Meckel-Gruber syndrome and the role of primary cilia in kidney, skeleton, and central nervous system development

Abstract

The ciliopathies are a group of related inherited diseases characterized by malformations in organ development. The diseases affect multiple organ systems, with kidney, skeleton, and brain malformations frequently observed. Research over the last decade has revealed that these diseases are due to defects in primary cilia, essential sensory organelles found on most cells in the human body. Here we discuss the genetic and cell biological basis of one of the most severe ciliopathies, Meckel-Gruber syndrome, and explain how primary cilia contribute to the development of the affected organ systems.

Keywords: :

• cilia
• ciliopathy
• signalling
• Hedgehog
• Wnt
• skeleton
• kidney
• brain
• Meckel-Gruber syndrome
• Joubert syndrome
• nephronophthisis

Introduction

The organization of cells into tissues and organs is fundamental to animal biology. To form such a complex series of tissues, cells from the appropriate lineage must grow, divide, migrate, differentiate, and adhere in a sequence of highly ordered events during embryogenesis. Many of the signals that control these events are relatively well-understood; however research over the last decade has revealed an unexpected link between transduction of key developmental signaling pathways and a once almost forgotten organelle: the primary cilium. In this review we discuss the current understanding of how primary cilia mediate development of the kidney, skeleton, and central nervous system, and some of the consequences when cilium functions in these tissues go awry.

Cilia are highly evolutionarily conserved organelles which project from the surface of virtually every cell type in the vertebrate body¹ (the major exceptions being leukocytes). The cilium consists of the microtubule-based core structure known as the axoneme, which is surrounded by the ciliary membrane.² Cilia have classically been categorised as either primary or motile. Motile cilia are present in large numbers in the tracheal epithelium, ependymal cells of the brain, and the oviduct, where multi-ciliated cells use a coordinated beating motion to generate fluid movement. In contrast, primary cilia are sensory organelles that act as “antennae” to sample the extracellular environment, and are involved with transducing extracellular signals to the interior of the cell.³ All cilia are extended from a basal body. In the case of primary cilia, this is derived from the older of the two centrioles that together form the centrosome and have important roles in cell division. The centrioles move to the cell membrane—apical or basal depending on the cell type—and dock into the cortical cytoskeleton,⁴ after which they are termed basal bodies, and for both motile and non-motile cilia, a specialized transport system termed intraflagellar transport (IFT) is then used to build the cilium.⁵

The primary cilium plays an important role both during development and in adult life. The distinction between motile and sensory cilia is blurred, however, as motile cilia can be sensory,⁶ even though the precise signaling pathways that are transduced through motile and primary cilia may be distinct.⁷

The Ciliopathies

In accordance with the wide range of functions cilia perform, a number of inherited diseases, known as ciliopathies, have been increasingly linked with defects in genes that affect cilia assembly or function over recent years (reviewed in refs. 8 and 9). The protein products of these genes all localize to the cilium or basal body.¹⁰ Ciliopathies share many clinical features, with renal, retinal, and hepatic involvement frequently observed alongside skeletal malformations and central nervous system developmental defects. The most severe are lethal in the early gestation or neonatal periods.¹¹ While many genes are associated with a single ciliopathy, mutation in others can give rise to a number of clinically different outcomes. For example, there is substantial overlap in the underlying genetic basis of the kidney disorder nephronophthisis (NPHP), the neurodevelopmental disorder Joubert syndrome (JBTS) and the lethal malformation disorder Meckel-Gruber syndrome (MKS) (Fig. 1) even though the clinical presentation of the three diseases varies considerably (Fig. 1B) and genotype-phenotype correlations are frequently unclear.¹² It is likely that loss of function alleles might often lead to milder disease and nulls to a more severe phenotype; however, there is emerging evidence that mutation in a second ciliary gene may affect clinical outcome,¹³ implicating modifier alleles or triallelic inheritence in the ciliopathies. One recent study on a large cohort of ciliopathy patients revealed extensive involvement of modifier alleles in conjunction with classical recessive inheritance.¹⁴ Here, we focus on MKS as the most severe of these three syndromes. Diagnosis of MKS relies on the classical diagnostic triad of cystic renal disease, central nervous system malformation (almost invariably occipital encephalocele), and polydactyly, although hepatic fibrosis is also extremely common.¹⁵ One of the most severe of the ciliopathies, MKS is lethal and death occurs in the perinatal period.

