| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Ataxic mouse mutants and molecular mechanisms of absence epilepsy
Introduction
Tottering (CACNA1A): Functional Diversity For The Main Subunit Of P/Q-Type CA2+ Channels
Lethargic (CACNB4): Mutation Induces Altered Usage Of Other Regulatory Subunits
STargazer (CACNG2): New Kid On The Block
Combining Subunits To Understand Swd Mechanisms?
The Next Steps
Acknowledgements
References
Ataxic mouse mutants and molecular mechanisms of absence epilepsy
Received June 7, 1999; Accepted July 13, 1999
Mouse genetic models for common human diseases have been studied for most of the 20th century. Although many polygenic strain differences and spontaneous single gene mutants have been extensively characterized over the years, knowing their innermost secrets ultimately requires the identity of the mutated genes. One group of neurological mutants, detected initially due to cerebellar dysfunction, was identified as models for epilepsy when they were unexpectedly found to have spike-wave seizures associated with behavioral arrest, a central feature of absence or petit-mal epilepsy. A further surprise was that recently identified defective genes encode different subunits of voltage-gated Ca2+ channels (VGCCs), implying common seizure mechanisms. In this review we first consider these spontaneous mutants with VGCC defects in the context of other mouse models for epilepsy. Then, from the new wave of genetic and functional studies of these mutants we discuss their prospects for yielding insight into the molecular mechanisms of epilepsy.
INTRODUCTION
Epilepsy is second to stroke as the most common neurological disorder, afflicting >1% of the population (1), and about half of the cases are thought to have a major genetic component. In the past decade, genes for several rare forms of epilepsy have been identified (1), including at least three for idiopathic epilepsy [benign familial neonatal convulsions-1 and -2 (2,3) and autosomal dominant nocturnal frontal lobe epilepsy (4)], each encoding ion channel subunits, potentially offering the kinds of cellular mechanism that are familiar to neurophysiologists. Despite this important progress, the gene defects associated with the most common human idiopathic epilepsies are completely unknown. Although optimistically the dedicated collection of families for linkage will continue to provide genes through positional cloning, the inheritance of human idiopathic epilepsy can be very complex (5) and may well preclude their identification using this approach. Murine idiopathic epilepsy models can also show complex inheritance (e.g. DBA/2 and EL). Such models, which have spontaneous seizures and/or low thresholds to induced seizures, are inbred strains (derived from non-epileptic progenitors) in which the balance of natural susceptibility and resistance alleles is shifted (either by chance or design) towards susceptibility (6-13). As in the human, causal gene defects in these genetically complex models have not yet been established.
The genetics revolution of the past decade has led to gene discovery of several spontaneous mouse models for epilepsy and has spun-off an increasing number of engineered mutants with seizures as a predominant phenotype. The spontaneous mutants are not ideal `models' for idiopathic epilepsy because they all have other neurological defects, but their genetic simplicity offers a good chance of correlating gene defect with phenotype. In the face of the complex genetics of human epilepsy, the mouse models provide candidate genes and mechanisms which can then be assessed directly in human sib pairs and sporadic cases to probe human gene defects and their downstream consequences. Together with the potential of facile identification of the defective proteins in many models, these developments give much hope that unprecedented insight into the molecular mechanisms of epilepsy is just around the corner.
Mutants with spontaneous convulsions plus other neurological conditions include jimpy, quaking, weaver, opisthotonus, jittery and a few others. The defective gene is known in at least four of these and in each case they have different functions. The complex etiology of epilepsy is even more evident from the study of engineered mutations, of which 2% have convulsive seizures as a prominent feature (for an earlier review see ref. 14). These genes encode everything from ion channels, such as Gabrb3 (Cl-) (15), Glur2 (Na+, Ca2+) (16) and Kcna1 (K+) (17), to neuromodulatory proteins involved in G-protein coupled receptor signaling, such as Htr2c (18), Npy (19), Plc1b (20) and Camk2a (21), to enzymes important for neurotransmitter metabolism, such as Gad2 (22) and Akp2 (23), to transcription factors, such as Otx1 (24), to proteins involved in basic cellular functions, such as PIMT (25). The frequency of seizing knockouts, which were, in general, randomly ascertained, suggests as many as 1000 seizure genes in mammals. Although this estimate may seem remarkable, it is consistent with the genetic heterogeneity of common epilepsy in humans (5).
