- Research article
- Open Access
Genome-wide Twist1 occupancy in endocardial cushion cells, embryonic limb buds, and peripheral nerve sheath tumor cells
© Lee et al.; licensee BioMed Central Ltd. 2014
- Received: 16 May 2014
- Accepted: 22 September 2014
- Published: 28 September 2014
The basic helix-loop-helix transcription factor Twist1 has well-documented roles in progenitor populations of the developing embryo, including endocardial cushions (ECC) and limb buds, and also in cancer. Whether Twist1 regulates the same transcriptional targets in different tissue types is largely unknown.
The tissue-specificity of Twist1 genomic occupancy was examined in mouse ECCs, limb buds, and peripheral nerve sheath tumor (PNST) cells using chromatin immunoprecipitation followed by sequencing (Chip-seq) analysis. Consistent with known Twist1 functions during development and in cancer cells, Twist1-DNA binding regions associated with genes related to cell migration and adhesion were detected in all three tissues. However, the vast majority of Twist1 binding regions were specific to individual tissue types. Thus, while Twist1 has similar functions in ECCs, limb buds, and PNST cells, the specific genomic sequences occupied by Twist1 were different depending on cellular context. Subgroups of shared genes, also predominantly related to cell adhesion and migration, were identified in pairwise comparisons of ECC, limb buds and PNST cells. Twist1 genomic occupancy was detected for six binding regions in all tissue types, and Twist1-binding sequences associated with Chst11, Litaf, Ror2, and Spata5 also bound the potential Twist1 cofactor RREB1. Pathway analysis of the genes associated with Twist1 binding suggests that Twist1 may regulate genes associated with the Wnt signaling pathway in ECCs and limb buds.
Together, these data indicate that Twist1 interacts with genes that regulate adhesion and migration in different tissues, potentially through distinct sets of target genes. In addition, there is a small subset of genes occupied by Twist1 in all three tissues that may represent a core group of Twist1 target genes in multiple cell types.
- Endocardial cushions
- Limb buds
- Peripheral neuron sheath tumor cells
Twist1 is a highly conserved basic-helix-loop-helix (bHLH) transcription factor, which regulates cell migration, proliferation, and differentiation in progenitor populations, such as embryonic neural crest cells, cranial mesoderm, limb buds, and endocardial cushions (ECCs) [1–4]. In addition to its role in developmental processes, human TWIST1 is highly expressed in diseased heart valves and in metastatic tumor cells [5–7]. Twist1 expression is normally silenced after embryonic development, but expression is reactivated during disease, including cancer. Twist1 controls gene expression through binding E-box (5’-CANNTG-3’) consensus sites as a homodimer or heterodimer with other bHLH transcription factors [4, 8]. However, a comparison of Twist1 direct transcriptional targets in developing and diseased tissues has yet to be reported. This study focuses on identifying and comparing Twist1 genomic occupancy in developing ECC, limb buds, and a cancer cell line.
In multiple structures in the developing embryo, including the heart, limb buds, cranial sutures, skeletal elements, and neural crest derivatives, Twist1 regulates cell migration and proliferation of progenitor populations . In the developing heart, Twist1 expression is robust in the mesenchymal cells of the ECC and is subsequently down-regulated during remodeling and differentiation of the valve leaflets [9, 10]. In the ECC, Twist1 promotes cell migration and proliferation of mesenchymal cells. Direct transcriptional targets of Twist1 include genes involved in cell migration (Tbx20, Cdh11, Sema3C) and proliferation (Rab39b and Gadd45a) [2, 5, 9]. In the developing limbs, Twist1 is broadly expressed in the limb bud mesenchyme and strictly regulated genetic dosage of Twist1 is required for limb bud outgrowth and patterning [1, 3, 11]. Although extensive genetic studies have been performed demonstrating an essential role for Twist1 in limb patterning and morphogenesis, the direct transcriptional targets underlying these functions are only recently being elucidated . In addition, it is not known if Twist1 regulates the same downstream targets in distinct progenitor populations or if there is tissue-specificity to Twist1 regulatory mechanisms in the developing embryo.
While Twist1 is largely absent from adult differentiated tissues, it is expressed in diseased heart valves and highly metastatic cancers such as breast, pancreatic, gastric, prostate, and malignant peripheral nerve sheath tumors (MPNST) [6, 7, 13, 14]. Human MPNST cells can harbor NF1 and p53 mutations, and murine NF1 and p53 mutations can cause similar nerve-associated sarcomas, peripheral nerve sheath tumors (PNST) [14–16]. siRNA-mediated knockdown of Twist1 abrogates the migratory activity of human MPNST cells in vitro [14, 16]. Twist1 also is expressed in invasive carcinoma and promotes endothelial-to-mesenchymal transformation (EMT) and metastasis of tumor cells through regulation of genes involved in differentiation, adhesion, and proliferation [6, 17]. It has not previously been determined if the downstream targets of Twist1 related to regulation of cell migration or proliferation in embryonic structures and cancer cells are shared or are cell-type specific.
