Reduction of mutant ATXN1 rescues premature death in a conditional SCA1 mouse model

Spinocerebellar ataxia type 1 (SCA1) is a progressive neurodegenerative disease with a typical age of onset in the 30s and initial symptoms consisting of cerebellar ataxia and poor motor coordination (1). This disease is caused by a trinucleotide (CAG) repeat expansion in the first coding exon of the gene ATXN1 (2, 3). The pathogenic polyglutamine–expanded ATXN1 is more stable than its WT counterpart (4). This enhanced stability further promotes its association with native interactors, such as the transcriptional repressor Capicua (CIC) and is therefore thought to, at least in part, explain its toxicity (5, 6).

The bulk of investigation in SCA1 has focused on mechanisms leading to cerebellar degeneration, particularly the degeneration of Purkinje neurons; however, other noncerebellar symptoms such as muscle wasting, breathing and swallowing difficulties, stiffness, sensory abnormalities, and cognitive decline remain understudied (7–9). Of particular interest to us are the pathophysiological mechanisms that promote death in SCA1. Individuals with SCA1 prematurely die from respiratory complications, such as aspiration pneumonia. Weakness in the muscles supporting breathing and safe swallowing is an established major contributor to the development of aspiration pneumonia in a diverse array of neurological disorders. Motor neurons regulate these activities, and when they degenerate, such as in amyotrophic lateral sclerosis (ALS), individuals suffer from premature death by the same respiratory complications as those seen in SCA1 (10).

We have previously hypothesized that premature death in SCA1 is due to motor neuron degeneration (11). Several pieces of evidence support that motor neuron degeneration occurs in SCA1; this evidence includes clinical signs on neurological exams, electrophysiology data and autopsy descriptions (12–17). We and others previously published that there is motor neuron disease in a SCA1 mouse model (11, 18). Furthermore, we demonstrated that this degeneration is time locked to neuromuscular respiratory dysfunction in SCA1 mice (11). In this manuscript, we describe a potentially novel conditional SCA1 mouse model that we created to investigate whether expression of the pathogenic polyglutamine ATXN1 restricted to motor neurons is causative for muscle wasting, neuromuscular respiratory dysfunction, and premature death. We started by crossing this conditional SCA1 line to a ubiquitous Cre line and found that the F1 offspring of this cross recapitulate all the major SCA1 features. However, these mice produced about 50% less of the polyglutamine-expanded ATXN1, resulting in a significant delay in disease onset and progression. Next, we investigated the F1 offspring of our conditional SCA1 mice crossed with a motor neuron–specific Cre line, as well as Purkinje neuron– and skeletal muscle–specific Cre lines to assess the cell-specific effects on survival. For each of these cell-specific crosses, with the reduced polyglutamine-expanded ATXN1 levels produced in our conditional model, we did not recapitulate premature death. We conclude that pathology in more than 1 cell type may be necessary to drive premature death in SCA1.

Generation of a conditional SCA1 mouse model. To investigate cell-specific contributions toward respiratory failure and premature death in SCA1, we set out to create a novel conditional mouse model, Atxn1^{154Q_flox_stop/+} (hereafter referred to as cSCA1). With a conditional SCA1 mouse model, we can probe the effects of the pathogenic polyglutamine protein expression in a cell-specific manner. Our design focused on manipulating the well-established SCA1–knock-in mouse model, Atxn1^154Q/+ (hereafter referred to as SCA1) (19). To generate cSCA1 mice, we utilized the CRISPR/Cas9 system to knock in a floxed stop cassette in the intron just upstream of the first coding exon of Atxn1 (Figure 1A). Due to a unique sequence found in the upstream intron of the CAG repeat–containing allele, but not in the WT allele, we were able to specifically target the CAG repeat–containing allele. Using this approach, we were able to identify 1 founder mouse, a female. We confirmed the correct knock-in sequence and location by Sanger sequencing. To investigate if expression of the CAG repeat–expanded Atxn1 allele was dependent on Cre recombinase, small clippings of the ear from our cSCA1 founder mouse and a control unmodified SCA1 mouse were collected, from which fibroblasts were extracted and grown in cell culture. These cultures were then electroporated with either a control vector or a vector expressing Cre recombinase. Forty-eight hours later, RNA was extracted from these cells, and reverse transcriptase PCR (RT-PCR) using primers that flank the CAG repeats was performed. The expected band size for WT Atxn1 is 306 bp, while that for the CAG repeat Atxn1 allele is 768 bp. We found that fibroblasts from the cSCA1 mouse expressed only the WT Atxn1 allele; however, when expressing Cre, it induced recombination and robust expression of the CAG repeat containing Atxn1 transcript, similar to that of the SCA1 mouse (Figure 1B). This established that our cSCA1 mouse model effectively silenced expression of the pathogenic Atxn1 allele in the absence of Cre recombinase. We next moved in vivo to validate that cSCA1 mice crossed with a ubiquitous Cre driver produced polyglutamine-expanded ATXN1.

