Spinocerebellar ataxia type 1 (SCA1) is a dominantly inherited neurodegenerative disease caused by expansion of a CAG repeat that encodes a polyglutamine tract in ATAXIN1 (ATXN1). Molecular and genetic data indicate that SCA1 is mainly caused by a gain-of-function mechanism. However, deletion of wild-type ATXN1 enhances SCA1 pathogenesis, whereas increased levels of an evolutionarily conserved paralog of ATXN1, Ataxin 1-Like, ameliorate it. These data suggest that a partial loss of ATXN1 function contributes to SCA1. To address this possibility, we set out to determine if the SCA1 disease model (Atxn1154Q/+ mice) and the loss of Atxn1 function model (Atxn1−/− mice) share molecular changes that could potentially contribute to SCA1 pathogenesis. To identify transcriptional changes that might result from loss of function of ATXN1 in SCA1, we performed gene expression microarray studies on cerebellar RNA from Atxn1−/− and Atxn1154Q/+ cerebella and uncovered shared gene expression changes. We further show that mild overexpression of Ataxin-1-Like rescues several of the molecular and behavioral defects in Atxn1−/− mice. These results support a model in which Ataxin 1-Like overexpression represses SCA1 pathogenesis by compensating for a partial loss of function of Atxn1. Altogether, these data provide evidence that partial loss of Atxn1 function contributes to SCA1 pathogenesis and raise the possibility that loss-of-function mechanisms contribute to other dominantly inherited neurodegenerative diseases.
Spinocerebellar Ataxia type 1 (SCA1) is one of nine neurodegenerative diseases caused by an increase in the number of the amino acid glutamine in their respective proteins. Genetic studies have pointed to the fact that the glutamine expansion in Ataxin-1 causes SCA1 by causing Ataxin-1 to gain some function(s). Here, we demonstrate that in addition to the toxic gain-of-function mechanism, partial loss of the normal functions of Ataxin-1 contributes to SCA1. Ataxin-1 forms protein complexes with Capicua, a protein that silences expression of other genes, and we found that in SCA1 mouse models the levels of these complexes are reduced, resulting in increased expression of some genes. We also demonstrate that increased levels of Ataxin-1-Like, a protein that is similar to Ataxin-1 and protects against mutant Ataxin-1 in mice, rescues molecular and behavioral defects in mice deficient in Ataxin-1. These results show that Ataxin-1-Like compensates for loss of Ataxin-1 and that Ataxin-1 and Ataxin-1-Like share some normal functions. Together, these findings suggest that rescue of SCA1 symptoms by Ataxin-1-Like could be partly due to restoration of lost normal functions of Ataxin-1 in mice that express the mutant polyglutamine-expanded Ataxin-1.
Citation: Crespo-Barreto J, Fryer JD, Shaw CA, Orr HT, Zoghbi HY (2010) Partial Loss of Ataxin-1 Function Contributes to Transcriptional Dysregulation in Spinocerebellar Ataxia Type 1 Pathogenesis. PLoS Genet 6(7): e1001021. doi:10.1371/journal.pgen.1001021
Editor: Gregory S. Barsh, Stanford University School of Medicine, United States of America
Received: September 14, 2009; Accepted: June 8, 2010; Published: July 8, 2010
Copyright: © 2010 Crespo-Barreto et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by NIH/NINDS grants NS27699 (to HYZ), NS22920 and NS45667 (to HTO), and the pre-doctoral NIH/NRSA grant 1F31NS052925 (to JC-B). This work was also in part supported by IDDRC Award P30HD024064 from the Eunice Kennedy Shriver National Institute Of Child Health & Human Development. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Polyglutamine diseases are caused by the expansion of an unstable translated CAG repeats that encode a polyglutamine tract in unrelated proteins , . There are nine dominantly inherited neurodegenerative disorders caused by expanded polyglutamine tracts: Huntington's disease (HD), spinobulbar muscular atrophy (SBMA), dentatorubropallidoluysian atrophy (DRPLA), and six spinocerebellar ataxias (SCA1–3, 6,7 and 17) –. Several genetic studies have revealed that loss of the involved proteins in humans and mice does not cause neurodegeneration, leading to the conclusion that the polyglutamine expanded protein causes disease by a dominant gain-of-function mechanism whereby it confers toxic properties to the host proteins –. The importance of protein context and sub-cellular localization has been highlighted in SCA1 pathogenesis because expansion of the polyglutamine tract is necessary but not sufficient to cause neurodegeneration. For example, overexpression of polyglutamine-expanded ATXN1 that has a single serine residue mutated to alanine (S776A) does not lead to Purkinje cell degeneration, and overexpression of polyglutamine-expanded ATXN1 lacking a functional nuclear localization signal or lacking the AXH domain is not toxic in mice –. These data revealed key domains in ATXN1 that are critical for SCA1 pathogenesis, and indicated that mutant ATXN1 must be localized in the nucleus to exert toxicity. Furthermore, these data suggest that perhaps normal interactions or functions of ATXN1 are relevant to SCA1 pathogenesis.
Several protein interactors of ATXN1 have been identified to date. Among these, there are various transcriptional regulators, including the Capicua homolog CIC, SMRTER, HDAC3, GFI-1 and RORα. Some of these factors modify the pathogenesis of SCA1 in mice and fly models , –. For instance, Rorα haploinsufficiency results in enhanced pathogenesis in SCA1 transgenic mice . Furthermore, SCA1 transgenic mice share common gene expression changes with the staggerer mice, which have a spontaneous mutation in the Rorα gene that leads to cerebellar defects and ataxia –.
Recent evidence shows that altered interactions of ATXN1 with its native partners contribute to SCA1 pathogenesis. Studies in the knock-in mouse model of SCA1, Atxn1154Q/+, show that polyglutamine-expanded Atxn1 prefers the formation of a protein complex with the RNA splicing factor RBM17 while concomitantly diminishing the formation of Atxn1 complexes with CIC, a transcriptional repressor . These data suggest an endogenous role of ATXN1 in transcriptional regulation that might be altered in SCA1.
Even though much of the genetic evidence suggests that SCA1 is mainly caused by a gain-of-function mechanism, additional data also suggest that partial loss of Atxn1 function might contribute to pathogenesis. We demonstrated that removing the wild-type copy of Atxn1 in the knock-in mouse model of SCA1 (Atxn1154Q/−) leads to worsened SCA1 phenotypes . Moreover, phenotypes and neuropathology in Atxn1154Q/+mice are partially suppressed by the mild overexpression of an evolutionarily conserved paralog, Ataxin-1-Like (Atxn1L) , . Atxn1L shares high homology with Atxn1, however it lacks the polyglutamine tract, and it interacts with all tested proteins that bind to Atxn1 , , . Together, these studies indicate that altering the levels of wild-type Atxn1 and its paralog Atxn1L results in strong modulation of SCA1 phenotypes.
Taken together, these data led us to hypothesize that SCA1 pathogenesis results from a gain-of-function mechanism, with partial loss of Atxn1 function potentially contributing to the disease. Since Cic protein levels are also reduced in the Atxn1−/− mouse model , we predict that loss of functional Atxn1-Cic complexes could lead to common transcriptional defects in Atxn1−/− and Atxn1154Q/+ mice. Given the predicted role of Atxn1 in transcriptional regulation, we set out to test this possibility by performing microarray analyses on Atxn1−/− and Atxn1154Q/+cerebella. We show that Atxn1−/− and Atxn1154Q/+mice share many common transcriptional and molecular phenotypes, some potentially involving alterations in Atxn1-Cic-mediated transcriptional repression. Loss of Atxn1 function also results in several cerebellar transcriptional changes in common with mice lacking Rorα, another transcription factor that genetically and physically interacts with ATXN1. Finally, we demonstrate that overexpression of the Atxn1-related gene Atxn1L can rescue some of the molecular and behavioral defects caused by loss-of-function of Atxn1, strongly suggesting that Atxn1L-mediated suppression of SCA1 neuropathology could be due to restoration of the partial loss of Atxn1 function component in the disease.
Atxn1−/− mice share cerebellar transcriptional alterations with Atxn1154Q/+mice
To identify gene expression changes in SCA1 that might be due to partial loss of function of Atxn1, we surveyed transcriptional changes in Atxn1−/− and Atxn1154Q/+mouse cerebella using the Affymetrix mouse Exon Array ST 1.0 and searched for shared expression alterations. These arrays potentially enable the detection of even small fold changes due to the existence of multiple probes sets for most transcripts. Early symptomatic (7-week-old) Atxn1154Q/+mice were used to reflect early changes in SCA1 pathogenesis (Figure 1A). Due to a lack of overt phenotypes in 7-week-old Atxn1 null mice , 16-week-old Atxn1−/− mice were chosen in order to maximize the potential number of gene expression changes detected. Each mutant allele was studied and compared to age-matched wild-type littermates controls.
Figure 1. Comparison of transcriptional profiles of Atxn1154Q/+ and Atxn1−/− cerebella.
