tested our hypothesis and show that a significant component of these phenotypic alterations are derived from a variety of RA-driven transcriptional regulatory effects. The present study is the first to examine the responses of these cells to RA exposure in vitro and to analyze large-scale gene expression profiles by microarray analysis. Time course analysis over 48 h of exposure enabled us to distinguish early and late targets acted on by RA, and hierarchical clustering shows several distinctive patterns of time-dependent changes in gene expression. These patterns inform the known biology as to relevant gene targets, illustrate the combination of signals that contribute to a teratogenic phenotype, and provide insight on cross talk between RA and developmental signaling pathways. The implication of this study is that neural crest cells appear to be highly susceptible to developmental reprogramming by RA.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Neural crest cultures and experimental design.
A breeding colony of ICR mice was maintained in the AAALAC-approved facility of the Cincinnati Children’s Hospital Research Foundation (CCHRF). All experiments were performed in accordance with guidelines of the American Physiological Society and with the approval of the CCHRF IACUC, protocol 2B11073. Animals were mated overnight and checked the following morning for evidence of a coital plug. Noon of that day was recorded as 0.5 days postcoitum (dpc). Embryos at 8.5 dpc with 8–13 somites were used to isolate neural crest as previously described (53). The dorsal neural tube was bisected along the midline with one half from each embryo used for control and the other half treated with 10–6 M RA dissolved in ethanol for 6, 12, 24, or 48 h. For 6-h exposures, the explants were cultured overnight on fibronectin-coated 35-mm dishes (Biocoat; Becton-Dickinson Labware, Bedford, MA) in DMEM with 10% horse serum, to allow for sufficient outgrowth of neural crest cells. The RA was added the following morning. A minimum of three independent experiments were conducted for each time point.

In preliminary studies (not shown), explants were untreated or exposed to 0.1% ethanol. We detected no difference in behavior or gene expression between ethanol-treated and untreated samples. Therefore, to enable characterization of the normal transcriptome of neural crest cells in culture (in preparation), all subsequent analyses and comparisons were made between RA-treated and untreated controls.

RNA isolation and amplification.
Explants were picked away from the crest cell outgrowth 2 h before the end of each incubation period, and the cultures were allowed to recover. The media was removed, and the cells were washed twice with sterile PBS containing calcium and magnesium. The cells were scraped into 50 µl of PBS and stored at –70°C. Total RNA was isolated using RNeasy Mini (Qiagen, Valencia, CA) according to the manufacturer’s directions and linearly amplified as previously described (5, 71). Briefly, a T7 promoter-dT primer was used to generate cDNA with SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA) and amplified by vitro transcription using T7 RNA polymerase. A second round of cDNA synthesis was performed using random hexamers as primers, and a final in vitro transcription using biotin-labeled nucleotides (Bioarray High Yield RNA; Enzo Life Sciences, Farmingdale, NY) yielded labeled cRNA for microarray hybridization.

Array hybridization and analysis.
Twenty-seven labeled RNA samples (13 RA-treated samples and 14 controls) were submitted to the CCHRF Affymetrix Microarray Core for hybridization to Affymetrix MG-U74Av2 GeneChip probe arrays using standardized protocols. Processed arrays were scanned using Microarray Suite 5.0 software (MAS 5.0; Affymetrix, Santa Clara, CA). GeneChip "*.cel" files from MAS 5.0 were subjected to RMA analysis (http://stat-www.berkeley.edu/users/bolstad/RMAExpress/RMAExpress.html) (6, 32) and loaded into GeneSpring software (Silicon Genetics, Redwood City, CA) for subsequent analysis. The RMA-normalized data from the 27 samples was first normalized to the median of the controls within each time point to select for genes that were retinoid dependent at each time. This depiction of the experiment (27_RA_NormToTime_RMA_Jan_04; see access information, below) allows for a gene behavior in which its expression may be increased or decreased at a given culture time. The statistical comparison of RA vs. control expression within each time point was done by applying a Student’s t-test ( P ≤ 0.05) to genes that had passed a 1.5-fold change filter in 2 or more samples within the time point. Results combined from the four time points identified 137 genes that were significantly regulated by RA over the median control population at each time. The original data set of 27 samples was alternatively normalized with respect to the median of the gene’s expression value in the untreated samples across all time points. This depiction of the experiment (27_RA_NTC_RMA_April_04; see access information, below) allows the visualization of the effects of RA at each time point, independent of whether there is normally a change in a gene’s expression between time points in the absence of RA. The same t-test and filtering criteria were applied, resulting in 158 RA-responsive genes (see Supplemental Tables S2 and S3, available at the Physiological Genomics web site).1 The union of these two lists yielded 205 genes, with 90 genes common to both lists. Both versions of the experiment are available via guest login at the Silicon Genetics GeNet web site (http://genet.cchmc.org/) in the "MColbert" directory of the U74Av2 genome (select "Gene Lists" folder, then "MColbert" folder) or accessible through the NCBI Gene Expression Omnibus (GEO) database (accession numbers GSM24606 and GSM27252GSM27277; http://www.ncbi.nlm.nih.gov/geo/). We then sought to analyze the gene expression patterns represented by these 205 genes under the second normalization schema where expression levels were normalized relative to the median expression value in the untreated samples across all time points (27_RA_NTC_RMA_April_04). Hierarchical cluster analysis (Pearson correlation) of the pooled 205 genes showed correlated groups of genes and their expression patterns across all time points. Additional clustering of all 27 GeneChips by a distance measurement was performed to determine the correlation of the biological replicates.

PCR analysis.
Samples of RNA (2 µg) from treated and control cultures were isolated as above and cDNA prepared using SuperScript II as outlined by the manufacturer (Invitrogen). Semiquantitative PCR was performed to verify changes in gene expression in response to RA. Gene-specific primers were designed and are listed along with annealing conditions in Supplemental Table S1. Amplification, using Taq polymerase (Invitrogen), was restricted to the linear phase, which we determined to be between 25–35 cycles for most genes. PCR products were resolved on 1.5% agarose gels and stained with ethidium bromide. Actin or GAPDH primers were used as controls. For quantitative PCR, 2 µl of cDNA was used in a 25.0-µl reaction consisting of Promega Master Mix (0.625 U of Taq, 200.0 µM each dNTP, 1.5 mM MgCl2 in a proprietary buffer at pH 8.5; Promega, Madison, WI), 1.0 µM of each primer, SYBR Green I (Molecular Probes, Eugene, OR) at 0.1x and H2O. Samples were run in an Opticon 2 DNA Engine (MJ Research), and actin or GAPDH was used to normalize the sample cDNA concentrations. Standard curves were generated using mouse whole embryo (9.5 dpc) cDNA. All samples were run in duplicate, and the average values were normalized by dividing by the corresponding average value of the controls. Fold changes for each time point were calculated by dividing the normalized value of the treated sample by the normalized value of the untreated sample at that time point.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Experimental design and analysis.
To examine the time course of the response of primary neural crest cultures to teratogenic levels of RA, we treated explants from dorsal neural tube with 10–6 M RA from 6–48 h. The neural tube was split along the roof plate, with one half of the explanted tissue treated and the contralateral half untreated, serving as control. Thus both treated and control samples were always from the same stage embryos. Cultures were exposed for 6 h to evaluate early responding genes and for 12, 24, or 48 h to determine late or secondary responses. A minimum of three independent biological replicates was analyzed at each time point. This design allowed for comparison of retinoid-mediated changes in gene expression within the context of kinetic changes in gene expression of the primary crest cultures. Because of the limited amount of tissue available, two rounds of in vitro RNA amplification were used prior to hybridization with Affymetrix MG-U74Av2 GeneChip probe arrays. We have used this technique previously, and found it to be reliable and highly reproducible (71).