The clinical relevance of a mouse model is difficult to assess without genetic evidence that it is at least partially homologous to human disorders. Currently investigators rely on phenotypic criteria, such as abnormal EEG discharges, associated behavioral abnormalities and profiles of pharmacological sensitivity. By these criteria several spontaneous single gene mutants that exhibit non-convulsive, generalized spike-wave discharges (SWDs) are considered models for human absence epilepsy. These mutants were detected initially because of moderate to severe ataxia and accordingly named ducky, lethargic, stargazer and tottering. Upon subsequent screening they were found to have 5-7 Hz SWD bursts associated with behavioral arrest which, importantly, respond to anti-absence drugs that are efficacious in human (26). Surprisingly, at least three of the four genes encode subunits of voltage-gated Ca2+ channels (VGCCs) (27-29). A more recent ataxic mutant, slow-wave epilepsy, exhibits both convulsions and SWD (30), but its 3 Hz SWD burst frequency is characteristically different from the other ataxia models and may more closely resemble common human absence epilepsies. Interestingly, the gene mutated in slow-wave epilepsy mice, Slc9a1, does not encode a VGCC but rather the ubiquitous Na+/H+ exchanger, Nhe1.
The 5-7 Hz SWD mutants remain an important class for further consideration. While hindsight reveals the striking ascertainment bias by which they were obtained (i.e. mutants with ataxia and 5-7 Hz SWD are selective for defects in VGCC subunits), this series of independent mutations nevertheless provides a rich resource for investigating SWD. First, because of the pleiotropy beyond the seizures these mutations can provide further insight into molecular mechanisms of SWD. Second, comparative examination of multiple mutations within one gene which give rise to different phenotypes could reveal a common pathological mechanism underlying SWDs. [Such correlative analysis has been successfully applied to mutations of the microphthalmia transcription factor; for example, only the dominant negative mutations give rise to the osteopetrosis phenotype (31).] Last, by contemporary genetic approaches (e.g. transgenics, knockouts and construction of double mutants) one can potentially test the hypothesis that independent mutations in different VGCC subunits (e.g. [alpha]1A, [beta]4 and [gamma]2) cause SWD through a common mechanism. In the remainder of this article we will review research on mouse epilepsy mutants of the VGCC/5-7 Hz SWD class and evaluate how comparative functional analyses are beginning to reveal seizure mechanisms.
TOTTERING (CACNA1A): FUNCTIONAL DIVERSITY FOR THE MAIN SUBUNIT OF P/Q-TYPE CA2+ CHANNELS
The tottering mouse was discovered as a spontaneous mutation at the Jackson Laboratory in 1957 (32) and was later found to encode the [alpha]1A pore-forming subunit of the P/Q-type VGCC: gene Cacna1a (27). This polypeptide contains four functional domains, each consisting of six transmembrane subunits, which cooperate to form the calcium pore. Expression of the protein is restricted to brain, with highest expression in cerebellum, hippocampus, cortex and olfactory bulb. The P/Q channel is known to have a significant role in fast neurotransmission, but is also expressed in cell bodies and dendrites. There are now several mouse alleles at Cacna1a, the best characterized of which include the original (Cacna1atg), leaner (Cacna1atg-la) and rolling Nagoya (Cacna1atg-rol). At least one additional spontaneous allele (Cacna1atg-4J) is presently under investigation, as is a gene-targeted null mutation. Although the gross behavioral phenotype of each allele is strikingly different, at least three of the four spontaneous alleles have absence seizure-like phenotypes in common.