Figure 1. Genetic and Phenotypic overlap in Ciliopathies. (A) Venn diagram to show the high degree of allelism in the genetic basis of Joubert syndrome, nephronophthisis, and Meckel-Gruber syndrome. (B) Table of clinical phenotypes in Joubert Syndrome, nephronophthisis, and Meckel-Gruber syndrome.

Meckel Gruber Syndrome Proteins in Cilium Formation and Function

MKS shows extensive genetic heterogeneity. Mutations in MKS1, TMEM216, TMEM67, CEP290, RPGRIP1L, CC2D2A, NPHP3, TCTN2, B9D1 and B9D2, and TMEM231 thus far have been shown to be causative of MKS in humans,^16-26 and likely pathogenic mutations in TMEM138, TMEM237, EVC2, C5orf42, and SEC8 have also been identified.^27,28 The first genes underlying MKS were identified in 2006^21,24 and, following the realization that MKS is linked to ciliary biology, initial efforts focused on localizing MKS proteins within cells and assessing the effect of deleting or depleting them on cilium formation and morphology. The picture is complicated and many MKS-linked proteins are found in multiple subcellular locations, including the basal body which templates the ciliary axoneme^29-31 and centrioles/centrosomes^31,32,33,34 in non-ciliated cells. Several MKS proteins are predicted to be transmembrane proteins, ^26,35,36but the best studied of these, TMEM67, is also found at the basal body and the actin cytoskeleton as well as the apical and ciliary membrane.^35,36 Similarly, TMEM216 is found at basal bodies, the ciliary membrane, and other microtubule-based cellular structures in addition to membranes.²⁶ Despite having different distribution patterns, TMEM67 interacts with both TMEM216 and MKS1 in mammalian cells^26,36and C. elegans.²⁹ There are conflicting data regarding their role in cilium assembly, which may reflect differences in experimental technique or model organism/tissue-specificity, with some studies showing defects in early stages of ciliogenesis and loss of cilia^26,36,37,38 and others observing increased cilium length^24,39 or multi-centriolar, multi-ciliated cells⁴⁰ on depletion of the protein, sometimes with alterations to cilium structure.⁴¹ As well as their mutual interaction, MKS1 and TMEM67 also interact with nesprins and filamins, important scaffold proteins for rearrangement of the actin cytoskeleton^35,42 and cells from Meckel-Gruber syndrome patients carrying mutations in either TMEM67 or TMEM216 display actin cytoskeletal defects as well as ciliogenesis defects.^26,35 Less is known about many of the other MKS proteins. Several are required for ciliogenesis in model organisms^{25,30,31,33,43,44} and mutation or loss can also give rise to defects in wider cellular processes as microtubule organization,^33,43 vesicle trafficking,⁴⁵ and signal transduction.^46,47

The ciliary transition zone and ciliary compartmentalization

Lately the focus has shifted away from basal body and axonemal functions of MKS proteins to the transition zone, where multiple ciliopathy proteins form complexes involved in ciliary compartmentalisation. The commonality of ciliopathy proteins in this context provides an exciting linkage between the various syndromes.

It is not enough simply to build a cilium. For cilium-generated signaling to occur correctly, the cilium must be home to a distinct subset of proteins which must be able to transfer in and out of the cilium in response to different signals, and maintaining this compartmentalization represents a unique challenge. The membrane overlying the ciliary axoneme is contiguous with the plasma membrane, yet has a distinct composition,⁴⁸ suggesting that there must be a mechanism which prevents free diffusion between the two compartments. Several different mechanisms restrict entry and exit, and it is unclear how these co-operate or whether there are organism-, tissue- or cell-specific differences. Indeed, one study found no evidence of a diffusion barrier inhibiting entry to the cilium at all, but did find that certain proteins were selectively retained.⁴⁹ Several recent studies have identified overlapping protein complexes in this region which, when abrogated, result in a loss of correct ciliary compartmentalisation.^47,50-52 The majority of the proteins found in these complexes are ciliopathy proteins implicated in NPHP, JBTS, and MKS, which has led to the attractive hypothesis that these ciliopathies are linked via these complexes. The current model suggests there are two main complexes; a large MKS/JBTS complex consisting of approximately 16 proteins, and a second smaller complex consisting of NPHP1, 4 and 5.² All the MKS proteins except NPHP3 are found in the MKS/JBTS complex; MKS7/NPHP3 localizes to the inversin compartment just distal to the TZ.⁵³ In C. elegans, at least, the two complexes are linked via RPGRIP1L.⁵²