A total of 27 independent GeneChips, 13 RA-treated and 14 from controls, were analyzed, and the data were normalized and baseline-referenced within the context of the experimental design. Two normalization strategies were implemented. In the first stage, we normalized all cultures at each time point to the median of the corresponding untreated crest cultures, (normalized to time). This allowed us to identify genes that had a reproducible, time-dependent response to RA. Since it was reasonable that some of these RA-affected genes might vary normally in their expression during the life of the cell culture, we generated a second normalization of the experiment series such that each gene’s relative expression was normalized to the median of its expression across all of the untreated sample times (normalized to control). Thus a gene that is normally induced or repressed at one time vs. another could be observed, and clusters that reflect this behavior could be obtained. The t-test ( P ≤ 0.05) and fold change analyses of the data normalized to time generated 137 genes that were significantly regulated by RA treatment compared with controls within the four time points (Supplemental Fig. S1, B). When the data were alternatively normalized to controls across all times, t-test and fold change analyses generated 158 genes that were significantly regulated by RA treatment (Supplemental Fig. S1, A). A 1.5-fold change value was chosen as cutoff rather than higher values (>2) because we considered the consistency of response more critical than the absolute magnitude of the change in expression level. To illustrate both the RA-dependent changes at each time point, as well as the RA effects on any time-dependent changes in control expression, the two lists were pooled, yielding 205 genes, with 90 genes common to both lists. Hierarchical cluster analysis of the 205 genes across the 27 different samples at the 4 different times using Pearson correlation revealed groups of genes that shared similar expression patterns (Fig. 1, left gene tree).



View larger version (61K):
[in this window]
[in a new window]
 
Fig. 1. Hierarchical cluster analysis of 205 retinoic acid (RA)-responsive genes. Expression was increased or decreased by 1.5-fold, P ≤ 0.05, in at least 2 experiments. Gene tree (Pearson correlation; left tree) shows correlated groups of genes and their expression patterns across all time points. Clustering by distance measurement (top tree) was performed to determine the correlation of the biological replicates. Lines were added to highlight the separation of the 6 h time samples, the clustered controls, and the clustered RA-treated samples. The numbers and lines outside of the tree structure identify five groups of genes. Color bar shows the range of expression, from blue = 0, to red = 3.

 
Analysis of the gene tree showed five distinct patterns of responsive genes based upon their behavior relative to controls, over the time course of the experiment. Group 1 contained transient, immediate early changes in gene expression, seen at 6 h (Fig. 1, groups 1u and 1d), whereas RA consistently induced high-level expression at all times in the second group (Fig. 1, group 2). Group 3 was characterized by suppression of a time-dependent increase in expression, normally seen in untreated cells (Fig. 1, group 3). For groups 4 and 5, RA increased or decreased expression of these gene sets, respectively over the course of the experiment (Figs. 1, groups 4 and 5).

Clustering of the 27 sample GeneChips into a condition tree by a distance measurement further demonstrated that, with the exception of the 6 h cohort, the RA-treated samples all clustered together as did the controls (Fig. 1, top gene tree). This early responding set at 6 h may constitute a category of gene regulation different from the later time points, which would be consistent with early vs. late responses to RA. Thus the reproducibility of the analysis and the response of neural crest cultures to RA were remarkably consistent in the face of the expected biological variability that exists in primary cultures taken from developmentally staged embryos.

Group 1: Immediate early changes in gene expression.
This group of genes contained significantly more candidates and more functional variety than any of the other groups, despite being held to a higher statistical filter, P = 0.01 (Fig. 2). Over 70 genes passed filtering; 30 showed increased expression (group 1u) and 43 decreased (group 1d) following 6 h of RA treatment (Tables 1 and 2). Group 1 focused on the immediate early responses of neural crest cells to RA as these changes in expression occurred mostly at 6 h, and then returned to control levels by 12 and 24 h (Fig. 3; genes with a sustained response to RA, group 2, were excluded from this category). However, expression patterns became more complex at 48 h, suggesting other mechanisms may be involved in regulation at later times.



View larger version (34K):
[in this window]
[in a new window]
 
Fig. 2. The classification of genes within groups 1–5. Categories ascribed to genes were determined and selected from the Gene Ontology (GO) listing that accompanied the LocusLink ID. The percent of genes within that category was then determined for the whole group and represented in the pie charts. ECM/ECS, extracellular matrix or space.

 

View this table:
[in this window]
[in a new window]
 
Table 1. Group 1: Genes upregulated by RA at 6 h

 

View this table:
[in this window]
[in a new window]
 
Table 2. Group 1: Genes downregulated by RA at 6 h

 


View larger version (28K):
[in this window]
[in a new window]
 
Fig. 3. Dynamic expression profiles of each group identified in Fig. 1. The average expression level of genes within each group is presented as a function of time point. Color bar is as in Fig. 1. Coloring of expression levels for genes in group 1 was determined at 6 h in control and RA-treated groups; all other groups were based upon expression levels at 48 h.

 
Two genes within this group were well-characterized targets for RA: cellular retinoic acid binding protein (Crabp2) and Nmyc, recognized as an early and rapid responder (3). Both contain DNA elements for RAR binding (19, 88). Other genes of interest included four associated with mitochondrial function (induced: Timm8a, Timm8b, Mrps21; and suppressed: cytochrome b5), several involved with different aspects of cell signaling (induced: Stmn2, Sema4b, and Cmkor1; and suppressed: Ogfr, Csnk1d, Sart1, Arrb1, Tyro3, and Cttn), and five genes that participate in ubiquitin-mediated protein degradation, all of which had decreased expression (Ube2i, Usp52, Psmd3, Ube1x, and Btrc). Unlike the following four groups, more transcriptional regulators as opposed to transcription factors were represented, and all of them were suppressed (Bat8, Ncor2, Nono, Hdac6, and Fliih). Interestingly, Ncor2, a transcriptional corepressor, and Fliih, a coactivator, both impinge on nuclear hormone receptor activated signaling (30, 50). Finally, 20% of the early responding genes were expressed sequence tags (ESTs), many more than in subsequent groups, which may represent important, novel genes associated with retinoid action.

Group 2: Consistently high levels of increased gene expression at all times.
The expression of genes represented by this group was not only increased in RA samples at all time points, but also consistently showed the highest levels of induction (Fig. 1, group 2; Fig. 3, and Fig. 4A). In MAS 5 analyses (not shown), many of the control samples were scored as absent. Semiquantitative PCR analysis was used to demonstrate negligible expression, suggesting the induction of several genes could be considered de novo. These include Cyp26a1, Hoxa5, and Meox1 (Fig. 4A). Furthermore, nearly one-third of the genes in this cluster have been previously identified as retinoid responsive (Table 3; see also Ref. 3 and references therein). The fact that we found so many known RA-responsive genes induced in our experiments provided "internal controls" and further supports the experimental paradigm as well as our results.



