The primary cilium at a glance

To build a primary cilium, the centrosome migrates towards the cell surface as a cell enters G₀. The mother centriole attaches to a Golgi-derived vesicle that then expands, with the axoneme of the primary cilium growing above the centriole within the ciliary membrane, which projects into the vesicle lumen (Sorokin, 1962). The axonemal microtubules polymerize at the growing tip of the projection, to which cargo is delivered by intraflagellar transport (IFT; see below). The vesicle is eventually exocytosed to expose the primary cilium at the cell surface, where ciliary growth continues for up to several micrometers to form the mature cilium. The docking of the mother centriole to the Golgi-derived membrane is thought to be mediated in part by the distal appendages of the centriole, also known as transition fibers, and the original attachment point to the membrane probably corresponds to the region known as the ciliary necklace (Gilula and Satir, 1972), the proposed barrier between the ciliary membrane and the general cell membrane. The transition fibers and necklace are thought to be part of a ‘ciliary pore complex’ through which only selected proteins are allowed passage into the ciliary compartment (Rosenbaum and Witman, 2002).

IFT is an evolutionarily conserved motility process that is required for growth and maintenance of both motile and primary cilia (Rosenbaum and Witman, 2002). IFT relies on the association of ciliary building blocks (e.g. tubulin, radial-spoke proteins and peripheral membrane proteins such as guanine-nucleotide exchange factors) with a scaffold of IFT-particle protein complexes, the components of which are orthologous from Chlamydomonas reinhardtii to humans. The IFT particles and their associated cargo proteins are transported along axonemal microtubules by kinesin 2 motor proteins in the anterograde (base-to-tip) direction, after which cargo is delivered to the growing tip, and by cytoplasmic dynein 2 in the retrograde (tip-to-base) direction. The process is schematized in the accompanying poster. Other reviews on IFT have been published recently (Cole and Snell, 2009; Pedersen and Rosenbaum, 2008).

The significance of primary cilia in signaling became clear on examination of a hypomorphic mouse mutant (Tg737^{orpk RpW}) (Lehman et al., 2008) that was deficient in the homolog of IFT88, a Chlamydomonas IFT protein (Pazour et al., 2000). The mouse mutant had been developed as a model for human autosomal recessive polycystic kidney disease (ARPKD), because the mice developed polycystic kidneys and many other pleiomorphic phenotypes before dying a few days after birth. Just as in a Chlamydomonas Ift88 mutant, primary cilia did not grow normally in kidney cells (and elsewhere) in the mutant mouse. It was proposed that polycystic kidneys developed because primary-cilium signaling was defective. This was confirmed when the polycystins 1 and 2 (PC1 and PC2, respectively) — which form a TRP Ca²⁺ channel protein complex and are defective in models for autosomal dominant polycystic kidney disease (ADPKD) — were found to be normally localized to the primary cilium (Pazour et al., 2002; Yoder et al., 2002), but were mislocalized or absent when ADPKD developed. Meanwhile, Praetorius and Spring demonstrated that the kidney primary cilium acts as a flow sensor; when the cilium is bent, Ca²⁺ enters the cell as a messenger molecule (Praetorius and Spring, 2001). This mechanoregulation of intracellular Ca²⁺ is impaired in the mutant mouse kidney (Liu et al., 2005). Ciliary mechanoregulation appears not to be limited to kidney epithelial cells, but might operate in a variety of cell types and tissues, either in response to fluid flow, such as on the embryonic node (McGrath et al., 2003), endothelial cells (Iomini et al., 2004; Nauli et al., 2008; AbouAlaiwi et al., 2009; Hierck et al., 2008) and cholangiocytes (Masyuk et al., 2006), or through direct interaction between the primary cilium and the extracellular matrix (ECM), such as in chondrocytes (McGlashan et al., 2006) and smooth muscle cells (Lu et al., 2008).

Analysis of mutants such as the Tg737^{orpk RpW} mouse has indicated that signaling through the primary cilium is essential for normal development and function, not only of the kidney, but also of many other tissues and organs. Consequently, ciliary dysfunction might lead to an array of developmental abnormalities and diseases (ciliopathies), including randomization of the left-right body axis, abnormalities in neural-tube closure and patterning, skeletal defects, cystic diseases, blindness, behavioral and cognitive defects, and obesity (Lehman et al., 2008; Quinlan et al., 2008; Veland et al., 2009). An overview of ciliopathies and syndromes caused by defects in assembly or function of primary cilia is presented in the poster.