Vesicle transport of proteins to the cilium is limited by size restrictions imposed by the transition fibers, which extend from the distal end of the basal body and dock to the plasma membrane. The periciliary region may also represent a barrier to protein entry maintained via the apical actin cystoskeleton⁴⁹ and a ring of Septins 2 and 7 at the ciliary base,⁵⁴ which also mediates correct localization of the TZ protein complexes.⁵⁰ Another barrier exists at the base of the cilium; possibly a nuclear pore-like mechanism, consisting of many of the same pore proteins and utilizing an importin-mediated transport system responding to a Ran-GTP gradient from within the cilium.^55-57

Several important questions remain to be addressed, perhaps most importantly how these complexes act as a barrier: are they a molecular sieve, or do they regulate some other mechanism contributing to ciliary entry and exit? The current model is complicated by the various organisms involved in the work; for example, in C. elegans at least two proteins must be abrogated—one from each complex—before a phenotype is seen except in the case of RPGRIP1L,⁵² whereas in mammalian models (and apparently in human disease) mutation of a single gene is sufficient. It is also important to remember that while these proteins may function as a complex at the TZ after cilium assembly, many ciliopathy proteins have multiple localizations and roles outside of this TZ complex, including extra-ciliary functions as observed for IFT proteins in non-ciliated immune cells.⁵⁸ In the next few years, we will hopefully come to understand exactly how these complexes form, their proper function and regulation, and the contribution of each component.

Rodent models of MKS

We are only just beginning to understand the molecular and cellular functions of MKS proteins, and most of the roles we have described have only been uncovered in the last 2–3 y. Many mouse models of ciliopathies have been studied over the last few years⁵⁹ and these have helped yield new insights into the phenotypic defects observed following mutation in distinct ciliary proteins. Mouse and rat models covering many of the MKS genes have been produced.^{17,19,24,33,37-39,47,50,60-66} Many of the phenotypes observed can be explained by dysregulation of cilium-dependent signaling in these animals, presumably due to problems in the regulation of ciliary permeability at the transition zone. MKS mutants exhibit phenotypes that apparently reflect alterations in major developmental signaling pathways including Hedgehog (Hh) and Wnt signaling (see below) and several studies have found alterations in Hh signaling in these animals.^{19,37,38,50,60,62,66,67} The role of cilia in Hh signaling is now generally accepted; however their role in mediating canonical Wnt and non-canonical Wnt/planar cell polarity signaling remains controversial.⁶⁸ Nevertheless, MKS1, TMEM67, CEP290, and RPGRIP1L mutant mice show perturbation of either the canonical or non-canonical/planar cell polarity Wnt pathways,^38,60,63-65 suggesting a link between some MKS proteins and Wnt signaling that merits further study. Several other important developmental signaling pathways have also been linked with cilia, including fibroblast growth factor (FGF),⁶⁹ platelet-derived growth factor (PDGF)⁷⁰ and Notch/Delta.⁷¹ These remain less well understood and the effect of mutation in MKS genes on transduction of these and other pathways has not been examined. Further research will be required to disentangle the complex phenotypes and rationalise some of the conflicting data in the literature to gain a comprehensive understanding of how each mutation leads to the clinical presentation of MKS. In the meantime, a large body of evidence provides insights into why ciliopathy patients frequently present with developmental anomalies of the kidney, skeleton, and brain.

Cilia in Development of the Kidney, Skeleton, and Central Nervous System

Cystic kidney disease

Cystic kidney disease is characterized by abnormal proliferation and differentiation of the epithelial cells lining the tubules of the kidney, resulting in formation of fluid-filled cysts that prevent proper organ function and can eventually lead to renal failure. Kidney cysts are relatively common in adults over the age of 50⁷² and can be symptomatic of various diseases including diabetes and high blood pressure. However, cystic kidney disease is the most common genetic cause of end-stage renal disease⁷³ and nephronophthisis is the main genetic cause of renal failure in patients under 30.⁷⁴ It is apparent that the kidney is highly susceptible to cyst formation. Numerous genes are linked to cystogenesis, the majority localizing to cilia and/or basal bodies, but the effector pathways involved are still not well understood. Renal cysts are a hallmark of the ciliopathies, leading to the suggestion that all genes causative of cystic kidney disease have a cilium-related function,⁷⁵ but cystic kidneys do occur in well-characterized diseases with no obvious ciliary role, such as Zellweger syndrome (a disorder of peroxisome biogenesis).