View larger version (34K):
[in this window]
[in a new window]
 
Fig. 4. Verification of changes in gene expression levels. Semiquantitative RT-PCR (A) or real-time RT-PCR (B) was used to evaluate changes in expression of selected genes at 48 h, with the exception of group 1, which was taken at 6 h. Actin or GAPDH levels were used as controls. The linear range for semiquantitative amplification was determined to be between 25 and 35 cycles for most genes, as indicated.

 

View this table:
[in this window]
[in a new window]
 
Table 3. Group 2: Genes induced by RA at all times

 
Of the RARE-containing genes, three directly regulated retinoid levels. Cyp26a1, responsible for metabolism and elimination of RA, Dhrs3, the retinal short-chain dehydrogenase-reductase that reversibly converts retinol to retinal, and Rbp1, also known as cellular retinol binding protein (9, 55, 79), were robustly induced across all treatment times. Thus neural crest cells evolved a strong and rapid protective mechanism coordinated to moderate retinoid levels.

More than half (54%) of these highly induced genes were associated with transcription and cell signaling (Table 3, Fig. 2). The Hox genes were the most prominently represented class of developmental transcription factors in this cohort, including the 3' members of the A and B clusters. So too was Pbx1, a cofactor for Hox activity (46, 66). Other strongly induced transcriptional regulators included Meox1, a mesenchyme homeobox gene, Bhlhb2, a basic helix-loop-helix transcription factor (Fig. 4A), and a T-box gene, Tbx2. Genes involved in cell signaling included two members of the Ras family (Rhpn2 and Rin2), a chemokine (Cxcl5), and a component of intracellular signal transduction (Stmn4) that were also highly induced (Fig. 4A).

Several genes associated with cell cycle and growth regulation were also affected (Table 3). Insulin-like growth factor 1 (Igf1) was included in this category, although it could also be considered with the cell signaling group. Of significance, two genes that influence cell cycle progression and G0-G1 growth arrest were also upregulated: Gas1 and the cyclin-dependent kinase inhibitor p15 (Cdkn2a), also validated by PCR (Fig. 4A).

Group 3: Suppression of increased control gene expression.
In untreated neural crest cultures, the expression of this group of genes, as a whole, increased with time in culture and reached peak expression at 48 h. However, the response to RA treatment was in sharp contrast to untreated expression patterns (Fig. 3). Continued exposure to RA reprogrammed expression patterns and prevented the time-dependent increase observed for this group of genes in control cultures. Although only 12 genes were affected, 75% were cell adhesion and integral membrane proteins or associated with the extracellular matrix or space (Table 4, Fig. 2). These findings are consistent with and highly relevant to the biological effects of RA on neural crest cell adhesion and cell migration (53, 80).


View this table:
[in this window]
[in a new window]
 
Table 4. Group 3: Genes suppressed by RA

 
Several important cell adhesion molecules were identified as affected in our analyses including Vcam1, another reported RA target gene (12, 77), laminin (Lama2), biglycan (Bgn), procollagen type VI {alpha}3 (Col6a3), and thrombospondin (Thbs1; Table 4). Since our criteria for change was set at 1.5-fold up or down, quantitative real-time PCR analysis was used to improve the sensitivity in detecting reduced expression in response to RA. Our results show that Vcam1 expression was markedly reduced compared with control cultures, by more than twofold (Figs. 3 and 4B). Similar analysis of Thbs1, however, detected very little if any difference in expression between control and RA-treated samples (not shown). This was one of only two instances where PCR failed to validate the microarray data.

Two additional genes from this group were of interest. The first was TGFß3, which again could be classified as either a growth factor or a signaling component. Like Vcam1, TGFß3 is a recognized target for RA, although no regulatory motif has been identified (64). The second is the nephroblastoma-overexpressed gene (Nov). Nov, a member of the CCN family, regulates a number of different biological processes, including proliferation, attachment, and migration (65). Thus both TGFß and Nov, whose expression is suppressed by RA, control multiple, key developmental functions.

Group 4: Increases in gene expression over time.
In this group of genes, the magnitude of their response to RA increases over time and could be further subdivided by the time of initial response. At 6 h, only 5% of the genes in this group were significantly induced (one >1.5 expression). By 12 h this number increased to 20%, by 24 h to 55%, and at 48 h all genes demonstrated significant and in some cases robust induction (50% had 2- to 12-fold increases; Fig. 3). This group therefore represented genes whose response to RA may be indirect or secondary and included insulin-like growth factor binding protein 6 (Igfbp6), epidermal growth factor receptor (Egfr), tissue plasminogen activator (Plat), and collagen VII (Col7a1; Table 5, Figs. 3 and 4A). These genes have been identified as RA regulated but frequently require protein synthesis, indicating they are indirect targets (3).


View this table:
[in this window]
[in a new window]
 
Table 5. Group 4: Genes with increased expression over time of exposure to RA

 
The group 4 genes were equally divided between transcription factors, cell cycle/growth regulation, cell signaling, and extracellular components (Table 5, Fig. 2). Genes regulating cell growth that were upregulated included Igfbp3 and Igfbp6 and schlafen 2 (Slfn2). Because of the inhibitory effects of RA on cell proliferation, increased expression of these genes was significant, as all negatively affect cell proliferation (17, 72). Both Igfbp3 and Igfbp6 were checked for expression by PCR. Unlike Igfbp6, we were unable to confirm any significant change in Igfbp3 compared with control samples by either seimquantitative (Fig. 4A) or real-time PCR (not shown) in neural crest cells, although Igfbp3 met the filtering criteria for increased expression and was a reputed secondary target for RA (64, 74). Again, it was difficult to determine the precise ontological category in which some genes should be placed, such as Egfr. It was included with cell cycle/cell growth; however, it fit just as well with cell signaling. Two members of the Wnt family, which participate in noncanonical Wnt signaling (87) fall in group 4: the Wnt 11 ligand and frizzled 4 receptor (Fzd4). The enhanced expression of both genes was verified by PCR (Fig. 4A).

Group 5: Decreases in gene expression over time.
Like the genes of group 4, the magnitude of the responses of group 5 genes increased with prolonged exposure to RA (Fig. 3); however, expression decreased with time. Suppressed expression also occurs secondarily although changes were dramatic only at later time points (Fig. 3).

A number of developmental transcription factors were affected in this category: two members of the distal-less family, expressed in arch mesenchyme (Dlx1 and Dlx5), two Sox family members (Sox2 and Sox11), and Egr2, also known as Krox 20, all of which play important roles in nervous system development (27, 89, 92), and the helix-loop-helix inhibitor Idb4 (Table 6). Combined with cell signaling, these two categories constitute 50% of this responding class. Dickkopf (Dkk1), another component of the Wnt pathway (61), was significant among the signaling molecules. This factor functions as an inhibitor of canonical Wnt signaling, and its suppression by RA would suggest relief of repression of Wnt. Quantitative PCR substantiated a marked reduction in expression of both Dkk and Sox2 at 48 h (Fig. 4B). In the case of Sox2, expression of this gene declined dramatically in control cultures, and RA treatment further exacerbated the decrease (Fig. 3, group 5).


