The Moonwalker Mouse: New Insights into TRPC3 Function, Cerebellar Development, and Ataxia

Mwk mice harbor a single non-synonymous point mutation in exon 7 of the Trpc3 gene located on chromosome 3 [18]. This gene encodes the TRPC3 channel, a member of the transient receptor potential (TRP) family of ion channels. The TRP superfamily constitutes one of the largest ion channel families and is involved in numerous physiological functions [20, 21]. The mammalian TRP channel proteins form six-transmembrane cation-permeable channels that are grouped into six subfamilies including the canonical TRPC family (TRPC 1–7). TRPC channels are widely expressed in different mouse tissues including the nervous system, where they function as non-selective calcium entry channels with distinct modes of activation [22, 23]. Microarray analysis of more than 80 different mouse cell lines and tissues shows that TRPC3 is highly expressed in the nervous system, particularly in the cerebellum (BioGPS gene portal) (http://​Biogps.​org). Notably, it is the alternatively spliced short isoform TRPC3c that is preferentially expressed in the cerebellum [24]. This TRPC3 isoform lacks part of the CIRB domain (see Fig. 1b), which is known to modulate TRPC channel activity via competitive binding of the inositol triphosphate (IP₃) receptor (IP₃R) and calmodulin (CaM), resulting in a channel with enhanced efficacy due to increased channel-opening frequency [24]. Within the cerebellum, TRPC3 is most highly expressed on the soma as well as the dendrites of Purkinje cells [25, 26] (Fig. 1a). In fact, TRPC3 was shown to be the most abundant TRP channel in Purkinje cells [25]. In addition to Purkinje cells, TRPC3 was recently found to be selectively expressed in type II unipolar brush cells (UBCs), which are excitatory glutamatergic interneurons in the cerebellum (Fig. 1a) [19]. Using a Trpc3 knockout mouse model, Hartmann et al. have demonstrated that TRPC3 mediates metabotropic glutamate receptor subtype 1 (mGluR1)-dependent synaptic transmission in Purkinje cells [25]. Similarly, activation of group I mGluR gates TRPC3 channels in type II UBCs [19]. Moreover, TRPC3 was recently suggested to be essential for the induction of long-term depression (LTD) in cerebellar Purkinje cells [27]. However, how loss of synaptic function in the Trpc3 mouse mutants affects LTD remains to be determined.

The Mwk mutation results in a threonine-to-alanine amino acid change (T635A; RefSeq# NP_062383)¹ in the highly conserved cytoplasmic S4–S5 linker region of the TRPC3 protein (Fig. 1b). The Mwk mutation does not alter the normal expression pattern of TRPC3. However, electrophysiological recordings from Mwk Purkinje cells revealed that mutant TRPC3 exhibits altered gating properties that promote channel opening under conditions of low mGluR1 activation [18]. Consistently, Mwk but not wild-type TRPC3 promotes increased calcium signaling and cell death upon overexpression in neuronal cell lines [18] (unpublished observations). The molecular mechanisms of how the Mwk mutation leads to abnormal TRPC3 channel gating remain to be fully elucidated. One possible mechanism might be the loss of an inhibitory phosphorylation by protein kinase C γ (PKCγ). PKCγ has been shown to inhibit TRPC3 channel activity in overexpression experiments in heterologous cell lines [28‐31]. Indeed, threonine 635, which is mutated in the Mwk mouse, has been shown to be phosphorylated by PKCγ in an in vitro kinase assay [18]. However, this inhibitory phosphorylation event has not yet been confirmed in vivo. Moreover, Nelson and Glitsch have argued that native TRPC3-dependent currents elicited in cerebellar Purkinje cells are unlikely to be targets of conventional PKC or PKG kinases in contrast to the findings observed in overexpression experiments [32]. Alternatively, the Mwk mutation in the S4–S5 linker might disturb intramolecular interactions within TRPC3 that are important for the gating mechanism of the channel. This mechanism might be similar to the established electromechanical coupling in the structurally related voltage-gated potassium channels, in which the S4–S5 linker is critical for gating [33]. Recently, the S4–S5 linker in TRPC4 and TRPC5 channels was identified as critical constituent of TRPC4/5 channel gating [34], suggesting a similar key function for the S4–S5 linker in TRPC3. Future structural studies on TRPC3 and related TRP channels will undoubtedly shed more light on the intramolecular interactions that are crucial for TRPC3 channel activation and function.