Figure 1

Generation of a conditional SCA1 mouse model. (A) Construct design illustrating the targeted location of the flox-stop-flox cassette knock-in using CRISPR/Cas9 tools. The location of the knock-in is in the intron upstream of the first coding exon (exon 7) in the endogenous mouse Atxn1 gene. Specifically, the allele encoding the expanded CAG repeat tract was targeted for knock-in. (B) Reverse transcriptase PCR from primary fibroblast–derived SCA1 and cSCA1 mice. Cells were either electroporated with empty vector or a vector expressing Cre recombinase. cSCA1 cells selectively expressed the expanded CAG repeat allele in the presence of Cre recombinase.

cSCA1 mice produce less polyglutamine-expanded ATXN1 compared with SCA1 mice. We crossed cSCA1 mice to the ubiquitous Cre driver that expresses Cre recombinase by E6.5, Sox2-Cre mice (20). We verified that the cSCA1 × Sox2-Cre mice had complete recombination of the floxed allele and contained the same number of CAG repeats as the SCA1 mice (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.154442DS1). Tissue was collected from WT, SCA1, and cSCA1 × Sox2-Cre mice at 3 weeks of age to quantitatively assess Atxn1 RNA and protein levels. We chose to collect at this time point based on previously published data, together with our own unpublished findings demonstrating that this age yields the highest amount of soluble polyglutamine ATXN1 on Western blot in adult mouse brains (19). Quantitative PCR (qPCR) demonstrated a 2-fold increase in Atxn1 mRNA levels in cSCA1 × Sox2-Cre mice, compared with WT or SCA1 mice (Figure 2A). However, this difference was only seen in the cerebellum, brainstem, and spinal cord — and not in skeletal muscle tissue. To further isolate the pathogenic allele, we crossed these 3 genotypes with Atxn1-KO mice (21). We again observed the same pattern of mRNA expression seen when both the WT and expanded repeat alleles were expressed (Supplemental Figure 2), indicating that cSCA1 × Sox2-Cre express higher mRNA levels of the pathogenic allele than SCA1 mice. At the protein level, however, the pattern flips. We observed that the pathogenic polyglutamine ATXN1 is expressed at half or less that amount in cSCA1 × Sox2-Cre mice compared with SCA1 mice (Figure 2B). Assessing whole brain levels of polyglutamine ATXN1 in cSCA1 × Sox2-Cre compared with SCA1 mice at P1 demonstrated a similar reduction as that seen at 3 weeks of age (Supplemental Figure 3). Cic is a transcriptional repressor and major protein interactor of ATXN1 (5). Previous literature has documented that CIC levels are unchanged in SCA1 mice (6). We assessed CIC levels in whole brain tissue from age- and sex-matched controls, SCA1, and cSCA1 × Sox2-Cre mice at P1 on Western blot. Despite the reduced levels of polyglutamine ATXN1 in cSCA1 × Sox2-Cre mice, there was no significant difference in CIC levels between genotypes (Supplemental Figure 3). These studies demonstrate a significant reduction in levels of the polyglutamine-expanded ATXN1 in the cSCA1 mice compared with SCA1 mice. The inverse relationship at the mRNA level may represent a compensatory upregulation of the pathogenic allele, given its low protein expression (22).

Figure 2

Mutant Atxn1 mRNA and protein levels at 3 weeks of age. Tissue was collected from F1 progeny of cSCA1 × Sox2-Cre mice at 3 weeks of age. Specifically, cerebellum, brainstem, spinal cord, and tibialis anterior skeletal muscle were harvested and divided in half, with one half being used to extract RNA and the other half for protein extraction. For each genotype (WT, SCA1, or cSCA1 × Sox2-Cre [Sox2 cSCA1]), between 3 and 5 biological replicates were used. (A) qPCR was used to quantify mRNA expression of the Atxn1 gene. Statistical assessment performed using 1-way ANOVA, followed by Tukey’s multiple comparisons test **P < 0.01 and ****P < 0.0001. (B) Western blot was utilized to quantify protein levels of either WT ATXN1(2Q) or the polyglutamine expanded ATXN1(154Q) using the 11750 antibody. A representative Western blot is provided for each tissue region, followed by the full quantification of all samples in the bar graph below. ATXN1(2Q) statistical assessment performed using 1-way ANOVA followed by Tukey’s multiple comparisons test *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. ATXN1[154Q] statistical analysis was carried out with Student’s t test.

