Primer and interviews: Diverse connections between primary cilia and Hedgehog signaling


  • Julie C. Kiefer

    Corresponding author
    1. Department of Neurobiology and Anatomy, University of Utah, Salt Lake City, Utah
    • Department of Neurobiology and Anatomy, 20 North 1900 East, 401 MREB, University of Utah, Salt Lake City, UT 84132
    Search for more papers by this author


On the surface, the Hedgehog (Hh) pathway and primary cilia make strange bedfellows. Hh is a dynamic regulator of a myriad of developmental processes, ranging from spinal cord and limb patterning to lung branching morphogenesis. By contrast, immotile primary cilia were long considered ancestral holdovers with no known function. Considering the disparate perceptions of these two phenomena, the relatively recent discovery that there is a symbiotic-like relationship between Hh and cilia was unexpected. This primer covers the basics of primary cilia and Hh signaling, highlighting variations in ways they are connected across species, and also discusses the evolutionary implications of these findings. Roles of cilia in signal transduction are analyzed further in an interview with Søren T. Christensen, PhD, and Andrew S. Peterson, PhD, in the A Conversation With the Experts section. Developmental Dynamics 239:1255–1262, 2010. © 2010 Wiley-Liss, Inc.


The Basics of Primary Cilia

Primary cilia were long considered vestigial curiosities, a surprising notion considering there are ample clues indicating that they must have some importance. First, cilia and their basic structural components are conserved from green algae to humans (Satir and Christensen,2008). Second, with rare exceptions, there is precisely one of these hair-like projections emanating from nearly every cell type in the vertebrate body. Third, cilia growth and maintenance are under tight control. In dividing cells, cilia are assembled at G1 and retracted upon entry into mitosis (Santos and Reiter,2008). In addition, cilia machinery vigilantly maintains primary cilia structural integrity, enabling a constant presence on non-dividing cells. Evolution usually has its reasons for devoting a cell's precious resources towards conservation of structure and mechanism.

Primary cilia are structurally similar to their motile cilia cousins, which have long been recognized for various roles such as directing extracellular flow along respiratory epithelia and in the female reproductive tract. The projection, or ciliary axoneme, is comprised of nine microtubule (MT) doublets that make the circumference of a cylinder. Motile cilia have an MT pair in the center (“9+2” MT doublet arrangement) that helps propel axoneme movement, while primary cilia lack the center pair (“9+0” arrangement). The axoneme grows from a modified centrosome, the basal body, at the apical cell surface and is enclosed by an extension of the cell membrane. Intraflagellar transport protein (IFT) complexes travel along axonemal MT tracks, transporting cargo required for dynamic growth and maintenance of the axoneme. These ciliary freight trains are propelled by adjoining kinesin-2 motors that direct cargo toward the cilia tip (anterograde) or dynein 2 motors that move toward the base (retrograde). While some functions of motile cilia can be inferred from their beating motions, the relevance of typical, immotile primary cilia are less obvious.

In contrast to typical primary cilia, the roles of atypical cell-type-specific primary cilia, for example in the nose and kidney, have been subjects of study for decades (Berbari et al.,2009). In the nose, each olfactory neuron has multiple, non-motile 9+2 cilia that house the G-protein-coupled receptor signaling cascade required for detecting odorants and transducing oderant signals. Renal primary cilia also have a sensory role: they act like mechanosensors that detect and transmit information about flow within kidney tubules. These findings combined with the fact that typical primary cilia are antenna-like structures that suggestively project into the extracellular environment, make it easy to imagine that the structures play a general role in sensory function.

Discovery of a Link Between Hedgehog and Primary Cilia

In the late 1990s, evidence started to mount that there is a connection between cilia and Hedgehog (Hh) signaling. Mice lacking the ciliogenesis genes, the kinesin-2 motor Kif3A, and those bearing a hypomorphic allele of Ift88, were each reported to have defects in left-right (L-R) patterning. Patterning phenotypes result from an absence of cilia in the embryonic node that ordinarily direct a leftward flow of molecules critical for establishing L-R asymmetry. Phenotypes consistent with aberrant Sonic hedgehog (Shh) signaling were also noted, but believed to be derived from abnormal differentiation of ventral node epithelium (Eggenschwiler and Anderson,2007).

