The blood-brain barrier (BBB) is composed of tightly bound endothelial cells (ECs) and perivascular astrocytes that regulate central nervous system (CNS) homeostasis. We showed that astrocytes secrete Sonic hedgehog and that BBB ECs express Hedgehog (Hh) receptors, which together promote BBB formation and integrity during embryonic development and adulthood. Using pharmacological inhibition and genetic inactivation of the Hh signaling pathway in ECs, we also demonstrated a critical role of the Hh pathway in promoting the immune quiescence of BBB ECs by decreasing the expression of proinflammatory mediators and the adhesion and migration of leukocytes, in vivo and in vitro. Overall, the Hh pathway provides a barrier-promoting effect and an endogenous anti-inflammatory balance to CNS-directed immune attacks, as occurs in multiple sclerosis.

In the spinal cord, sonic hedgehog (Shh) is secreted by the floor plate to control the generation of distinct classes of ventral neurons along the dorsoventral axis. Genetic and in vitro studies have shown that Shh also later acts as a midline-derived chemoattractant for commissural axons. However, the receptor(s) responsible for Shh attraction remain unknown. Here we show that two Robo-related proteins, Boc and Cdon, bind specifically to Shh and are therefore candidate receptors for the action of Shh as an axon guidance ligand. Boc is expressed by commissural neurons, and targeted disruption of Boc in mouse results in the misguidance of commissural axons towards the floor plate. RNA-interference-mediated knockdown of Boc impairs the ability of rat commissural axons to turn towards an ectopic source of Shh in vitro. Taken together, these data suggest that Boc is essential as a receptor for Shh in commissural axon guidance.

Serum-to-2i interconversion of mouse embryonic stem cells (mESCs) is a valuable in vitro model for early embryonic development. To assess whether 3D chromatin organization changes during this transition, we established Capture Hi-C with target-sequence enrichment of DNase I hypersensitive sites. We detected extremely long-range intra- and inter-chromosomal interactions between a small subset of H3K27me3 marked bivalent promoters involving the Hox clusters in serum-grown cells. Notably, these promoter-mediated interactions are not present in 2i ground-state pluripotent mESCs but appear upon their further development into primed-like serum mESCs. Reverting serum mESCs to ground-state 2i mESCs removes these promoter-promoter interactions in a spatiotemporal manner. H3K27me3, which is largely absent at bivalent promoters in ground-state 2i mESCs, is necessary, but not sufficient, to establish these interactions, as confirmed by Capture Hi-C on Eed(-/-) serum mESCs. Our results implicate H3K27me3 and PRC2 as critical players in chromatin alteration during priming of ESCs for differentiation.

The mammalian HoxD cluster lies between two topologically associating domains (TADs) matching distinct enhancer-rich regulatory landscapes. During limb development, the telomeric TAD controls the early transcription of Hoxd genes in forearm cells, whereas the centromeric TAD subsequently regulates more posterior Hoxd genes in digit cells. Therefore, the TAD boundary prevents the terminal Hoxd13 gene from responding to forearm enhancers, thereby allowing proper limb patterning. To assess the nature and function of this CTCF-rich DNA region in embryos, we compared chromatin interaction profiles between proximal and distal limb bud cells isolated from mutant stocks where various parts of this boundary region were removed. The resulting progressive release in boundary effect triggered inter-TAD contacts, favored by the activity of the newly accessed enhancers. However, the boundary was highly resilient, and only a 400-kb deletion, including the whole-gene cluster, was eventually able to merge the neighboring TADs into a single structure. In this unified TAD, both proximal and distal limb enhancers nevertheless continued to work independently over a targeted transgenic reporter construct. We propose that the whole HoxD cluster is a dynamic TAD border and that the exact boundary position varies depending on both the transcriptional status and the developmental context.