One of the first links between cilia and renal cysts was the discovery that polycystin-1 (PC1/Pkd1) and 2 (PC2/Pkd2), causative of Autosomal Dominant Polycystic Kidney Disease (PKD), localize to the primary cilium. The polycystins are not essential for ciliogenesis⁷⁶ but form a mechanosensory ion channel at the cilium which responds to fluid flow through the tubule of the nephron by inducing a calcium influx possibly involved in cell proliferation and differentiation.⁷⁷ Mutations in a third gene, fibrocystin (Pkhd1), cause the autosomal recessive form of the disease, ARPKD. Fibocystin localizes to primary cilia and basal bodies and interacts directly with polycystin-2.⁷⁸ However, loss of mechanosensation alone is insufficient for cyst formation,⁷⁹ and disease progression differs according to age of onset, suggesting an additional level of complexity. Disruption of IFT88 or Kif3a (both required for ciliogenesis) during development leads to acute cystic kidney disease shortly after birth, while abrogation of gene function in adult mice results in slow disease progression despite a comparable loss of cilia.^80,81 Acute injury exacerbates this slow development of cysts in the adult.⁸⁰ It should also be noted that in at least two cases, cilia appear to be present in non-cystic kidney tissue but are lost after initiation of cystogenesis.^82,83 Both polycystins also play roles in cell-cell adhesion and several signaling pathways, some of which are implicated in cystogenesis (for review see refs. 84 and 85).

The mTOR pathway, which integrates growth factor input with cell metabolism and proliferation, has risen to prominence in recent years; increased mTOR activation is observed in both ADPKD and ARPKD patients. There may be direct links to mTOR downstream of the polycystins,^86-88 though other evidence shows calcium-independent links between the primary cilium and mTOR.⁸⁹ Initiation of cystogenesis occurs independently of mTOR in a Pkd1 mouse model.⁹⁰ Despite this, rapamycin treatment reduces cyst formation and prevents loss of kidney function in a variety of animal models, including at least one nephronophthisis model (reviewed in ref. 91), suggesting that the various mechanisms implicated in the different renal phenotypes might eventually converge upon mTOR,⁹² possibly due to activation of renal repair mechanisms.⁹³ Unfortunately, clinical trials in ADPKD patients have thus far shown little efficacy.^94,95 Two studies using cyclin-dependent kinase (CDK) inhibitors have shown promise in mouse models of ADPKD.^96,97 These inhibitors prevent the excess cell proliferation which contributes to cyst enlargement and decrease the number of apoptotic cells – a feature also implicated in cyst formation.⁹⁸

The role of Wnt signaling in the kidney: Proliferation vs. polarity

The renal disease in the ciliopathies may vary in cyst location, age of onset and gross morphology—for example, the kidneys of patients with ADPKD, MKS or infantile nephronophthisis are grossly enlarged while the kidneys of patients with juvenile or adolescent nephronophthisis are of normal or reduced size. This age-dependent variation in phenotype provides further evidence that mechanisms other than the mechanosensory contribute to cystogenesis, and that correctly regulated cilium-dependent signaling and compartmentalisation are key. The Wnt signaling pathway has emerged as a key player in the development of cystic kidney disease. This complex family of secreted glycoproteins is implicated in cell growth, polarity, and cell fate determination.⁹⁹

Wnt signaling is usually broken down into three distinct pathways: canonical/ β-catenin dependent, and two β-catenin-independent pathways - planar cell polarity (PCP, also called Wnt/JNK) and Wnt/calcium. However, while it is convenient to categorize in this way, these pathways are not independent and recent evidence suggests that cellular responses are more complex, resulting in a move toward an integrative approach.^100,101 The role of the cilium in Wnt signaling is controversial (discussed in ref. 68) but mutation of numerous ciliopathy genes results in perturbations in Wnt signaling,^102,103 including MKS genes,^38,60,64,65 though the exact nature of the changes varies. Inversin/NPHP2, for example, acts as a molecular switch from canonical to non-canonical signaling,¹⁰³ while AHI1/Jouberin is a positive regulator of canonical Wnt signaling and transports the effector β-catenin to the nucleus.¹⁰⁴

Canonical Wnt signaling induces the stabilization of β-catenin, which then translocates to the nucleus where it activates Wnt-specific target genes. Canonical Wnt signaling is linked to cell proliferation and is often hyperactivated in cancer.¹⁰⁵ Hyperactivation of canonical Wnt signaling also seems to play a central role in kidney cyst development; persistent β-catenin activation during development yields renal cysts similar to those seen in ADPKD.^106,107 However, downregulation of canonical Wnt signaling also results in cyst formation.¹⁰⁸ There is evidence that polycystin-1 may both stabilize β-catenin¹⁰⁹ and inhibit its activity,¹¹⁰ suggesting complex effects on canonical Wnt signaling in the pathogenesis of ADPKD. Together, these data imply that canonical Wnt signaling must be very carefully controlled through the development of the kidney, and that misregulation in either direction may lead to cyst development.