A direct link between cilia and Shh was clinched through an unbiased mutagenesis screen for embryonic patterning mutants (Huangfu et al.,2003). Mice with mutations in each of Ift172 and IFt88 bear signs of disturbed Shh signaling, including an absence of ventral neural tube cell types; the mice also exhibited randomized L-R patterning. Through epistasis experiments, Huangfu et al. (2003) demonstrated that neural patterning phenotypes stem from a defect in Shh signaling downstream of the receptor Patched1 (Ptch1) within neural cells, and thus is independent of node defects. Since then, researchers have also discovered that that PDGFα, and non-canonical and canonical Wnt pathways also intersect with cilia to varying degrees (Schneider et al., 2005; Park et al., 2006; Gerdes et al., 2007) although the latter finding is controversial (Huang and Schier, 2009; Ocbina et al., 2009). The once-overlooked organelle has successfully captured the attention of scientists worldwide.

The Core Hh Pathway

The ways in which the Hh pathway and cilia are intertwined hint at why their relationship is successful. Shh is the most widely expressed and best-studied ligand of the Hh family, which also includes Desert hedgehog (Dhh) and Indian hedgehog (Ihh). Hh ligands signal via the 12-pass transmembrane (TM) receptor Ptc, the 7-pass TM transducer Smoothened (Smo), and downstream effectors, the Gli zinc-finger transcription factors, Gli1, Gli2, and Gli3 (Fig. 1). In the absence of Hh, Ptc resides on the cilia membrane and inhibits ciliary Smo accumulation. In addition, Gli, predominantly Gli3, are proteolytically processed to repressor forms (GliR) that inhibit transcription of target genes. Basal body proteins are implicated in Gli3 processing, and ciliary proteins in Gli2 processing and function (Gerdes et al.,2009). When Hh is present, Ptc binds the ligand and exits from cilia, allowing for Smo accumulation. Transport of Smo to cilia in vitro is dependent on formation of a complex with the kinesin-2 anterograde motor, Kif3a. Activated Smo triggers a signal transduction cascade that culminates in modification of full-length Glis to an active form (GliA) that promotes target gene activation. For more detailed descriptions of the Hh pathway, the reader is referred to two of many excellent reviews (Jiang and Hui,2008; Simpson et al.,2009). By employing cilia and cilia machinery for its own purposes, Hh signal transduction seamlessly integrates with cilia architecture.

Figure 1.

Schematic of the Drosophila Hh pathway in the (A) absence and (B) presence of Hh ligand. Schematic of the mouse Shh pathway in the (C) absence and (D) presence of Shh ligand. See text for details. A, Gli activator; MT, microtubules; R, Gli repressor; Sp, Spop.

Why Cilia?

What advantages do localization of Hh signal transduction machinery to cilia offer? One possible answer stems from the observation that as a morphogen, Shh specifies cell fates in a concentration-dependent manner. In the neural tube, for example, floor plate identity is dependent on high levels of Shh, and interneurons of the intermediate spinal cord on relatively low levels. Like a weathervane that is mounted on a pole to better measure the wind, the Ptc receptor may be perched on cilia is so that it can accurately sense varying amounts of circulating ligand. A second advantage stems from the cilium's role as a subcellular compartment. As such, cilia may serve to organize and concentrate signal transduction components into a defined space, thereby facilitating a rapid, efficient response. These hypotheses may also hold true for relationships between cilia and other molecular pathways, suggesting cilia may generally serve as a hub for signal transduction.


Drosophila: Hh Signaling in the Absence of Cilia

Discovery of the intimate relationship between primary cilia and Hh signaling in mice was particularly surprising considering that intense investigation of the Hh pathway in Drosophila had revealed nothing of the sort. As it turns out, there is a good explanation. Unlike in vertebrates, the only ciliated cell types in Drosophila are sensory neurons and sperm, and unciliated cell types respond to Hh signaling. Therefore, Drosophila must transduce Hh in the absence of cilia.

