Spinocerebellar Ataxia Type 1: Molecular Mechanisms of Neurodegeneration and Preclinical Studies
Spinocerebellar ataxia type 1 (SCA1) is an adult-onset, inherited disease that leads to degeneration of Purkinje cells of the cerebellum and culminates in death 10–30 years after disease onset. SCA1 is caused by a CAG repeat mutation in the ATXN1 gene, encoding the ATXN1 protein with an abnormally expanded polyglutamine tract. As neurodegeneration progresses, other brain regions become involved and contribute to cognitive deficits as well as problems with speech, swallowing, and control of breathing. The fundamental basis of pathology is an aberration in the normal function of Purkinje cells affecting regulation of gene transcription and RNA splicing. Glutamine-expanded ATXN1 is highly stable and more resistant to degradation. Moreover, phosphorylation at S776 in ATXN1 is a post-translational modification known to influence protein levels. SCA1 remains an untreatable disease managed only by palliative care. Preclinical studies are founded on the principle that mutant protein load is toxic and attenuating ATXN1 protein levels can alleviate disease. Two approaches being pursued are targeting gene expression or protein levels. Viral delivery of miRNAs harnesses the RNAi pathway to destroy ATXN1 mRNA. This approach shows promise in mouse models of disease. At the protein level, kinase inhibitors that block ATXN1-S776 phosphorylation may lead to therapeutic clearance of unphosphorylated ATXN1.
KeywordsNeurodegeneration Polyglutamine Therapeutic approaches
Spinocerebellar ataxia type 1 (SCA1) is a fatal neurodegenerative disorder with an autosomal dominant inheritance. The genetic basis of SCA1 is a CAG repeat expansion mutation in the coding region of the Ataxin-1 (ATXN1) gene [1, 2]. As with other trinucleotide repeat disorders, SCA1 shows genetic anticipation where the unstable repeats expand over generations, and disease onset occurs earlier and is more severe. Unaffected individuals harbor 19–36 CAG repeats, with CAT interruptions in tracts over 21 repeats. SCA1 patients have 43–81 pure CAG repeats . The gene product, ATXN1, is a predominantly nuclear protein that harbors the polyglutamine expansion mutation in its amino-terminal region. While ATXN1 is widely expressed in the brain, cerebellar Purkinje cells present more vulnerability to mutant ATXN1 with their degeneration underlying the ataxic phenotype. As disease progresses, brain stem nuclei become involved and affected individuals succumb due to chronic lung infections and respiratory failure.
6.1 Molecular Mechanisms of Neurodegeneration
Mutant ATXN1 toxicity is not solely dictated by the polyglutamine tract mutation. The C-terminal region of ATXN1 harbors several motifs that are required for the protein to be pathogenic. The nuclear localization signal (NLS) directs ATXN1 translocation into the nucleus. Mice expressing mutant ATXN1[82Q] with a nonfunctional NLS do not develop disease, which proves ATXN1 exerts its toxicity in the nuclear compartment . Once inside the nucleus, ATXN1 interacts with cellular proteins involved in regulation of gene expression, both at the level of transcription and RNA processing. The ATXN1 AXH domain allows its homodimerization and interaction with regulators of transcription such as the transcriptional repressor Capicua (Cic) . ATXN1 and Cic together exist in large protein complexes where mutant ATXN1 is thought to affect Cic repressor activity . In addition to regulating Cic function, stability of Cic depends on ATXN1 binding. ATXN1 immunodepletion results in loss of Cic and Sca1 null mice have reduced Cic levels.