To investigate the tissue-specificity of Twist1 genomic occupancy, we performed chromatin immunoprecipitation (ChIP) with anti-Twist1 followed by NextGen Sequencing (ChIP-seq) in mouse embryonic day (E)12.5 ECCs, E10.5 forelimb buds, and cultured PNST cells. The specificity of Twist1 binding was confirmed by ChIP qPCR in each tissue. Potential Twist1 cofactors and intersecting signaling pathways also were identified.
Analysis of Twist1 genomic occupancy in ECCs, limb buds, and PNST cells
Tissue-specific and shared gene targets of Twist1 in ECCs, limb buds, and PNST cells corresponding to binding regions within 50 Kb of TSS were analyzed . Binding region overlap determination was performed before and after peaks were compared to gene expression profiles (Figure 1B). These analyses produced datasets containing binding regions that are unique to a specific tissue type or shared among multiple tissue types. Groups of binding regions were classified as ECC only, Limb only, PNST only, ECC-Limb, ECC-PNST, Limb-PNST, or ECC-Limb-PNST. E12.5 ECCs and E10.5 limb buds share the most binding regions, with 1,237 shared peaks in the ECC-Limb group before comparison to mRNA expression data, and 205 ECC-Limb binding regions associated with expressed genes after comparison with mRNA expression. Similarly the Limb-PNST group contains 722 shared binding regions, before the expression filter, with only 8 shared binding regions corresponding to expressed genes. The ECC-PNST group contains 657 total shared binding regions, with 10 shared binding regions corresponding to expressed genes. Surprisingly, only 6 binding regions associated with expressed genes were occupied by Twist1 in all three cell types (ECC-Limb-PNST), indicating a high degree of tissue-specificity in Twist1 genomic occupancy. Genomic sequences of the shared binding regions associated with expressed genes were analyzed for presence of an E-box consensus site on the sense or antisense strand (5’-CANNTG-3’), which would indicate a potential Twist1 binding site. Of the 205 ECC-Limb shared binding regions associated with expressed genes, 138 contain an E-box consensus site, corresponding to 108 genes. In addition, E-box consensus sites are present in 7/10 ECC-PNST shared binding regions, 6/8 of the Limb-PNST binding regions, and 5/6 of the ECC-Limb-PNST shared binding regions. Collectively, these ChIP-seq experiments demonstrate extensive Twist1 binding to genomic sequences and indicate that Twist1 bound predominantly in a tissue-specific manner, with much smaller numbers of shared binding regions detected among ECCs, limb buds, and PNST cells.
In order to examine Twist1-DNA binding distribution across the genome and determine if Twist1 is bound to specific genomic regions relative to gene organization, location analysis was performed. Twist1 binding regions located within 50 Kb of a TSS for genes expressed in each tissue were subjected to genomic location and peak distribution analysis using Genomatix RegionMiner (Additional file 6). Twist1 binding regions associated with expressed genes were analyzed in tissue-specific gene groups (ECC only, Limb only, and PNST only), and also in groups of binding regions shared between tissue types (ECC-Limb, ECC-PNST, and Limb-PNST). Distribution of the tissue-specific binding regions and ECC-Limb shared peaks is largely within intergenic and intronic regions. However, ECC-PNST and Limb-PNST shared binding regions are more evenly distributed in genomic regions. This lack of specificity in peak distribution relative to gene organization is consistent with other transcription factors, including Twist1, in other cell types [22, 23]. These data suggest that Twist1 occupies genomic sequences associated with expressed genes located in intergenic and intragenic genomic regions and does not bind exclusively to promoter regions.
Functional classification of genes associated with Twist1 binding peaks
Validation of Twist1 genomic occupancy in ECC, limb buds and PNST cells
Identification of potential Twist1 cofactors in ECCs, limb buds, and PNST cells
The presence of tissue-specific Twist1 cofactors was examined as a potential mechanism underlying specificity of target occupancy. Twist1 binding regions associated with expressed genes for each tissue type (Additional files 7, 8 and 9) were analyzed for transcriptional modules containing a transcription factor-binding sequence within +/- 40 base pairs of the E-box consensus site using Genomatix RegionMiner (Additional file 10). In ECC, CP2-erythrocyte factor related to drosophila Elf1 (CP2) sites and heat shock factor sites are located in proximity to E-boxes in approximately 10% of the Twist1 occupied sequences. Retinoid X receptor (RXR) sites (Figure 5E) are present in 43% of the Limb-only Twist1 occupied sites and also in 44% of the ECC-Limb shared Twist1 occupied sites. Interestingly, binding sites for Ras-related element binding protein (RREB1) sites (Figure 5F) are present in 4/5 E-box-containing peaks in ECC-Limb-PNST shared Twist1 binding regions. To validate binding occupancy of Twist1 and candidate cofactors to Twist1 binding regions identified by ChIP-seq, sequence-specific ChIP was performed with antibodies directed against RXR and RREB1 [24, 25]. These analyses confirm that Twist1 and RXR can occupy Twist1 binding regions associated with Wnt5b and Ror2 (Ror2-a) in limb buds (Figure 5C). RREB1 binding to shared Twist1 binding region sequences associated with Spata5, Ror2 (Ror2-a), Litaf, and Chst11 was detected in PNST cells, and RREB1 bound to Spata5 and Litaf Twist1 binding regions in limb buds (Figure 5B,D). However, RREB1 occupancy at shared binding regions was not detected in ECCs. Together these data provide evidence that RXR can also occupy Twist1 binding regions in limb buds, and that RREB1 can occupy Twist1 binding regions in both limb buds and PNST cells, but not ECC. Therefore, RXR is a candidate cofactor for Twist1 in limb buds, and RREB1 is a candidate cofactor for Twist1 in limb buds and PNST cells.