While there are many similarities between the core Hh pathways in Drosophila and vertebrates, key variations lie between Smo and Gli/Cubitus interruptus (Ci), which could explain how the Drosophila pathway functions without cilia (Table 1, Fig. 1). In the absence of Hh, Ptc localizes to the plasma membrane and prevents Smo from accumulating there (Jiang and Hui,2008). Cells in this state harbor a complex containing the single Gli homolog, Ci, the kinesin-like protein Costal2 (Cos2), the Ser/Thr kinase Fused (Fu), and the novel protein, Supressor of fused (Sufu), which is bound to microtubules (MT) via Cos2. MT localization of the so-called cytoplasmic Hh signaling complex (HSC) is essential for targeting of Ci for phosphorylation and subsequent proteolytic cleavage to generate the repressor form, CiR, which then enters the nucleus and inhibits target gene expression. In the presence of high Hh, Ptc relieves its repression of Smo, permitting accumulation of Smo on the plasma membrane. Cos2 and Fu translocate from MTs to Smo at the plasma membrane, freeing full-length Ci to enter the nucleus and mature to its active form, CiA, which induces expression of target genes. The Drosophila MT-bound HSC that controls transduction between Smo and Ci may comprise a subceullular compartment that functionally substitutes for cilia's role as an organizer of signal transduction components.

Table 1. Links Between Hh and Ciliogenesis
  • a

    sufu is not essential for the Drosophila Hh pathway.

  • b

    fused and kif7 morpholinos affect node cilia; for kif7, ciliogenesis function assayed indirectly.

  • c

    Vertebrate kif7 and kif27 likely arose from duplication of ancestral cos2.

  • d

    Functional role for kif27 in ciliogenesis remains to be tested.

IFTHh signaling++
sufuHh signaling+++/−a+
fusedHh signaling++
 Ciliogenesis (motile)++b+
cos2/kif27/Kif7cHh signalingkif7kif7cos2
 Ciliogenesis (motile)kif27?dkif7bkif7

Functional Diversity: The Key to Adaptability

Although mice harbor orthologs to Drosohpila HSC genes, some have assumed different cellular roles (Table 1, Fig. 1). This functional diversification sheds light on how Hh signal transduction has been adapted to work in ciliated and non-ciliated cells.

Mice have two Cos2 orthologs, Kif7 and Kif27, suggesting there was duplication of an ancestral Cos2 in the vertebrate lineage. Evidence suggests that over time, the kinesins have each become functionally distinct. Mouse Kif27, but not Kif7, binds Fu, which has no apparent role in the Hh pathway. Instead Fu regulates generation of the central pair apparatus in axonemes of motile cilia (Wilson et al., 2009). Whether Kif27 works with Fu to regulate motile ciliogenesis remains to be determined. However, such a finding would imply that when cilia functionally substitutes for an HSC-like complex, proteins that would otherwise reside within the complex, like Fu and Cos2 orthologs, can become dispensable for Hh signaling.

Like Cos2, Kif7 participates in Hh signaling (Cheung et al.,2009; Endoh-Yamagami et al.,2009; Liem et al.,2009). However, unlike Cos2, Kif7 activity is dependent on cilia. Kif7 binds Gli in vitro, and is required for proteolytic processing of Gli3 in the absence of Hh. Kif7 is also required for accumulation of Glis at the cilia tip in response to Hh. These data implicate a model whereby Kif7 is critical for Gli localization, a role not unlike that of Drosophila Cos2, only modified to function in a ciliary framework.

Mouse Sufu also retains a role in Hh signal transduction, but its role within the pathway has shifted compared with Drosophila Sufu. In Drosophila, Sufu was originally identified as a suppressor of fu mutations. Fu mutations inhibit Ci proteolysis, thus favoring CiA and target gene activation, while Sufu retains Ci in the cytoplasm and prevents target gene activation (Jiang and Hui,2008). However, sufu mutants are viable, suggesting that sufu is dispensable for Drosophila Hh signaling. In contrast, mouse Sufu is a potent regulator of the Hh pathway. Although enriched in cilia, mouse Sufu appears to work in a cilia-independent manner to inhibit a Gli-degrading protein, Spop (Chen et al.,2009). In this way, Sufu maintains a pool of full-length Gli proteins from which GliA and GliR can be derived, allowing for a dynamic range of Hh responsiveness. One possibility is that without ties to a Fu/Cos2 complex, Sufu is free to take on another role within the mouse Hh pathway.