View this table:
[in this window]
[in a new window]
 
Table 6. Group 5: Genes that decrease expression over time of exposure to RA

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
This study was initiated to describe the changes in the cranial/cardiac neural crest transcriptome in response to teratogenic levels of RA and determine how these changes in expression relate to the biological effects of RA on neural crest cells. We have previously shown that treatment of neural crest cultures with RA has a dose-dependent effect on neural crest migration and proliferation (53). Furthermore, by evaluating the time-dependent changes in gene expression that accompany high-dose exposure, we sought to distinguish between early and late effects of RA on gene expression. Our data indicate that RA-mediated cross talk impinges upon multiple transcriptional regulators and signaling pathways that impact upon the described biology of retinoids. The overall changes in gene expression patterns suggest a large scale reprogramming of normal differentiative behavior of neural crest in vitro.

Early changes.
Our findings show that reprogramming is apparent as early as 6 h following RA treatment, in the induction or repression of over 70 different genes (Fig. 1, Tables 1 and 2). This large number of changes was observed even with the ANOVA set to a more stringent P value of 0.01. Most of the early response genes we have identified, including a large number of ESTs, were not previously linked to retinoid-mediated effects. Interestingly, the repression of several genes associated with ubiquitination and protein turnover may provide additional insight into the effects of retinoid biology on neural crest. Ube1x, an ubiquitin-activating enzyme, represents the first step in targeting proteins for degradation. Ube2i acts in combination with an E3 ubiquitin ligase during the second step in ubiquitination. There are over 30 distinct ubiquitin-conjugating (E2) enzymes that can complex with an even larger number of E3 proteins, which recognize distinct and different signals for target degradation (see also Refs. 28 and 37 for reviews). An equally large family of ubiquitin-specific proteases, including Usp52, have also been identified, some of which interact with Hdac6 (73). Decreased expression of these genes could enhance the half-life of specific sets of proteins in neural crest. This has been reported in neuroblastoma, where RA enhanced stability of p27kip1 by reducing ubiquitination and proteasome-dependent degradation (7). Similarly, RA inhibits JNK phosphorylation/activation and growth in human bronchial epithelial cells (49). This inhibition required a posttranslational RAR-dependent mechanism that enhanced the stability of the dual specificity MAP kinase phosphatase 1 (MKP1) (48). Our analyses identified multiple members of the ubiquitin-signaling pathway affected by RA following limited exposure. Thus effects on posttranscriptionally regulated proteins could contribute to the retinoid-mediated inhibition of both cell proliferation and JNK activation in neural crest cultures.

Increased expression.
The sustained expression of genes in group 2 (Fig. 3) further strengthens the case for reprogramming of neural crest cells following RA treatment. Over half of these genes represent transcription factors or components of signaling pathways (Table 3). Many of the genes in this group most likely mediate the subsequent regulation of transcription and the biological responses of the neural crest cells following prolonged exposure to RA, that is, the secondary effects of RA. It is possible that the genes represented in group 4 (increase in expression over time, Fig. 3) include candidate targets for genes of group 2. This hypothesis is compelling for the following reason. Group 2 contains many members of the Hox family, well-known targets for RA (58, 59). Combined with the sustained induction of Pbx2, a cofactor for Hox gene activity (46, 66), the stage is now set for induction of potential Hox downstream target genes. Few, if any, recognized targets of Hox genes have been identified (43), and neural crest cultures therefore provide a platform from which to further explore candidates for Hox activation.

Among the induced genes in groups 2 and 4 are several that negatively affect cell cycle and proliferation. Our previous studied showed that RA treatment reduced BrdU incorporation by 45%, essentially blocking S-phase progression (53). Signals that promote or suppress growth balance cell proliferation rates and converge on G1 and the G1/S transition, a checkpoint in the cycle. During G1, the cell must decide whether to withdraw from the cell cycle into G0, to differentiate, to die, or to initiate DNA synthesis and divide (for review see Ref. 82). Each outcome has been reported in association with retinoid-mediated teratogenesis; therefore, this decision point appears to be a prominent target for RA. Although expression of genes such as Igf1 and the Egfr, which enhance cell proliferation, were induced, so too were Gas-1, Cdkn2b (p15), Igfbp3, Igfbp6, and Slfn2 (Tables 3 and 5). Each of these gene products inhibits growth at this checkpoint (15, 22, 69, 72). Although we do not know whether one option is favored over another in neural crest, one reasonable pathway would be to accelerate differentiation (see Sox2 below; Refs. 21, 33, 34, 52, 86, 95). Enhanced expression of any one of these genes may not be sufficient in and of itself to affect crest cell proliferation; however, the combined, cumulative influence may have a larger impact overall. Thus reprogramming may occur by modestly altering expression of a number of different genes in the same pathway.

Decreased expression.
The constellation of genes affected in groups 3 and 5 dramatically illustrates the kinetic changes in gene expression imposed by RA upon the normal program of neural crest differentiation and supports a combinatorial contribution to reprogramming. RA suppressed expression (Fig. 3), particularly increases that normally occur over the life of the culture. Although the overall effect is similar between groups 3 and 5, the genes targeted fall into distinctly different categories influencing different processes. Most of genes in group 3 (59%) are associated with cell adhesion or the extracellular compartment (Table 4, Fig. 2). Failure to increase expression of adhesion molecules, such as laminin, biglycan, and collagen, and gelsolin, which regulates cytoskeletal interactions with the extracellular matrix (1, 81) contributes to decreased substrate attachment and cell migration. Furthermore, the marked suppression of Vcam1 (Figs. 3 and 4) has two potential consequences. Vcam may play a role in neural crest cell-cell attachment and network organization (83), and decreased expression could contribute to failure of these interactions. Vcam may also mark crest cells destined to differentiate into vascular smooth muscle (78). Neural crest, a major contributor of smooth muscle to embryonic vessels in the head and neck, is vital for vessel persistence and subsequent developmental remodeling (90) (see also Ref. 20 and references therein). Thus retinoids may differentially affect emerging subsets of neural crest cells, inhibiting differentiation of some while promoting differentiation of others.

The majority of genes suppressed in group 5 are developmental transcription factors and components of cell signaling pathways (50%; Table 6, Fig. 2), and may represent repressed secondary targets for RA. The expression of two members each of the SRY-box (Sox2 and Sox11) and distal-less homeobox (Dlx1 and Dlx5) families decreased over time. For Sox2, declining expression occurs in untreated neural crest normally but is further accelerated and amplified by exposure to RA (Figs. 3 and 4). During development Sox2 serves as an early marker of neural cells, and in chick, expression is downregulated as the neural crest leaves the neural tube. Terminal neural differentiation and Sox2 expression are thought to be mutually exclusive (see Refs. 89 and 92). We interpret the rapid and sustained loss of Sox2 expression to support the possibility of premature differentiation subsequent to cell cycle dysregulation. Early differentiation may limit the multipotential character of neural crest and reduce its capacity to contribute to subsequent tissue types. The second set of transcription factors, the distal-less genes (Dlx), regulate proximodistal patterning and are extensively expressed in neural crest-derived ectomesenchyme of the pharyngeal arches. Loss of Dlx1 expression changes the fate of crest-derived mesenchyme, and ablation of Dlx5 affects olfactory bulb and craniofacial development (68, 84). Thus the reduced expression of these developmental regulators has consequences for subsequent cell fate decisions and patterning events, which fit into an overarching blueprint for reprogramming cell behavior.

