Summary Directed differentiation of human embryonic stem cells (ESCs) into cardiovascular cells provides a model for studying molecular mechanisms of human cardiovascular development. Though it is known that chromatin modification patterns in ESCs differ markedly from those in lineage-committed progenitors and differentiated cells, the temporal dynamics of chromatin alterations during differentiation along a defined lineage have not been studied. We show that differentiation of human ESCs into cardiovascular cells is accompanied by programmed temporal alterations in chromatin structure that distinguish key regulators of cardiovascular development from other genes. We used this temporal chromatin signature to identify regulators of cardiac development, including the homeobox gene MEIS2. We demonstrate using the zebrafish model that MEIS2 is critical for proper heart tube formation and subsequent cardiac looping. Temporal chromatin signatures should be broadly applicable to other models of stem cell differentiation to identify regulators and provide key insights into major developmental decisions.

Fibroblast growth factor (Fgf) proteins are important regulators of pharyngeal arch development. Analyses of Fgf8 function in chick and mouse and Fgf3 function in zebrafish have demonstrated a role for Fgfs in the differentiation and survival of postmigratory neural crest cells (NCC) that give rise to the pharyngeal skeleton. Here we describe, in zebrafish, an earlier, essential function for Fgf8 and Fgf3 in regulating the segmentation of the pharyngeal endoderm into pouches. Using time-lapse microscopy, we show that pharyngeal pouches form by the directed lateral migration of discrete clusters of endodermal cells. In animals doubly reduced for Fgf8 and Fgf3, the migration of pharyngeal endodermal cells is disorganized and pouches fail to form. Transplantation and pharmacological experiments show that Fgf8 and Fgf3 are required in the neural keel and cranial mesoderm during early somite stages to promote first pouch formation. In addition, we show that animals doubly reduced for Fgf8 and Fgf3 have severe reductions in hyoid cartilages and the more posterior branchial cartilages. By examining early pouch and later cartilage phenotypes in individual animals hypomorphic for Fgf function,we find that alterations in pouch structure correlate with later cartilage defects. We present a model in which Fgf signaling in the mesoderm and segmented hindbrain organizes the segmentation of the pharyngeal endoderm into pouches. Moreover, we argue that the Fgf-dependent morphogenesis of the pharyngeal endoderm into pouches is critical for the later patterning of pharyngeal cartilages.

Skeletal muscle fibers are classified into fiber types, in particular slow twitch versus fast twitch. Muscle fiber types are generally defined by the particular myosin heavy chain isoforms that they express, but many other components contribute to a fiber’s physiological characteristics. Skeletal muscle fiber type can have a profound impact on muscle diseases, including certain muscular dystrophies and sarcopenia, the aging-induced loss of muscle mass and strength. These findings suggest that some muscle diseases may be treated by shifting fiber type characteristics either from slow to fast, or fast to slow phenotypes, depending on the disease. Recent studies have begun to address which components of muscle fiber types mediate their susceptibility or resistance to muscle disease. However, for many diseases it remains largely unclear why certain fiber types are affected. A substantial body of work has revealed molecular pathways that regulate muscle fiber type plasticity and early developmental muscle fiber identity. For instance, recent studies have revealed many factors that regulate muscle fiber type through modulating the activity of the muscle regulatory transcription factor MYOD1. Future studies of muscle fiber type development in animal models will continue to enhance our understanding of factors and pathways that may provide therapeutic targets to treat muscle diseases.

Deletion of a subset of the D4Z4 macrosatellite repeats in the subtelomeric region of chromosome 4q causes facioscapulohumeral muscular dystrophy (FSHD) when occurring on a specific haplotype of 4qter (4qA161). Several genes have been examined as candidates for causing FSHD, including the DUX4 homeobox gene in the D4Z4 repeat, but none have been definitively shown to cause the disease, nor has the full extent of transcripts from the D4Z4 region been carefully characterized. Using strand-specific RT-PCR, we have identified several sense and antisense transcripts originating from the 4q D4Z4 units in wild-type and FSHD muscle cells. Consistent with prior reports, we find that the DUX4 transcript from the last (most telomeric) D4Z4 unit is polyadenylated and has two introns in its 3-prime untranslated region. In addition, we show that this transcript generates (i) small si/miRNA-sized fragments, (ii) uncapped, polyadenylated 3-prime fragments that encode the conserved C-terminal portion of DUX4 and (iii) capped and polyadenylated mRNAs that contain the double-homeobox domain of DUX4 but splice-out the C-terminal portion. Transfection studies demonstrate that translation initiation at an internal methionine can produce the C-terminal polypeptide and developmental studies show that this peptide inhibits myogenesis at a step between MyoD transcription and the activation of MyoD target genes. Together, we have identified new sense and anti-sense RNA transcripts, novel mRNAs and mi/siRNA-sized RNA fragments generated from the D4Z4 units that are new candidates for the pathophysiology of FSHD.