(A) Venn diagram reporting the number of significant gene expression changes in Atxn1−/− and Atxn1154Q/+cerebella. A total of 197 transcripts were significantly altered in both Atxn1−/− and Atxn1154Q/+mice, using a p-value of less than 0.01, and a minimal fold change of |±0.1| log2. (B) The majority of shared changes (68% of genes or 135 out of 197) went in the same direction.doi:10.1371/journal.pgen.1001021.g001
We observed a highly significant rate of concordance and overlap between cerebellar expression profiles of Atxn1−/− and Atxn1154Q/+ mice (z = 11.9485, P-value<2.2e-16, Kendall-τ test). Applying a very stringent cutoff value (false discovery rate (FDR)–corrected P<0.01, Fold change ≥|±0.1| log2), we identified 197 transcripts that are significantly dysregulated in both mouse models (Figure 1A). These 197 transcriptional changes account for 22.1% of all gene expression changes detected in Atxn1154Q/+cerebella (Figure 1A and Table S1). Remarkably, a majority of the differentially regulated transcripts (135 out of 197, 68.5%) are in the same direction in both mouse models. (Figure 1B). A preponderance of down-regulated genes in both models is observed (90 out of 197), followed by genes up-regulated in both models (45 out of 197), with only less than a third (62 out of 197) altered in an opposite direction (Figure 1B). These data suggest that at least 15.1% (135 out of 892) of the cerebellar transcriptional changes found in the SCA1 knock-in model (Atxn1154Q/+) could be attributed to loss of Atxn1 function.
To verify the microarray results, we selected 15 of these shared gene expression changes for independent validation by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) using cerebellar RNA samples from an independent set of 16-week-old Atxn1−/− and Atxn1154Q/+mice, and their respective littermates. Ten out of 15 gene expression changes, or 66.6%, were positively validated in Atxn1−/− mice, while 13 out of 15 genes, or 86.7%, were validated in Atxn1154Q/+cerebella (Table 1). Most importantly, of the 10 genes validated in Atxn1−/− cerebella, 9 were also validated in Atxn1154Q/+cerebella (Table 1). Taken together, the comparison of cerebellar microarray studies of Atxn1 null and polyglutamine-expanded Atxn1 knock-in mouse models demonstrate that a significant amount of transcriptional changes are shared between these models, supporting the notion that some loss of Atxn1 endogenous function contributes to disease.
Table 1. Real-time qRT–PCR validation of genes commonly altered in Atxn1154Q/+ and Atxn1−/− cerebella.doi:10.1371/journal.pgen.1001021.t001
In order to gain insight into the potential molecular pathways commonly affected in Atxn1−/− and Atxn1154Q/+ mice, we performed Gene Ontology (GO) analysis and pathway analysis based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) (Figures S1, S2, S3, S4, S5, S6, and Table S2). GO analysis revealed some biological functions that are enriched in commonly dysregulated genes. For example, cell junction and synapse, guanyl-nucleotide exchange factor activity and GTPase activity categories were enriched in the common set of up-regulated genes (Figures S1, S2, S3). In the down-regulated gene set, genes encoding for calcium ion binding were the most significantly affected both in Atxn1154Q/+ and Atxn1−/− cerebella (Figures S4, S5, S6). Using KEGG pathway analysis, among the top enriched categories for genes commonly down-regulated in both Atxn1−/− and Atxn1154Q/+ mice are the phosphatidylinositol and calcium signaling, Long Term Depression (LTD) associated genes, and Alzheimer's disease pathways (Table S2). These results strongly suggest that loss of Atxn1 results in transcriptional changes that are potentially pathogenic, since in addition to the enrichment for genes involved in neurodegenerative disease, the phosphatidylinositol and calcium signaling pathways are also known to be dysregulated in SCA1 models –. Interestingly, commonly up-regulated genes in Atxn1−/− and Atxn1154Q/+ mice include genes that are involved in cancer pathways (Table S2). This could potentially point to a novel function of Atxn1 that remains to be clarified in proliferating cells. The categories enriched for genes going in opposite directions between Atxn1−/− and Atxn1154Q/+ mice involve citrate cycle and ubiquitin-mediated proteolysis. These genes going in opposite directions are of interest, since they might reflect potential differences reflecting the main gain of function mechanism in SCA1 pathogenesis.
Atxn1−/− cerebella exhibit key pathological gene expression changes present in staggerer mice
Given the significant overlap between the transcriptional profiles of Atxn1−/− and Atxn1154Q/+cerebella, we were interested in examining whether these transcriptional changes are due to Atxn1 function being affected in both mouse lines and not simply due to cerebellar dysfunction. Previous studies showed that ATXN1 and RORα interact genetically and biochemically, and deficiency of Rorα enhances phenotypes in a transgenic model of SCA1 . Interestingly, we observed a significant overlap between the genetic profiles of Atxn1−/− mice, and those common between staggerer mice and SCA1 transgenic mice (Table 2) , . Most of these shared changes between staggerer and SCA1 mice are not present in Sca7266Q/+ cerebella, a knock-in mouse model for SCA7 , suggesting that most of these changes are not due to cerebellar dysfunction or polyglutamine disease in general (Table 2). We were able to confirm some of these changes by real-time quantitative RT-PCR in cerebella from 16-week-old Atxn1−/− mice, validating 4 out of 6 transcriptional changes tested (Figure S7). Given the known genetic and physical interaction between ATXN1 and RORα , these results could potentially indicate that Rorα -dependent transcriptional regulation is altered by loss of Atxn1 function. In contrast to SCA1 transgenic mice, it is important to note that Rorα levels are not changed in Atxn1−/− mice . Together, these findings suggest that loss of endogenous function of Atxn1 results in transcriptional changes that could potentially contribute to cerebellar pathogenesis.
Table 2. Rorα-responsive genes altered in SCA1 mouse models and Atxn1−/− mice.doi:10.1371/journal.pgen.1001021.t002
Atxn1 and Capicua associate at the promoter regions of genes altered in Atxn1−/− and Atxn1154Q/+ cerebella
ATXN1 forms stable complexes in vivo with the Capicua homolog (CIC), a transcriptional repressor that exhibits reduced levels in Atxn1154Q/+ mice , . To test the hypothesis that reduced Atxn1-Cic complexes lead to dysregulated gene expression in Atxn1154Q/+mice, we searched our microarray data for up-regulated genes that are direct targets of Cic. Microarray analysis revealed at least 3 significantly up-regulated genes (False Discovery Rate-corrected P<0.05) in Atxn1−/− cerebella that have been identified as direct targets of Cic-mediated repression, namely Etv1, Etv5 , and Ccnd1 (Fryer and Zoghbi, unpublished data). Interestingly, the microarray data show that Etv5 and Ccnd1 are also significantly up-regulated in Atxn1154Q/+mice cerebella. These results could reflect the fact that Cic protein levels are diminished both in Atxn1−/− and Atxn1154Q/+mice , .
Since Atxn1 and Cic form stable complexes in vivo, we rationalized that both proteins should bind the promoter regions of target genes if they mediate transcriptional repression together as a complex. To test this possibility, we performed chromatin immunoprecipition analysis, followed by PCR for the promoter regions of Etv5 and Ccnd1 (ChIP-PCR). ChIP-PCR analysis using antibodies against Cic confirmed that Cic is present on the promoters of Etv5 and Ccnd1 in vivo (Figure 2A). To determine if Atxn1 binds to the promoters of Cic targets and if the binding is altered due to polyglutamine expansion, we prepared cross-linked chromatin from mice expressing one wild-type Atxn1 allele (Atxn1+/−) and compared it to mice expressing one polyglutamine-expanded Atxn1 allele (Atxn1154Q/−). We used Atxn1−/− mice as a negative control for testing Atxn1 antibody specificity. As predicted, in chromatin extracts prepared from Atxn1+/− cerebella and immunoprecipitated using Atxn1 antibody, wild-type Atxn1 was detected on the promoters of Etv5 and Ccnd1 (Figure 2B). In contrast, we could not detect any specific signal for Atxn1[154Q] above background levels in Atxn1154Q/− cerebellar chromatin immunoprecipitations using Atxn1 antibodies (compared to Atxn1−/− and pre-immune sera controls) (Figure 2B). These data suggest that there is minimal association of polyglutamine-expanded Atxn1[154Q] to the promoters of target genes in vivo. Alternatively, it is possible that a conformational change in Atxn1[154Q] renders it inaccessible for immunoprecipitation, resulting in reduced signal.
Figure 2. Chromatin immunoprecipitation (ChIP) reveals co-occupancy at the promoters of Cic target genes by Atxn1 and Cic.