Twist1 DNA occupancy is associated with genes in the Wnt signaling pathway
Tissue-specificity of Twist1 binding to distinct subsets of gene sequences was observed by ChIP-seq and then confirmed by ChIP and sequence-specific qPCR in ECC, limb buds, and PNST cells. The tissue-specific binding regions occupied by Twist1 are consistent with known functions of Twist1 in each cell type, although the individual genes are distinct. Twist1 binds to DNA sequences associated with genes that regulate migration and other biological functions in each tissue, but few shared binding regions were detected between samples. Among the six sequences occupied by Twist1 in all three tissues are genes involved in cell adhesion-migration that are also bound by the Ras-related element binding protein (RREB1). Moreover, Twist1 binding regions are associated with several Wnt pathway genes and pathway components. These data suggest that Twist1 binds target genomic sequences predominantly in a tissue-specific manner and provide initial evidence that Twist1 occupies sequences of genes acting in the Wnt signaling pathway. However, the mechanisms that dictate the tissue-specificity of Twist1 binding to distinct transcriptional targets in different cell types remain to be determined.
Analysis of Twist1 and its downstream target genes in a variety of cell types have revealed roles in cell proliferation, adhesion/migration, and ECM protein expression. In the ECC mesenchymal cells of the developing heart valves, alterations in Twist1 expression affect cell migration and expression of Tbx20, Cadherin-11, and Periostin [2, 9]. Additionally, Twist1 binding to genomic regions associated with genes that regulate cell proliferation and migration, including Tbx20, Cdh11, Rab39b, Sema3C, and Gadd45a, have been identified . Likewise, expression analysis of AVC of Twist1-null embryos relative to WT mice revealed reduced expression of cell migration and ECM genes . Interestingly, Ror2 expression is decreased in Twist1-null ECC and also was found to contain a Twist1 binding region in this study . In synovial sarcoma cells, Twist1 direct downstream targets are involved in maintenance of mesenchymal differentiation , and Twist1 targets promote cell migration and proliferation in thyroid cancer cells . Interestingly, in the current study only one binding region, Ror2, overlaps with these 3 other Twist1 ChIP-seq studies. These studies further support the ability of Twist1 to regulate different networks of downstream genes depending on the specific cellular context. However, predominant roles for Twist1 downstream target genes in cell migration and adhesion are highly reproducible.
Potential Twist1 cofactors were identified based on the presence of transcription factor binding sites in Twist1 binding regions and were confirmed by ChIP-qPCR in the tissues of interest. The retinoic acid (RA) receptor RXR was identified as a potential Twist1 cofactor in limb buds. Both RA signaling and Twist1 have been implicated in limb bud patterning, and the binding of both factors to the same genomic target sequence supports RXR as a candidate cofactor for Twist1 [4, 28]. In skeletal muscle, the bHLH transcription factor MyoD physically interacts with RXR to promote cell-type specific gene regulation, suggesting the possibility of a related or intersecting regulatory mechanism with Twist1 . RREB1 binding sites are present in proximity to Twist1 occupied sites detected in ECC, limb buds, and PNST cells. RREB1 is expressed in developing tissues and cancer where it acts downstream of Ras and MAPK signaling Twist1 is required for K-Ras mediated progression of lung tumors, supporting a role for RREB1 in this regulatory interaction [29, 30]. In addition, RREB1 has been shown to interact and bind with bHLH factor NeuroD to enhance transcriptional activity [25, 30]. Thus Twist1 and RREB1 may act together in transcriptional regulation of genes activated by Ras-MAPK signaling in multiple tissue types. Together, these studies have uncovered candidate cofactors, previously demonstrated to interact with related bHLH factors, which may contribute to the cell-type specificity of Twist1 genomic occupancy and target gene regulation.