Non-canonical Wnt signaling has at least two branches; the planar cell polarity (PCP) pathway and Wnt/calcium. Wnt/Calcium results in a Wnt-induced rise in intracellular calcium levels, which activates several calcium-sensitive enzymes including Protein Kinase C (PKC), calcineurin, and calcium–calmodulin-dependent kinase II (CamKII) which in turn activate NFAT, NFκB, and CREB transcription factors.^111,112 This pathway is also linked to a decrease in cyclic guanosine monophosphate (cGMP) levels via PDE6, a cGMP-specific phosphodiesterase. Wnt/calcium may be required for proper development of the kidney,¹¹³ but as yet its roles in cystogenesis are not clear, and the majority of work has focused on the PCP pathway. The PCP pathway signals via the Rho family GTPases (including Rho, Rac, Cdc42) to control cell polarity and cytoskeletal arrangement. Loss of core PCP proteins results in cystogenesis,¹¹⁴ suggesting this pathway is fundamental to kidney homeostasis. PCP is involved in oriented cell division (OCD), which controls the orientation of mitosis within the plane of an epithelial cell layer. In kidney tubules, division along the axis of the tubule results in tubule elongation, while division perpendicular to that axis results in tubule dilation, an early step in cyst formation. Mice lacking Kif3a lose OCD, implicating the cilium as a key organelle in PCP signaling in the kidney.⁸⁰ OCD is commonly lost in cystic kidneys, but the degree to which it is a cause or an effect of cystogenesis is controversial. Some models such as the pck rat (a model of ARPKD) and a Pkd1 mouse model show loss of OCD prior to cystogenesis.^115,116 However, another study of Pkd1 and Pkd2 mutant mice show that tubule dilation, the first step in cystogenesis, precedes the loss of OCD, and the Pkhd1^del4/del4 mouse model (ARPKD) loses OCD without ever forming cysts.¹¹⁷ Therefore proliferation rather than polarity may be the main culprit in cystogenesis. In contrast, loss of OCD initiates cystogenesis prior to changes in proliferation in the inversin/NPHP2 mouse model of nephronophthisis.¹¹⁸

It is clear that Wnt signaling plays an important role in the development of cystic kidneys, but the exact contribution of canonical and non-canonical Wnt signaling pathways in the initiation of cystogenesis remains complex, and either up- or down-regulation of the canonical Wnt signaling pathway leads to cyst development.^104,106 It is not yet clear how Wnt signaling links to either mechanosensation or to mTOR activation, but it is apparent that a delicate balance of signaling is required for the correct formation and homeostasis of the kidney, and any perturbation of this balance results in cystogenesis.

Cilium-dependent Hedgehog signaling and skeletal patterning

Polydactyly has been a characteristic symptom of aberrant Hedgehog signaling for many years, but it was only in 2003 that it was finally discovered that cilia play a fundamental role in the Hh signaling pathway in vertebrates.¹¹⁹ As such, polydactyly is perhaps the best-understood of the ciliopathy symptoms. There are three Hh ligands in mammals—Indian, Desert, and Sonic—which signal via similar pathways. Sonic Hedgehog signaling patterns the vertebrate limb in a concentration-dependent manner; deletion of Sonic Hedgehog (Shh) results in only a single digit, while overexpression results in formation of additional digits.¹²⁰ Shh is normally only expressed in the posterior portion of the limb bud, which specifies digits 3, 4, and 5; digit 2 does not express Shh but receives signal from the adjoining region.¹²⁰ As such, the single digit observed when Shh is lost represents digit 1 (in humans, the thumb). If the Shh signal is extended throughout more of the limb bud, additional digits are formed. Hh signaling is likewise essential for proper bone development and skeletal patterning, and therefore is also responsible for the various other skeletal defects common throughout the ciliopathies such as craniofacial malformations and short limbs.

