Primary cilia mediate intercellular signaling pathways which are involved in the regulation of cellular processes and the formation and maintenance of all organs and structures within the human body. Cancer is characterized by uncontrolled cell division as well as an impaired ability to undergo apoptosis [4] and because it develops as a result of altered intra- and intercellular signaling, disturbances of cilia-mediated signaling pathways can result in tumor formation [5–7]. While it seems as if canonical WNT signaling is restricted by cilia [8–10], various publications have shown cilia-dependent mediation of sonic hedgehog (SHH), platelet-derived growth factor receptor-? (PDGFR?), NOTCH, transforming growth factor (TGF)-?, and non-canonical WNT signaling (Fig. 1a–e) [8, 11–18].
Cilia-mediated signaling pathways whose proper regulation is dependent on the proteasome and the structure of the proteasome. a–e SHH, PDGFR?, NOTCH, TGF?, and canonical WNT signaling is transduced by primary cilia. a In the absence of the ligand SHH, SMO remains in cytoplasmic vesicles and is inhibited by PTCH1. As a result, GLI2 and GLI3 (forming a complex with SUFU) are phosphorylated most likely within the cilium and subsequently get proteolytically processed to their repressor forms (GLI2/3-R) by the proteasome at the ciliary base. In turn, GLI2/3-R translocate into the nucleus and represses the expression of SHH target genes. Importantly, GLI3 is the predominant repressor. When SHH binds to its receptor PTCH1, the SHH/PTCH1 complex leaves the cilium and PTCH1 is not able to inhibit the action of SMO any longer. Thereupon, SMO is transported into the cilium and converts the full-length forms of GLI2 and GLI3 (GLI2/3-FL) into their activator forms. In the course of this conversion process, SUFU dissociates from the complex enabling the GLI2 and GLI3 activator forms to induce SHH target gene expression. b In the ciliary membrane, PDGFR? is bound by its ligand PDGF-AA and subsequently becomes dimerized and phosphorylated. The phosphorylation of PDGFR? induces the activation of the MEK 1/2-ERK 1/2 and AKT/PKB signaling pathways. c Initiating NOTCH signaling, the extracellular domain of a NOTCH ligand (JAGGED or DELTA) binds to the NOTCH receptor which is located in the ciliary membrane. As a result, the NOTCH receptor undergoes a three-step cleavage and finally releases the NOTCH intracellular domain (NIC). NIC enters the nucleus and activates NOTCH target genes. d The receptors of the TGF? pathway, TGF?-RI and TGF?-RII, are located at the ciliary base. When the TGF? ligand binds to the receptors a heterotetrameric receptor complex composed of TGF?-RI and TGF?-RII is formed and activated. This activation results in the phosphorylation and activation of SMAD2 and SMAD3. The phosphorylated SMADs 2 and 3 associate with a co-SMAD called SMAD4. Afterwards, the complex consisting of SMAD2, 3, and 4 enters the nucleus and activates TGF? target genes. e In the inactive state of the canonical WNT pathway, a destruction complex consisting of APC and AXIN triggers the phosphorylation of ?-catenin by GSK3. After this phosphorylation event, ?-catenin gets ubiquitinated and finally degraded. In the active state, WNT ligands bind to FRIZZLED and LRP receptors leading to the activation of DSH. DSH recruits the destruction complex to the plasma membrane, thereby interfering phosphorylation of ?-catenin. Afterwards, ?-catenin translocates into the nucleus and activates canonical WNT target gene expression. Primary cilia restrict canonical WNT signaling because the ciliary protein KIF3A is able to inhibit the phosphorylation of DSH. f The proteasome consists of the catalytic 20S subunit and two regulatory 19S subunits. The 20S subunit displays a cylindrical arrangement of four stacked heptameric rings. Each ring is composed of seven ? and ? subunits, respectively. Only three subunits (PSMB8-10) display a proteolytic activity equipping the proteasome with trypsin-like, chymotrypsin-like, and caspase-like abilities. The 19S subunit can be subdivided into two subcomplexes: a base complex (being constituted of six ATPases [PSMC1-6] and three non-ATPases [PSMD1, 2 and 4]) and a lid complex (consisting of nine non-ATPases [PSMD3, 6-8, 11-14, and SHFM1])