Increasing evidence implicates the Wnt signaling pathway as a key pathway active in developing heart valves and limb buds and also in disease states, such as cancer . Twist1 genomic occupancy was observed in association with genes in the Wnt pathway in ECC and limb buds. In ECCs, Wnt/β-catenin signaling is required for EMT of the endothelial cells, and for proliferation of mesenchymal cells [32, 33]. In Twist1-null AVC and OFT, genes associated with the Wnt signaling pathway are significantly differentially regulated relative to control embryos . Likewise, conditional deletion of the Twist1 target gene Tbx20 in endothelial cells leads to reduced expression of Wnt pathway genes Sfrp1 and Fzd4 in ECCs . The current study identified genomic sequences associated with Sfrp1 and Fzd4 occupied by Twist1 in ECCs, supporting an intersecting regulatory network for Twist1, Tbx20, and Wnt signaling in heart valve development. In developing limb buds, Wnt signaling is critical for limb bud outgrowth, while Wnt and RA signaling are required for proximal-distal limb patterning [35–37]. Here, we identify Wnt5b associated sequences bound by Twist1 and RXR in limb buds. In limb bud cartilage progenitors, RA and Wnt signaling pathways intersect, and Wnt5b promotes cell migration and inhibition of cell adhesion mediated by cadherins, which are often regulated by Twist1 [38, 39]. Additionally, mutations in Wnt/β-catenin signaling pathway genes have been identified in patients with neurofibromatosis type 1, and in vitro experiments in MPNST cells demonstrate that Wnt signaling promotes cell proliferation [40, 41]. Furthermore, Wnt signaling through Ror2 is linked with cell invasion and EMT in human osteosarcoma cells, and the same binding region, Ror2-b, was recently identified as a Twist1 target in both E10.5 AVC of the heart and developing limb buds [12, 42]. The identification of Twist1 binding regions in proximity to multiple genes associated with Wnt signaling suggests that Twist1 may regulate components of the Wnt signaling pathway in ECCs, limb buds, and PNST cells.
This study examined Twist1 genomic occupancy in two murine developmental tissues (E12.5 ECCs and E10.5 limb buds) and a murine cancer cell line (PNST cells). Interestingly, little overlap between binding regions was observed in ECCs, limb buds, and PNST cells, although the predominant functions of Twist1 occupied genes, including cell adhesion and migration are similar in the three tissues. The specificity of Twist1 binding was confirmed for tissue-specific targets, while Twist1 binding in multiple tissues was detected for shared target genes, including four genes that represent core targets of Twist1 in multiple cell types. The molecular basis for the tissue-specificity of Twist1 binding remains to be determined. Interestingly RREB1 binding sites are located in proximity to Twist1 sites in the occupied regions detected in multiple tissues providing initial evidence for a core regulatory interaction. In addition, we provide evidence that Twist1 interacts with genes in the Wnt signaling pathway in ECCs, limb buds, and PNST cells. A limitation of our analysis is the focus on Twist1 genomic occupancy of expressed genes, which is indicative of Twist1 acting as a transcriptional activator, not repressor. However, the identification of tissue-specific targets and DNA binding mechanisms of genes potentially activated by Twist1 will aid in our understanding of its role in transcriptional hierarchies active in development and disease.
All experiments involving animals were performed using experimental protocols and procedures reviewed and approved by the Cincinnati Children’s Hospital Medical Center Biosafety Committee and Institutional Animal Care and Use Committee.
Chromatin immunoprecipitation (ChIP) sample isolation
FVBN (Taconic) wild-type mouse litters were generated from timed matings, where presence of the copulation plug was considered E0.5. Atrioventricular canal (AVC) ECC tissue from E12.5 embryos was dissected in 1X phosphate buffered saline (1XPBS) with tungsten needles and pooled from 10-12 individuals for each ChIP experiment [43, 44]. Forelimb bud tissue from 10-12 E10.5 embryos was isolated and pooled in 1XPBS. Mouse peripheral nerve sheath tumor (PNST) cells were cultured in DMEM medium (Gibco) containing 10% fetal bovine serum (FBS, Hyclone) and 1% penicillin-streptomycin (pen-strep, Invitrogen) [15, 45]. PNST cells were cultured in 100mm tissue culture plates, and cells from 1 plate were utilized for each experiment when they reached 80-90% confluency. Genomic DNA from ECCs, limb buds, and PNST cells was cross-linked and isolated according to manufacturer’s protocols (Magnify ChIP kit, Invitrogen). Crosslinked DNA complexes were sheared by sonication (Virsonic 60, Virtis, output of 5) 6 times for 5 seconds with a 2-minute refractory period on ice. Samples were sheared to a size of 150-250 basepairs (bp). ChIP was then performed according to the manufacturer’s protocol (Magnify ChIP Kit, Invitrogen). Immunoprecipitation was performed with anti-Twist1 (Sigma, T6451, 5 μg) [5, 33], anti-RREB1 (Santa Cruz, sc-138055, 5 μg) , and anti-RXR (Santa Cruz, sc-774, 5 μg)  or with the appropriate IgG control (Magnify ChIP Kit, Invitrogen, 492024, 5 μg).
Approximately 5 litters of FVB/N AVC ECCS and limb buds and 2, 100 mm culture plates of PNST cells were isolated and pooled for ChIP-seq analysis. Each ChIP-seq sample yielded approximately 10 ng of DNA for sequencing with Illumina HiSeq2000. A single read genomic DNA library was generated using an IntegenX robot (Apollo 324) and PrepX DNA library kit (IntegenX 400040). NextGen high-throughput sequencing was performed on ChIP DNA fragments using an Illumina HiSeq2000 with 10 million raw reads per sample following the manufacturer’s protocols. Peaks were called using Model-based Analysis of ChIP-seq (MACS, v1.4.2) using the following parameters, band width = 300, and p-value cut off of 1.00e-05 relative to a whole mouse genome data set generated from wild type FVB/N mouse genomic DNA as a control. 50bp sequence reads were mapped to the mouse genome (NCBI Build 37, mm9) with Illumina gapped aligner Eland v2 and Bowtie to identify the nearest transcriptional start site upstream and downstream of the peak . ChIP-seq datasets of unaligned reads and reads aligned to the genome for E12.5 ECCs, E10.5 limb buds, and PNST cells are available on the Gene Expression Omnibus (GEO, accession #GSE50111).