(A) ChIP using Cic antisera confirmed Cic binding at the promoter of two direct targets of Capicua, Ccnd1 and Etv5, that were up-regulated in the Atxn1−/− and Atxn1154Q/+ cerebella. (B) ChIP using Atxn1 anti-sera in cerebellar extracts from Atxn1+/−, Atxn1154Q/−, and Atxn1−/− mice reveals a signal in mice expressing only wild-type (Atxn1+/−) but not polyQ-Atxn1 (Atxn1154Q/−) compared to negative controls (pre-immune sera, and Atxn1−/−) (C) ChIP as in (B), this time using Cic antibody. In contrast to (B), Cic is present at comparable levels at the target promoters in Atxn1+/−, Atxn1154Q/−, and Atxn−/− mice. All ChIP assays were repeated three times on independent samples, representative results shown. (D) ChIP followed by quantitative PCR (ChIP-qPCR) on the promoter of Etv5 confirms that the proportion of immunoprecipitated DNA by Cic antibody is comparable in all three genotypes (as seen in C). A region containing two Capicua binding sites (promoter-CBS) of Etv5 is more enriched by Cic antibody than a region lacking CBSs (promoter-no CBS) compared to preimmune sera. (E) ChIP-qPCR on the promoter of Ccnd1 also shows similar enrichment of immunoprecipitated DNA by Cic antibody in all three genotypes (as seen in C). ChIP-qPCR using primers designed for a region within 100 bps of the CBS in the Ccnd1 promoter (which is highly conserved across species) show more Cic binding than primers designed for a poorly conserved region further downstream (~400 bps) of the CBS at the promoter of Ccnd1, compared to preimmune sera. ChIP-qPCR experiments in (D) and (E) were performed in triplicate on three independent sets of samples (3 cerebella per genotype) using SYBR Green Dye. N.S. = not significant, ** p<0.01, *** p<0.005.doi:10.1371/journal.pgen.1001021.g002
Given that Atxn1[154Q]-specific signal is not detected on the promoters of Etv5 and Ccnd1, we next asked if Cic protein can be detected at these promoters in Atxn1154Q/− mice using Cic antisera. We detected Cic binding to the promoter regions of Etv5 and Ccnd1 (Figure 2C) on all genotypes tested (Atxn1+/−, Atxn1154Q/− and Atxn1−/−), despite the fact that Cic protein levels are reduced in Atxn1−/− and Atxn1154Q/− mice , . To better quantify potential differences in Cic binding between the different genotypes, we performed ChIP followed by quantitative PCR. Primer sets designed for conserved regions in the promoters of Etv5 and Ccnd1 that contain or are adjacent to Cic binding sites (CBS) (TGAATGAA or TGAATGGA) were able to amplify with no significant difference in all three genotypes (Atxn1+/−, Atxn1154Q/− and Atxn1−/−), both for Etv5 and Ccnd1 (Figure 2D–2E). To test for the specificity of Cic-binding to its consensus sequences, primers were designed for regions lacking predicted Cic-binding sites sequences, either upstream or downstream of the CBS-containing regions in Etv5 and Ccnd1 promoters (Figure 2D–2E). As expected, quantitative PCR for these regions show less binding relative to the corresponding positive regions of Etv5 and Ccnd1 by quantitative PCR. These findings suggest that although Cic still binds the promoter, its function in repression is less efficient in the absence of wild-type Atxn1 or that Atxn1L partially compensates for the loss of Atxn1 at the promoters. The gene expression and ChIP-PCR data suggest that polyglutamine-expanded Atxn1 is less efficient in Cic-dependent repression. More importantly, they provide evidence for a model in which loss of Atxn1/Cic function can result in transcriptional dysregulation, contributing to SCA1 pathogenesis.
Atxn1Ldp/+ rescues Cic levels in Atxn1−/− mice by forming functional Atxn1L-Cic complexes
Given that mild overexpression of the Atxn1 paralog, Atxn1L, can displace both wild-type and polyglutamine-expanded Atxn1 from large native complexes in a dose-dependent manner, we wondered if Atxn1L could partially replace Atxn1 as a Cic binding partner, especially because Atxn1L interacts with Cic in wild-type cerebellum , . This led us to propose that in addition to competing with Atxn1[154Q] in the large native complexes , increased Atxn1L levels can suppress a putative loss of Atxn1 function in Atxn1154Q/+ mice by directly substituting for Atxn1 in Cic-containing complexes. To test this possibility, we generated Atxn1−/− and Atxn1−/−; Atxn1Ldp/+ littermates, and performed western blot analysis for Cic. As shown in Figure 3A, Atxn1L overexpression results in restoration of Cic levels back to wild-type levels in Atxn1−/− cerebellum. We then tested if overexpression of Atxn1L increased formation of Atxn1L-Cic complexes, thus stabilizing Cic. For this, we performed co-immunoprecipitation studies using Cic antibody on cerebellar extracts of Atxn1−/− mice with or without the Atxn1Ldp allele, followed by western blot analysis for Atxn1L. As expected, Atxn1L co-immunoprecipitates with Cic in wild-type cerebella (Figure 3B). However, despite the reduced levels of Cic protein in Atxn1−/− cerebella (Figure 3A and ), the relative fraction of Atxn1L co-immunoprecipitated with Cic in Atxn1−/− protein extracts was greater than in wild-type cerebella (Figure 3B). Atxn1L overexpression further increased the levels of Atxn1L-Cic co-immunoprecipitation in Atxn1−/−; Atxn1Ldp/+ mice (Figure 3B). These results suggest that Atxn1L overexpression enhances the formation of Atxn1L-Cic complexes in Atxn1−/− mice, thus stabilizing Cic protein levels.
Figure 3. Atxn1Ldp stabilizes Cic protein levels in Atxn1−/− mice by enhancing Atxn1L-Cic complex formation.
(A) Western blot analysis shows that overexpression of Atxn1L rescues the reduced Cic levels in Atxn1−/− mice (* p<0.05). (B) Co-immunoprecipitation of Atxn1L using Cic antibodies show that, despite the reduced levels of Cic protein in Atxn1−/− cerebella, the relative fraction of Atxn1L co-immunoprecipitated with Cic in Atxn1−/− protein extracts was greater than in wild-type cerebella. Atxn1L overexpression further increased the levels of Atxn1L-Cic co-immunoprecipitation in Atxn1−/−; Atxn1Ldp/+ mice. Images show representative blots and the quantification of three independent experiments. Error bars in graphs represent +/−SEM. * p<0.05. (C) To determine if Atxn1L-Cic complexes are functional, we measured the transcriptional effect of Cic, Atxn1 and Atxn1L on the expression of a luciferase reporter construct containing a tandem array of Cic binding sites (CBS). Luciferase activity is expressed as fraction of the activity when the reporter is expressed alone (100%). Co-transfection of Cic and Atxn1[2Q]) results in synergistic co-repression of this luciferase reporter (Cic+Atxn1[2Q]). Interestingly, co-transfection of constructs expressing Cic and Atxn1L (Cic+Atxn1[2Q]) results in synergistic repression of the reporter similar to co-transfection of Cic and wild-type Atxn1. These results suggest that Atxn1L-Cic complexes are functional in Capicua-dependent repression. Assays were performed in duplicate in 5 independent experiments. Error bars in graph represent +/− SEM, *p<0.05,***p<0.005.doi:10.1371/journal.pgen.1001021.g003
We further asked whether Atxn1L-Cic complexes were proficient in Cic-dependent transcriptional repression. For this, we co-transfected HEK293T cells with a luciferase reporter construct containing a tandem array of Cic binding sites , , along with Atxn1L and Cic-expressing plasmids. Consistent with previous studies , co-transfection of Cic and Atxn1 resulted in synergistic repression of the Cic-responsive luciferase reporter (Figure 3C). Interestingly, co-transfection of constructs expressing Cic and Atxn1L resulted in synergistic repression of the reporter similar to co-transfection of Cic and wild-type Atxn1 (Figure 3C). These results strongly suggest that Atxn1L-Cic and Atxn1-Cic complexes are functionally redundant in Cic-dependent transcriptional repression. Thus, we conclude that mild overexpression of Atxn1L in Atxn1−/− mice might partially rescue a loss of Atxn1 endogenous function related to reduced Atxn1-Cic complexes.
Atxn1Ldp/+ restores some gene expression alterations in Atxn1−/− mice
Given that mild Atxn1L overexpression rescues Cic levels in Atxn1−/−; Atxn1Ldp/+ mice, we predicted that Atxn1L should rescue some of the transcriptional changes in Atxn1−/− cerebella if it can functionally substitute for Atxn1. To test this, we performed qRT-PCR on cerebellar RNA isolated from Atxn1−/−; Atxn1Ldp/+ mice and Atxn1−/− littermates at 16 weeks of age. We focused our qRT-PCR analysis on the 9 genes that are commonly dysregulated in Atxn1154Q/+ and Atxn1−/− cerebella (Table 1), and also on the 4 down-regulated Rorα targets validated by qRT-PCR in Atxn1−/− cerebellum (Figure S7). Interestingly, 5 out of 13 genes tested (Ccnd1, Igfbp5, Apba2bp, Robo1 and Grid2) were partially or completely restored back to wild-type levels in Atxn1−/−; Atxn1Ldp/+, compared to Atxn1−/− cerebella (Figure 4A–4E). Ccnd1, one of the up-regulated genes that is potentially a direct target of Atxn1-Cic complexes (Figure 4A), and Igfbp5, an early key pathogenic marker in SCA1 disease (Figure 4B), were among the genes rescued by Atxn1L overexpression in Atxn1−/− mice. Additionally, the Rorα target Grid2 is also significantly rescued in Atxn1−/−; Atxn1Ldp/+ mice. Thus, mild overexpression of Atxn1L in vivo results in partial rescue of several transcriptional changes related to SCA1 pathogenesis in a loss-of-function model of Atxn1. Since Atxn1L binds to transcriptional regulators that interact with ATXN1, such as CIC and SMRT-NCoR, it is possible that Atxn1L rescues transcriptional changes by functionally replacing polyglutamine-expanded Atxn1 in these transcriptional complexes.
Figure 4. Mild overexpression of Atxn1L restores expression levels of some genes altered in Atxn1−/− cerebella.
Real-time quantitative RT-PCR on RNA extracted from wild-type (n = 8), Atxn1−/− (n = 10) and Atxn1−/−; Atxn1Ldp/+ cerebella (n = 6) for (A) Ccnd1, (B) Igfbp5, (C) Apba2bp, (D) Robo1, and (E) Grid2 reveals that these transcripts were restored close to wild-type levels in Atxn1−/−; Atxn1Ldp/+ cerebellum. Real-time qRT-PCR experiments were performed in triplicate for each sample. Error bars represent +/−SEM, *p<0.05.doi:10.1371/journal.pgen.1001021.g004
Partial rescue of behavioral defects in Atxn1−/− mice by mild overexpression of Atxn1L
Although Atxn1−/− mice do not exhibit overt ataxia phenotypes or progressive neurodegeneration, they do exhibit a variety of neurological deficits ( and Figures S8, S9, S10, S11). Interestingly, Atxn1−/− mice exhibit deficits in spatial learning and memory, and in motor learning and coordination, phenotypes shared with Atxn1154Q/+mice ( and Figures S10 and S11). This raises the possibility that partial loss of Atxn1 function of could also contribute to these phenotypes in Atxn1154Q/+mice. The fact that the Atxn1Ldp allele can suppress behavioral phenotypes in Atxn1154Q/+mice  also raises the possibility that Atxn1L suppresses SCA1 pathogenesis by functionally replacing those Atxn1 functions altered by polyglutamine-expanded Atxn1.
