Primary Cilia and Mammalian Hedgehog Signaling

Abstract

It has been a decade since it was discovered that primary cilia have an essential role in Hedgehog (Hh) signaling in mammals. This discovery came from screens in the mouse that identified a set of genes that are required for both normal Hh signaling and for the formation of primary cilia. Since then, dozens of mouse mutations have been identified that disrupt cilia in a variety of ways and have complex effects on Hedgehog signaling. Here, we summarize the genetic and developmental studies used to deduce how Hedgehog signal transduction is linked to cilia and the complex effects that perturbation of cilia structure can have on Hh signaling. We conclude by describing the current status of our understanding of the cell-type-specific regulation of ciliogenesis and how that determines the ability of cells to respond to Hedgehog ligands.

Figure 1.

The backbone of the Hedgehog (Hh) signal transduction pathway. The core of the Hh signaling pathway is conserved between Drosophila and vertebrates. In the absence of ligand, the Hh receptor Patched (PTCH1) keeps the pathway off by inhibiting the activity of the seven transmembrane-domain protein Smoothened (SMO). When SMO is inactive, the GLI/Ci (glioblastoma/Cubitus interruptus) transcription factors are proteolytically processed to make a transcriptional repressor that binds to Hh target genes and blocks their transcription. Binding of Hh to PTCH1 inhibits its activity, relieving the repression of SMO, which promotes conversion of full-length GLI/Ci into a transcriptional activator. In vertebrates, cilia are required for the production of both GLI-repressor and GLI-activator.

Figure 2.

Mutations in cilia genes alter Hedgehog (Hh)-dependent neural patterning. Schematics of the spatial distribution of neural cell types in the developing lumbar neural tube in mutants with abnormal cilia. Dorsal up, ventral down; the notochord is the small oval below the neural tube. SHH ligand released from the notochord and floorplate patterns the dorsoventral axis of the neural tube. The highest levels of Shh signaling specify floorplate cells (pink). Dorsal to the floorplate are V3 neural progenitors, which also require high levels of Shh signaling (magenta). Motor neurons (green) require Shh but are specified at lower concentrations of ligand. V2, V1, and V0 interneurons (orange) require even lower levels of Shh activity. Dorsal progenitors (yellow) are specified by default, and high levels of Shh signaling inhibit the specification of dorsal progenitors. Smo mutant embryos fail to specify ventral neural subtypes and all neural progenitors follow a dorsal fate. In Patched1 mutant embryos, all neural progenitors follow a ventral fate and express markers of the floorplate and V3 neural progenitors. Mouse embryos that lack a core intraflagellar transport (IFT-B) protein or kinesin-II lack primary cilia and therefore are unable to respond to Hh signaling; these mutants lack floorplate, V3 interneurons, and most motor neurons, V0-V2 interneurons and dorsal cell types extend ventrally, but V0-V2 interneurons are specified. Dync2h1 mutants, which lack the heavy chain of the dynein retrograde IFT motor and strong loss of IFT-A mutant embryos, have very short, bulged primary cilia and cannot transduce Hh signals efficiently; these mutants lack floor plate V3 interneurons and some motor neurons. In combination with Shh, Smo, or Patched1, double mutant embryos resemble single dynein or strong IFT-A mutants indicating that like IFT-B these components of the cilium assembly machinery are required downstream from SMO and PTCH1. Surprisingly, IFT-A mutant embryos that have a milder effect on cilia morphology (e.g., with cilia of near-normal length and bulged tips) have the opposite effect on neural patterning, with expanded ventral neural cell types. Cilia in Arl13b^hnn/hnn mutant embryos have structural defects in the microtubule axoneme, and the mutants fail to specify both the most ventral and most dorsal neural cell types. Double mutants that lack both a cilia component and Smo or Patched1 have phenotypes similar to the cilia single mutants, indicating that the cilia machinery is required downstream from Ptch1 and Smo.

Figure 3.

Hedgehog (Hh) signal transduction in the primary cilium. The primary cilium has several spatially distinct regions that promote normal Hh signal transduction. Entry into the cilium is gated at the transition zone; the EVC zone facilitates pathway activation in some cell types; the cilia tip compartment, defined by KIF7, is the site of enrichment and activation of the GLI/SUFU complex; and proteolytic processing to make GLI repressors that may occur at the base where protein kinase A (PKA) is localized and leads to formation of Gli3R. In the absence of Hh signal, PTCH1 and Gpr161, both negative regulators of the pathway, are present in the cilia membrane. Gpr161 trafficking into the cilium depends on TULP3 and intraflagellar transport (IFT)-A; GPR161 appears to activate Gαs, which activates adenylyl cyclase, which increases the levels of cAMP, thereby activating PKA. Activated PKA phosphorylates sites on GLI3 that promote partial proteolysis by βTrCP/Cul1 and the proteasome, generating Gli3 repressor (Gli3R), which moves to the nucleus and represses expression of Hh target genes. GLI2 and GLI3 are trafficked to the tip of the cilium in the absence of ligand in a complex with SuFu, and processing of GLI3 depends on it having transited the cilium. Binding of Hh to PTCH1 triggers its removal from the cilium, allowing Smoothened (SMO) to translocate into the cilium where it can activate downstream signaling. Gpr161 also exits the cilium after exposure to ligand. Binding of EvC to SMO near the base of the cilium promotes SMO activity. In the presence of ligand, the GLI/SUFU complex accumulates to high levels at the tip of the cilium, where dissociation of the complex allows formation of the activator form of GLI2.

Figure 4.

Lineage-dependent cilia formation in the early mouse embryo. (A) Three lineages defined in the preimplantation embryo persist in the postimplantation embryo (E8.0 embryo shown). Two extraembryonic lineages, the trophectoderm (blue) and visceral endoderm (red) contribute to the placenta and yolk sac respectively; these two lineages lack primary cilia. Nearly all nonmitotic cells of the third lineage, the epiblast (black), are ciliated; this lineage gives rise to almost all cells of the embryo proper, as well as to mesoderm-derived components of the placenta and yolk sac. (B) Paracrine Hh signaling in the yolk sac. Upper drawing represents the two layers of the E8.0 yolk sac (lassoed in panel A). The unciliated visceral endoderm layer produces IHH, but does not activate Hh target genes. The adjacent ciliated extraembryonic mesoderm responds to the IHH produced by the visceral endoderm and produces BMP4, which leads to formation of the blood islands (lower panel). (Image based on data in Baron 2001.) (C) A blood vessel in the E14.5 yolk sac from an embryo expressing a Centrin2-GFP, a marker for the basal body, and ARL13b-mCherry, a marker of the cilia membrane. The lineage relationships are preserved from earlier development—the visceral endoderm (cell layer at the top of the image) has GFP⁺ centrioles, but lacks cilia (arrowhead), whereas the extraembryonic mesoderm that surrounds the vessel filled with round blood cells is ciliated (arrow). (Image from Bangs et al. 2015; modified, with permission, from the authors.)