At first glance, the primary cilium appears to be radially symmetrical, with each microtubule doublet equivalent to every other doublet; however, this is probably not the case because, in motile cilia, every doublet can be specifically numbered with respect to the effective stroke, whose direction (arbitrarily defined as cellular left) is identified by a basal foot. Similarly to motile cilia, primary cilia often have a basal foot (Wheatley, 1982) but, with or without this structure, basal bodies and primary cilia seemingly ‘know’ left from right (Bell et al., 2008). In conjunction with other factors, orientation of the basal body, and therefore orientation of cilia, translates into left-right cell orientation and nodal flow and then to body-axis determination at the embryonic node (Hirokawa et al., 2006). In a flow gradient (as in the kidney tubule) or in a chemotactic gradient (as in fibroblast migration during wound healing) (Christensen et al., 2008), all primary cilia align in a single direction, and cell orientation seems to be determined by this. In the kidney, the direction in which the cilia bend with flow depends on their orientation. The flow, and therefore the bend, is always in the anterior-posterior direction so the gradient of Ca²⁺ concentration or other signaling molecules along the cellular axes should be identically displayed in the anterior-posterior direction in every cell. Because the gradient is uniform from cell to cell, we can probably infer that this determines a consistent anterior-posterior orientation of the mitotic spindle for cell division. In the migrating fibroblasts at a wound edge, orientation of primary cilia ensures that the direction of cell migration is uniform. However, when the primary cilium is missing or defective, orientation is random, which translates into randomization of left-right asymmetry in body form, randomization of migration direction (Schneider et al., 2010) or randomization of mitotic-spindle orientation (Fischer et al., 2006).

The signaling pathways coordinated by primary cilia are quite diverse, and depend on the cell type. As indicated in the accompanying poster, the pathways include signaling through Ca²⁺, receptor tyrosine kinases (RTKs), Hedgehog (Hh), Wingless (Wnt), neuronal and purinergic receptors, as well as through communication with the ECM (Christensen et al., 2007; Christensen et al., 2008; Eggenschwiler and Anderson, 2007; Kiprilov et al., 2008; Gerdes et al., 2009; Knight et al., 2009; Masyuk et al., 2008; Praetorius and Leipziger, 2009; Wong and Reiter, 2008; Jensen et al., 2004). A single primary cilium can be set up for several different kinds of signaling and can respond, for example, to mechanical strain as well as to several morphogens, hormones or growth factors. Different receptors or channels can be present in the same cilium at the same time or at different times. The diversity and dynamics of important ciliary membrane proteins and their effector molecules in the cilium-centrosome axis has only begun to be catalogued, and the list of which pathways are important for specific tissues, especially during development, is likely to be expanded in the future.

Signaling via the primary cilium is of paramount importance during development, and probably remains so in stem-cell populations in various tissues. In the adult, primary cilia might still function in fibroblast cell-cycle control and/or cell migration during tissue regeneration and wound healing (Schneider et al., 2005; Schneider et al., 2009; Schneider et al., 2010). Most other differentiated, non-dividing cells of the adult body, including neurons and kidney cells, possess primary cilia. The primary cilia in these tissues might be necessary for maintenance of the differentiated state and suppression of cyst formation or oncogenesis. Davenport and colleagues used an inducible system to disrupt IFT and remove primary cilia from all adult mouse tissues or from specific tissues (Davenport et al., 2007). Even when all tissues were affected, the devastating abnormalities and lethality seen after embryonic loss of IFT did not occur, although PKD eventually developed after about a year, presumably because cell division was greatly reduced and cystogenesis in the adult is slow.