To investigate whether Atxn1L and Atxn1 are functionally redundant in vivo, we tested if the Atxn1Ldp allele can ameliorate behavioral defects in Atxn1−/− mice. We focused on two characteristic phenotypes clinically associated with SCA1 disease: cognitive deficits (learning and memory) and deficits in motor coordination and balance , . We performed the conditioned fear paradigm on Atxn1−/− mice carrying the Atxn1Ldp allele and compared them to Atxn1−/− littermates. As shown in Figure 5A, and in agreement with previous data , Atxn1−/− mice exhibited significant deficits, as determined by reduced freezing behavior, in the contextual fear-conditioning test compared to wild-type and Atxn1+/− littermates. Wild-type and Atxn1+/− mice expressing the Atxn1L dp allele performed similarly to the controls (Figure 5A). Interestingly, the duplication of Atxn1L rescued the freezing behavior due to loss of Atxn1 (compare Atxn1−/− mice to Atxn1−/−; Atxn1Ldp/+ mice; Figure 5A).
Figure 5. Atxn1Ldp suppresses the behavioral deficits in Atxn1−/− mice.
We assessed the effects of mild overexpression of Atxn1L on Atxn1−/− phenotypes (A) Mice of the following genotypes: Atxn1+/+ (n = 5), Atxn1+/− (n = 12), Atxn1−/− (n = 9), Atxn1+/+; Atxn1Ldp/+(n = 10), Atxn1+/−; Atxn1Ldp/+(n = 17), and Atxn1−/−; Atxn1Ldp/+(n = 12), were tested at 8 weeks for contextual conditioned fear. Atxn1−/− mice exhibited significantly reduced freezing behavior in the contextual fear-conditioning test compared to wild-type and Atxn1+/− littermates. However, Atxn1−/− mice carrying the Atxn1L duplication performed significantly better than Atxn1−/− littermates in this task (p<0.05). Wild-type and Atxn1+/− mice expressing the Atxn1Ldp allele performed similarly to wild-type and Atxn1+/− mice without the Atxn1L duplication. (B–E) An independent cohort of mice was generated to test the effects of Atxn1L duplication on the motor deficits of Atxn1−/− mice. Atxn1+/− (n = 16), Atxn1+/−; Atxn1Ldp/+(n = 23), Atxn1−/− (n = 14), and Atxn1−/−; Atxn1Ldp/+ mice (n = 17) were tested on the dowel and wire hang paradigms at 8 weeks. The latency of Atxn1−/− mice to reach the side for the first time was increased compared to Atxn1+/− and Atxn1+/−; Atxn1dp/+ littermates on the rod (B); they also walked off the rod fewer times in the 120 s interval (C). In contrast, Atxn1−/−; Atxn1Ldp/+ mice took less time to walk off the dowel (B), and they also crossed the dowel more times in 120 s than Atxn1−/− littermates (C). In the wire hang test, Atxn1−/− mice showed increased latency to reach the sides of the wire compared Atxn1+/− and Atxn1+/−; Atxn1Ldp/+controls (D). Additionally, Atxn1−/− mice reached the sides fewer times than control littermates (E). In contrast, Atxn1−/− mice overexpressing Atxn1L exhibited marked reduction in the time for the first touch (D), and increased number of side touches in the 120 s interval, when compared to Atxn1−/− mice (E). Error bars in graphs represent +/− SEM, *p<0.05.doi:10.1371/journal.pgen.1001021.g005
Having shown that mild overexpression of Atxn1L rescues Pavlovian contextual learning in Atxn1−/− mice, we then wanted to assess the effects of increased Atxn1L levels on motor coordination and balance deficits in Atxn1−/− mice, another phenotype common to SCA1 mouse models. In order to discriminate between motor learning impairments and motor coordination deficits, we chose to use the dowel rod and wire hang tests. These paradigms do not rely on consecutive daily training, as does the rotating rod test, therefore we selected them in order to discern motor coordination and balance defects from cerebellar learning deficits , . We generated an independent cohort of Atxn1−/− mice and Atxn1−/−; Atxn1Ldp/+ mice, with Atxn1+/− and Atxn1+/−; Atxn1dp/+ littermates as controls, and tested them at 8 weeks of age. We measured the latency to reach the side (first touch) and frequency of walking off the rod (number of side touches in 120 seconds) (Figure 5B and 5C). All genotypes tested remained on the dowel for the maximum time (data not shown), but Atxn1−/− mice moved much less than Atxn1+/− and Atxn1+/−; Atxn1dp/+ littermates on the rod. The latency of Atxn1−/− mice to reach the side for the first time was increased (Figure 5B), and consequently they walked off the rod fewer times (Figure 5C). It is noteworthy that Atxn1−/− mice are active and travel the same distance as wild-type littermates the open field analysis (Figure S5). Thus the hesitancy to move on the dowel suggests that in addition to learning and memory deficits, Atxn1−/− mice have balance or motor coordination impairments, evident at an early age (8-week-old). In contrast, Atxn1−/− mice carrying the Atxn1Ldp allele take less time to walk off the dowel (Figure 5B), and they also crossed the dowel more times in 120 s than Atxn1−/− littermates (Figure 5C). Thus, mild Atxn1L overexpression partially rescues the dowel phenotype caused by loss of Atxn1.
The wire-hang paradigm assesses motor coordination and grip strength . In this paradigm, mice are hanging onto the center of an elevated wire from their forepaws, and they need to get to the sides for relief, which requires coordination and normal strength. We measured the time and frequency to reach the sides in a 120 s interval, and found no significant falling off the wire in Atxn1−/− mice compared to the control littermates, suggesting that they have reasonable grip strength (data not shown). However, Atxn1−/− mice showed increased latency to reach the sides for the first time compared Atxn1+/− and Atxn1+/−; Atxn1Ldp/+ controls (Figure 5D). Additionally, Atxn1−/− mice reached the sides fewer times than control littermates (Figure 5E). In contrast, Atxn1−/− mice overexpressing Atxn1L exhibited marked reduction in the time for the first touch and increased number of side touches in the 120 s interval, when compared to Atxn1−/− mice (Figure 5D and 5E). Taken together, these behavioral data demonstrate that a 50% increase in the levels of AtxnlL is sufficient to partially rescue several behavioral deficits caused by loss of Atxn1 function. Furthermore, this rescue correlates with the molecular data, demonstrating that Atxn1L is a functional homolog of Atxn1 in vivo.
Recently, we proposed the possibility that in addition to toxic gain-of-function due to polyglutamine-expanded ATXN1, a concomitant partial loss of ATXN1 function might contribute to SCA1 pathogenesis . It is challenging to establish the extent of the contribution of a potential loss-of-function mechanism to SCA1 pathogenesis in models carrying the mutant protein, since the severe gain-of-function effects might mask any subtle loss-of-function component, thus confounding the interpretation of the results. In the present study, we sought to distinguish between gain- and loss-of-function mechanisms by focusing on transcriptional defects in Atxn1−/− mice, and comparing them to the knock-in model of SCA1, Atxn1154Q/+ mice. Using this approach, we identified several molecular changes that could be attributable to loss of ATXN1 function in SCA1.
We found that loss-of-function of Atxn1 in mice is sufficient to cause many transcriptional changes common to the Atxn1154Q/+ knock-in mice, a model of SCA1 that faithfully replicates many features of the disease, and with SCA1[82Q] transgenic mice. It has been reported that ATXN1 interacts with several factors involved in transcriptional regulation, including CIC, SMRTER, HDAC3, Gfi1 and RORα , –. Therefore, these shared expression changes might be indicative of altered transcriptional functions of ATXN1 in SCA1 pathogenesis. Furthermore, we showed that a majority of the shared transcriptional changes go in the same direction in both Atxn1−/− and Atxn1154Q/+ mice, strongly arguing that part of the transcriptional dysregulation in SCA1 might be explained by a partial loss-of-function of Atxn1. The up-regulation of direct targets of Atxn1-Cic provides evidence for this concept. Another important finding of this study is that there are many transcriptional changes that are unique to the Atxn1154Q/+ model, which could potentially be related to toxicity of polyglutamine-expanded Atxn1.
We propose that a combination of toxic gain-of-function and mild loss-of-function mechanisms contribute to SCA1 pathogenesis, with the partial loss-of-function of ATXN1 being sufficient to cause some transcriptional changes that are pathogenic in the cerebellum. Previous studies using microarray analysis reported on the down-regulation of the dopamine receptor D2 (Drd2) in Atxn1−/− mouse cerebella . However, with the exception of a couple of genes (e.g. Pafahb3, Sp1), we were unable to find extensive overlap between the changes reported by Goold et al. and the microarray analysis presented in this study. The differences in genetic background, microarray platform, and the age of the Atxn1−/− animals (5-week-old) in the Goold et al. studies , compared to 16 weeks in our studies, are likely to contribute to the minimal overlap in gene expression changes in the two studies.