Zebrafish: Variations on a Theme

The connection between primary cilia and Hh signaling is likely a vertebrate-wide phenomenon that has also been documented in chick, Xenopus, and zebrafish. Work in zebrafish, the most intensely studied vertebrate after mice, has turned up some intriguing functional variation among genes orthologous to those in the Drosophila HSC. Comparisons between Drosophila, zebrafish, and mice yield additional clues as to how Hh pathway genes function in different cellular contexts.

Zebrafish have Kif7 but lack a Kif27 ortholog, suggesting that duplication of ancestral Cos2 took place after divergence of the fish lineage. In line with this idea, zebrafish Kif7 functions like a hybrid of mouse Kif7 and Kif27: it participates in both ciliogenesis and Hh signaling (Wilson et al.,2009). zFu also participates in both pathways and binds zKif7 in vitro, indicating that the two work together. Importantly, in fu- or kif7-deficient zebrafish, effects on Hh signaling and ciliogenesis are independent of one another, suggesting zFu/zKif7 regulates two separate pathways. Intriguingly, mouse Fu can rescue both Hh and ciliogenesis phenotypes in fu-deficient zebrafish, implying that mouse Fu has the capacity to participate in the Hh pathway, but has somehow been uncoupled from this process.

sufu-deficient zebrafish display a mild increase in Hh signaling, thus the importance of its role in the Hh pathway appears to lie somewhere between its dispensability in Drosophila and its critical role in mouse. Although Sufu is required to varying degrees in the three organisms, a role in Gli processing is conserved. In fact, Drosophila and zebrafish Sufu can restore Gli protein levels in Sufu −/− mouse embryonic fibroblasts (MEFs) in vitro, indicating that some cilium-independent processes, like those mediated by Sufu, may be conserved across lineages (Chen et al.,2009).


Evolution of relationships between cilia and Hh across species can most easily be interpreted in one of two ways. One possibility is that in response to the widespread presence of cilia, vertebrates adapted Hh signal transduction to fit within a ciliary framework, perhaps because it conferred certain advantages. In this case, zebrafish, in which genes like Fu regulate both ciliogenesis and Hh signaling, can be seen as an intermediate state between Drosophila and mice where genes are largely dedicated to one process or the other. Another possibility is that a relationship between cilia and Hh is an ancient one that has been lost to varying degrees in different groups. Analysis of an organism that bears Hh signal transduction and cilia, and evolved independently of Drosophila and vertebrates, could shed light on the origins of Hh/cilia connections. Although ciliogenesis and cilia as sensory organelles have been studied extensively in C. elegans, this popular model organism lacks Hh signaling and obvious orthologs to key genes in question like Fu and Cos2, and so is not a suitable model in which to address this question.

The question of whether there are ancestral connections between Hh and cilia was addressed in the invertebrate planaria, Schmidtea mediterranea. The flatworm bears cilia, a functional Hh pathway, as well as a Fu homolog, and a single Cos2 ortholog, Smed-kif27 (Rink et al.,2009). As in vertebrates, planarian Fu and Kif27 regulate motile ciliogenesis (and as in mice, they do not participate in Hh signal transduction). The authors present these data as evidence that the connection between cilia and Hh components predates the vertebrate lineage. They hypothesize that a HH/cilia connection arose when an ancestral Hh pathway incorporated cilia machinery, perhaps for the purpose of transporting core components to and from subcellular compartments, one of which was cilia. Further, disappearance of a cilia/Hh connection in Drosophila may have followed a loss of cilia, or adoption of a new motor or subcellular compartment that brought components away from cilia. Confirmation of these hypotheses awaits discovery of an ancestral organism with direct molecular links between Hh and primary cilia.