A knockin SCA1 mouse model that expresses 154Q at the endogenous Sca1 mouse locus (Sca1 154Q/2Q ) develops an ataxia phenotype as well as other features of the human disease, namely wasting and brainstem degeneration-dependent lethality . Sca1 154Q/2Q mice having partial (50%) genetic ablation of Cic show improvement in motor coordination, weight, survival, and neuropathology. Microarray analysis revealed that many Cic target genes that were repressed in Sca1 154Q/2Q were restored to near wild type levels with Cic+/−. This finding argues for Atxn1 toxicity acting through a gain of function mechanism, whereby mutant Atxn1 affects the normal function of an Atxn1 binding partner (Cic) and exacerbates its normal gene-repressor function in vivo. This is consistent with previous in vitro Cic-mediated transcription assay that show Atxn1 enhances Cic repressor activity and this requires Atxn1-Cic interaction . Interestingly, genes normally repressed by Cic were overexpressed in Sca1 154Q/2Q and partial Cic ablation restored their levels. Thus, mutant Atxn1 also results in Cic loss of function, revealing complexity in Atxn1 effects. Partial ablation of Cic in Sca1154Q/2Q resulted in an improvement in disease, suggesting Atxn1 exerts its toxicity via Cic gain of function and hyper-repression of Cic target genes. Relieving Atxn1-Cic interactions may prove beneficial. Interestingly, nonpathogenic ATXN1[82Q]-S776A does not associate with large protein complexes containing Cic . These results indicate that mutant Atxn1 pathogenicity acts through aberrant interactions with its native binding partners and not via novel interactions.
Another key site at the C-terminus of ATXN1 is the serine at amino acid position 776 (S776). A large body of evidence supports that S776 phosphorylation is critical for ATXN1 toxicity. While SCA1 mice expressing wild type ATXN1[30Q] in Purkinje cells are not affected, ATXN1[30Q]-S776D phosphomimetic mutation drives disease. Furthermore, D776 mutation enhances toxicity in ATXN1[82Q]-D776 mice . Notably, mice expressing ATXN1[82Q]-S776A phospho-resistant mutation do not develop ataxia and largely resemble wild type mice, despite harboring a mutant polyQ tract . What does S776 phosphorylation do to ATXN1 that changes it so drastically? Two main properties S776 phosphorylation confers to ATXN1 are greater stability and enhanced protein-protein interactions. Abnormally elevated levels of ATXN1, even in the wild type form, are pathogenic when overexpressed in both Drosophila and mice, suggesting ATXN1 protein load itself is toxic [10, 11]. The nonpathogenic ATXN1-A776 protein is unstable in vivo and its lowered levels are thought to be an important factor contributing to averting toxicity .
One mechanism by which phosphorylated pS776-ATXN1 is stabilized is by interaction with the molecular chaperone 14-3-3. Co-expression of ATXN1[82Q] and d14-3-3ε in Drosophila retina enhances 82Q-induced eye degeneration . In mice, 14-3-3ε haploinsufficiency (Sca1 154Q/2Q ; 14-3-3ε +/− ) rescues motor phenotype and PC numbers . In vivo 14-3-3ε haploinsufficiency results in lower Atxn1 levels in the cerebellum ; in vitro using a peptide to competitively disrupts 14-3-3/ATXN11 interaction or siRNA-mediated knockdown of 14-3-3 prompts ATXN1-S776 dephosphorylation and increased clearance of ATXN1 . While 14-3-3/pS776-ATXN1 binding protects ATXN1 clearance by blocking dephosphorylation, it seems that the phosphorylation itself is what makes ATXN1 more stable: ATXN1-D776 does not bind 14-3-3 and is as stable as phosphorylated ATXN1, suggesting the phosphorylation alone stabilizes the protein . Further studies exploring 14-3-3ε/pS776-ATXN1 interactions showed that perhaps this interaction is unique to the cerebellum. Although 14-3-3ε +/− results in lowered Atxn1 (154Q and 2Q) levels in the cerebellum, this was not the case in the brainstem. Whether a different 14-3-3 isoform mediates this stabilizing interaction in the brainstem remains unexplored. Perhaps 14-3-3ε +/− does not affect Atxn1 levels in the brainstem as it does in the cerebellum due to distinct protein interactions in these different brain regions. Yet, biochemical studies showed a shift in the incorporation of mutant Atxn1 from large (toxic) to small protein complexes in the cerebellum, but this was not observed in the brainstem. Protein complex composition in the brainstem differs from that in the cerebellum. This may explain why motor deficits related to the cerebellum were rescued in Sca1 154Q/2Q ; 14-3-3ε +/− mice, even though these mice continued to display brainstem-related deficits, including weight loss, premature death, and respiratory dysfunction. Differences in Atxn1-binding complexes reveal potentially diverging mechanisms by which mutant Atxn1 exerts toxicity in different regions of the brain.