In order confirm that E-box binding consensus sequences were enriched throughout the ChIP-seq datasets for E12.5 ECCs, E10.5 limb buds, and PNST cells de-novo motif analysis was performed using PscanChIP, which utilizes the TRANSFAC database, and MEME-ChIP, which utilizes the JASPAR database [18, 19]. Chromosomal coordinates for binding peaks following MACS analysis were used as input for PscanChIP analysis, and sequences of binding peaks following MACS analysis were used as input for MEME-ChIP.
Gene expression data
Mouse gene expression datasets generated by Mouse Genome 430 2.0 Array (Affymetrix) and raw data files (.cel) were obtained from the GEO database for developing heart valves (E12.5 ECCs and E17.5 valves [GSE11040]), and developing forelimb buds (E10.5 forelimb buds and E13.5 forelimb buds [GSE30138]) [10, 20]. Raw Affymetrix microarray expression data (.cel) files were processed with GeneSpring v12.1 software with robust multichip average (RMA) analysis [5, 10]. Pairwise analysis was performed on datasets from the same GEO submission, E12.5 ECCs versus E17.5 valves, E10.5 forelimbs versus E13.5 forelimbs. Developmentally expressed genes were identified as genes present in E12.5 ECCs (≥1 fold expression in E12.5 ECCs v. E17.5 limb buds) or in E10.5 limb buds (≥1 fold expression in E10.5 limb buds v. E13.5 limb buds). These lists of probe sets associated with genes include genes expressed at both developmental stages or preferentially at the earlier stage, but do not include the genes expressed preferentially at later stages of valve or limb maturation. Genes enriched in human MPNST and mouse PNST cells, as compared to normal nerves for each species, were identified as described [14, 16]. These probe set expression lists were used for comparison to genomic sequences bound by Twist1, as determined by ChIP-seq.
Genes with Twist1 ChIP-seq peaks located within 50 Kb of the transcriptional start site (TSS) were compared to E12.5 ECC, E10.5 limb bud, and human MPNST/mouse PNST gene expression microarray lists using GeneSpring v12.1 software (Agilent). Genes associated with Twist1 ChIP-seq peaks for ECC, limb bud, and PNST cells within 50 Kb of a TSS were imported into GeneSpring for conversion of each gene into corresponding probe sets on the Mouse Genome 430 2.0 (Affymetrix) Array platform. Probe sets from ChIP-seq peak associated gene lists and corresponding tissue-specific probe set expression lists (ECCs, limb buds, and shared human MPNST/mouse PNST cells) were compared. Overlapping expression was represented in Venn diagrams and shared probe sets were identified. The genes associated with shared ChIP-seq probe sets and tissue-specific expressed probe sets were analyzed for Gene Ontology (GO) enrichment using ToppGene and PANTHER databases [26, 48].
Analysis of ChIP-seq datasets
ECC, limb bud, and PNST Twist1 ChIP-seq peaks associated with genes that are expressed in each cell type were then compared with each other to identify peaks shared between datasets (ECC-Limb, ECC-PNST, Limb-PNST, and ECC-Limb-PNST) using IntersectBed from Nebula with a minimum overlap of 5% of the nucleotides in shared peaks . Genomic sequences for each of the binding peak sequences (~50 bp to 750 bp in length) were obtained using Genomatix Software Suite v2.7. RegionMiner (Genomatix Software Suite v2.7, Release 4.7) was used to determine the enriched transcription factor-binding sites (TFBS) present in each data set. Peaks associated with expressed genes in ECC only, limb bud only, PNST only, ECC-Limb shared, ECC-PNST shared, and sequences shared in the three cell types (ECC-Limb-PNST) were extended +/- 200 bp. Transcription modules containing an E-box binding site plus another binding site with a z-score ≤ -2.0 within 40 bp of an E-box consensus site were identified. Lists of genes associated with Twist1 binding peaks unique to each cell type or shared among multiple cell types were examined for pathway identification and interactions using ToppGene Suite and GeneMANIA (http://www.genemania.org) [26, 49].
Quantitative real-time PCR (qPCR) analysis of ChIP
Relative abundance of specific genomic sequences in eluted DNA from ChIP experiments was quantified by qPCR compared to normal rabbit IgG control as previously described . qPCR amplification began with an initial denaturation step of 94°C for 3 minutes followed by 40 cycles of 94°C for 20 sec, annealing step for 30 seconds (annealing temperature was 55°C for all primer sets), 72°C for 30 seconds. Fold-enrichment was calculated relative to IgG control, which was set to 1.0. Primer sequences, distance of the binding peak from the nearest transcriptional start site, and annealing temperatures are listed in Additional file 11.
Availability of supporting data
The data sets supporting the results of this article are available in the GEO database, [GSE50111].