Tottering (Cacna1atg) mice were originally detected based on mild to moderate cerebellar ataxia and intermittent episodes of so-called `motor seizures', which are not actually epilepsy but rather paroxysmal dyskinesis, a type of movement disorder. These episodes involve stereotyped, progressive involvement of all limbs that occurs over 20-30 min. Involvement of brainstem nuclei and cerebellum have been demonstrated (33,34). At the cellular level, histological examination has not detected any gross cytoarchitectural defects in the cerebellum or brainstem, although global hyperinnervation by noradrenergic fibers originating in the locus coeruleus has been documented (35). In addition, tyrosine hydroxylase (TH), a key enzyme in the noradrenergic biosynthesis pathway, is persistently expressed in tottering cerebellar Purkinje cells (36).
Until recently, it was not clear how these cerebellar phenotypes hint at SWD mechanisms that are more likely to be associated with thalamocortical dysfunction. However, with identification of the Cacna1atg mutation as an amino acid substitution (P->L) near the pore lining region of the second domain of [alpha]1A, assessment of its function in the cerebellum has begun and two hypotheses emerge. The first is based on electrophysiology and reveals significantly reduced current density in mutant Purkinje cells and minor effects on P/Q current kinetics, including changes in the voltage dependence of inactivation (37). Assuming that other neuronal types also have similar defects, aberrant P/Q-type channel function could directly account for the reduced evoked EPSCs seen in thalamic VB neurons oftottering mice (38). There is a precedent, at least in the cerebellum, for association of different VGCC currents with excitatory versus inhibitory synapses (39) and it is conceivable that the thalamic circuit, if physically intact, may be unbalanced if excitatory and inhibitory inputs are differentially affected by altered current density. The second hypothesis concerns the indirect effects of [alpha]1A mutation on the expression or assembly of other VGCC channels, which themselves may play a more proximate role in SWD generation. For tottering, although no major compensatory changes in other VGCC currents were detected electrophysiologically (37), a modest increase in expression of the [alpha]1C subunit of L-type channels was reported in Purkinje cells (40). Interestingly, this overexpression seems to have causal implications for at least the paroxysmal dyskinesis: induction of movement disorder is prevented by treatment with antagonists selective for L-type channels (40). These results suggest the possibility that [alpha]1A defects indirectly influence the expression and function of other VGCCs, such as [alpha]1C, accounting for SWD by an unknown mechanism(s). Neither hypothesis is mutually exclusive and perhaps some combination of both direct and indirect effects is required for SWD.
Structural and functional data from the growing Cacna1a allelic series contribute to both hypotheses. Homozygous Cacna1atg-la (leaner) mice, which have SWD, have severe ataxia and slowly progressive cerebellar degeneration. The molecular defect is a splice donor mutation that causes truncation of the open reading frame and deletion of the C-terminus of the protein. Purkinje cells of leaner mice also show reduced current density and changes in voltage dependence of activation and inactivation (37). It has also been shown that the region deleted in leaner encodes a calmodulin-binding domain that apparently moderates regulation of the [alpha]1A current by calcium (41).Homozygosity for Cacna1atg-rol confers an intermediate ataxic phenotype and may cause hypoplasia of the anterior cerebellum (42), although there are conflicting reports in the literature (43). Electrophysiological analysis has not yet been reported for Cacna1atg-rol. Although data on these mice have been published, they have not been made widely available to researchers and therefore direct phenotypic comparison with other alleles is difficult. Cacna1atg-4J homozygous mice have moderate cerebellar ataxia and paroxysmal dyskinesia, but no obvious signs of cell death. Cacna1atg-4J encodes a minor amino acid substitution in the pore-lining region of the second domain, not far from the Cacna1atg mutation (W.N. Frankel, D. Hosford and C. Lutz, unpublished data). Last, a null allele of [alpha]1A has been engineered (C.F. Fletcher, N.A. Jenkins, N.G. Copeland and L. Tessarollo, unpublished data). The mice are severely ataxic and often do not survive past weaning. Heterozygous mice have reduced protein levels but exhibit normal behavior and do not have SWD. These results suggest that a partial reduction in [alpha]1A levels is not, in itself, sufficient to cause SWD.