Perhaps the main reason the role of cilia in Hh signaling came as such a surprise was that the majority of studies had been performed in Drosophila, where Hh signaling is cilia-independent. Likewise, the popular model organism C. elegans has a divergent Hh pathway compared with vertebrates, with many “key” pathway members absent.¹²¹ It was not until studies of skeletal malformations in mice linked polydactyly with a variety of IFT genes that the importance of cilia in Hh signaling became apparent¹¹⁹ and the role of cilia in Hh signaling has since then been relatively well-characterized, though some questions remain.

Hh ligands interact with the Patched-1 (Ptch-1) receptor on the primary cilium membrane, which stimulates translocation of Ptch-1 out of the primary cilium, allowing entry of the signaling protein Smoothened (Smo).¹²² Activated Smo prevents processing of Gli transcription factors into transcriptional repressors, and thus allows transcription of Hh target genes. As the cilium is required for this response, it could be assumed that in the absence of a cilium, Shh signals are blocked; however, Shh absence results in loss rather than gain of digits, suggesting that the situation is more complex. In the absence of signal, the Gli transcription factors are normally processed to repressor forms (GliRs), which maintain the “off” state of Hh target genes. However, Hh ligands prevent this processing, leading to production of activated forms of Gli (GliAs). IFT proteins are required downstream of Ptch-1 and Smo for processing of Gli into both activator and repressor forms.^123-125 As such, ciliary malfunction can lead to phenotypes characteristic of both hyper- and hypo-activation of Hh signaling; indeed, ciliopathies commonly present with both polydactyly in the limb, combined with brain abnormalities due to a weakened Hedgehog response in the neural tube.

The key role of IFT proteins in Hh signaling may be due less to a direct modulatory effect and more due to their roles in trafficking; large portions of the Hh signaling pathway are linked with dynamic relocalization of various proteins. Activation of Gli transcription factors relies on their relocalization to the ciliary tip;^126-129 the current model suggests that here, Smo is able to release suppression by their binding partner, Suppressor of Fused (SuFu).^130,131 In the absence of signal, Sufu sequesters Gli proteins in the cytoplasm^128,132; both Sufu and Gli are trafficked to the ciliary tip via a Protein Kinase A-dependent mechanism shortly after Hh stimulation.^128,133 Notably, Sufu abrogration in cells devoid of cilia leads to full activation of Hh signaling^134,135; this is also the case in cells lacking Smo,¹³⁵ indicating that the Smo-mediated suppression of Sufu is key to pathway regulation, and that this suppression is cilium-dependent.

Recent evidence linking a number of ciliopathy proteins with ciliary compartmentalisation via the transition zone, together with the reliance of the Hh pathway on trafficking of signaling intermediates into and out of the primary cilium, suggests that many ciliopathy symptoms attributable to aberrant Hh signaling are caused by changes in ciliary permeability. At least one study has shown that Hh intermediates are not properly localized in response to Shh activation when TZ complexes are disrupted.⁴⁷ However, we do not yet understand the nature of the interaction between IFT complexes and transition zone complexes, nor the interactions of individual ciliopathy proteins with IFT-related processes. It is also essential that we develop a better understanding of the recently-uncovered “non-canonical” Hedgehog signaling mechanisms, defined as those which do not require the Gli effector proteins and which therefore may or may not require ciliary involvement.¹³⁶

Cilia in development of the CNS

The role of cilia within the nervous system is less well defined than within the kidney and skeleton but, given that many ciliopathies present with CNS malformations, it is clear that these organelles are of critical importance for brain development.^137,138 While most cell types in the brain have a primary cilium, including all neurons, motile cilia are restricted to the ependymal epithelia that line the brain ventricles and direct cerebrospinal fluid (CSF) flow through the ventricles. Non canonical Wnt signaling is required for correct organization and function of these cilia,^139,140 and defects lead to hydrocephalus. However, hydrocephalus can also develop prior to formation of a ciliated ependyma¹⁴¹ due to misregulation of primary cilium-dependent ion transport and neuropeptide signaling within the choroid plexus cells,^141,142 which produce the CSF.¹⁴² Taken together, these studies indicate that primary cilia on choroid plexus cells act as chemosensors to regulate CSF production and transport into the brain ventricles, where the motile cilia are used to regulate CSF flow.