Bioinformatics analyses of the genes commonly altered in Atxn1−/− and Atxn1154Q/+ cerebella show enrichment for categories associated with pathological pathways involved in neurodegeneration (Alzheimer's disease), and also pathways previously implicated in pathogenesis both in knock-in and transgenic SCA1 mouse models, such as the phosphatidylinositol and calcium signaling pathways –. These results strongly suggest that Atxn1−/− mice have dysfunctional cerebella due to a loss of endogenous Atxn1 function. We also found that Atxn1−/− mice share significant overlap in cerebellar transcriptional profiles with staggerer mice, which have a spontaneous loss-of-function mutation in the gene encoding the transcription factor Rorα. Rorα-regulated genes involved in calcium signaling (Itpr1 and Calb1) and glutamatergic signaling (Grm1 and Slc1a6) are significantly down-regulated in Atxn1−/− cerebellum, as determined by microarray analysis and real-time qRT-PCR. It is noteworthy that loss-of-function mutations in several of these genes result in ataxic phenotypes (e.g. Itpr1, Slc1a6, and Grm1) –, raising the possibility that simultaneous down-regulation of several of these genes could contribute to the motor coordination impairments observed in Atxn1−/− mice. Rorα mRNA transcript and protein levels appear normal in Atxn1−/− cerebellum , ruling out that changes in Rorα targets are due to reduced Rorα protein levels in Atxn1−/− mice. Since ATXN1 and Rorα physically interact via Tip60, it is conceivable that loss of Atxn1 affects Rorα-dependent transcription directly . Altogether, these data support two important conclusions: first, that Atxn1−/− cerebellum exhibits pathological molecular changes, even in the absence of progressive neurodegeneration, and second, that transcriptional changes in the loss-of-function model of Atxn1 could identify endogenous pathways that might also be altered by the expression of mutant Atxn1.
We previously described a reduction of Atxn1-Cic complexes in Atxn1154Q/+cerebella, with Atxn1[154Q] favoring the formation of enhanced toxic gain-of-function complexes with RBM17 . It is interesting that among the genes up-regulated both in Atxn1−/− and Atxn1154Q/+cerebella, we identified two potential direct targets of Cic-dependent repression, Ccnd1 and Etv5, the genes encoding for Cyclin D1 and Ets variant 5, respectively ( and J. Fryer, unpublished data). We demonstrated that wild-type Atxn1 and Cic are bound to the promoter regions of Ccnd1 and Etv5. Interestingly, we failed to detect mutant Atxn1[154Q] on these promoters in mice only expressing expanded Atxn1 (Atxn1154Q/−). One interpretation of this result is that Atxn1[154Q] has diminished association to the promoters, resulting in reduced Atxn1-Cic dependent repression. Alternatively, it is possible that polyglutamine-induced conformational changes make the Atxn1[154Q]-Cic complexes less accessible for antibody recognition, resulting in reduced chromatin immunoprecipitation. Irrespective of the basis of the inability to detect Atxn1[154Q] binding on these promoters, the data strongly suggest that polyglutamine-expanded Atxn1 and Cic have reduced transcriptional repression function on these specific promoters in vivo. These results provide a mechanistic explanation on how diminished Atxn1-Cic function can contribute to transcriptional defects in SCA1.
Mild overexpression of the evolutionarily conserved gene Atxn1L partially suppresses the neuropathology caused by polyglutamine-expanded ATXN1 in flies and mice , . Increased Atxn1L levels induce the sequestration of polyglutamine-expanded Atxn1 into nuclear inclusions, leading to a proposed model in which Atxn1L suppresses toxicity by displacement of mutant Atxn1 from its major endogenous complexes that contain Cic . In the present study, we show an additional mechanism contributing to this rescue, by demonstrating that mild overexpression of Atxn1L suppresses several molecular and behavioral phenotypes in Atxn1−/− mice, potentially by replacing Atxn1 in Cic-containing complexes (Figure 6). The motor coordination and learning deficits suppressed by Atxn1L are common to both Atxn1−/− and polyglutamine-expanded Atxn1154Q/+ mouse models. Therefore, these data provide evidence for an additional mechanism in which Atxn1L can functionally compensate for a partial loss of Atxn1 function to suppress SCA1 pathogenesis. Although our studies demonstrate that Atxn1L is a functional homolog of Atxn1 in Cic-mediated transcriptional repression, we cannot rule out that Atxn1L overexpression can restore other yet to be determined Atxn1-related functions not addressed in these studies. In sum, based on our previous data and this study, we propose that partial loss of ATXN1 function actively contributes to SCA1 pathogenesis as part of a two-pronged mechanism, in which enhanced toxic gain-of-function of polyglutamine-expanded ATXN1 leads to neurodegeneration, while a simultaneous loss-of-function of other stable endogenous protein complexes, Atxn1-Cic, contributes to the SCA1 phenotypes (Figure 6B).
Figure 6. Model of Atxn1L rescue through Cic stabilization in Atxn1−/− mice.
(A) In wild-type mice, Atxn1-Cic and Atxn1L-Cic complexes bind the promoters of target genes and repress them effectively. (B) In mice expressing polyglutamine-expanded Atxn1, mutant Atxn1 associates preferentially with Rbm17, while the decrease in Atxn1-Cic complexes destabilizes Cic and reduces its levels at the promoters, thus leading to de-repression of its target genes. (C) A similar Cic destabilization occurs in the absence of wild-type Atxn1 in Atxn1−/− mice, also resulting in increased expression of target genes. (D) When Atxn1L is moderately overexpressed in Atxn1−/− mice, it stabilizes Cic levels by forming functional Atxn1L-Cic complexes that can substitute for Atxn1-Cic at the promoters, thus rescuing target gene repression. We propose that this mechanism might also act to rescue loss-of-function of Atxn1 in Atxn1154Q/+; Atxn1Ldp/+ mice.doi:10.1371/journal.pgen.1001021.g006
It was previously reported that reduction of normal functions of genes involved in other polyglutamine diseases results in enhanced pathology, providing evidence for concomitant gain- and loss-of-function mechanisms in polyglutamine disorders –. In SBMA transgenic mouse models expressing polyglutamine-expanded androgen receptor (AR), loss of endogenous AR protein resulted in accelerated motor neuron degeneration . These studies suggested two independent pathways contributing to SBMA pathogenesis: gain-of-function due to mutant AR nuclear toxicity and loss of AR trophic effects on motor neurons . Conditional deletion of Htt in mouse forebrain leads to several features reminiscent of Huntington disease, including motor deficits, tremors and progressive degeneration of the striatum and cortex, hinting that loss of htt function could contribute to these phenotypes in HD . Moreover, loss of wild-type Htt function leads to enhanced neurodegeneration in transgenic models expressing polyglutamine-expanded Htt, while overexpression of wild-type Htt reduces toxicity caused by mutant htt –. The mechanism by which loss of Htt contributes to HD is unclear at this time; it might involve either its anti-apoptotic properties, its role in BDNF-mediated neuroprotection, both, or some other yet to be determined function –. Our studies comparing Atxn1−/− and SCA1 knock-in mice pinpoint Cic-dependent transcriptional repression as one of the molecular pathways mediating the partial loss-of-function component in SCA1 pathogenesis.
Loss of normal endogenous function of mutant proteins may also play a role in other dominant neurodegenerative diseases caused by gain-of-function mutations. The Parkinson's disease model mice overexpressing the mutant A53T SNCA gene lacking endogenous alpha-synuclein, exhibit worsened synucleinopathy when compared to littermates carrying wild-type Snca alleles . In Alzheimer disease (AD), increased aggregation of the amyloid beta peptides induced by AD-related presenilin mutations is thought to be a consequence of a dominant gain-of-function mechanism . However, loss of function of presenilin in the mouse brain results in phenotypes strikingly reminiscent of AD (progressive memory loss and neurodegeneration) in the absence of beta-amyloid deposition , . These results suggest that altered pathways leading to Alzheimer disease can be caused from a combination of dominant gain of function and/or loss of function mechanisms. The potential prevalence of mutations that lead to both loss- and gain-of-function in human neurological diseases underscores the importance of understanding the endogenous functions of causative genes through the careful analysis of loss-of-function models, which may uncover critical pathways leading to pathogenesis.
Materials and Methods
All mouse lines used in this study have been previously described , , . Atxn1154Q/+ and Atxn1−/− mice have been backcrossed into the C75Bl/6J strain for over ten generations. In the case of Atxn1L duplication (Atxn1Ldp) mice, all experiments described in this study were done in mice backcrossed for at least 7 generations into the C75Bl/6J strain. Atxn1−/− mice were bred to Atxn1Ldp/+ mice, and Atxn1−/− mice carrying the Atxn1L duplication locus were generated by intercrossing the progeny. Mouse experiments followed protocols approved by the Baylor College of Medicine Institutional Animal Care and Use Committee (IACUC).
Total RNA isolation from adult mouse cerebella was performed as described elsewhere . Briefly, RNA was extracted from cerebella of 16-week-old mice using TRIzol reagent (Invitrogen Corporation, Carlsbad, CA), DNaseI-treated, and purified using the RNeasy mini kit according to the manufacturer's protocol (Qiagen, Valencia, CA).