In summary, the tottering allelic series shows qualitative and quantitative diversity of phenotype, commensurate with the central role and complexity of [alpha]1 subunits in VGCC function. Continued collection of new alleles, coupled with phenotypic screens for SWD, should be of great value. It is also relevant to mention that there are several dominant human alleles of [alpha]1A (reviewed in ref. 44) which result in ataxia or migraines, but not absence epilepsy. It remains to be determined whether the 5-7 Hz SWD phenotype is predominantly associated with recessive alleles or if its expression is species specific.
LETHARGIC (CACNB4): MUTATION INDUCES ALTERED USAGE OF OTHER REGULATORY SUBUNITS
Thelethargic mouse mutant (gene Cacnb4lh) contains a mutation in the [beta]4 VGCC subunit (28). The four [beta] subunits are cytoplasmic proteins that bind to [alpha]1 proteins, together with additional regulatory subunits [alpha]2[delta] and [gamma]. The [beta] subunits affect the kinetics and membrane targeting of [alpha]1 proteins. [beta]4 is widely expressed in brain, spinal cord, kidney, gonads and skeletal muscle and is associated with L-, N-, P/Q- and R-type VGCCs. The Cacnb4lh splice donor mutation destabilizes the mRNA, causes exon skipping and a translational frameshift and was predicted to produce a functionally null peptide (28). The lack of [beta]4 protein in Cacnb4lh mice was demonstrated subsequently (45).
The behavioral phenotype of lethargic mice is very similar to that of tottering, with the same degree of ataxia, overexpression of TH and paroxysmal dyskinesis as well as 5-7 Hz SWD. The Cacnb4lh mutation might be expected to have more widespread phenotypic consequences for VGCC function because [beta]4 is not exclusively associated with any particular [alpha]1 subunit. Although peripheral motor nerves show reduced conduction velocity and prolonged distal latency (46), no pathological changes are found in the central nervous system (CNS) or in skeletal muscle, perhaps due to compensation by other [beta] subunits. In fact, at least two different reports of altered [beta] subunit expression and association with VGCC [alpha]1 subunits in Cacnb4lh mice have been published (45,47). Both describe increased [beta]1 expression, but differ in their assessment of effects on the [beta]2 and [beta]3 subunits as well as [alpha]1B, the major determinant of N-type channels. Despite this compelling evidence for some fashion of altered [beta] subunit usage in Cacnb4lh mice and the suggestion that these alterations represent adult persistence of channel types or mechanisms of VGCC assembly typically found in early development (45,47), major changes in P/Q currents were not detected in Cacnb4lh Purkinje cells (47). Smaller effects and a direct assessment of the consequences of compensatory subunit expression may require analysis of large numbers of cells in specific neuronal populations. Therefore, although there is some evidence for developmental abnormalities in Cacnb4lh mice, perhaps pointing to an `indirect' mechanism for generation of SWD, it is still possible that acute alteration of P/Q-type VGCC function is responsible.
STARGAZER (CACNG2): NEW KID ON THE BLOCK
Thestargazer locus encodes a novel protein related to the skeletal muscle-specific VGCC [gamma]1 subunit, about which little is known, and was proposed as the first neuronal [gamma] subunit, [gamma]2 (gene Cacng2) (29). Expression of Cacng2 is restricted to, but widespread within, the CNS. The Cacng2stg mutation, a retrotransposon insertion in the second intron, results in a vast reduction in normal mRNA, although the small amount remaining suggests that the mutation is not null. Co-expression of Cacng2 in BHK cells with other VGCC subunits causes a negative shift in the voltage dependence of [alpha]1A channel inactivation, reminiscent of the original [gamma]1 subunit. Predictions from these initial in vitro observations have not yet been tested in cells from Cacng2stg mutants.