RA also affects the Wnt signaling pathway. Wnt-frizzled signaling plays a critical role in the initial induction of neural crest (94). In neural crest cultures, RA severely decreased dickkopf (Dkk1) expression, an inhibitor of canonical Wnt signaling (24, 39) (Fig. 4B), whereas Wnt 11, which participates in noncanonical Wnt signaling (reviewed in Ref. 87), and the frizzled 4 receptor are upregulated (Fig. 4A). Loss of Dkk1 by targeted deletion prevents the development of all anterior embryonic structures (61). Although the reduction of anterior structures following RA exposure is most often attributed to alteration of Hox gene expression (see Ref. 58 for review), the repression of Dkk may be a novel unrecognized target in RA teratogenesis. Furthermore, noncanonical Wnt signaling plays an important role in the directed migration of coherent groups of cells during early development (87), a movement pattern well characterized in migrating neural crest cells. Overexpression of Wnt 11 in Xenopus impairs cell movements without changing cell fates (18); evidence also suggests the Fzd4 receptor may activate noncanonical signaling in some cases (40, 75). More significantly, deletion of Fzd4 is associated with hypoplastic thymic development and cardiovascular patterning defects (16). Although this study focused on a loss of function mutation, too much or too little RA produces remarkably similar dysmorphogenetic phenotypes. Thus the reprogramming of Wnt signaling by RA in neural crest may have a direct effect on crest cell migration and contribute to aspects of RA embryopathy in humans.

RA regulation and RAREs.
Approximately 10% of the genes whose expression is altered in neural crest by RA have been previously recognized as RA regulated. Nearly half (11 of 20) are rapidly induced by RA (within 6 h) and expression is sustained at all times (Table 3, group 2). These genes also show the highest level of induction; many of them have characterized RAREs suggesting expression is directly activated by RA (3). The canonical RARE is composed of a direct repeat of six nucleotides (DR) separated by five bases (DR5), by two bases (DR2), or single nucleotide separation (DR1) (4, 11). All types of RAREs are represented within these early responding genes. The Hox A and Hox B clusters both contain DR5 elements located 3' of the clusters (45). While these elements are similar, the Hox B enhancer requires additional conserved elements for RA responsiveness, including a DR2 as well as an additional DR5 element (25). The RBP I gene is regulated by a DR2 (31). Conservation of promoter elements across species has been reported for the DR5 of cytochrome P-450RA suggesting that other elements may also be critical in regulating RA inducibility (55). In contrast, the human CRABPII is regulated by a DR5, while the mouse gene contains both a DR1 and DR2 (2). Thus, despite this commonality of early and/or high-level sustained expression, there does not appear to be any correlation with the type of RARE involved in RA regulation and the type of response we observe in neural crest.

In summary, we have identified over 200 genes that are affected by RA in neural crest cells with distinctive kinetic behaviors and altered expression, representing both early and late targets. Multiple transcription factors and signaling pathways are affected that facilitate the pleiotropic consequences of RA teratogenesis. Although many of these changes in expression are modest over the first 48 h, the combined, cumulative influence on specific pathways may have a larger overall effect and contribute to the final outcome: the reprogramming of the normal development of the neural crest. The implication of this study is that neural crest cells appear to be highly susceptible to developmental reprogramming by RA, and this is likely to be closely entwined with the critical capacity of neural crest cells to undergo multipotential differentiation during embryogenesis.


    GRANTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
This work was supported by National Institutes of Health Grants ES-11747 and PO1-HL-36059 (to M. C. Colbert).


    ACKNOWLEDGMENTS
 
The technical assistance of Linping Wang, Jing Fang, and Liming Sun was greatly appreciated.

A. S. Greene served as the review editor for this manuscript submitted by Editor B. J. Aronow.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: M. C. Colbert, 3333 Burnet Ave., Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229 (E-mail: Melissa.Colbert{at}cchmc.org).