Microarray analysis with exon arrays
We used the Affymetrix Mouse Exon 1.0 ST microarray, which carries 1.2 million probe-sets covering one million exon clusters, with an average of 40 probes per gene. The exon array data were analyzed as previously described . Briefly, raw data were processed in the R statistical programming environment using locally developed methods and the exonmap package. RMA normalization was applied, and linear models were calculated to analyze genotype effects for each gene. Genomic annotations were obtained from UCSC (http://genome.ucsc.edu). The normalized probe level data were then averaged within each exon to produce exon-level data for each gene for each animal. A two-way ANOVA with main effects for genotype and exonic region was calculated for each gene. The ANOVA model was fit using weighted least squares analysis where the weights were determined according to the probe counts within each exon. Since separate wild-type (WT) control littermates were used in the Atxn1−/− and Atxn1154Q/+ experiments, a separate linear model was estimated for each gene in each model (one fit for each WT background). A linear contrast was calculated comparing the WT and mutant cross-exon means for each gene. The cut-off rule for determining genes was a fold change threshold of +/−0.1 in both experiments, and a linear step up false discovery rate (FDR) of less than 0.01 value for the T-statistic corresponding to the linear contrast comparing each WT strain with its corresponding mutant. The gene set determined by this fold change and FDR multiplicity corrected cutoff, corresponds to a median raw marginal p-value of less than 0.00015 for the underlying T-statistics.
We performed Gene Ontology (GO) analysis on the obtained data using locally developed software and methods . Briefly, the gene ontology vocabulary and current mouse annotations were obtained from the GO website (9/1/2007 build). The mouse exon array was mapped to Entrez identifiers, and these identifiers were mapped to the GO data structure using the available annotations. Using our local ontology analysis system (OntologyTraverser), we tabulated the genes annotated at or below each GO node for the entire exon array. We then used a hypergeometric sampling model to examine the statistical representation of each GO node for genes in our gene sets. In order to make comparisons between sets, we took differences between the standardized scores determined for each gene set. Because of the extreme overlapping structure of the GO, many GO nodes report duplicate or redundant information. To avoid this problem, we calculated the GO covariance structure and used this estimate to compute de-correlated GO scores. For the KEGG pathway analysis, the overlapping gene list containing the 197 most significant genes was uploaded and analyzed using the web-based Functional Annotation Bioinformatics Microarray Analysis DAVID 6.7 (National Institute of Allergy and Infectious Diseases NIH, david.abcc.ncifcrf.gov/).
Real-time quantitative reverse transcriptase-polymerase chain reaction (qRT–PCR)
For the independent validation of the exon array data, real time qRT-PCR assays were performed on more than eight mice of each genotype. cDNA was synthesized from 1 µg of RNA using the RT2 First Strand Kit (SuperArray Bioscience Corporation, Frederick, MD). Quantitative real-time PCR reactions were performed on 10 ng of cDNA using RT2 SYBR Green/ROX PCR master mix and commercially available primers (SuperArray Bioscience Corporation, Frederick, MD). All RNA samples were analyzed in triplicate and normalized relative to Gapdh levels.
For validation of Rorα targets, cDNA was synthesized from RNA isolated from seven Atxn1−/− mice and control littermates, using Superscript III (Invitrogen). Quantitative real-time PCR reactions were performed on 10 ng of cDNA using the SYBR Green PCR Master Mix (Applied Biosystems) using the following primer sets:
Itpr1 fw 5′- GGGCCAACAGCACTACAGATG-3′
Grid2 fw 5′- CCCGCATTGAGAGCTCCAT-3′
rv 5′- GCCATAAGGGATATCTGTTTGCTT-3′
Inpp5a fw 5′- TGGTCAAGAAAAGGCTTCATCA-3′
rv 5′- CCAAGTCGAAGGCACAGTCA-3′
Grm1 fw 5′- TCCTCTGACCTGAGACCAATAGC-3′
rv 5′- CGCGTTAGTGGCCATAAGCT-3′
Chromatin Immunoprecipitation-polymerase chain reaction analysis (ChIP-PCR)
ChIP was performed as previously described . Cerebella were dissected from mice of each genotype at 7–8 weeks of age and incubated in 1% formaldehyde for 10 minutes at room temperature to cross-link DNA to associated proteins. Chromatin was treated with micrococcal nuclease and sheared by sonication to generate fragments with an average length of ~100–200 bps, as determined by agarose gel electrophoresis. For immunoprecipitation, 200 µl of chromatin was diluted 1:10 in ChIP dilution buffer (Millipore Corporation, Billerica, MA) and 1% of the diluted sample was saved as input. The sample was first precleared with protein A Dynabeads (Invitrogen Corporation, Carlsbad, CA), then incubated overnight with protein A Dynabeads that were pre-blocked with salmon sperm DNA and coupled to 20 µl of affinity purified rabbit anti-Atxn1 (11NQ) antibody, or 4 µl of guinea pig polyclonal anti-Cic antibody. Mock immunoprecipitations using nonspecific preimmune sera for each antibody were included as negative controls. After immunoprecipitation, the beads were washed at room temperature with low salt buffer, followed by high salt buffer, LiCl buffer (Millipore Corporation, Billerica, MA), and TE buffer (10 mM Tris-HCl pH 7.4, 1 mM EDTA pH 8.0). Elution was performed twice in 250 µl of fresh elution buffer (1% SDS, 0.1 M NaHCO3) for 15 minutes at room temperature. The eluates were combined, the crosslinks were reversed, and DNA was purified with Qiagen PCR cleanup kit (Qiagen, Valencia, CA) and recovered in 30 µl of 10 mM Tris-HCl pH 8.0. One or two µl of DNA were used for each PCR reaction using primer pairs for the promoter regions of the following genes:
Etv5 fw 5′-GGGGAAGCTTAGCTGAGTCAGTGAA-3′
rv 5′- GTTTCTGTGTGTGGAATGACGAATTC-3′
Ccnd1 fw 5′-GGTTAACTGAATGGACTCCTAAGTTT-3′
Capicua binding to the promoters was also analyzed using ChIP followed by quantitative real-time PCR (SYBR green). Quantitative real-time PCR experiments were performed in triplicate on three independent sets of samples. Relative amounts of immunoprecipitated DNA were determined based on the threshold cycle (Ct) value for each PCR reaction. In order to control for variation between ChIP fractions, for every gene promoter studied, a ΔCt value was calculated for each sample (Atxn1+/−, Atxn1154Q/+, Atxn1−/−) by subtracting the Ct value for the input (CtInput) from the Ct value for the immunoprecipitated sample (Ctantibody or Ctpreimmune). Since the input DNA fraction represents only 1% of the total material, the CtInput value was first adjusted for this dilution factor by subtracting 6.644 cycles (Log2 of 100), then substracted from the immunoprecipitated samples using the following formula:
Differences between the specific immunoprecipitation and the preimmune serum background were then determined and plotted as fold enrichment over the preimmune serum (for each genotype sample: ΔCtantibody/ΔCtpreimmune).
Primer and sequences for the promoter regions used were as follows:
Etv5 qA (highly conserved; two predicted Cic-binding sites):
fw 5′- TTGCTCCTGATCACACATGC -3′
rv 5′- GCTGGAACCTCGTGAATGAT -3′
Etv5 qB (conserved; ~1 Kb upstream of positive region Etv5 qA)
fw 5′- AATCAGCACCGGCTTGTTTA-3′
rv 5′- CTAAGCTTCCCCCTCAGGTC-3′
Ccnd1 qA (highly conserved; ~150 bp upstream of predicted Cic-binding site)
fw 5′- AAATTTGCATGAGCCAATCC-3′
rv 5′- GCAGAGCTCAACGAAGTTCC-3′
Ccnd1 qB (poorly conserved; ~400bp downstream of predicted Cic-binding site)
fw 5′- GGTCGTGGTTAACTGAATGGA -3′
rv 5′- AGGTGGTGGAACCGCTTTAT-3′
Cerebella from three mice of each genotype were dissected, and dounce-homogenized in 1 ml of ice-cold TST buffer (10 mM Tris–HCl, pH 7.5, 0.9% NaCl, 0.05% Tween 20). Protein was adjusted to 0.5 mg in 200 µl with TST buffer (10 mM Tris–HCl, pH 7.5, 0.9% NaCl, 0.05% Tween 20) and then added 800 ml cold PBS. Fifteen microlitters of the extract were saved as input. The extract was pre-cleared using protein A Sepharose beads. BSA-blocked protein A Sepharose beads were coupled to 4 µl of polyclonal guinea pig anti-Cic serum or pre-immune serum for 1 hr at room temperature, and the diluted extract was incubated with the antibody-coupled beads overnight at 4°C. Four washes were carried out using ice-cold TST buffer and the pellet was resuspended in sample buffer. Input and pellets were analyzed by SDS-PAGE and protein blot.
Luciferase reporter assays for Cic-dependent repression were performed as previously described . HEK293T cells in 24-well plates were co-transfected using Lipofectamine 2000 (Invitrogen) with 50 ng of the pGL3-Promoter (Promega) containing six copies of CIC binding sites (TGAATGAA or TGAATGGA), 10 ng of pRL-TK, and 10 ng of expression plasmids for Atxn1[2Q] (wild-type), Atxn1L and Cic-expressing plasmids as indicated. All constructs have also been previously described , . The total amount of DNA transfected was kept constant by adding pcDNA3.1(-) (Invitrogen). Luciferase activities were measured using the dual luciferase reporter assay system (Promega).
Contextual fear conditioning
For this task, the mice were trained in a novel environment with a neutral stimulus (a tone) paired with a foot shock. The mice are placed in a different context with the same tone (cued test) or back in the same environment without a tone (contextual test) 24 h later. Mice of each genotype were placed in a Med Associates/Actimetrics chamber system where a 30 second tone was followed by a 2 second foot shock at 1.5 mA. The tone–foot shock were repeated at 2 min. Twenty-four hours later, the mice were tested for freezing in the same chamber with no tone to evaluate contextual fear. Scoring for freezing is automated in this system. Analysis was performed using ANOVA and t-test analysis.