Like tottering and lethargic, stargazer has 5-7 Hz SWD, but was detected initially for overt defects including cerebellar ataxia (without paroxysmal dyskinesis) and vertical head-tossing (hence `stargazer'), possibly due to inner ear defects. The cerebellar defects of stargazer and waggler mice have been characterized further. These alleles have reduced expression of the neurotrophic factor BDNF (confined to the cerebellum only) and defective post-translational modification of several signal transduction molecules in the TrkB pathway (48,49). In addition, the stargazer cerebellar granule cell has an immature GABAA receptor profile. These molecular abnormalities are associated with delayed maturation of the cerebellum, including persistence of immature granule cells in adult mutants and consequently a severe impairment in the acquisition of classical eyeblink conditioning. In contrast, these aspects of cerebellar function appear to be relatively normal in tottering mice (49), although it may be relevant that GABAA receptor expression is altered in the tottering cortex (50). The relationship between these events and SWD is unclear, except to speculate that for a similar delay of maturation, synaptic transmission may be impaired in thalamocortical neurons.
In addition to the original stargazer and waggler alleles, a third allele, Cacng2stg-3J, recently arose at the Jackson Laboratory and its mutation was characterized (V. Letts, W.N. Frankel, B. Harris and K. Johnson, unpublished data). The effect of the mutation is similar to the others: reduced wild-type mRNA. Interestingly, despite the similar mechanism, the allelic range of phenotype is surprisingly wide. Waggler mice, while showing moderate to severe cerebellar ataxia and the same types of cerebellar abnormality as stargazer, do not have the same degree of head-tossing and a much lower incidence of SWD. Gross behavioral abnormalities of Cacng2stg-3J homozygotes are milder still. Although more data are needed, e.g. comparing the effects of these mutations on the same genetic background, careful quantitation of Cacng2 mRNA and assessment of SWD in Cacng2stg-3J, the observations suggest that the range of variation in the stargazer allelic series represents a threshold effect.
COMBINING SUBUNITS TO UNDERSTAND SWD MECHANISMS?
The striking observation that at least three of the four published 5-7 Hz SWD models have gene defects in VGCC subunits suggests that there is a discrete, unifying mechanism between them. (Although there are phenotypic differences between mutants, some varying functional redundancy is to be expected based on different tissue distributions of mutated subunits and their relatives.) Indeed, from both tottering and lethargic mice there is sufficient evidence for a hypothesis that invokes a direct effect of impaired channel function on phenotypic outcome. In addition to the results described above, both mutants have similar effects on EPSCs in thalamocortical VB neurons (38) and reduced Ca2+ uptake in thalamic and cortical synaptosomes (51; D. Hosford, personal communication). It is therefore conceivable that they both act through a common biophysical mechanism, for example, decreased P/Q current or poor regulation of channel kinetics. However, it is equally likely that both mutant SWD phenotypes could occur by an indirect mechanism, e.g. decreased P/Q current altering the fractional contribution of N-type current. Indeed, on considering the secondary mechanisms, significant further caveats requiring further explanation begin to creep in, especially given that more distal events have been observed, such as derangement of neuronal differentiation and maturation. In fact, inasmuch as the SWD phenotype is determined as much by circuit architecture as by molecular composition, the apparent 5-7 Hz specificity may merely reflect the cellular and synaptic architecture of the circuit. The subtle SWD differences amongst mutants may be relevant here: stargazer has longer and more frequent bursts and lethargic is uniquely sensitive to the anti-absence effects of the GABAB receptor antagonist CGP35348 (52,53). Furthermore, ablation of NE fibers by 6-OHDA treatment ameliorates seizures in tottering but not in stargazer (54). Therefore, it is conceivable that each primary mutation may have an `acute' effect on the function of molecules in the circuit which at the same time synergizes with a distal malformation (e.g. altered cellular composition) to cause SWDs. It could ultimately prove to be very complicated.