The authors would like to thank Anil Jegga for technical support with bioinformatics and the members of the Yutzey laboratory for critical analysis and comments during manuscript preparation.
- Chen ZF, Behringer RR: Twist is required in head mesenchyme for cranial neural tube morphogenesis. Genes Dev. 1995, 9 (6): 686-699.PubMedView ArticleGoogle Scholar
- Shelton EL, Yutzey KE: Twist1 function in endocardial cushion cell proliferation, migration, and differentiation during heart valve development. Dev Biol. 2008, 317 (1): 282-295.PubMed CentralPubMedView ArticleGoogle Scholar
- Firulli BA, Redick BA, Conway SJ, Firulli AB: Mutations within helix I of Twist1 result in distinct limb defects and variation of DNA binding affinities. J Biol Chem. 2007, 282 (37): 27536-27546.PubMed CentralPubMedView ArticleGoogle Scholar
- Barnes RM, Firulli AB: A twist of insight - the role of Twist-family bHLH factors in development. Int J Dev Biol. 2009, 53 (7): 909-924.PubMed CentralPubMedView ArticleGoogle Scholar
- Lee MP, Yutzey KE: Twist1 directly regulates genes that promote cell proliferation and migration in developing heart valves. PLoS One. 2011, 6 (12): e29758-PubMed CentralPubMedView ArticleGoogle Scholar
- Qin Q, Xu Y, He T, Qin C, Xu J: Normal and disease-related biological functions of Twist1 and underlying molecular mechanisms. Cell Res. 2012, 22 (1): 90-106.PubMed CentralPubMedView ArticleGoogle Scholar
- Wirrig EE, Hinton RB, Yutzey KE: Differential expression of cartilage and bone-related proteins in pediatric and adult diseased aortic valves. J Mol Cell Cardiol. 2011, 50 (3): 561-569.PubMed CentralPubMedView ArticleGoogle Scholar
- Bildsoe H, Loebel DA, Jones VJ, Hor AC, Braithwaite AW, Chen YT, Behringer RR, Tam PP: The mesenchymal architecture of the cranial mesoderm of mouse embryos is disrupted by the loss of Twist1 function. Dev Biol. 2013, 374 (2): 295-307.PubMedView ArticleGoogle Scholar
- Chakraborty S, Wirrig EE, Hinton RB, Merrill WH, Spicer DB, Yutzey KE: Twist1 promotes heart valve cell proliferation and extracellular matrix gene expression during development in vivo and is expressed in human diseased aortic valves. Dev Biol. 2010, 347 (1): 167-179.PubMed CentralPubMedView ArticleGoogle Scholar
- Chakraborty S, Cheek J, Sakthivel B, Aronow BJ, Yutzey KE: Shared gene expression profiles in developing heart valves and osteoblast progenitor cells. Physiol Genomics. 2008, 35 (1): 75-85.PubMed CentralPubMedView ArticleGoogle Scholar
- Krawchuk D, Weiner SJ, Chen YT, Lu BC, Costantini F, Behringer RR, Laufer E: Twist1 activity thresholds define multiple functions in limb development. Dev Biol. 2010, 347 (1): 133-146.PubMed CentralPubMedView ArticleGoogle Scholar
- Vrljicak P, Cullum R, Xu E, Chang AC, Wederell ED, Bilenky M, Jones SJ, Marra MA, Karsan A, Hoodless PA: Twist1 transcriptional targets in the developing atrio-ventricular canal of the mouse. PLoS One. 2012, 7 (7): e40815-PubMed CentralPubMedView ArticleGoogle Scholar
- Ansieau S, Morel AP, Hinkal G, Bastid J, Puisieux A: TWISTing an embryonic transcription factor into an oncoprotein. Oncogene. 2010, 29 (22): 3173-3184.PubMedView ArticleGoogle Scholar
- Miller SJ, Rangwala F, Williams J, Ackerman P, Kong S, Jegga AG, Kaiser S, Aronow BJ, Frahm S, Kluwe L, Mautner V, Upadhyaya M, Muir D, Wallace M, Hagen J, Quelle DE, Watson MA, Perry A, Gutmann DH, Ratner N: Large-scale molecular comparison of human schwann cells to malignant peripheral nerve sheath tumor cell lines and tissues. Cancer Res. 2006, 66 (5): 2584-2591.PubMedView ArticleGoogle Scholar
- Zhu Y, Ghosh P, Charnay P, Burns DK, Parada LF: Neurofibromas in NF1: Schwann cell origin and role of tumor environment. Science. 2002, 296 (5569): 920-922.PubMed CentralPubMedView ArticleGoogle Scholar
- Jessen WJ, Miller SJ, Jousma E, Wu J, Rizvi TA, Brundage ME, Eaves D, Widemann B, Kim MO, Dombi E, Sabo J, Hardiman Dudley A, Niwa-Kawakita M, Page GP, Giovannini M, Aronow BJ, Cripe TP, Ratner N: MEK inhibition exhibits efficacy in human and mouse neurofibromatosis tumors. J Clin Invest. 2013, 123 (1): 340-347.PubMed CentralPubMedView ArticleGoogle Scholar