1 The Supplementary Material for this article (Supplemental Tables S1–S3 and Supplemental Fig. S1) is available online at http://physiolgenomics.physiology.org/cgi/content/full/00136.2004/DC1. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Akisaka T, Yoshida H, Inoue S, and Shimizu K. Organization of cytoskeletal F-actin, G-actin, and gelsolin in the adhesion structures in cultured osteoclast. J Bone Miner Res 16: 1248–1255, 2001.[ISI][Medline]
  2. Astrom A, Pettersson U, Chambon P, and Voorhees JJ. Retinoic acid induction of human cellular retinoic acid-binding protein-II gene transcription is mediated by retinoic acid receptor-retinoid X receptor heterodimers bound to one far upstream retinoic acid-responsive element with 5-base pair spacing. J Biol Chem 269: 22334–22339, 1994.[Abstract/Free Full Text]
  3. Balmer JE and Blomhoff R. Gene expression regulation by retinoic acid. J Lipid Res 43: 1773–1808, 2002.[Abstract/Free Full Text]
  4. Bastien J and Rochette-Egly C. Nuclear retinoid receptors and the transcription of retinoid-target genes. Gene 328: 1–16, 2004.[CrossRef][Medline]
  5. Baugh LR, Hill AA, Brown EL, and Hunter CP. Quantitative analysis of mRNA amplification by in vitro transcription. Nucleic Acids Res 29: E29, 2001.
  6. Bolstad BM, Irizarry RA, Astrand M, and Speed TP. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19: 185–193, 2003.[Abstract/Free Full Text]
  7. Borriello A, Pietra VD, Criscuolo M, Oliva A, Tonini GP, Iolascon A, Zappia V, and Ragione FD. p27Kip1 accumulation is associated with retinoic-induced neuroblastoma differentiation: evidence of a decreased proteasome-dependent degradation. Oncogene 19: 51–60, 2000.[CrossRef][ISI][Medline]
  8. Bouillet P, Oulad-Abdelghani M, Vicaire S, Garnier JM, Schuhbaur B, Dolle P, and Chambon P. Efficient cloning of cDNAs of retinoic acid-responsive genes in P19 embryonal carcinoma cells and characterization of a novel mouse gene, Stra1 (mouse LERK-2/Eplg2). Dev Biol 170: 420–433, 1995.[CrossRef][ISI][Medline]
  9. Cerignoli F, Guo X, Cardinali B, Rinaldi C, Casaletto J, Frati L, Screpanti I, Gudas LJ, Gulino A, Thiele CJ, and Giannini G. retSDR1, a short-chain retinol dehydrogenase/reductase, is retinoic acid-inducible and frequently deleted in human neuroblastoma cell lines. Cancer Res 62: 1196–1204, 2002.[Abstract/Free Full Text]
  10. Chambery D, de Galle B, and Babajko S. Retinoic acid stimulates IGF binding protein (IGFBP)-6 and depresses IGFBP-2 and IGFBP-4 in SK-N-SH human neuroblastoma cells. J Endocrinol 159: 227–232, 1998.[Abstract/Free Full Text]
  11. Chambon P. A decade of molecular biology of retinoic acid receptors. FASEB J 10: 940–954, 1996.[Abstract/Free Full Text]
  12. Chen J, Maltby KM, and Miano JM. A novel retinoid-response gene set in vascular smooth muscle cells. Biochem Biophys Res Commun 281: 475–482, 2001.[CrossRef][ISI][Medline]
  13. Chen S and Gardner DG. Retinoic acid uses divergent mechanisms to activate or suppress mitogenesis in rat aortic smooth muscle cells. J Clin Invest 102: 653–662, 1998.[Abstract/Free Full Text]
  14. Creazzo TL, Godt RE, Leatherbury L, Conway SJ, and Kirby ML. Role of cardiac neural crest cells in cardiovascular development. Annu Rev Physiol 60: 267–286, 1998.[CrossRef][ISI][Medline]
  15. Del Sal G, Ruaro ME, Philipson L, and Schneider C. The growth arrest-specific gene, gas1, is involved in growth suppression. Cell 70: 595–607, 1992.[ISI][Medline]
  16. DeRossi C, Laiosa MD, Silverstone AE, and Holdener BC. Mouse fzd4 maps within a region of chromosome 7 important for thymus and cardiac development. Genesis 27: 64–75, 2000.[CrossRef][ISI][Medline]
  17. Dokmanovic M, Chang BD, Fang J, and Roninson IB. Retinoid-induced growth arrest of breast carcinoma cells involves co-activation of multiple growth-inhibitory genes. Cancer Biol Ther 1: 24–27, 2002.[ISI][Medline]
  18. Du SJ, Purcell SM, Christian JL, McGrew LL, and Moon RT. Identification of distinct classes and functional domains of Wnts through expression of wild-type and chimeric proteins in Xenopus embryos. Mol Cell Biol 15: 2625–2634, 1995.[Abstract]
  19. Durand B, Saunders M, Leroy P, Leid M, and Chambon P. All-trans and 9-cis retinoic acid induction of CRABPII transcription is mediated by RAR-RXR heterodimers bound to DR1 and DR2 repeated motifs. Cell 71: 73–85, 1992.[ISI][Medline]
  20. Etchevers HC, Couly G, and Le Douarin NM. Morphogenesis of the branchial vascular sector. Trends Cardiovasc Med 12: 299–304, 2002.[CrossRef][ISI][Medline]
  21. Freemantle SJ, Kerley JS, Olsen SL, Gross RH, and Spinella MJ. Developmentally-related candidate retinoic acid target genes regulated early during neuronal differentiation of human embryonal carcinoma. Oncogene 21: 2880–2889, 2002.[CrossRef][ISI][Medline]
  22. Fuxe J, Akusjarvi G, Goike HM, Roos G, Collins VP, and Pettersson RF. Adenovirus-mediated overexpression of p15INK4B inhibits human glioma cell growth, induces replicative senescence, and inhibits telomerase activity similarly to p16INK4A. Cell Growth Differ 11: 373–384, 2000.[Abstract/Free Full Text]
  23. Geelen JAG. Hypervitaminosis A induced teratogenesis. CRC Crit Rev Toxicology: 351–375, 1979.
  24. Glinka A, Wu W, Delius H, Monaghan AP, Blumenstock C, and Niehrs C. Dickkopf-1 is a member of a new family of secreted proteins and functions in head induction. Nature 391: 357–362, 1998.[CrossRef][ISI][Medline]
  25. Gould A, Itasaki N, and Krumlauf R. Initiation of rhombomeric Hoxb4 expression requires induction by somites and a retinoid pathway. Neuron 21: 39–51, 1998.[ISI][Medline]
  26. Groves AK and Bronner-Fraser M. Neural crest diversification. Curr Top Dev Biol 43: 221–258, 1999.[ISI][Medline]
  27. Hargrave M, Wright E, Kun J, Emery J, Cooper L, and Koopman P. Expression of the Sox11 gene in mouse embryos suggests roles in neuronal maturation and epithelio-mesenchymal induction. Dev Dyn 210: 79–86, 1997.[CrossRef][ISI][Medline]
  28. Hershko A and Ciechanover A. The ubiquitin system. Annu Rev Biochem 67: 425–479, 1998.[CrossRef][ISI][Medline]
  29. Honein MA, Paulozzi LJ, and Erickson JD. Continued occurrence of Accutane-exposed pregnancies. Teratology 64: 142–147, 2001.[CrossRef][ISI][Medline]
  30. Horlein AJ, Naar AM, Heinzel T, Torchia J, Gloss B, Kurokawa R, Ryan A, Kamei Y, Soderstrom M, Glass CK, and Rosenfeld MG. Ligand-independent repression by the thyroid hormone receptor mediated by a nuclear receptor co-repressor. Nature 377: 397–404, 1995.[CrossRef][ISI][Medline]
  31. Husmann M, Hoffmann B, Stump DG, Chytil F, and Pfahl M. A retinoic acid response element from the rat CRBPI promoter is activated by an RAR/RXR heterodimer. Biochem Biophys Res Commun 187: 1558–1564, 1992.[ISI][Medline]