In support of this pathogenic model are studies using SCA1 animal models, which consistently implicate aberrations in gene expression as an early part of the disease process. PCR-based cDNA subtractive hybridization first revealed abnormal gene expression in transgenic SCA1 animals . A set of neuronal genes highly expressed by Purkinje cells were downregulated in SCA1 cerebella, as early as P11, just one day after transgene expression and weeks before histological or behavioral deficits are detectable. These changes were also reflected with immunostaining of SCA1 patient sections. Over the years, studies have focused in groups of genes affected in disease Ca2+ handling/homeostasis , glutamate signaling [20, 21], dopamine receptor , VEGF angiogenic factor , and synaptic proteins . Early expression of mutant ATXN1 contributes to disease at least in part by affecting RORα-mediated gene expression important for Purkinje cell maturation and cerebellar development . That disease onset begins by affecting development is a possibility suggested by several SCA1 mouse models [25, 26].
Recent studies show disease onset and progression are distinct processes, best illustrated by transgenic mice expressing ATXN1[30Q]-D776 in Purkinje cells. The S776D mutation confers a phospho-mimicking change that converts nonpathogenic ATXN1[30Q] to a toxic species. ATXN1[30Q]-D776 mice develop an ataxia that closely resembles that seen in ATXN1[82Q] transgenic mice, but their disease does not progress to Purkinje cell death . Recent RNAseq analysis of cerebellar tissue at early, mid, and late stage of disease was performed using ATXN1[82Q] mice (with a progressive disease) and ATXN1[30Q]-D776 mice (having a nonprogressive disease) to uncover pathways that underlie progressive disease as well as those that might prevent disease progression . Weighted Gene Coexpression Network Analysis (WGCNA) of the RNAseq data across all ages and genotypes identified two modules, designated Magenta and Lt Yellow, as being associated with ataxia. Interestingly, the Magenta module is enriched for genes specifically expressed by Purkinje cells and a significant proportion have upstream Cic binding sites. This is in line with established role of aberrations in Cic target gene expression in SCA1 models discussed above. Moreover, the course of gene changes in the Magenta module track with the progressive phenotypic changes that occur in ATXN1[82Q] disease.
Analysis of RNAseq data further revealed cholecystokinin (Cck) as being uniquely overexpressed in ATXN1[30Q]-D776 and downregulated in ATXN1[82Q] cerebellar RNA . To test whether Cck upregulation protects ATXN1[30Q]-D776 from disease progression, ATXN1[30Q]-D776 mice were crossed to Cck −/− mice, producing ATXN1[30Q]-D776; Cck −/− mice. Loss of Cck resulted in a progressive disease with molecular layer atrophy and Purkinje cell loss comparable to that seen in ATXN1[82Q] model. This raises the possibility that Cck overexpression in the ATXN1[30Q]-D776 model activates a protective pathway that prevents progression of disease. The activation is thought to be via autocrine activation of Purkinje cell CckR1 since ATXN1[30Q]-D776; CckR1 −/− (but not ATXN1[30Q]-D776; CckR2 −/− ) mice demonstrated a similar progressive pathology.