From early postnatal days, Mwk mouse mutants display growth retardation and remain about 60 % the size of their wild-type littermates throughout life. Homozygous mutants (Mwk/Mwk) are embryonically lethal. Heterozygous Mwk mouse mutants display motor and coordination defects from about 3 weeks of age [18] (Fig. 2). Specifically, Mwk mice exhibit retropulsion and a wider and “shuffling” gait compared to wild-type littermates. Mwk mice are also severely impaired in their ability to maintain their balance in the static rod test. Following their initial behavioral characterization, Mwk mice are being extensively characterized as part of the EuroPhenome project [35]. In this large-scale phenotyping effort, Mwk mutants were found to display significant hyperactivity, tail elevation, and hind limb grasping behavior, indicative of extra-cerebellar signs. Many other behavioral tests in the EuroPhenome pipelines are still ongoing.

Notably, the Trpc3 knockout mice also exhibit an ataxic phenotype, albeit much milder than the Mwk mice [25]. The fact that both loss and gain of TRPC3 function lead to cerebellar ataxia has been previously attributed as “conflicting” [36]. However, it is commonly observed in other ataxic mouse mutants as well as human ataxia patients that both decreased and increased ion channel activities can lead to Purkinje cell dysfunction and consequently cerebellar ataxia [6, 37]. This established concept that functionally different mutations in the same ion channel gene can cause the same disease phenotype highlights the importance of the associated biological mechanisms such as maintenance of calcium homeostasis for proper Purkinje cell function.

Histopathological analysis of the Mwk brain revealed a slow but progressive loss of Purkinje cells starting at 4 months of age [18] (Fig. 2). Purkinje cell degeneration is particularly pronounced in the lateral hemispheres of the cerebellum. Moreover, type II UBCs are dramatically reduced by 1 month of age [19]. The mode of cell death in the Mwk cerebellum remains unclear. No evidence of DNA fragmentation of activation of pro-apoptotic signaling molecules was found, suggesting a cell death mechanism other than apoptosis. Given the underlying gain-of-function mutation in TRPC3 and observed increase in calcium signaling, cell death is likely to occur due to calcium overload. This mechanism might also explain the fact that type II UBCs are lost much earlier compared to Purkinje cells in the Mwk cerebellum. Purkinje cells have an extensive calcium-buffering capacity due to the presence of high concentrations of calcium-binding proteins including parvalbumin and calbindin. The latter has been estimated to comprise more than 15 % of total protein in Purkinje cells [38]. Hence, the ability to handle an excessive influx of calcium is likely to be much more rapidly exceeded in UBCs compared to Purkinje cells. It would be interesting to further dissect the relative contribution of UBC loss and Purkinje cell dysfunction with regard to the ataxic Mwk phenotype in a conditional mouse mutant. Type II UBCs are predominantly found in the vestibulocerebellum, suggesting that their loss in the Mwk mouse might contribute to the profound balance impairment of the mutant.

Significant cerebellar TRPC3 expression starts within the second postnatal week and peaks at 3 weeks postnatally [18] (see also Cerebellar Development Transcriptome Database) (http://​www.​cdtdb.​neuroinf.​jp) (Fig. 2), suggesting a role for this non-selective cation channel during Purkinje cell development. Interestingly, Mwk mice exhibit ataxic behavior long before loss of Purkinje cells is observed (Fig. 2), indicating that the ataxic phenotype is caused by the dysfunction rather than the loss of Purkinje cells. Indeed, electrophysiological alterations are evident in the Mwk Purkinje cells as early as 3 weeks of age [18, 19]. Cell recordings showed that only a small fraction of Mwk Purkinje cells are spontaneously active and suggested that Mwk Purkinje cells are generally depolarized [19]. Consistent with these observations, Mwk Purkinje cells responded differently to stimulation of mGluR1, with either no inward current or an inward current with significantly smaller spikes at a higher frequency [18]. Together, these findings suggest that Purkinje cells in the developing Mwk cerebellum are profoundly impaired in their intrinsic and evoked electrophysiological properties.