  32. Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, and Speed TP. Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31: e15, 2003.[Abstract/Free Full Text]
  33. Kelley MW, Turner JK, and Reh TA. Retinoic acid promotes differentiation of photoreceptors in vitro. Development 120: 2091–2102, 1994.[Abstract/Free Full Text]
  34. Kelly DL and Rizzino A. DNA microarray analyses of genes regulated during the differentiation of embryonic stem cells. Mol Reprod Dev 56: 113–123, 2000.[CrossRef][ISI][Medline]
  35. Kirby ML, Gale TF, and Stewart DE. Neural crest cells contribute to normal aorticopulmonary septation. Science 220: 1059–1061, 1983.[ISI][Medline]
  36. Kirby ML and Waldo KL. Neural crest and cardiovascular patterning. Circ Res 77: 211–215, 1995.[Free Full Text]
  37. Kornitzer D and Ciechanover A. Modes of regulation of ubiquitin-mediated protein degradation. J Cell Physiol 182: 1–11, 2000.[CrossRef][ISI][Medline]
  38. Krull CE. Segmental organization of neural crest migration. Mech Dev 105: 37–45, 2001.[CrossRef][ISI][Medline]
  39. Krupnik VE, Sharp JD, Jiang C, Robison K, Chickering TW, Amaravadi L, Brown DE, Guyot D, Mays G, Leiby K, Chang B, Duong T, Goodearl AD, Gearing DP, Sokol SY, and McCarthy SA. Functional and structural diversity of the human Dickkopf gene family. Gene 238: 301–313, 1999.[CrossRef][ISI][Medline]
  40. Kuhl M, Sheldahl LC, Malbon CC, and Moon RT. Ca2+/calmodulin-dependent protein kinase II is stimulated by Wnt and Frizzled homologs and promotes ventral cell fates in Xenopus. J Biol Chem 275: 12701–12711, 2000.[Abstract/Free Full Text]
  41. Kuratani SC and Kirby ML. Initial migration and distribution of the cardiac neural crest in the avian embryo: an introduction to the concept of the circumpharyngeal crest. Am J Anat 191: 215–227, 1991.[ISI][Medline]
  42. Kuratani SC and Kirby ML. Migration and distribution of circumpharyngeal crest cells in the chick embryo. Formation of the circumpharyngeal ridge and E/C8+ crest cells in the vertebrate head region. Anat Rec 234: 263–280, 1992.[ISI][Medline]
  43. Kwon Y, Ko JH, Kim B, and Kim MH. Analysis of plausible downstream target genes of Hoxc8 in F9 teratocarcinoma cells. Putative downstream target genes of Hoxc8. Mol Biol Rep 30: 141–148, 2003.[CrossRef][ISI][Medline]
  44. Lammer EJ, Chen DT, Hoar RM, Agnish ND, Benke PJ, Braun JT, Curry CJ, Fernhoff PM, Grix AJ Jr, Lott IT, Richard JM, and Sun SC. Retinoic acid embryopathy. N Engl J Med 313: 837–841, 1985.[Abstract]
  45. Langston AW, Thompson JR, and Gudas LJ. Retinoic acid-responsive enhancers located 3' of the Hox A and Hox B homeobox gene clusters. Functional analysis. J Biol Chem 272: 2167–2175, 1997.[Abstract/Free Full Text]
  46. LaRonde-LeBlanc NA and Wolberger C. Structure of HoxA9 and Pbx1 bound to DNA: Hox hexapeptide and DNA recognition anterior to posterior. Genes Dev 17: 2060–2072, 2003.[Abstract/Free Full Text]
  47. LeDouarin NM and Kalcheim G. The Neural Crest. Cambridge: Cambridge Univ. Press, 1999.
  48. Lee HY, Sueoka N, Hong WK, Mangelsdorf DJ, Claret FX, and Kurie JM. All-trans-retinoic acid inhibits Jun N-terminal kinase by increasing dual-specificity phosphatase activity. Mol Cell Biol 19: 1973–1980, 1999.[Abstract/Free Full Text]
  49. Lee HY, Walsh GL, Dawson MI, Hong WK, and Kurie JM. All-trans-retinoic acid inhibits Jun N-terminal kinase-dependent signaling pathways. J Biol Chem 273: 7066–7071, 1998.[Abstract/Free Full Text]
  50. Lee YH, Campbell HD, and Stallcup MR. Developmentally essential protein flightless I is a nuclear receptor coactivator with actin binding activity. Mol Cell Biol 24: 2103–2117, 2004.[Abstract/Free Full Text]
  51. Lee YM, Osumi-Yamashita N, Ninomiya Y, Moon CK, Eriksson U, and Eto K. Retinoic acid stage-dependently alters the migration pattern and identity of hindbrain neural crest cells. Development 121: 825–837, 1995.[Abstract/Free Full Text]
  52. Li A, Zhu X, Brown B, and Craft CM. Gene expression networks underlying retinoic acid-induced differentiation of human retinoblastoma cells. Invest Ophthalmol Vis Sci 44: 996–1007, 2003.[Abstract/Free Full Text]
  53. Li J, Molkentin JD, and Colbert MC. Retinoic acid inhibits cardiac neural crest migration by blocking c-Jun N-terminal kinase activation. Dev Biol 232: 351–361, 2001.[CrossRef][ISI][Medline]
  54. Locascio A and Nieto MA. Cell movements during vertebrate development: integrated tissue behaviour versus individual cell migration. Curr Opin Genet Dev 11: 464–469, 2001.[CrossRef][ISI][Medline]
  55. Loudig O, Babichuk C, White J, Abu-Abed S, Mueller C, and Petkovich M. Cytochrome P450RAI (CYP26) promoter: a distinct composite retinoic acid response element underlies the complex regulation of retinoic acid metabolism. Mol Endocrinol 14: 1483–1497, 2000.[Abstract/Free Full Text]
  56. Ma Y, Koza-Taylor PH, DiMattia DA, Hames L, Fu H, Dragnev KH, Turi T, Beebe JS, Freemantle SJ, and Dmitrovsky E. Microarray analysis uncovers retinoid targets in human bronchial epithelial cells. Oncogene 22: 4924–4932, 2003.[CrossRef][ISI][Medline]
  57. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M, Chambon P, and Evans RM. The nuclear receptor superfamily: the second decade. Cell 83: 835–839, 1995.[ISI][Medline]
  58. Marshall H, Morrison A, Studer M, Popperl H, and Krumlauf R. Retinoids and Hox genes. FASEB J 10: 969–978, 1996.[Abstract/Free Full Text]
  59. Marshall H, Studer M, Popperl H, Aparicio S, Kuroiwa A, Brenner S, and Krumlauf R. A conserved retinoic acid response element required for early expression of the homeobox gene Hoxb-1. Nature 370: 567–571, 1994.[CrossRef][ISI][Medline]
  60. Miyagawa-Tomita S, Waldo K, Tomita H, and Kirby ML. Temporospatial study of the migration and distribution of cardiac neural crest in quail-chick chimeras. Am J Anat 192: 79–88, 1991.[ISI][Medline]
  61. Mukhopadhyay M, Shtrom S, Rodriguez-Esteban C, Chen L, Tsukui T, Gomer L, Dorward DW, Glinka A, Grinberg A, Huang SP, Niehrs C, Belmonte JC, and Westphal H. Dickkopf1 is required for embryonic head induction and limb morphogenesis in the mouse. Dev Cell 1: 423–434, 2001.[ISI][Medline]
  62. Mulder GB, Manley N, and Maggio-Price L. Retinoic acid-induced thymic abnormalities in the mouse are associated with altered pharyngeal morphology, thymocyte maturation defects, and altered expression of Hoxa3 and Pax1. Teratology 58: 263–275, 1998.[CrossRef][ISI][Medline]
  63. Nau H, Chahoud I, Dencker L, Lammer EJ, and Scott WJ. Teratogenicity of vitamin A and retinoids. In: Vitamin A in Health and Disease, edited by Blumhoff R. New York: Dekker, 1994, p. 615–664.
  64. Nugent P, Ma L, and Greene RM. Differential expression and biological activity of retinoic acid-induced TGFß isoforms in embryonic palate mesenchymal cells. J Cell Physiol 177: 36–46, 1998.[CrossRef][ISI][Medline]
