Craniofacial morphogenesis involves the cranial neural crest (NC) cells, a vertebrate-specific multipotent cell population that provides most of the head skeletogenic mesenchyme. Cranial NC cells delaminate from different points along the developing neural tube and migrate into distinct facial and pharyngeal arch processes, where they give rise to distinctly shaped cartilage and bone elements, which in turn assemble into a harmonious face. How do distinct cranial NC cell subpopulations acquire their regional identity, allowing them to generate the specific subsets of craniofacial elements appropriate to their position? Premigratory NC cells that contribute to frontonasal, maxillary, or mandibular processes share similar patterning potential, because they can replace each other in building a whole craniofacial skeleton. Such plasticity is maintained during and after migration until subpopulation-specific transcriptional identity and positional patterning programs are established as a result of interactions with their local surrounding environment. We asked how chromatin regulation may allow cranial NC cells to maintain broad patterning competence through migration while being poised to respond to local cues and induce position-specific transcriptional subprograms.

We used genome-wide RNA sequencing (RNA-seq), chromatin immunoprecipitation followed by sequencing (ChIP-seq), and assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-seq) and integrated the information to propose a model to explain how cranial NC subpopulations maintain broad patterning competence through chromatin epigenetic regulation and how transcription factor–dependent responses to local cues can modify the chromatin pattern to establish unique subpopulation-specific transcriptional subprograms. To this aim, we microdissected the Hox-free frontonasal, maxillary, mandibular, and the Hox-expressing second pharyngeal arch processes of E10.5 mouse embryos. We isolated the NC cell subpopulations from each of these prominences by cell sorting and analyzed their transcriptional state, as well as the H3K27me3, H3K4me2, and H3K27Ac histone modification and chromatin accessibility profiles at promoters and enhancers. We then compared these data sets with the transcriptional, histone mark, and chromatin accessibility profiles of the Hox-free NC premigratory progenitors and of E10.5 frontonasal, maxillary, mandibular, and second pharyngeal arch NC cell subpopulations in which we conditionally inactivated the Polycomb H3K27 methyltransferase gene enhancer of zeste homolog 2 (Ezh2cKO mutants).

Early postmigratory NC subpopulations contributing to distinct mouse craniofacial structures displayed similar chromatin accessibility patterns yet differed transcriptionally. The differentially expressed genes (positional genes) displayed accessible and H3K27me3⁺/H3K4me2⁺ bivalent enhancers and promoters, and were embedded in large Ezh2-dependent Polycomb domains, in the NC cell subpopulations in which they were silenced, indicating transcriptional poising. These postmigratory chromatin domains of poised gene regulation were inherited from NC premigratory progenitors. At Polycomb domains, H3K27me3 antagonized H3K4me2 deposition, which was restricted to accessible promoter and enhancer elements, preventing ectopic activation at inappropriate positions.

Our findings explain how cranial NC cell plasticity is maintained through migration until postmigratory stages. We propose that an Ezh2-dependent poised chromatin organization underlies the …