Most recent research has focused on the role of cilia in neural development. The vertebrate brain starts out as a sheet of neuroepithelial cells that migrate and differentiate to form the plethora of cell types found in the adult.¹⁴³ A complex network of signaling cascades regulates this process¹⁴⁴ and following induction, a series of cell shape changes and intercalation movements transform the sheet into the neural tube, which ultimately forms the brain and spinal cord.¹⁴⁵ Closure of the neural tube begins in the middle and works outwards. Failure to close the tube at any point leads to neural tube defects including spina bifida, anencephaly (a lethal condition caused by anterior neural tube closure defects and consequent degeneration of the forebrain and skull), and encephalocele (characterized by protrusion of the brain, cerebrospinal fluid and surrounding membrane through the skull). The neural tube defect seen in MKS (and more rarely in JBTS) is almost invariably encephalocele.

Cell migration, polarization and differentiation within the forming neural tube are regulated by several signaling pathways linked to primary cilia. Neural tube development is specified by opposing dorso-ventral gradients of Wnt (and bone morphogenetic proteins) and Shh, which regulate each other’s expression in a negative feedback loop (Fig. 2). During development, Shh is produced by the notochord to specify the ventral neural tube and form neuronal stem cells,¹⁴⁶ a process which requires the primary cilium.¹⁴⁷ Links between cilia and Wnt signaling in the brain are less clear-cut,⁶⁸ but it is clear that the various branches of Wnt signaling are critically important for brain development. During neural tube development, Wnts are expressed dorsally and regulate neural crest development and cellular differentiation.¹⁴⁸ Primary cilia may inhibit the canonical Wnt pathway: cells where cilium assembly or function is compromised have upregulated canonical Wnt signaling^149,150 and the ciliopathy protein Inversin is proposed to act as a switch between canonical and non-canonical Wnt signaling at the basal body.¹⁰³ As such, loss of the primary cilium is likely to upregulate canonical Wnt signaling and downregulate non-canonical signaling, with downstream consequences for brain development. The non-canonical PCP pathway is vital to neural tube development and has also been linked to ciliogenesis in the brain,^151-153 providing another potential level of cross-talk between Wnt and Hh. Therefore it is unsurprising that many ciliopathies present with neural tube defects.¹⁵⁴

Figure 2. Hedgehog and Wnt Signaling in Kidney, Limb, and Neural Tube. A cartoon to show the effects of the Wnt signaling pathway on renal cystic disease, of the Hedgehog (Hh) pathway in the developing limb, and of both pathways in the neural tube. In the kidney, the balance between canonical (β-catenin dependent) and non-canonical Wnt signaling must be maintained; dysregulation of either branch, whether up- or downregulation, leads to development of cysts. In the limb bud, a gradient of Hh signaling is required to pattern the digits; where upregulation of Hh occurs, the gradient is expanded and additional digits are specified whereas when Hh is downregulated, the gradient becomes restricted and a reduced number of digits is specified. In the neural tube, a Wnt (and, to perhaps a greater extent, BMP) gradient is involved in dorsal cell fate determination; the gradient is counteracted by a ventral Hh gradient. Up- or downregulation of either pathway perturbs this balance and affects axis specification in the developing neural tube. Wnt signaling also has important effects on neuronal migration in the developing brain, leading to numerous other CNS effects.

Following neural tube closure, a population of multipotent neural crest cells de-adhere from the epithelium at the top of the neural tube and migrate until they encounter cues that regulate their differentiation into various cell types including neurons and glia of the peripheral nervous system, craniofacial cartilage and bone, smooth muscle, and pigment cells.¹⁵⁵ A subset of these cells form a primary cilium which, unusually, is formed on the basolateral rather than the apical cell surface, and the number of ciliated cells increases in frequency as neurogenesis proceeds, suggesting that the two may be linked.¹⁵⁶ The function of this cilium is unknown, but given its protrusion into the basement membrane underlying the neural crest cells it is tempting to speculate that it might be involved in sensing the signals that regulate changes from adhesive to migratory behavior that are key to early brain development. Wnt and Hh signaling are essential for neural crest cell migration^157,158 as well as neural tube patterning, and when this migration fails in ciliopathy patients, characteristic craniofacial defects are observed,^158,159 suggesting that cilium dysfunction leads to wide-spread defects in developmental signaling required for neural crest cell migration and patterning.