Developmental abnormalities in the Mwk cerebellum are also evident in the altered dendritic morphology of the mutant Purkinje cells. The postnatal increase in TRPC3 expression coincides with the most intensive phase of dendritic arborization of Purkinje cells, suggesting a role for TRPC3 in this process. Consistent with this idea, Mwk Purkinje cell dendritic arbors were found to be significantly less elaborate compared to wild-type Purkinje cells [18] (Fig. 2). This phenomenon was particularly apparent in organotypic slice cultures of the developing cerebellum, where Mwk Purkinje cells exhibit a profoundly reduced dendritic arbor. This dendritic phenotype in the Mwk mice that have increased cerebellar calcium signaling is consistent with other studies showing reduced Purkinje cell dendritic growth upon activation of mGluR1 [39, 40] and PKC [41‐43]. Recently, voltage-gated calcium channels (VGCCs) were implicated in the regulation of Purkinje cell dendritic growth after chronic mGluR1 or PKC activation [44]. Interestingly, genetic knockout of Trpc3 or pharmacological inhibition of TRPC3 did not alter Purkinje cell dendritic arborization [44]. Together, these findings suggest that it is the excessive calcium influx during the critical period of dendritic development in Mwk mice that causes inhibition of dendritic growth. The inhibition of Purkinje cell dendritic arborization by increased calcium signaling through activation of mGluR1, TRPC3, and PKC might be a powerful negative feedback mechanism to limit the size of the dendritic arbor after the establishment of a sufficient number of granule cell parallel fiber contacts. This model is consistent with earlier findings that electric activity controls Purkinje cell dendritogenesis [45]. In further support of this, Purkinje cells of the Lurcher mouse with a gain-of-function mutation in the δ2 glutamate receptor (GluD2), which changes the receptor into a leaky membrane channel resulting in chronic depolarization, also exhibit profoundly impaired dendritic growth [46].

An extensive literature supports a role for calcium signaling in the dendritic morphogenesis of neurons including the local regulation of dendritic branch dynamics as well as the global control of gene transcription (reviewed in [47, 48]). However, surprisingly, little is known about the specific calcium-activated signaling mechanisms that regulate Purkinje cell dendritic arborization downstream of mGluR1, TRPC3, and VGCCs. PKCs have emerged as one of the kinases that regulate Purkinje cell dendritic arborization in response to increased calcium influx (reviewed in [43, 49]). It appears that both PKC isoforms, PKCα and PKCγ, contribute to Purkinje cell dendritic development, depending on the strength of the stimulus. Furthermore, calmodulin-dependent kinases (CaMKs) have been implicated in Purkinje cell dendritic growth. Pharmacological block of CaMKs has been reported to reduce the number of Purkinje cell primary dendrites in cerebellar cultures [50]. This effect was only observed between 5 and 15 days in vitro but not later, suggesting that CamKs might be involved in the early phases of Purkinje cell dendritogenesis and before a critical function of TRPC3 in this process. The specific substrates of PKCs and CamKs that regulate Purkinje cell dendritic development remain largely unknown. The only reported downstream effector of CamKII signaling in Purkinje cells controlling local dendritic branching is the microtubule destabilizer stathmin, which is phosphorylated by CamKII upon activation of VGCCs and mGluR1 [51]. Thus, further studies are needed to dissect the molecular mechanisms that control the calcium-dependent Purkinje cell dendritic growth during development. The Mwk mouse provides an excellent model system to study the transcription-dependent and transcription-independent effector pathways that regulate the developmental dendritic arborization of Purkinje cells. Efforts are underway to identify the genes and pathways that mediate Purkinje cell dendritic growth upon TRPC3 activation. Importantly, as Purkinje cells constitute significantly less than 1 % of the total cerebellar cell population [52], Purkinje cell-specific transcriptional and biochemical profiling requires enriched cell preparations obtained by cell sorting or laser microdissection [53, 54].

Mouse knockout studies have demonstrated a key role for TRPC3 in mediating mGluR1-dependent synaptic transmission in Purkinje cells [25]. In the molecular layer, Purkinje cells receive major glutamatergic input from granule cell parallel fibers. Glutamate released from the parallel fiber presynaptic terminals binds to AMPA receptors and mGluR1 at the parallel fiber-Purkinje cell synapse (Fig. 3). Influx of sodium ions through AMPA receptor channels leads to a rapid postsynaptic depolarization, whereas mGluR1s mediate a local dendritic calcium signal and a prolonged depolarization known as slow excitatory postsynaptic current (slow EPSC) (reviewed in [55]). Glutamate binding to mGluR1 activates phospholipase C (PLC) via G protein Gα_q to produce IP₃ and diacylglycerol (DAG). IP₃ subsequently activates calcium-permeable IP₃R1 receptor channels on the endoplasmic reticulum (ER) membrane, resulting in a characteristic local dendritic calcium response that is required for the induction of LTD [56]. The mGluR1-dependent slow EPSC is mediated by TRPC3 [25]. The exact gating of TRPC3 following of mGluR1 activation remains to be elucidated but might involve Rho GTPase-dependent activation of phospholipase D1 (PLD1) [57] as well as IP₃R and calmodulin [58] and others (reviewed in [59, 60]). TRPC3 activity has been proposed to be negatively regulated through phosphorylation by PKCγ and perhaps other kinases [28‐31, 59], but the exact nature of this inhibitory phosphorylation remains controversial [32].