The vertebrate inner ear develops from the otic placode, an ectodermal thickening that forms adjacent to the presumptive hindbrain. Previous studies have suggested that competent ectodermal cells respond to signals from adjacent tissues to form the placode. Members of the Fgf family of growth factors and the Dlx family of transcription factors have been implicated in this signal-response pathway. We show that compromising Fgf3 and Fgf8 signaling blocks ear development; only a few scattered otic cells form. Removal of dlx3b, dlx4b and sox9a genes together also blocks ear development, although a few residual cells form an otic epithelium. These cells fail to form if sox9b function is also blocked. Combined loss of Fgf signaling and the three transcription factor genes, dlx3b,dlx4b and sox9a, also completely eliminates all indications of otic cells. Expression of sox9a but not dlx3b, dlx4b or sox9b requires Fgf3 and Fgf8. Our results provide evidence for Fgf3-and Fgf8-dependent and -independent genetic pathways for otic specification and support the notion that Fgf3 and Fgf8 function to induce both the otic placode and the epithelial organization of the otic vesicle.

The basic helix-loop-helix (bHLH) transcription factor Myod directly regulates gene expression throughout the program of skeletal muscle differentiation. It is not known how a Myod-driven myogenic program is modulated to achieve muscle fiber-type-specific gene expression. Pbx homeodomain proteins mark promoters of a subset of Myod target genes, including myogenin (Myog); thus, Pbx proteins might modulate the program of myogenesis driven by Myod. By inhibiting Pbx function in zebrafish embryos, we show that Pbx proteins are required in order for Myod to induce the expression of a subset of muscle genes in the somites. In the absence of Pbx function, expression of myog and of fast-muscle genes is inhibited, whereas slow-muscle gene expression appears normal. By knocking down Pbx or Myod function in combination with another bHLH myogenic factor, Myf5, we show that Pbx is required for Myod to regulate fast-muscle, but not slow-muscle, development. Furthermore, we show that Sonic hedgehog requires Myod in order to induce both fast-and slow-muscle markers but requires Pbx only to induce fast-muscle markers. Our results reveal that Pbx proteins modulate Myod activity to drive fast-muscle gene expression, thus showing that homeodomain proteins can direct bHLH proteins to establish a specific cell-type identity.

A prominent region of the vertebrate hindbrain is subdivided along the anterior-posterior axis into a series of seven segments, or rhombomeres. The identity of each rhombomere is specified by the expression of conserved transcription factors, including Krox-20, vHnf1, Val (Kreisler, Mafb) and several Hox proteins. Previous work has shown that retinoic acid (RA) signaling plays a critical role in regulating the expression of these factors and that more posterior rhombomeres require higher levels of RA than more anterior rhombomeres. Models to account for RA concentration dependency have proposed either a static RA gradient or increasing time periods of RA exposure. Here, we provide evidence against both of these models. We show that early zebrafish rhombomere-specification genes, including vhnf1 in r5-r6 and hoxd4a in r7, initiate expression sequentially in the hindbrain, each adjacent to the source of RA synthesis in paraxial mesoderm. By knocking down RA signaling, we show that progressively more posterior rhombomeres require increasingly higher levels of RA signaling, and vhnf1 and hoxd4a expression are particularly RA-dependent. RA synthesis is required just at the time of initiation, but not for maintenance, of vhnf1 and hoxd4a expression. Furthermore, a premature RA increase causes premature activation of vhnf1 and hoxd4a expression. Our results support a new model of dynamic RA action in the hindbrain, in which a temporally increasing source of RA is required to sequentially initiate progressively more posterior rhombomere identities.

Facioscapulohumeral dystrophy (FSHD) is a progressive muscular dystrophy caused by decreased epigenetic repression of the D4Z4 macrosatellite repeats and ectopic expression of DUX4, a retrogene encoding a germline transcription factor encoded in each repeat. Unaffected individuals generally have more than 10 repeats arrayed in the subtelomeric region of chromosome 4, whereas the most common form of FSHD (FSHD1) is caused by a contraction of the array to fewer than 10 repeats, associated with decreased epigenetic repression and variegated expression of DUX4 in skeletal muscle. We have generated transgenic mice carrying D4Z4 arrays from an FSHD1 allele and from a control allele. These mice recapitulate important epigenetic and DUX4 expression attributes seen in patients and controls, respectively, including high DUX4 expression levels in the germline, (incomplete) epigenetic repression in somatic tissue, and FSHD–specific variegated DUX4 expression in sporadic muscle nuclei associated with D4Z4 chromatin relaxation. In addition we show that DUX4 is able to activate similar functional gene groups in mouse muscle cells as it does in human muscle cells. These transgenic mice therefore represent a valuable animal model for FSHD and will be a useful resource to study the molecular mechanisms underlying FSHD and to test new therapeutic intervention strategies.