In addition to informing mechanisms of development and evolution, links between Hh and cilia have important implications for human health. Ciliopathies are a class of human syndromes that arise from defects in ciliary and basal body proteins (Badano et al.,2006). Given that motile or primary cilia are present in most cell types and that they perform multiple functions (i.e., mechanosensory, signal transduction, and tissue-specific functions), it is not surprising that ciliopathies are pleiotropic. Many phenotypes overlap between ciliopathies, such as polydactyly, retinal degeneration, anosmia, situs inversus, and polycystic kidney disease, aiding in disease diagnosis and etiology. Detection of signature phenotypes in patients predicts that the gene(s) behind them are cilia associated. For example, as regulators of limb patterning, Shh pathway genes are candidates for causing polydactyly. It will be particularly informative to identify causes of ciliopathy phenotypes for which candidates are less apparent, like obesity and mental retardation.

The intersection of Hh and cilia also suggests novel courses of action for diagnosing and treating certain cancers. Nearly all basal cell carcinomas (BCC), the most commonly diagnosed cancer in North America, and nearly 30% of medulloblastomas, the most common malignant brain tumor in children, show signs of dysregulated Hh signaling. More specifically, a subset of individuals with BCC harbor a mutant activated form of Smo that is constitutively localized to cilia. To investigate the role of cilia in oncogenesis, mouse models of BCC and medublastomas were generated by targeting constitutively active Smo to skin cells and cerebellar granule neuron precursors, respectively (Han et al.,2009; Wong et al.,2009). Strikingly, when Kif3a or Ift88 were genetically ablated in the altered cell types, tumor formation was inhibited. Moreover, it was found that medulloblastomas with high Hh or Wnt activation had primary cilia, whereas the vast majority without those changes did not. These findings have two main clinical implications. First, the presence of primary cilia may prove a useful biomarker for tumor diagnosis. Second, cilia ablation may be a course of treatment for a subset of cancers. The effectiveness of these approaches awaits further testing. Nevertheless, they are an example of how an enhanced understanding of the molecular mechanisms of ciliogenesis, and of connections between Hh signaling and cilia, can inspire novel therapies to treat cancer and disease.


Featured below is a discussion with Hh/cilia experts Søren T. Christensen, PhD, and Andrew S. Peterson, PhD (Fig. 2) about current topics in the field.

Figure 2.

Left: Søren T. Christensen, PhD, Associate Professor, Department of Biology, University of Copenhagen, Denmark. Right: Andrew S. Peterson, PhD, Associate Director, Department of Molecular Biology, Genentech, South San Francisco, CA.

Developmental Dynamics: What is your lab's research focus?

Andrew Peterson: Our entry point into biological areas has always been the phenotype of a mutant mouse. A number of years ago, we carried out several screens using random mutagenesis and began studying mutants with provocative patterns of developmental defects in the forebrain. One of these turned out to be a mutation in the retrograde motor for cilia transport, Dnchc2. As we worked out the mechanism, we realized that several other mutations that we had in our hands probably affected cilia processes as well. One of those was a mutation in Rfx4, a transcription factor whose role in development is at least in part to regulate cilia proteins. This has provided a very rich source of starting points and provided a way to screen for and analyze protein complexes that are required for signal transduction within the cilia.

Søren Christensen: Our lab investigates the mechanisms by which cilia assemble and disassemble, and how the primary cilium coordinates signaling pathways in cell-cycle entry, migration, and differentiation, which are important in, for example, tumorigenesis. A great deal of our work is based on cell and molecular biological studies of cultures of embryonic and cancer stem cells and a variety of different cell lines from human and mouse tissues. One main line of research is focused on how ciliary signaling pathways, such as the PDGFRα, Hedgehog, and Wnt pathways, coordinate the proliferation, migration, and positioning of cells during development and in tissue regeneration and how these pathways regulate the differentiation of stem cells such as during heart development. My partner, Lotte B. Pedersen, and I are fortunate to have a lab workforce of very talented students and technicians that deliver the goods and bring about new ideas, so we never run out of new projects to pursue.

Dev Dyn: What initially provoked your interest in this field?