- Feng MY, Wang K, Shi QT, Yu XW, Geng JS: Gene expression profiling in TWIST-depleted gastric cancer cells. Anat Rec. 2009, 292 (2): 262-270.View ArticleGoogle Scholar
- Zambelli F, Pesole G, Pavesi G: PscanChIP: Finding over-represented transcription factor-binding site motifs and their correlations in sequences from ChIP-Seq experiments. Nucleic Acids Res. 2013, 41 (Web Server issue): W535-W543.PubMed CentralPubMedView ArticleGoogle Scholar
- Machanick P, Bailey TL: MEME-ChIP: motif analysis of large DNA datasets. Bioinformatics. 2011, 27 (12): 1696-1697.PubMed CentralPubMedView ArticleGoogle Scholar
- Taher L, Collette NM, Murugesh D, Maxwell E, Ovcharenko I, Loots GG: Global gene expression analysis of murine limb development. PLoS One. 2011, 6 (12): e28358-PubMed CentralPubMedView ArticleGoogle Scholar
- Boeva V, Lermine A, Barette C, Guillouf C, Barillot E: Nebula–a web-server for advanced ChIP-seq data analysis. Bioinformatics. 2012, 28 (19): 2517-2519.PubMedView ArticleGoogle Scholar
- Hutchins AP, Poulain S, Miranda-Saavedra D: Genome-wide analysis of STAT3 binding in vivo predicts effectors of the anti-inflammatory response in macrophages. Blood. 2012, 119 (13): e110-e119.PubMedView ArticleGoogle Scholar
- Lee KW, Lee NK, Ham S, Roh TY, Kim SH: Twist1 is essential in maintaining mesenchymal state and tumor-initiating properties in synovial sarcoma. Cancer Lett. 2014, 343 (1): 62-73.PubMedView ArticleGoogle Scholar
- Froeschle A, Alric S, Kitzmann M, Carnac G, Aurade F, Rochette-Egly C, Bonnieu A: Retinoic acid receptors and muscle b-HLH proteins: partners in retinoid-induced myogenesis. Oncogene. 1998, 16 (26): 3369-3378.PubMedView ArticleGoogle Scholar
- Ray SK, Nishitani J, Petry MW, Fessing MY, Leiter AB: Novel transcriptional potentiation of BETA2/NeuroD on the secretin gene promoter by the DNA-binding protein Finb/RREB-1. Mol Cell Biol. 2003, 23 (1): 259-271.PubMed CentralPubMedView ArticleGoogle Scholar
- Chen J, Bardes EE, Aronow BJ, Jegga AG: ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Res. 2009, 37 (Web Server issue): W305-W311.PubMed CentralPubMedView ArticleGoogle Scholar
- Di Maro G, Orlandella FM, Bencivenga TC, Salerno P, Ugolini C, Basolo F, Maestro R, Salvatore G: Identification of targets of twist1 transcription factor in thyroid cancer cells. J Clin Endocrinol Metab. 2014, 99 (9): E1617-E1626.PubMedView ArticleGoogle Scholar
- Tickle C: Limb development: an international model for vertebrate pattern formation. Int J Dev Biol. 2000, 44 (1): 101-108.PubMedGoogle Scholar
- Tran PT, Shroff EH, Burns TF, Thiyagarajan S, Das ST, Zabuawala T, Chen J, Cho YJ, Luong R, Tamayo P, Salih T, Aziz K, Adam SJ, Vicent S, Nielsen CH, Withofs N, Sweet-Cordero A, Gambhir SS, Rudin CM, Felsher DW: Twist1 suppresses senescence programs and thereby accelerates and maintains mutant Kras-induced lung tumorigenesis. PLoS Genet. 2012, 8 (5): e1002650-PubMed CentralPubMedView ArticleGoogle Scholar
- Kent OA, Fox-Talbot K, Halushka MK: RREB1 repressed miR-143/145 modulates KRAS signaling through downregulation of multiple targets. Oncogene. 2013, 32 (20): 2576-2585.PubMedView ArticleGoogle Scholar
- Clevers H: Wnt/beta-catenin signaling in development and disease. Cell. 2006, 127 (3): 469-480.PubMedView ArticleGoogle Scholar
- Liebner S, Cattelino A, Gallini R, Rudini N, Iurlaro M, Piccolo S, Dejana E: Beta-catenin is required for endothelial-mesenchymal transformation during heart cushion development in the mouse. J Cell Biol. 2004, 166 (3): 359-367.PubMed CentralPubMedView ArticleGoogle Scholar
- Alfieri CM, Cheek J, Chakraborty S, Yutzey KE: Wnt signaling in heart valve development and osteogenic gene induction. Dev Biol. 2010, 338 (2): 127-135.PubMed CentralPubMedView ArticleGoogle Scholar
- Cai X, Zhang W, Hu J, Zhang L, Sultana N, Wu B, Cai W, Zhou B, Cai CL: Tbx20 acts upstream of Wnt signaling to regulate endocardial cushion formation and valve remodeling during mouse cardiogenesis. Development. 2013, 140 (15): 3176-3187.PubMed CentralPubMedView ArticleGoogle Scholar