Of all the investigated associations between primary cilia and signaling pathways, the relationship between primary cilia and SHH signaling is the best studied. In SHH signaling, the 12-pass transmembrane protein patched1 (PTCH1) is located in the ciliary membrane of vertebrates (Fig. 1a). When the SHH ligand binds to its receptor PTCH1, the SHH/PTCH1 complex leaves the cilium. As a consequence, the seven-transmembrane protein smoothened (SMO) is allowed to accumulate in the ciliary membrane and to invoke glioblastoma (GLI) transcription factors. Three GLI isoforms exist in vertebrates—GLI1, 2, and 3. The GLI proteins regulate the expression of SHH target genes and thereby cell proliferation, differentiation, survival and growth [19, 20]. While GLI1 exclusively functions as a constitutive transcriptional activator [21, 22], GLI2 and GLI3 can serve as an activator or a repressor [23]. In the presence of SHH, full-length GLI2 (GLI2-185) and GLI3 (GLI3-190) proteins are converted into a transcriptional activator (GLI2-A and GLI3-A, respectively) most likely by modifications [24, 25]. In the absence of SHH, the full-length proteins can be proteolytically processed into transcriptional repressors (GLI2-R, also known as GLI2-78, and GLI3-R, also known as GLI3-83, respectively) [26]. It was reported that GLI3-R is the predominant repressor of SHH target gene transcription [26]. The ratio of activator and repressor forms regulates cellular processes dependent on SHH signaling.
Similar to SHH signaling, activated PDGF receptors control cellular processes like proliferation, anti-apoptosis, migration, differentiation, actin reorganization, and cell growth [27–29]. The receptor PDGFR? localizes to cilia and undergoes dimerization and phosphorylation after being bound by its ligand PDGF-AA [14] (Fig. 1b). Stimulation of PDGFR? provokes the activation of signal transduction through the MEK 1/2-ERK 1/2 and AKT/PKB pathways. In the absence of cilia, PDGFR? signaling is inhibited [14]. Additionally, PDGFR? signaling is restricted by the mammalian target of rapamycin (mTOR) signaling pathway [30–32], which is also associated with cilia-mediated signaling. LKB1, a negative regulator of mTOR, localizes to cilia and its action leads to an accumulation of phosphorylated AMPK at the basal body [33]. In turn, the phosphorylation of AMPK results in the inhibition of mTOR signaling via a mechanism that is only poorly understood. Interestingly, deregulation of mTOR signaling has been described in many cancer types [34–36]. Previously, it has been demonstrated that NOTCH signaling depends on primary cilia [16, 17] (Fig. 1c). NOTCH signaling starts when the extracellular domain of a NOTCH ligand, e.g., delta-like1–4 or jagged1–2, binds to the NOTCH receptor (NOTCH1–4) [37]. A ciliary localization was shown for NOTCH1 and NOTCH3 [16, 17]. After the binding event, the NOTCH receptor undergoes a three-step cleavage and finally releases the NOTCH intracellular domain (NIC). Following this, NIC enters the nucleus and interacts with its DNA-binding cofactor RBP-J/CBF1/CSL thereby activating NOTCH target genes. NOTCH signaling controls among other proliferation and differentiation [38].
Moreover, TGF? signaling relates to cilia [18] (Fig. 1d). Both receptors of the pathway, TGF?-RI and TGF?-RII, are located at the base of primary cilia. The ligand-induced formation and activation of a heterotetrameric receptor complex composed of TGF?-RI and TGF?-RII results in the phosphorylation and activation of the SMAD2 and SMAD3 proteins which are present at the ciliary base [18]. The phosphorylated SMADs 2 and 3 associate with a co-SMAD called SMAD4 that is also detectable at the base of cilia. Subsequently, the complex consisting of SMAD2, 3, and 4 enters the nucleus and activates TGF? target genes. TGF? target genes control cellular processes like proliferation, differentiation, morphogenesis, tissue homeostasis, and regeneration [39].