AP: Kathryn Anderson published a paper in Nature in 2003 describing the requirement for intraflagellar transport for Hh signaling in the mouse. That sentence sums up the information in the paper pretty well and at the time it wasn't possible to integrate the observation into anything we knew about signal transduction. I put the idea aside but certainly didn't forget it.

SC: I think it all began in 1999 when Peter Satir and I started to look at the sensory function of motile cilia in the protozoan ciliate, Tetrahymena. In these cells, we cloned a ciliary protein kinase that linked extracellular signaling to regulation of cell survival and chemotaxis. We then took this a step further to show that PDGFRα signaling is coordinated by the fibroblast primary cilium to regulate cell-cycle entry and migration. Another person who provoked my interest in primary cilia was Denys Wheatley, who actually was the first to introduce me to primary cilia. Then of course, the seminal work by Joel Rosenbaum and colleagues on IFT was a blockbuster for all in this new and exciting field of cell biology.

Dev Dyn: Which papers have most impacted your research?

AP: Clearly Kathryn's 2003 paper (see Huangfu et al.,2003).

A paper from Jonathan Scholey's lab has been very important in our thinking about how transport of signal transduction complexes to the tip of the cilia might work (see Ou et al.,2005).

Beyond those two it is hard to pick out single papers. The work of Joel Rosenbaum to describe the basic processes of intraflagellar transport was the absolutely critical foundation for the field but they were described in a number of papers.

SC: This is a difficult question to answer. Certainly, the work by Joel Rosenbaum and colleagues on IFT and the follow-up on mechanosensing and polycystic kidney disease by Gregory Pazour, Bradley Yoder, Helle Prætorius, and others were an eye-opener to the entire field on primary cilia.

My initial interest in primary cilia, however, actually came from a series of original, and in some cases speculative, papers before any molecular evidence on the function of primary cilia had emerged. For example, Anthony Poole and co-workers suggested that the primary cilium functions as a “cybernetic probe” that senses chemical and physical changes in the extracellular environment (see Poole et al.,1985), and Denys Wheatley later published a provocative paper on the implications of dysfunction or agenesis of primary cilia, which he predicted to cause major disorders in the body (see Wheatley,1995). These and other similar papers in the field had a great influence on my choice of career.

Dev Dyn: Do you think all Shh signaling is mediated through primary cilia in vertebrates?

AP: The evidence is very strong that proper Shh signal transduction occurs in the cilia. Does that mean that nothing happens if cilia are disrupted or are not present? No, I think that some events, for example processing of Gli3 to its repressor form, can still occur. That by itself does not provide very strong evidence that signal transduction takes place outside of the cilia in any meaningful way during normal development. If you dam a stream, the water will find a way around it but its normal course is straight downhill.

SC: I concur with Andy. All available data strongly support the conclusion that the primary cilium plays a key role in coordination of vertebrate Hh signaling. However, we have just begun to sort out the significance of the cilium in regulating the individual steps in, e.g., Gli processing and other molecular mechanisms that eventually lead to altered gene expression, turning the Hh signaling pathway on and off.

Dev Dyn: Why do you think cilia are used for Hh signaling and other modes of signal transduction?

AP: Long ago, the first sensory systems must have been to find or respond to locally higher concentrations of nutrients. Major sensory systems that are most clearly derived from such a beginning, such as vision and smell, use specialized cilia structures. I would guess that Hedgehog signaling is derived from a nutrient sensing system but was co-opted to respond to the internal rather than the external milieu. That pushes the question back to one of why cilia should be used to sense the external milieu. In C. elegans, cilia are present at “ports” in the cuticle armor that protects the organism from the outside. Olfactory reception is organized at those sites and cilia provide an excellent means of limiting exposure of the cell to the outside environment whilst still sampling it.

To turn the question around, I think it is much more difficult (for me at least) to try to understand what evolutionary forces caused Drosophilia to abandon cilia for signal transduction in the internal milieu. It seems like the ancestral state is one that has worked well for us, so why not for them?