  65. Perbal B. NOV (nephroblastoma overexpressed) and the CCN family of genes: structural and functional issues. Mol Pathol 54: 57–79, 2001.[Abstract/Free Full Text]
  66. Phelan ML, Rambaldi I, and Featherstone MS. Cooperative interactions between HOX and PBX proteins mediated by a conserved peptide motif. Mol Cell Biol 15: 3989–3997, 1995.[Abstract]
  67. Robertson JA, Martinez LP, Gallegos S, Leen-Mitchell MJ, Garcia V, Neuman J, and Carey JC. Accutane cases: a teratogen information service’s approach. Teratology 66: 1–2, 2002.[ISI][Medline]
  68. Robledo RF, Rajan L, Li X, and Lufkin T. The Dlx5 and Dlx6 homeobox genes are essential for craniofacial, axial, and appendicular skeletal development. Genes Dev 16: 1089–1101, 2002.[Abstract/Free Full Text]
  69. Roninson IB and Dokmanovic M. Induction of senescence-associated growth inhibitors in the tumor-suppressive function of retinoids. J Cell Biochem 88: 83–94, 2003.[CrossRef][ISI][Medline]
  70. Salvarezza S and Rovasio R. Exogenous retinoic acid decreases in vivo and in vitro proliferative activity during the early migratory stages of neural crest cells. Cell Prolif 30: 71–80, 1997.[CrossRef][ISI][Medline]
  71. Schwab K, Patterson LT, Aronow BJ, Luckas R, Liang HC, and Potter SS. A catalogue of gene expression in the developing kidney. Kidney Int 64: 1588–1604, 2003.[CrossRef][ISI][Medline]
  72. Schwarz DA, Katayama CD, and Hedrick SM. Schlafen, a new family of growth regulatory genes that affect thymocyte development. Immunity 9: 657–668, 1998.[ISI][Medline]
  73. Seigneurin-Berny D, Verdel A, Curtet S, Lemercier C, Garin J, Rousseaux S, and Khochbin S. Identification of components of the murine histone deacetylase 6 complex: link between acetylation and ubiquitination signaling pathways. Mol Cell Biol 21: 8035–8044, 2001.[Abstract/Free Full Text]
  74. Shang Y, Baumrucker CR, and Green MH. Signal relay by retinoic acid receptors alpha and beta in the retinoic acid-induced expression of insulin-like growth factor-binding protein-3 in breast cancer cells. J Biol Chem 274: 18005–18010, 1999.[Abstract/Free Full Text]
  75. Sheldahl LC, Park M, Malbon CC, and Moon RT. Protein kinase C is differentially stimulated by Wnt and Frizzled homologs in a G-protein-dependent manner. Curr Biol 9: 695–698, 1999.[CrossRef][ISI][Medline]
  76. Shenefelt RE. Morphogenesis of malformations in hamsters caused by retinoic acid: relation to dose and stage at treatment. Teratology 5: 103–118, 1972.[ISI][Medline]
  77. Sheppard AM, McQuillan JJ, Iademarco MF, and Dean DC. Control of vascular cell adhesion molecule-1 gene promoter activity during neural differentiation. J Biol Chem 270: 3710–3719, 1995.[Abstract/Free Full Text]
  78. Sheppard AM, Onken MD, Rosen GD, Noakes PG, and Dean DC. Expanding roles for alpha 4 integrin and its ligands in development. Cell Adhes Commun 2: 27–43, 1994.[ISI][Medline]
  79. Smith WC, Nakshatri H, Leroy P, Rees J, and Chambon P. A retinoic acid response element is present in the mouse cellular retinol binding protein I (mCRBPI) promoter. EMBO J 10: 2223–2230, 1991.[Abstract]
  80. Smith-Thomas L, Lott I, and Bronner-Fraser M. Effects of isotretinoin on the behavior of neural crest cells in vitro. Dev Biol 123: 276–281, 1987.[ISI][Medline]
  81. Spinardi L, Rietdorf J, Nitsch L, Bono M, Tacchetti C, Way M, and Marchisio PC. A dynamic podosome-like structure of epithelial cells. Exp Cell Res 295: 360–374, 2004.[CrossRef][ISI][Medline]
  82. Stein GS, Baserga R, Giordano A, and Denhardt DT. The Molecular Basis of Cell Cycle and Growth Control. New York: Wiley-Liss, 1999, p. 389.
  83. Testaz S, Delannet M, and Duband J. Adhesion and migration of avian neural crest cells on fibronectin require the cooperating activities of multiple integrins of the ß1 and ß3 families. J Cell Sci 112: 4715–4728, 1999.[Abstract/Free Full Text]
  84. Thomas BL, Tucker AS, Qui M, Ferguson CA, Hardcastle Z, Rubenstein JL, and Sharpe PT. Role of Dlx-1 and Dlx-2 genes in patterning of the murine dentition. Development 124: 4811–4818, 1997.[Abstract/Free Full Text]
  85. Thorogood P, Smith L, Nicol A, McGinty R, and Garrod D. Effects of vitamin A on the behaviour of migratory neural crest cells in vitro. J Cell Sci 57: 331–350, 1982.[Abstract]
  86. Truckenmiller ME, Vawter MP, Cheadle C, Coggiano M, Donovan DM, Freed WJ, and Becker KG. Gene expression profile in early stage of retinoic acid-induced differentiation of human SH-SY5Y neuroblastoma cells. Restor Neurol Neurosci 18: 67–80, 2001.[ISI][Medline]
  87. Veeman MT, Axelrod JD, and Moon RT. A second canon. Functions and mechanisms of beta-catenin-independent Wnt signaling. Dev Cell 5: 367–377, 2003.[ISI][Medline]
  88. Wada RK, Seeger RC, Reynolds CP, Alloggiamento T, Yamashiro JM, Ruland C, Black AC, and Rosenblatt JD. Cell type-specific expression and negative regulation by retinoic acid of the human N-myc promoter in neuroblastoma cells. Oncogene 7: 711–717, 1992.[ISI][Medline]
  89. Wakamatsu Y, Endo Y, Osumi N, and Weston JA. Multiple roles of Sox2, an HMG-box transcription factor in avian neural crest development. Dev Dyn 229: 74–86, 2004.[CrossRef][ISI][Medline]
  90. Waldo KL, Kumiski D, and Kirby ML. Cardiac neural crest is essential for the persistence rather than the formation of an arch artery. Dev Dyn 205: 281–292, 1996.[CrossRef][ISI][Medline]
  91. Warkany J, Roth CB, and Wilson JG. Multiple congenital malformations: a consideration of etiologic factors. Pediatrics 1: 462–471, 1948.[ISI]
  92. Wegner M. From head to toes: the multiple facets of Sox proteins. Nucleic Acids Res 27: 1409–1420, 1999.[Abstract/Free Full Text]
  93. Wei LN. Retinoid receptors and their coregulators. Annu Rev Pharmacol Toxicol 43: 47–72, 2003.[CrossRef][ISI][Medline]
  94. Wu J, Saint-Jeannet JP, and Klein PS. Wnt-frizzled signaling in neural crest formation. Trends Neurosci 26: 40–45, 2003.[CrossRef][ISI][Medline]
  95. Yuza Y, Agawa M, Matsuzaki M, Yamada H, and Urashima M. Gene and protein expression profiling during differentiation of neuroblastoma cells triggered by 13-cis retinoic acid. J Pediatr Hematol Oncol 25: 715–720, 2003.[CrossRef][ISI][Medline]
  96. Zhuang Y, Faria TN, Chambon P, and Gudas LJ. Identification and characterization of retinoic acid receptor ß2 target genes in F9 teratocarcinoma cells. Mol Cancer Res 1: 619–630, 2003.[Abstract/Free Full Text]




This Article
Abstract
Full Text (PDF)
Supplemental Tables and Figure
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Search for citing articles in:
ISI Web of Science (2)
Google Scholar
Articles by Williams, S. S.
Articles by Colbert, M. C.
Articles citing this Article
PubMed
PubMed Citation
Articles by Williams, S. S.
Articles by Colbert, M. C.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online
Copyright © 2004 by the American Physiological Society.