Dowel and wire hang paradigms
Mice were placed on the center of a 0.9 cm wooden dowel suspended between two platforms. If the mouse walked off of the dowel onto the platform, the mouse was placed back in the center. The test lasted 2 min beginning when the mouse was placed on the dowel. Data were analyzed using ANOVA and Student's t-test.
Cellular component gene ontology for genes commonly up-regulated in Atxn1−/− and Atxn1154Q/+ cerebella. Gene ontology categories shown were significantly enriched (positive z score, green) or depleted (negative z score, green) with the de-correlated z score for enrichment plotted in the x-axis. Only gene ontology categories with more than one gene represented and a z score>|+/−2| are represented.
(0.24 MB TIF)
Biological process gene ontology for genes commonly up-regulated in Atxn1−/− and Atxn1154Q/+ cerebella. Gene ontology categories shown were significantly enriched (positive z score, green) or depleted (negative z score, green) with the de-correlated z score for enrichment plotted in the x-axis. Only gene ontology categories with more than one gene represented and a z score>|+/−2| are represented.
(0.53 MB TIF)
Molecular function gene ontology for genes commonly up-regulated in Atxn1−/−and Atxn1154Q/+ cerebella. Gene ontology categories shown were significantly enriched (positive z score, green) or depleted (negative z score, green) with the de-correlated z score for enrichment plotted in the x-axis. Only gene ontology categories with more than one gene represented and a z score>|+/−2| are represented.
(0.51 MB TIF)
Cellular component gene ontology for genes commonly down-regulated in Atxn1−/− and Atxn1154Q/+ cerebella. Gene ontology categories shown were significantly enriched (positive z score, green) or depleted (negative z score, green) with the de-correlated z score for enrichment plotted in the x-axis. Only gene ontology categories with more than one gene represented and a z score>|+/−2| are represented.
(0.38 MB TIF)
Biological process gene ontology for genes commonly down-regulated in Atxn1−/− and Atxn1154Q/+ cerebella. Gene ontology categories shown were significantly enriched (positive z score, green) or depleted (negative z score, green) with the de-correlated z score for enrichment plotted in the x-axis. Only gene ontology categories with more than one gene represented and a z score>|+/−2| are represented.
(0.74 MB TIF)
Molecular function gene ontology for genes commonly down-regulated in Atxn1−/− and Atxn1154Q/+ cerebella. Gene ontology categories shown were significantly enriched (positive z score, green) or depleted (negative z score, green) with the de-correlated z score for enrichment plotted in the x-axis. Only gene ontology categories with more than one gene represented and a z score>|+/−2| are represented.
(0.73 MB TIF)
Real time qRT-PCR validation of Rorα targets in Atxn1−/− cerebella. Four out of six tested genes were significantly down-regulated in Atxn1−/− cerebella. Error bars in graph represent +/− SEM *p<0.05.
(0.30 MB TIF)
Open Field Analysis to measure activity in pure C57Bl/6J Atxn−/− mice. (A) Atxn1−/− mice at 10–11 weeks of age spend more time in the center of the field compared to wild-type littermates, suggesting they are less anxious. (B) Atxn1−/− mice and wild-type littermates traveled similar total distances. (C) Atxn1−/− mice had decreased rearing, as measured by vertical activity in the open field. Since Atxn1−/− mice show increased center/total distance, their reduced vertical activity probably reflects motor defects affecting rearing, and not increased anxiety. Error bars represent +/− SEM, **p<0.005.
(0.31 MB TIF)
Light/Dark box and Elevated Plus maze to test for anxiety in pure C57Bl/6J Atxn1−/− mice. Atxn1−/− mice (n = 10) and wild-type controls (n = 9) were tested at 12 weeks of age (A) Atxn1−/− mice had a small trend to spend more time in the light side than wild-type controls, although this trend does not reach significance (p = 0.26). (B) Both wild type and Atxn1−/− mice make the same number of transitions between the light and dark side of the box. (C) In the elevated plus maze, Atxn1−/− mice spent more time in the open arms than controls, and (D) also made more arm entries than controls. This shows that pure C57Bl/6J Atxn1−/− mice are not hypoactive, as reported before for Atxn1−/− mice in a mixed C57Bl/6J/129svEv background. Error bars represent +/− SEM, ***p<0.00001.
(0.44 MB TIF)
Conditioned Fear Analysis to measure Pavlovian learning in Atxn1−/− mice. Seven mice of each genotype were tested at 12 weeks of age. (A) In the contextual conditioned fear test, mice were exposed to a tone paired to a foot shock and 24 hrs later, placed in the same chamber and the amount of freezing behavior is recorded, (B) In the cued test, only the tone is administered in a different chamber and freezing is recorded. Atxn1−/− mice show less freezing both in the contextual and cued test (A and B), indicating amygdala and hippocampal deficits in learning and memory. Error bars represent +/− SEM, *p<0.05,***p<0.001.
(0.27 MB TIF)
Dowel and wire hang analysis to measure gross motor ability in Atxn1−/− mice. In the Dowel test for motor coordination, mice are placed on a 0.9 cm rod and the time that it takes to reach the side and number of side touches in 2 minutes is recorded. Atxn1−/−mice (n = 11) and wild-type littermates (n = 9) were tested at 12 weeks. (A) Atxn1−/− mice performed poorly in the dowel, with increased latency for to reach the sides for the first time. (B) Atxn1−/−mice also made less number of side touches in 2 min. (C and D) The wire hang test is similar to the dowel, except a wire is used instead of a rod. Atxn1−/− mice had a trend to reach the sides fewer times in 2 minutes than the controls (D), albeit not significant (p<0.06) Error bars +/− SEM, *p<0.05,**p<0.01.
(0.52 MB TIF)
List of 197 commonly altered genes in Atxn1−/− and Atxn1154Q/+ cerebella. (p<0.01 Fold Change |+/−0.1| log2 scale).
(0.37 MB DOC)
KEGG pathway analysis of the 197 commonly altered genes in Atxn1−/− and Atxn1154Q/+ cerebella.
(0.06 MB DOC)
We would like to thank Aaron Bowman for providing the Atxn1Ldp mice and guidance on the experimental approach, Maria Chahrour for advice on the chromatin immunoprecipitation assays, and all members of the Zoghbi lab and Marco Sardiello from Andrea Ballabio's lab for their input and valuable discussions. We also thank Lisa White, Laura Liles, and Hillary Lewis from IDDRC Microarray Core Facility at Baylor College of Medicine for their help with the microarray data collection. We also appreciate the help and advice provided by Richard Paylor and Corinne Spencer of IDDRC Behavioral Core at Baylor College of Medicine on the behavioral assays.
Conceived and designed the experiments: JCB JDF HYZ. Performed the experiments: JCB JDF. Analyzed the data: JCB JDF CAS HTO HYZ. Contributed reagents/materials/analysis tools: CAS. Wrote the paper: JCB. Provided substantial input on the manuscript: HYZ JDF HTO.
- 1. Orr HT, Zoghbi HY (2007) Trinucleotide repeat disorders. Annu Rev Neurosci 30: 575–621.
- 2. Gatchel JR, Zoghbi HY (2005) Diseases of unstable repeat expansion: mechanisms and common principles. Nat Rev Genet 6: 743–755.
- 3. Zuhlke C, Riess O, Bockel B, Lange H, Thies U (1993) Mitotic stability and meiotic variability of the (CAG)n repeat in the Huntington disease gene. Hum Mol Genet 2: 2063–2067.
- 4. Keckarevic D, Culjkovic B, Savic D, Stojkovic O, Kostic V, et al. (2000) The status of SCA1, MJD/SCA3, FRDA, DRPLA and MD triplet containing genes in patients with Huntington disease and healthy controls. J Neurogenet 14: 257–263.
- 5. Lorenzetti D, Bohlega S, Zoghbi HY (1997) The expansion of the CAG repeat in ataxin-2 is a frequent cause of autosomal dominant spinocerebellar ataxia. Neurology 49: 1009–1013.
- 6. Sobue G (1995) X-linked recessive bulbospinal neuronopathy (SBMA). Nagoya J Med Sci 58: 95–106.
- 7. Koide R, Ikeuchi T, Onodera O, Tanaka H, Igarashi S, et al. (1994) Unstable expansion of CAG repeat in hereditary dentatorubral-pallidoluysian atrophy (DRPLA). Nat Genet 6: 9–13.
- 8. Yoo SY, Pennesi ME, Weeber EJ, Xu B, Atkinson R, et al. (2003) SCA7 knockin mice model human SCA7 and reveal gradual accumulation of mutant ataxin-7 in neurons and abnormalities in short-term plasticity. Neuron 37: 383–401.
- 9. Watase K, Weeber EJ, Xu B, Antalffy B, Yuva-Paylor L, et al. (2002) A long CAG repeat in the mouse Sca1 locus replicates SCA1 features and reveals the impact of protein solubility on selective neurodegeneration. Neuron 34: 905–919.
- 10. Watase K, Barrett CF, Miyazaki T, Ishiguro T, Ishikawa K, et al. (2008) Spinocerebellar ataxia type 6 knockin mice develop a progressive neuronal dysfunction with age-dependent accumulation of mutant CaV2.1 channels. Proc Natl Acad Sci U S A 105: 11987–11992.
- 11. Yeh S, Tsai MY, Xu Q, Mu XM, Lardy H, et al. (2002) Generation and characterization of androgen receptor knockout (ARKO) mice: an in vivo model for the study of androgen functions in selective tissues. Proc Natl Acad Sci U S A 99: 13498–13503.