SC: Another intriguing aspect is the potential link between cilia and coordination of extracellular signaling events in the context of evolution. Protozoan cilia have receptors and signal transduction components that perceive environmental stimuli essential for, e.g., mating and motility, and in C. elegans, signaling via dendritic endings of sensory neurons is vital for developmental, physiological, and behavioral responses. Further, in many cases there are strong homologies in ciliary signaling components across the eukaryotic taxa. As an example, Maureen Barr and colleagues originally showed that homologues of the mammalian polycystic kidney-disease genes, polycystins-1 and -2, are part of the signaling machinery of sensory neurons in C. elegans. Further, Huang et al. showed that a flagellar polycystin-2 homologue is involved in the mating process of Chlamydomonas. As far as I know, neither organism has kidneys, and it is therefore tempting to speculate that coordination of signaling pathways, such as polycystin signaling, by ciliated structures was a precondition for the evolution of metazoan life as we know it and ultimately for development of the complexity of tissues and organs in vertebrate organisms, such as the kidneys.

Dev Dyn: Currently there is evidence for relatively few signaling pathways being linked to primary cilia. Do you think there are more connections that have yet to be discovered?

SC: Obviously, a large number of signaling pathways still need to be investigated in relation to primary cilia. However, up to now a panoply of diverse signaling systems have already been linked to primary cilia, including ion channels and osmolyte transporters, receptor tyrosine kinases, Wnt and Hh signaling, purinergic receptor signaling, neurotransmission and neuronal regulation, as well as interactions with extracellular matrix proteins. And the list is growing as we speak because an increasing number of people from different areas of life sciences are now entering the field of primary cilia.

AP: I think that there are undiscovered signal transduction events that occur in cilia but they are likely to be more isolated parts of a signaling network, and therefore more difficult to discern. Hh signaling has a set of core events that take place in cilia and they take place there in all Hh-responsive cells. The developing choroid plexus is responsive to the ionic balance of the cerebrospinal fluid (CSF) in a fashion that is dependent on cilia. That implies that there are critical signal transduction events that take place there but the molecular events remain unexplored. The choroid plexus is a case where the consequences of defective signal transduction, hydrocephaly, are easily apparent but I have little doubt that there are other cell type–specific examples that are not so easily recognized.

Dev Dyn: Do you think cell or tissue type–specific cilia modifications like those found in photoreceptors and olfactory sensory neurons may be more common than we realize?

SC: I think that the composition of ciliary signaling systems in some cases may mirror the functionality of the cell type. For example, primary cilia situated deep inside tissues and organs may in some cases comprise signaling pathways different from those that protrude from the apical surface into a lumen, such as on epithelial cells, to carry out diverse functions at various time points during development and in tissue homeostasis. Further, the composition of ciliary signal systems may fluctuate as part of the dynamic process that controls cell differentiation to determine cell fate and function during development. On top of this, there are several examples of modified primary cilia such as photoreceptors and olfactory sensory neurons that carry out highly specialized functions, which also reflect the diversity of ciliary signaling systems.

AP: Photoreceptors are an example of a highly derived version of a primary cilium. Such an extreme specialization is very unlikely to be found elsewhere but given that we can easily distinguish morphological differences using electron microscopy (EM) between cilia present in different tissues of the body makes it seem certain that there are specializations of cilia in different tissues. The specializations of photoreceptors are easily identifiable but if the specialization is at the molecular level, such as cell-specific IFT components, rather than at the utrastructural level, it will take much more work to ferret out. Given all of that, I would guess that specialization is probably more the rule than the exception.

Dev Dyn: Do you think like primary cilia, motile cilia may similarly be linked to certain signaling pathways?