THE NEXT STEPS
Having identified the molecular defects, it should be possible to characterize the primary neuropathic consequences of VGCC subunit mutations in the thalamocortical circuit. However, additional fundamental characterization may be needed to fully comprehend the results. For example, the complex spatial and temporal expression patterns of this large gene family suggest that different compensation mechanisms occur in different sites, requiring a more detailed characterization of VGCC subunit expression. In addition, the likely existence of additional subunit genes needs to be pursued.
Together with this information, contemporary genetic techniques can be employed to perturb the expression of VGCC genes and test various hypotheses for the generation of SWD (Fig. 1). Engineered null alleles can test whether residual function or loss of function underlies SWD. Null and transgenic overexpressing lines can investigate whether threshold/relative expression levels are important. Inducible conditional mutations can give a true estimate of the time between expression of the mutation and the onset of SWD, for example, to distinguish whether SWDs are a primary result of mutation or arise from distal events and to determine whether there is a developmental window in which the mutations act. Tissue-specific conditional mutations can identify the relevant neural structures. The classical test for epistasis, by construction of double mutants, can be employed to test the hypothesis of a common mechanism. For example, a tottering/stargazer double mutant strain could be examined by EEG recording to determine whether one locus dominated the SWD phenotype or whether the mutations had a synergistic effect. Finally, these efforts should not only include [alpha]1A, [beta]4 and [gamma]2 but also the other VGCC genes, in particular the newly cloned T-type channels, which have been predicted to play a key role in SWD.
Figure 1. Models for generation of 5-7 Hz SWDs in ataxic VGCC mutants.
In the end, even those of us who are enamored of the VGCC mutants must admit that the most appropriate mouse models for human epilepsy do not yet exist and will likely come from human epilepsy mutations knocked-in to mice and phenotype-driven mutagenesis screens (55-57). The ultimate aim of refining the diagnosis of epileptic disorders to the molecular level offers the promise of developing more effective treatments for this devastating disease.
ACKNOWLEDGEMENTS
We thank Verity Letts and Gregory Cox for comments on this manuscript and David Hosford for past discussions. W.N.F. was supported by grants from the National Institutes of Health. C.F.F. was supported by the National Cancer Institute, DHHS, under contract with ABL.
REFERENCES
+To whom correspondence should be addressed. Tel: +1 207 288 6354; Fax: +1 207 288 6077; Email: wnf{at}jax.org
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Batista-Brito, R. Machold, C. Klein, and G. Fishell Gene Expression in Cortical Interneuron Precursors is Prescient of their Mature Function Cereb Cortex, February 3, 2008; (2008) bhm258v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Imbrici, S. L. Jaffe, L. H. Eunson, N. P. Davies, C. Herd, R. Robertson, D. M. Kullmann, and M. G. Hanna Dysfunction of the brain calcium channel CaV2.1 in absence epilepsy and episodic ataxia Brain, December 1, 2004; 127(12): 2682 - 2692. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Kullmann The neuronal channelopathies Brain, June 1, 2002; 125(6): 1177 - 1195. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Pfaff Precision in mouse behavior genetics PNAS, May 3, 2001; (2001) 101128598. [Full Text]
|
|
|
|
 |
|
 A. J. Hunter, P. M. Nolan, and S. D.M. Brown Towards new models of disease and physiology in the neurosciences: the role of induced and naturally occurring mutations Hum. Mol. Genet., April 1, 2000; 9(6): 893 - 900. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Pfaff Precision in mouse behavior genetics PNAS, May 22, 2001; 98(11): 5957 - 5960. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||