Primary cilia are also connected to WNT signaling [40], which can be classified as canonical (?-catenin dependent) or non-canonical (?-catenin independent). In the inactive state of the canonical WNT pathway, a destruction complex consisting of adenomatous polyposis coli (APC) and AXIN triggers the phosphorylation of ?-catenin by casein kinase 1 (CK1) and glycogen synthase kinase 3 (GSK3) (Fig. 1e). Afterwards, ?-catenin gets phosphorylated, ubiquitinated, and finally degraded [41]. The WNT/?-catenin pathway becomes initiated by binding of WNT ligands to frizzled (FZ) receptors and low density lipoprotein-related proteins 5/6 (LRP 5/6) and leads to the activation of the cytoplasmatic phosphoprotein disheveled (DSH). Subsequently, DSH recruits the destruction complex to the plasma membrane, thereby inhibiting phosphorylation of ?-catenin. This operation of DSH enables ?-catenin to translocate into the nucleus for activating target gene transcription. Several processes are controlled by canonical WNT signaling: cell fate determination, migration, proliferation, tumor suppression, and self-renewal of stem and progenitor cells [42, 43].
In contrast to canonical WNT signaling, the non-canonical WNT pathway is less well understood. Hence, it is unknown, if ?-catenin-independent WNT pathways function as different distinct pathways or if these pathways form a large signaling network [44]. Like the canonical WNT pathway, it starts with a WNT ligand binding to the FZ receptor, but does not require the presence of LRP co-receptors or ?-catenin. Non-canonical WNT signals are mediated through intracellular Ca2+ levels and involvement of RHO A, ROCK, and JNK kinase. These factors play an important role in the regulation and remodeling of the cytoskeleton and are greatly involved in the control of planar cell polarity (PCP). PCP is established by intercellular communication that regulates the composition of cells polarizing structures within the plane of a tissue, i.e., stereocilia bundle orientation in the inner ear [45]. In addition to managing cytoskeleton organization, non-canonical WNT signals regulate proliferation and migration [46].
The restriction of canonical WNT signals by cilia is likely, since DSH is constitutively phosphorylated in Kif3a-negative mice which are unable to assemble cilia [47]. However, non-canonical WNT signaling seems to be mediated by primary cilia [8–10]. One core PCP gene product, van gogh-like 2 (VANGL2), was found in cilia [48]. The ciliary presence of VANGL2 [48] and the finding that VANGL2 is essential for the transduction of WNT5a-induced signals to establish PCP [49] suggest that non-canonical WNT signaling might be mediated by cilia. This hypothesis is supported by data showing that disruption of BBS protein function leads to ciliary dysfunction along with perturbation of PCP [48] and that ciliopathy genes interact genetically with VANGL2 [48, 50]. In summary, these data suggest that primary cilia mediate non-canonical WNT signals and limit canonical WNT signaling [51].
Dysregulation of any of these pathways could lead to oncogenesis. In many cases, upregulation of their target gene expressions led to an increased cell proliferation, which in turn caused tumorigenesis [52–56]. One of the best studied oncogenic signaling pathways is the SHH pathway which was already analyzed in combination with cilia in cancer cells [57, 58]. In 2009, Han et al. and Wong et al. [59, 60] described the role of primary cilia in the development of medulloblastomas and basal cell carcinomas. In regard to SHH signaling, both groups showed that the absence of cilia can protect against tumorigenesis and, in addition, that the presence of cilia can be necessary for the induction of tumors. First, they induced tumorigenesis via a cell type-specific expression of an activated SMO protein. Then, they performed the experiments in mice that were unable to form cilia in the particular cell type for the formation of either medulloblastomas or basal cell carcinomas. In both cases, ciliary deficiency protected against SMO-induced tumorigenesis [59, 60].
Second, the same groups investigated the consequences of constitutively active GLI2 on tumorigenesis [59, 60]. In case of basal cell carcinoma development, constitutively active GLI2 was sufficient to induce carcinogenesis [60], while, in case of medulloblastoma development, constitutively active GLI2 did not give rise to carcinogenesis [59]. Importantly, the combination of constitutively active GLI2 and loss of cilia led to the formation of medulloblastomas [59] giving circumstantial evidence that the additional decreased amount of GLI3-R caused by ciliary absence might be necessary to induce oncogenesis. Accordingly, the activation of SHH target gene expression alone is not strong enough for driving the development of some cancer types, but in combination with an inhibited repression of SHH target gene expression by reducing the amount of GLI3-R, activation of SHH target gene expression is sufficient to induce oncogenesis. Possibly, the reason for these differences is that the importance of GLI3-R is different in diverse cancer types. Perhaps, it is even the case that the efficiency of GLI3 processing is different in different cancer types and the amount of GLI3-R varies. A decisive factor for the proteolytic processing of GLI3 is the proteasome.