In many tissues, aberrant activation or absence of ciliary signaling is correlated with uncontrolled cell division and cancer (Christensen et al., 2008; Kuehn et al., 2007; Mans et al., 2008; Michaud and Yoder, 2006; Nielsen et al., 2008; Plotnikova et al., 2008; Wong et al., 2009; Han et al., 2009). However, an immediate effect of ciliary removal, either of all adult primary cilia or specifically of cilia on neurons of the hypothalamus in adult mice, is hyperphagia (compulsive eating), which leads to obesity. Secondarily, obesity leads to defects that resemble type II diabetes. These effects do not occur if the feeding of the knockout mice is restricted (Davenport et al., 2007). Specific hormone receptors associated with feeding behavior localize to cilia of the hypothalamus, including somatostatin sst 3 receptor (Sst3R) (Handel et al., 1999) and melanin-concentrating hormone receptor 1 (Mchr1) (Berbari et al., 2008) in neuronal primary cilia, as well as leptin receptor (LepR) in olfactory cilia (Baly et al., 2007). The Sst3R and Mchr1 receptors are mislocalized in neurons of mice that have mutations in proteins that correspond to those observed in individuals with the syndromic obesity condition Bardet-Biedl syndrome (Berbari et al., 2008; Seo et al., 2009). Furthermore, type 3 adenylyl cyclase, which is also associated with obesity in mice and humans, is also specifically localized to hypothalamic neurons (Wang et al., 2009). A ‘yin-yang’ relationship, in which activation between the ciliary receptors in hypothalamic neurons alternates to stop and start feeding behavior, might be involved in the satiety response (Satir, 2007).

Obesity might also be linked to primary cilia in adipose tissue. Adipogenic differentiation and fat accumulation is associated with transient formation of the primary cilium, containing Wnt and Hh signaling components, such that adipocytes in culture derived from dermal fibroblasts of individuals with Bardet-Biedl syndrome exhibit a higher predisposition for fat accumulation and a higher secreted leptin level than control cells (Marion et al., 2009). Zhu and colleagues demonstrated that the primary cilium and its basal body form an organized signaling pathway for the IGF-1 receptor to induce adipocyte differentiation in confluent 3T3-L1 preadipocytes (Zhu et al., 2009). In addition, childhood obesity and type II diabetes in Alström syndrome patients is caused by mutations in Alström syndrome 1 protein (ALMS1), which localizes to the base of primary cilia (Hearn et al., 2005) and is regulated during adipogenesis (Romano et al., 2008). Evidently, multiple ciliary signaling pathways, involving Wnt, Hh and RTK signaling take part in the regulation of adipogenic differentiation.

Because of its near-ubiquity on cells of the human body, and because of the multiple signaling pathways that require it both during development and in the adult, the primary cilium has moved from being nearly forgotten to a position of considerable importance in biomedicine. This is amply demonstrated for primary cilia in the kidney and there are important data that indicate how signaling pathways involving the cilium might affect other tissues.

One persistent unanswered question about ciliary function is why certain receptors and channels are concentrated more or less exclusively in the membrane of the primary cilium. Clearly, signaling molecules or second messengers that leave the cilium are initially spatially localized at the basal body or centrosome, which would not be true of signals arising from receptors or channels at the leading edge of the cell or dispersed in the cell membrane. Signals from the cilium might therefore interact with, activate or inactivate specific centrosomal proteins to control trafficking to the Golgi, to the leading edge of a migrating cell, to cell junctions or, in the case of transcription factors, to the nucleus. Ciliary orientation might impose a gradient of second messengers or effector molecules within the cytoplasm to help determine positioning of organelles and the mitotic spindle.

In certain cases, the amplitude of the signal or the concentration of signaling molecules arising from the cilium might be compared at the centrosome to signals arising from elsewhere in the cell to determine a specific physiological outcome, such as entry into the cell cycle and resorption of the cilium. At present, there are only hints of how this computation might be performed. As we learn more about IFT-complex assembly and IFT cargo, the role of activation of vesicular trafficking and exocytosis in building the cilium, and targeting processes in the cell in general, we might come to understand reasons for sequestration within the primary cilium more completely. In turn, we might be able to understand why different receptors are sequestered in different cilia, why there are ‘yin-yang’ pairs of ciliary receptors and why sequestration of ciliary receptors and effectors is so dynamic.

We apologize to those authors whose work has not been cited because of space limitations. P.S. is partially funded by the NIH (NIDDK). S.T.C. and L.B.P. are partially funded by grants from the Lundbeck Foundation (R9-A969), the Novo Nordisk Foundation, and the Danish Natural Science Research Council (272-070530). We thank Bradley K. Yoder for valuable comments on the manuscript and the poster. Deposited in PMC for release after 12 months.