SC: Absolutely, there is no reason to think otherwise. Evidently, classic type motile cilia with 9+2 microtubules are sensory organelles that control behavioral responses in, e.g., protozoan organisms, and I do not see why sensory capacity of motile cilia should have been lost through evolution; on the contrary! There are already a number of papers showing the unique localization of signaling pathways in motile cilia in mammals. Our group previously identified a number of signaling components in motile cilia lining the epithelium of the mammalian oviduct, including TRP ion channels as well as progesterone and angiopoetin receptors. We do not know yet the precise function of these ciliary ion channels and receptors, but we do know that that the oviduct is subjected to spectacular changes in the physiochemical milieu during the estrous cycle. And since the ciliary localization of the receptors is increased upon ovulation, we suspect that they take part in the coordination of a variety of signaling events that control, e.g., transport of the ovulated oocyte after adhesion to the ciliated infundibulum and priming of the ampulla region for reception and finally fertilization of the oocyte. Further, motile cilia of the airway system sense toxins or noxious compounds that regulate ciliary motility and help protect the lungs. There are other reports on receptors in airway cilia, such as FGF receptors, and I am confident that future research will show that motile cilia are as important sensory organelles as primary cilia when it comes to the detection of the external environment and the coordination of signaling events. And not only in the regulation ciliary beat frequency per se, but also in developmental processes and tissue homeostasis in other organs and tissues such as on the ovarian surface epithelium and in ependymal and choroid plexus epithelia.

AP: I agree with Søren completely and he gave some nice examples. Hard evidence will need to come from a sufficient understanding of the signaling and motility roles of a particular type of motile cilia that they can be disentangled. The example of the choroid plexus that I referred to above may be a particularly amenable situation. The role of cilia in regulating ionic balance in the CSF precedes the acquisition of motility during development and Brad Yoder's group has shown that the signaling mechanism involves regulation of cAMP levels. The role of cilia beating in keeping the CSF moving is straightforward so the situation is pretty ripe for distinguishing the two roles.

Dev Dyn: What are some important questions that remain to be answered?

SC: There are still a million questions to answer and, evidently, more questions will come forward as we learn more about the primary cilium in human health and disease. We have only begun to unravel the mysteries at the tip of the iceberg, and my bet is that we will be surprised to see how big this iceberg actually is. My gut feeling also tells me that many more signaling pathways than those previously described are regulated directly or indirectly by the cilium, and it will be important to understand how these signaling pathways are transmitted via the centrosomal region to change the gene expression profile of the cell. In addition, the monitoring and organization of appropriate signaling events seem to depend highly on appropriate trafficking of the signaling systems to and within the cilium. In a way, the various signaling pathways and their partners in ciliary assembly and trafficking can be regarded as the world-wide web. There are so many signaling pathways translated through the cilium and somehow they must be connected, some kind of network to determine when to turn signals on and off. Ultimately, new information about these events will lead us a step further towards the treatment of ciliopathies, such as cancer, developmental defects, and behavioral disorders.

AP: One of the most exciting areas right now is the regulation of specific cargo transport events in cilia. For example work from CC Hui's, from Kathryn Anderson's, and from my own laboratory identified a role for the Kif7 in Hh signaling. Kif7 is localized to the base of the cilia in the absence of Hh and moves to the tip along with Gli proteins in the presence of Hh. Kif7 seems likely to be an anterograde motor so it is an intriguing question as to whether it functions as a cargo adapter that regulates Gli transport into cilia or perhaps is itself a cilia motor. In either case, it does not appear to be required for other cilia transport events and so is an example of a regulator of molecularly specific transport and signal transduction within the cilia.

The question of how the cilia respond to the cell cycle is also an intriguing one. We know that the cilia are disassembled and centrioles are replicated and do their job of providing a framework for the segregation of chromosome during metaphase. What are the signals that trigger disassembly? What are the underlying molecular events involved in disassembly? How are signaling events that take place in the cilia affected during these events? Is this used as a means of allowing or vetoing certain signaling events during the cell cycle?

In a different vein, the question of how signaling complexes are formed in the cilia is a fascinating one and fundamental to its role in signal transduction. Signal transduction components need to enter the cilia in a regulated fashion and the downstream events need to be communicated to the cytoplasm or nucleus. Could this be as simple as regulating entry and exit of specific molecules into the cilia?


Many thanks to Andrew Peterson and Søren Christensen for their time, cooperation, and insightful comments, and to Pao-Tien Chuang for critically evaluating the manuscript. The author regrets that she was unable to directly credit each of the papers that has so far shaped this field. To find them, please see referenced reviews.