Neurons, glial cells and oligodendrocytes of the CNS arise from several populations of neural stem cells¹⁶⁰ and newborn neurons migrate long distances from their birthplace to reach their final destinations. This requires them to integrate and respond to a complex series of guidance cues that orchestrate the intricate migration patterns that are needed to form the brain.^161,162 Primary cilia are required for neuron formation in the developing brain^67,163-165 but the role they play in neuronal migration is less clear, with one study showing that cilia develop only on post-migratory cortical neurons¹⁶⁶ while other studies show that cilia are critical for migration and positioning of developing neurons.^167,168 Thus, while it seems clear that primary cilia are needed at early stages of brain development for neural tube closure and neural crest cell migration, further work is required to clarify their role in neuronal migration. Nevertheless, several studies have shown that Hh signaling through primary cilia is needed for brain patterning, with mutations leading to loss or reduction of IFT components resulting in a variety of CNS defects.^165,169,170 Cilia-mediated Shh signaling is key to cerebellum development and regulates proliferation of neonatal granule cells, with loss of IFT function or Shh removal leading to cerebellar hypoplasia.^171,172 The role of Wnt signaling is more difficult to resolve, as there is extensive crosstalk between the three branches of Wnt signaling during neuronal migration.¹⁷³ Disentangling the role of cilia in this process will require further investigation.

Finally, the role of primary cilia in the CNS is not restricted to embryogenesis and there is a growing body of evidence that indicates that they are also important in the postnatal and adult brain, where they mediate synapse formation in adult-born neurons¹⁷⁴ and are involved in appetite regulation and obesity.^81,175,176 Moreover, defects in brain primary cilia have been linked to the development of neurodegenerative disease¹⁷⁷ and cancer.^178,179 The picture is complicated and there is much to learn, but it is clear that, in common with primary cilia in the kidney and skeleton, primary cilia of the CNS may be regarded as antennae to receive signals from the extracellular environment and transmit them to the cell. Failure to build a cilium or to regulate cilium-dependent signaling can result in a multitude of developmental and adult brain defects.

Perspectives

It seems clear that the majority of symptoms found in ciliopathies are less to do with the loss of a cilium and more to do with defects in ciliary signaling—either developmental signaling or environment sensing. The linkage of many ciliopathy proteins together at the ciliary transition zone, and their role in ciliary compartmentalisation, is exciting. However, these are early days and the picture is likely to be much more complicated as the TZ is not the only mechanism contributing to ciliary signaling. Further, developmental signaling pathways such as Wnt and Hh are tremendously complicated, and there is extensive cross-talk, so it is important to consider the total signaling landscape rather than focusing on a single transduction cascade when assessing phenotypes. Moreover, it has only just become clear that the ciliary antenna can transmit signals as well as receive them¹⁸⁰; what it transmits and the contribution to ciliary function and organ development remains to be seen. However, it is important not to simply focus on developmental signaling pathways. These are largely restricted to metazoan organisms yet ciliopathy proteins play roles across eukaryotic evolution, implying more fundamental roles in ciliary biology. It will be essential to establish the function of ciliopathy proteins in a range of model organisms, and to integrate this information to provide a full picture.

Perhaps one of the key questions still remaining in the field is one of the hardest to answer—given the complexity of the interactions between the various ciliopathy genes and other ciliary and non-ciliary proteins, what are the relative contributions of perturbing a single protein vs. perturbation of an entire ciliary network? The intricacy of this ciliary network makes selection of appropriate controls extremely difficult—yet these are critical for distinguishing gene-specific roles from the role of the organelle itself. Furthermore, there is mounting evidence that “ciliary” proteins play extra-ciliary roles, and therefore dissecting out cilium-dependent vs. cilium-independent functions of these genes is essential for our understanding of ciliopathy. The picture may be far from complete, but many of the pieces of the jigsaw puzzle are now in place and the next few years will see some exciting advances in ciliopathy research that will advance our understanding of these devastating diseases.

Abbreviations:
JBTS	=	Joubert syndrome
NPHP	=	Nephronophthisis
MKS	=	Meckel-Gruber syndrome
ARPKD	=	autosomal recessive polycystic kidney disease
ADPKD	=	autosomal dominant polycystic kidney disease
CNS	=	central nervous system
Hh	=	Hedgehog
Shh	=	Sonic Hedgehog
PCP	=	planar cell polarity
OCD	=	oriented cell division
IFT	=	Intraflagellar transport
TZ	=	transition zone

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We would like to thank Marie Harrisingh (University of Edinburgh) for helpful comments on the manuscript. Work in our laboratory is funded by the Medical Research Council.

10.4161/org.27375