6.2 Preclinical Studies
To date, two general strategies are being employed to mitigate SCA1 pathogenesis preclinically in animal models. One strategy is to restore expression of a gene product downregulated in SCA1, This was first tested by genetic and pharmacologic treatments of Sca1 154Q/2Q mice with VEGF . Sca1 154Q/2Q cerebella present with a dramatic downregulation of Vegfa mRNA, likely due to increased occupancy by Atxn1 of the Vegfa promoter and consequent gene repression, in a histone acetylation dependent manner. VEGF is an angiogenic and neurotrophic factor. In the cerebellum it is expressed by neurons, glia, and endothelial cells [28, 29]. Sca1 154Q/2Q mice crossed to transgenic mice that overexpress VEGF from an embryonic stage showed improved rotarod motor performance in adulthood, with apparent full motor recovery by 6 months. This was coupled to only a slight but statistically significant improvement in molecular layer thickness of the cerebellar cortex. To test the therapeutic potential of exogenous VEGF delivery, intracerebroventricular administration of recombinant mouse VEGF was continuously delivered by osmotic pump into the lateral ventricles for two weeks starting at 11 weeks of age. This short period of treatment produced a complete recovery in rotarod motor performance in treated animals. It is unclear whether VEGF benefits come from the effects on microvasculature improving nutrient and oxygen delivery to the brain parenchyma and/or activation of neurotrophic pathways. Regardless, replenishing levels of a gene product downregulated in disease attenuates mutant Atxn1 toxicity.
A more recent study sought to reinstate expression of another cellular component deficient in Sca1 154Q/2Q cerebella . Mass-spectrometry analysis of Sca1 154Q/2Q cerebella revealed a reduction in Homer-3 levels in synapses, likely caused by defective mTOR signaling. Blocking mTORC1 signaling depleted Homer-3 levels and worsened Sca1 154Q/2Q pathology. While enhanced neurotransmission and activation of mTORC1 signaling enriched expression of the Purkinje cell scaffold protein Homer-3 in wild type cerebella, it failed to do so in Sca1 154Q/2Q . Thus, in order to improve Homer-3 synaptic levels, AAV-Homer-3 was injected into the ventricles of neonate brains. Behavioral examination at P40–P200 showed incomplete but significant persistent improvement in rotarod motor performance. Other parameters improved with AAV-Homer-3 gene treatment included Purkinje cell spine density, Purkinje cell numbers, and vesicle pool occupancy. Thus, replacement of Homer-3 levels during development restores synaptic deficits and partially improve motor performance in Sca1 154Q/2Q mice.
Perhaps the most promising preclinical therapeutic approaches under investigation are predicated on the finding that reducing the amount of mutant protein expressed in the cerebellum can alleviate its toxicity. Conditional SCA1 mice were generated to manipulate ATXN1[82Q] expression in Purkinje cells by Doxycycline administration to turn off ATXN1[82Q] expression and reduce mutant protein levels. Strikingly, this results in reversal of the motor deficits and histopathology. Moreover, the earlier the gene is turned off, the more pronounced disease reversal . These findings reveal two important points. First, progression of this inherited neurodegenerative disease can be reversed. Early intervention is key to preserve and recover the remaining tissue. If the disease is targeted early enough the cells are still present and able to recover. In fact, later studies have shown that there are differential circuit-specific changes that progress with the course of disease and precede histopathology . Second, there is a therapeutic opportunity in targeting the mutant protein.