- 12. Zeitlin S, Liu JP, Chapman DL, Papaioannou VE, Efstratiadis A (1995) Increased apoptosis and early embryonic lethality in mice nullizygous for the Huntington's disease gene homologue. Nat Genet 11: 155–163.
- 13. Matilla A, Roberson ED, Banfi S, Morales J, Armstrong DL, et al. (1998) Mice lacking ataxin-1 display learning deficits and decreased hippocampal paired-pulse facilitation. J Neurosci 18: 5508–5516.
- 14. Kiehl TR, Nechiporuk A, Figueroa KP, Keating MT, Huynh DP, et al. (2006) Generation and characterization of Sca2 (ataxin-2) knockout mice. Biochem Biophys Res Commun 339: 17–24.
- 15. Tsuda H, Jafar-Nejad H, Patel AJ, Sun Y, Chen HK, et al. (2005) The AXH domain of Ataxin-1 mediates neurodegeneration through its interaction with Gfi-1/Senseless proteins. Cell 122: 633–644.
- 16. Emamian ES, Kaytor MD, Duvick LA, Zu T, Tousey SK, et al. (2003) Serine 776 of ataxin-1 is critical for polyglutamine-induced disease in SCA1 transgenic mice. Neuron 38: 375–387.
- 17. Klement IA, Skinner PJ, Kaytor MD, Yi H, Hersch SM, et al. (1998) Ataxin-1 nuclear localization and aggregation: role in polyglutamine-induced disease in SCA1 transgenic mice. Cell 95: 41–53.
- 18. Lam YC, Bowman AB, Jafar-Nejad P, Lim J, Richman R, et al. (2006) ATAXIN-1 interacts with the repressor Capicua in its native complex to cause SCA1 neuropathology. Cell 127: 1335–1347.
- 19. Mizutani A, Wang L, Rajan H, Vig PJ, Alaynick WA, et al. (2005) Boat, an AXH domain protein, suppresses the cytotoxicity of mutant ataxin-1. EMBO J 24: 3339–3351.
- 20. Serra HG, Duvick L, Zu T, Carlson K, Stevens S, et al. (2006) RORalpha-mediated Purkinje cell development determines disease severity in adult SCA1 mice. Cell 127: 697–708.
- 21. Gold DA, Baek SH, Schork NJ, Rose DW, Larsen DD, et al. (2003) RORalpha coordinates reciprocal signaling in cerebellar development through sonic hedgehog and calcium-dependent pathways. Neuron 40: 1119–1131.
- 22. Jetten AM, Joo JH (2006) Retinoid-related Orphan Receptors (RORs): Roles in Cellular Differentiation and Development. Adv Dev Biol 16: 313–355.
- 23. Lim J, Crespo-Barreto J, Jafar-Nejad P, Bowman AB, Richman R, et al. (2008) Opposing effects of polyglutamine expansion on native protein complexes contribute to SCA1. Nature 452: 713–718.
- 24. Bowman AB, Lam YC, Jafar-Nejad P, Chen HK, Richman R, et al. (2007) Duplication of Atxn1l suppresses SCA1 neuropathology by decreasing incorporation of polyglutamine-expanded ataxin-1 into native complexes. Nat Genet 39: 373–379.
- 25. Carlson KM, Melcher L, Lai S, Zoghbi HY, Brent Clark H, et al. (2008) Characterization of the Zebrafish atxn1/axh Gene Family. J Neurogenet 1–11.
- 26. Serra HG, Byam CE, Lande JD, Tousey SK, Zoghbi HY, et al. (2004) Gene profiling links SCA1 pathophysiology to glutamate signaling in Purkinje cells of transgenic mice. Hum Mol Genet 13: 2535–2543.
- 27. Gatchel JR, Watase K, Thaller C, Carson JP, Jafar-Nejad P, et al. (2008) The insulin-like growth factor pathway is altered in spinocerebellar ataxia type 1 and type 7. Proc Natl Acad Sci U S A 105: 1291–1296.
- 28. Lin X, Antalffy B, Kang D, Orr HT, Zoghbi HY (2000) Polyglutamine expansion down-regulates specific neuronal genes before pathologic changes in SCA1. Nat Neurosci 3: 157–163.
- 29. Kawamura-Saito M, Yamazaki Y, Kaneko K, Kawaguchi N, Kanda H, et al. (2006) Fusion between CIC and DUX4 up-regulates PEA3 family genes in Ewing-like sarcomas with t(4;19)(q35;q13) translocation. Hum Mol Genet 15: 2125–2137.
- 30. Zoghbi HY (1995) Spinocerebellar ataxia type 1. Clin Neurosci 3: 5–11.
- 31. Wells RD, Warren ST, Sarmiento M (1998) Genetic instabilities and hereditary neurological diseases. San Diego, Calif.: Academic Press. 829 p.
- 32. Burright EN, Clark HB, Servadio A, Matilla T, Feddersen RM, et al. (1995) SCA1 transgenic mice: a model for neurodegeneration caused by an expanded CAG trinucleotide repeat. Cell 82: 937–948.
- 33. Lalonde R, Bensoula AN, Filali M (1995) Rotorod sensorimotor learning in cerebellar mutant mice. Neurosci Res 22: 423–426.
- 34. Crawley JN (2007) What's wrong with my mouse? : behavioral phenotyping of transgenic and knockout mice. Hoboken, N.J.: Wiley-Interscience.
- 35. Goold R, Hubank M, Hunt A, Holton J, Menon RP, et al. (2007) Down-regulation of the dopamine receptor D2 in mice lacking ataxin 1. Hum Mol Genet 16: 2122–2134.
- 36. Matsumoto M, Nagata E (1999) Type 1 inositol 1,4,5-trisphosphate receptor knock-out mice: their phenotypes and their meaning in neuroscience and clinical practice. J Mol Med 77: 406–411.
- 37. Huang YH, Dykes-Hoberg M, Tanaka K, Rothstein JD, Bergles DE (2004) Climbing fiber activation of EAAT4 transporters and kainate receptors in cerebellar Purkinje cells. J Neurosci 24: 103–111.
- 38. Aiba A, Kano M, Chen C, Stanton ME, Fox GD, et al. (1994) Deficient cerebellar long-term depression and impaired motor learning in mGluR1 mutant mice. Cell 79: 377–388.
- 39. Thomas PS Jr, Fraley GS, Damian V, Woodke LB, Zapata F, et al. (2006) Loss of endogenous androgen receptor protein accelerates motor neuron degeneration and accentuates androgen insensitivity in a mouse model of X-linked spinal and bulbar muscular atrophy. Hum Mol Genet 15: 2225–2238.
- 40. Van Raamsdonk JM, Pearson J, Rogers DA, Bissada N, Vogl AW, et al. (2005) Loss of wild-type huntingtin influences motor dysfunction and survival in the YAC128 mouse model of Huntington disease. Hum Mol Genet 14: 1379–1392.
- 41. Van Raamsdonk JM, Pearson J, Murphy Z, Hayden MR, Leavitt BR (2006) Wild-type huntingtin ameliorates striatal neuronal atrophy but does not prevent other abnormalities in the YAC128 mouse model of Huntington disease. BMC Neurosci 7: 80.
- 42. Dragatsis I, Levine MS, Zeitlin S (2000) Inactivation of Hdh in the brain and testis results in progressive neurodegeneration and sterility in mice. Nat Genet 26: 300–306.
- 43. Auerbach W, Hurlbert MS, Hilditch-Maguire P, Wadghiri YZ, Wheeler VC, et al. (2001) The HD mutation causes progressive lethal neurological disease in mice expressing reduced levels of huntingtin. Hum Mol Genet 10: 2515–2523.
- 44. Gauthier LR, Charrin BC, Borrell-Pages M, Dompierre JP, Rangone H, et al. (2004) Huntingtin controls neurotrophic support and survival of neurons by enhancing BDNF vesicular transport along microtubules. Cell 118: 127–138.
- 45. Zuccato C, Ciammola A, Rigamonti D, Leavitt BR, Goffredo D, et al. (2001) Loss of huntingtin-mediated BDNF gene transcription in Huntington's disease. Science 293: 493–498.
- 46. Zuccato C, Tartari M, Crotti A, Goffredo D, Valenza M, et al. (2003) Huntingtin interacts with REST/NRSF to modulate the transcription of NRSE-controlled neuronal genes. Nat Genet 35: 76–83.
- 47. Graham RK, Deng Y, Slow EJ, Haigh B, Bissada N, et al. (2006) Cleavage at the caspase-6 site is required for neuronal dysfunction and degeneration due to mutant huntingtin. Cell 125: 1179–1191.
- 48. Cabin DE, Gispert-Sanchez S, Murphy D, Auburger G, Myers RR, et al. (2005) Exacerbated synucleinopathy in mice expressing A53T SNCA on a Snca null background. Neurobiol Aging 26: 25–35.
- 49. Van Broeck B, Van Broeckhoven C, Kumar-Singh S (2007) Current insights into molecular mechanisms of Alzheimer disease and their implications for therapeutic approaches. Neurodegener Dis 4: 349–365.
- 50. Chen Q, Nakajima A, Choi SH, Xiong X, Tang YP (2008) Loss of presenilin function causes Alzheimer's disease-like neurodegeneration in the mouse. J Neurosci Res 86: 1615–1625.
- 51. De Strooper B (2007) Loss-of-function presenilin mutations in Alzheimer disease. Talking Point on the role of presenilin mutations in Alzheimer disease. EMBO Rep 8: 141–146.
- 52. Chahrour M, Jung SY, Shaw C, Zhou X, Wong ST, et al. (2008) MeCP2, a key contributor to neurological disease, activates and represses transcription. Science 320: 1224–1229.