- Kawakami Y, Capdevila J, Buscher D, Itoh T, Rodriguez Esteban C, Izpisua Belmonte JC: WNT signals control FGF-dependent limb initiation and AER induction in the chick embryo. Cell. 2001, 104 (6): 891-900.PubMedView ArticleGoogle Scholar
- Kawakami Y, Marti M, Kawakami H, Itou J, Quach T, Johnson A, Sahara S, O'Leary DD, Nakagawa Y, Lewandoski M, Pfaff S, Evans SM, Izpisua Belmonte JC: Islet1-mediated activation of the beta-catenin pathway is necessary for hindlimb initiation in mice. Development. 2011, 138 (20): 4465-4473.PubMed CentralPubMedView ArticleGoogle Scholar
- Cooper KL, Hu JK, ten Berge D, Fernandez-Teran M, Ros MA, Tabin CJ: Initiation of proximal-distal patterning in the vertebrate limb by signals and growth. Science. 2011, 332 (6033): 1083-1086.PubMed CentralPubMedView ArticleGoogle Scholar
- Yasuhara R, Yuasa T, Williams JA, Byers SW, Shah S, Pacifici M, Iwamoto M, Enomoto-Iwamoto M: Wnt/beta-catenin and retinoic acid receptor signaling pathways interact to regulate chondrocyte function and matrix turnover. J Biol Chem. 2010, 285 (1): 317-327.PubMed CentralPubMedView ArticleGoogle Scholar
- Bradley EW, Drissi MH: Wnt5b regulates mesenchymal cell aggregation and chondrocyte differentiation through the planar cell polarity pathway. J Cell Physiol. 2011, 226 (6): 1683-1693.PubMedView ArticleGoogle Scholar
- Rahrmann EP, Watson AL, Keng VW, Choi K, Moriarity BS, Beckmann DA, Wolf NK, Sarver A, Collins MH, Moertel CL, Wallace MR, Gel B, Serra E, Ratner N, Largaespada DA: Forward genetic screen for malignant peripheral nerve sheath tumor formation identifies new genes and pathways driving tumorigenesis. Nat Genet. 2013, 45 (7): 756-766.PubMed CentralPubMedView ArticleGoogle Scholar
- Mo W, Chen J, Patel A, Zhang L, Chau V, Li Y, Cho W, Lim K, Xu J, Lazar AJ, Creighton CJ, Bolshakov S, McKay RM, Lev D, Le LQ, Parada LF: CXCR4/CXCL12 mediate autocrine cell- cycle progression in NF1-associated malignant peripheral nerve sheath tumors. Cell. 2013, 152 (5): 1077-1090.PubMed CentralPubMedView ArticleGoogle Scholar
- Ren D, Minami Y, Nishita M: Critical role of Wnt5a-Ror2 signaling in motility and invasiveness of carcinoma cells following Snail-mediated epithelial-mesenchymal transition. Genes Cells. 2011, 16 (3): 304-315.PubMedView ArticleGoogle Scholar
- Lee KW, Kim JH, Han S, Sung CO, Do IG, Ko YH, Um SH, Kim SH: Twist1 is an independent prognostic factor of esophageal squamous cell carcinoma and associated with its epithelial-mesenchymal transition. Ann Surg Oncol. 2012, 19 (1): 326-335.PubMedView ArticleGoogle Scholar
- Shelton EL, Yutzey KE: Tbx20 regulation of endocardial cushion cell proliferation and extracellular matrix gene expression. Dev Biol. 2007, 302 (2): 376-388.PubMed CentralPubMedView ArticleGoogle Scholar
- Vogel KS, Klesse LJ, Velasco-Miguel S, Meyers K, Rushing EJ, Parada LF: Mouse tumor model for neurofibromatosis type 1. Science. 1999, 286 (5447): 2176-2179.PubMed CentralPubMedView ArticleGoogle Scholar
- Nielsen R, Pedersen TA, Hagenbeek D, Moulos P, Siersbaek R, Megens E, Denissov S, Borgesen M, Francoijs KJ, Mandrup S, Stunnenberg HG: Genome-wide profiling of PPARgamma:RXR and RNA polymerase II occupancy reveals temporal activation of distinct metabolic pathways and changes in RXR dimer composition during adipogenesis. Genes Dev. 2008, 22 (21): 2953-2967.PubMed CentralPubMedView ArticleGoogle Scholar
- Langmead B, Trapnell C, Pop M, Salzberg SL: Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 2009, 10 (3): R25-PubMed CentralPubMedView ArticleGoogle Scholar
- Mi H, Muruganujan A, Thomas PD: PANTHER in 2013: modeling the evolution of gene function, and other gene attributes, in the context of phylogenetic trees. Nucleic Acids Res. 2013, 41 (Database issue): D377-D386.PubMed CentralPubMedView ArticleGoogle Scholar
- Zuberi K, Franz M, Rodriguez H, Montojo J, Lopes CT, Bader GD, Morris Q: GeneMANIA Prediction Server 2013 Update. Nucleic Acids Res. 2013, 41 (Web Server issue): W115-W122.PubMed CentralPubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.