The RNAi pathway is a robust gene silencing mechanism through which mRNAs are cleaved and destroyed in the cytoplasm. In recent years, studies have taken advantage of this powerful tool to reduce mutant Atxn1 expression and test the therapeutic potential of this approach (Fig. 6.2). The first series of experiments focused on viral-mediated delivery of RNAi species targeting transgenic ATXN1[82Q] selectively overexpressed in murine Purkinje cells. AAV-shRNA injection was targeted to midline cerebellar lobules for expression in Purkinje cells. AAV-shSCA1 delivery at 7 weeks of age achieved complete preservation of molecular layer thickness (examined at 16 weeks of age) and partial improvement in rotarod motor performance (as late as 21 weeks of age) . Advances in RNAi technology led to development of the more refined use of endogenous miRNA backbones for in vivo expression of mature siRNA sequences [33, 34]. Benefits include improved siRNA processing with lower off-target potential, as well as greater safety profile. Indeed, delivery of AAV-miRNA targeting hATXN1[82Q] in SCA1 transgenic mice (AAV-miSCA1) achieved robust (70%) hATXN1 mRNA knockdown without toxicity . In these newer studies AAV-miSCA1 injections were targeted to deep cerebellar nuclei (DCN) for uptake and expression by Purkinje cells. With AAV injection at 5 weeks, motor recovery was observed out to 37 weeks of age, including ledge test, hindlimb clasping, and rotarod test for coordination. Histological measures of cerebellar integrity such as molecular layer thickness, total Purkinje cell numbers, and ectopic Purkinje cell numbers, were more modestly improved, consistent with the accumulating notion that full rescue of histopathological changes may not be required for complete recovery of motor performance.
More recently an upgraded AAV-miSCA1 was tested in the Sca1 154Q/2Q knockin model to examine its therapeutic effect beyond the Purkinje cell related pathology . As expected, this resulted in Atxn1 knockdown and fully rescued histopathology and motor impairments. A novel finding was that AAV-miSCA1 injection into the cerebellar DCN was able to transduce brainstem nuclei and knock down local Atxn1 levels. One strategic injection site (the DCN) could be sufficient to deliver the AAV-miSCA1 to reach the brain regions most critically affected in SCA1 (the cerebellum and brainstem). It will be important to see whether future studies demonstrate rescue in brainstem related phenotype, notably survival studies. Another important finding in this study is that a single surgical injection was sufficient to produce therapeutic benefit for as long as 40 weeks. Moreover, expression analysis of select genes affected in Sca1 154Q/2Q mice showed restoration of their expression levels comparable to wild type. Of note is the preservation of Vegfa levels, previously discussed.
Mutant ATXN1 toxicity is the primary insult in SCA1 disease. Downstream gene changes implicate diverse molecular pathways that contribute to the disease course. It is therefore not surprising that targeting these pathways incompletely rescue pathology. The most efficient therapeutic avenue for SCA1 is to target ATXN1 directly. Harnessing the RNAi pathway to knockdown ATXN1 gene expression is efficient, without measurable toxicity, and has long lasting therapeutic benefit. Another encouraging technology targeting mRNA degradation are ASOs (anti-sense oligonucleotides), which act via the RNAse-H-dependent pathway in the nucleus (Fig. 6.2). ASOs show therapeutic efficacy in treating disease-causing proteins, including Hungtintin and SOD1, in preclinical studies with clinical trials underway [37, 38]. Therapeutic ASOs could also be tested to target ATXN1 in SCA1 studies in the future.
SCA1 is a fatal trinucleotide repeat neurodegenerative disorder to which there is no cure. This chapter has covered how mouse models and biochemical studies have aided our understanding of the molecular mechanisms underlying this disease. Chiefly, the fundamental basis of pathology is an aberration in the normal function of Purkinje cells affecting regulation of gene transcription and RNA splicing. Factors that promote ATXN1 stabilization lead to abnormal protein load, toxic protein-protein interactions, and enhanced pathology. Preclinical studies are focused on alleviating protein burden by targeting mRNA levels. Other therapeutic opportunities underway ahead involve kinase inhibitors to block toxic S776 phosphorylation to promote ATXN1 protein clearance.
- 17.Masuda A, Takeda JI, Ohno K (2016) FUS-mediated regulation of alternative RNA processing in neurons: insights from global transcriptome analysis. Wiley Interdiscip Rev RNA 2–5Google Scholar