Importantly, disruption of any component of the mGluR1-triggered signaling cascade results in cerebellar dysfunction and disease. Similar to the Mwk mouse phenotype, knockout mice deficient in the genes encoding mGluR1 [61, 62], Gα_q [63], IP₃R1 [64], TRPC3 [25], and PKCγ [65] all exhibit cerebellar ataxia. Moreover, auto-antibodies against mGluR1 are associated with neoplastic as well as subacute cerebellar ataxia in human patients [66‐68]. Also, deletions in the ITPR1 gene encoding IP₃R1 cause SCA15 [69]. Gain-of-function missense mutations in the PRKCG gene encoding PKCγ cause SCA14 [70]. Interestingly, SCA14-associated PKCγ mutants were shown to fail to phosphorylate TRPC3, resulting in a sustained calcium influx that might be central to the SCA14 pathogenesis [30]. Thus, the Mwk mouse might represent a relevant model for SCA14. Overexpression of SCA14-associated PKCγ mutations in developing Purkinje cells also disturbs their dendritic development [71], which is consistent with the Mwk developmental phenotype.

Recently, activation of mGluR1 was shown to also trigger gating of the postsynaptic GluD2 receptor [72]. In another study, GluD2 was found to associate with a mGluR1-TRPC3-PKCγ signaling complex and to regulate mGluR1-mediated synaptic transmission in cerebellar Purkinje neurons [73]. Interestingly, in this study, genetic loss of GluD2 was found to increase extra-synaptic mGluR1 and TRPC3 expression in the hotfoot mutant mice, resulting in abnormal synaptic transmission. Both loss-of-function and gain-of-function mutations in the Grid2 gene encoding GluD2 result in defects in Purkinje cell dendritic arborization and cerebellar ataxia [6]. Recently, the first mutations in the human GRID2 gene have been reported in patients with cerebellar ataxia [12, 13].

In addition to the direct links between disrupted glutamatergic signaling and cerebellar ataxia described above, increasing evidence points towards a role of perturbed TRPC3 signaling in other genetic forms of cerebellar ataxia. For example, the expression of TRPC3 and other genes involved in calcium homeostasis was shown to be downregulated in a mouse model of SCA1 [74]. Importantly, these expression changes occurred early and before any observed pathological changes in SCA1, suggesting that they might be a causal factor in the pathogenesis of SCA1. Consistent with these findings, mGluR1-TRPC3-mediated synaptic signaling is disrupted in homozygous staggerer mice, an extreme model of SCA1 [75]. Both expression of mGluR1 and TRPC3 were found to be dramatically reduced in these mice, and consequently, slow EPSCs were absent. Similarly, mGluR1-mediated synaptic transmission is completely absent in a transgenic mouse model of SCA3 [76].

Collectively, these studies underscore that TRPC3 is a central player in a glutamate-triggered signaling cascade that is vital for Purkinje cell function and, when disrupted, results in cerebellar disease in mice and men (Fig. 3). Moreover, these findings suggest that TRPC3 itself might be a promising candidate gene for human cerebellar ataxia. Supporting this, a methylation-regulating TRPC3 promoter polymorphism was found to be enriched in patients with idiopathic cerebellar ataxia [77]. So far, candidate screening of the TRPC3 gene did not identify potential mutations in patients with sporadic late-onset cerebellar or episodic ataxia, suggesting that TRPC3 mutations are unlikely to be a common cause of cerebellar ataxia [78]. However, TRPC3 mutations might still contribute to cerebellar ataxia in specific subtypes of the disease that have not been screened yet. The increasing use of next-generation sequencing approaches for the genetic diagnosis of cerebellar ataxia is poised to fully elucidate the role that TRPC3 mutations may play in human cerebellar ataxia [79].