Spatial resolution in expression of pathogenic polyglutamine–expanded ATXN1. Next, we assessed whether our cSCA1 mice crossed with specific Cre drivers induced the polyglutamine-expanded ATXN1 exclusively in a cell type of interest, and we therefore crossed cSCA1 mice to either a ubiquitous (Sox2), motor neuron–specific (ChAT), skeletal muscle–specific (Ckm), or Purkinje neuron–specific (Pcp2) Cre mouse line (20, 23–25). Purkinje neuron and skeletal muscle Cre drivers were chosen to address the alternative hypotheses that pathology within cell types other than motor neurons drives premature death. We surveyed for intranuclear ATXN1 aggregate formation in aged mice as a proxy for polyglutamine-expanded ATXN1 expression. Previous work demonstrated that nuclear aggregates of ATXN1 begin to form in motor neurons around 12 weeks of age and in Purkinje neurons around 24 weeks of age; therefore, we chose to use aged mice to ensure robust presence of aggregates (11, 19). We harvested whole brain, spinal cord, and skeletal muscle (tibialis anterior) tissue in mice between 36 and 48 weeks of age and performed IHC staining on sections of tissue using an ATXN1 polyclonal antibody (11NQ) (Figure 3). When we assessed CA1 hippocampal neurons, Purkinje neurons, and motor neurons, we found ATXN1 nuclear aggregates in all cell types in the F1 of cSCA1 × Sox2-Cre. As expected, we only detected aggregate formation in motor neurons in the cSCA1 × ChAT-Cre F1s. In the cSCA1 × Pcp2-Cre F1s, we saw intranuclear aggregates restricted to only the Purkinje neurons. Intriguingly, we were not able to detect intranuclear aggregates in skeletal muscle tissue from the F1s of cSCA1 mice crossed with any of the Cre lines, including Ckm-Cre. Quantification of these results demonstrate that, in Purkinje neurons, 68% and 80% of the nuclei had ATXN1 aggregates for cSCA1 × Pcp2-Cre and cSCA1 × Sox2-Cre mice, respectively (Supplemental Figure 4), while 81% of hippocampal CA1 nuclei had aggregates in cSCA1 × Sox2-Cre mice. Furthermore, in motor neurons, 43% and 65% of the nuclei had ATXN1 aggregates for cSCA1 × ChAT-Cre and cSCA1 × Sox2-Cre mice, respectively (Supplemental Figure 4).

Figure 3

Cell type–specific restriction of the pathogenic polyglutamine–expanded ATXN1 protein defined by Cre driver. Tissue was harvested between 36 and 48 weeks of age from cSCA1 mice and F1 offspring from cSCA1 mice crossed with various Cre lines: Sox2 (ubiquitous), ChAT (motor neuron), Ckm (skeletal muscle), and Pcp2 (Purkinje neuron). Specifically, hippocampal CA1 neurons, cerebellar Purkinje neurons, cervical spinal cord motor neurons, and tibialis anterior skeletal muscle cells were imaged after staining with the ATXN1 11NQ antibody. Yellow arrows indicate nuclei with ATXN1 aggregate formation. Nuclear inclusions of ATXN1 protein occurs only in the presence of the polyglutamine expansion; thus, these inclusions are used as a proxy for expression of the recombined pathogenic allele. All images were taken with a 40× objective and then digitally zoomed to allow for optimal visualization of the nuclear inclusions. Scale bars: 50 μm.

As a complementary approach to confirm tissue specificity, we collected cerebellum, brainstem, spinal cord, and tibialis anterior skeletal muscle from the F1s of our cSCA1 mice crossed with the above-mentioned Cre drivers to assess for polyglutamine-expanded ATXN1 via Western blot (Supplemental Figure 5A). SCA1 mice displayed the appropriately sized band for the polyglutamine-expanded ATXN1 in all 4 tissues. As expected, there were reduced levels of the polyglutamine-expanded ATXN1 in the tissues from F1s of cSCA1 × Sox2-Cre mice. However, levels of the polyglutamine-expanded ATXN1 were undetectable in all tissues from F1s of our cSCA1 mice crossed with either ChAT-Cre, Ckm-Cre, or Pcp2-Cre. Given that the ChAT and Pcp2 Cre drivers turn on expression of the pathogenic allele in less than 5% of the total cell populations in these tissues, it’s not surprising that we found undetectable polyglutamine-expanded ATXN1 levels on Western blot. In muscle tissue, the predominate cell type is skeletal muscle; therefore, to confirm the pathogenic allele expression in this tissue from cSCA1 mice crossed with either Sox2-Cre or Ckm-Cre, we performed RT-PCR (Supplemental Figure 5B). We have concluded that the detection of aggregated ATXN1 confirms the expression of the mutant allele in the neuronal cell types assessed and that the RT-PCR confirms its expression in skeletal muscle.

Reduction of polyglutamine-expanded ATXN1 levels delays onset of muscle wasting, respiratory dysfunction, and lethality. Having established that cSCA1 mice make less mutant protein from an early age, we wanted to know how this might affect survival and disease progression, especially within the context of the cell-specific lines. SCA1 mice fail to gain weight beginning as early as 8 weeks of age, and this persists throughout their entire lives (19). In addition, these mice develop muscle wasting with neurogenic changes seen in their skeletal muscle, reduced lifespan, and neuromuscular respiratory weakness (11, 19). We aged SCA1 mice and F1 mice from cSCA1 mice crossed with the various Cre lines, and we measured their total body weight on a weekly basis. As expected, we found SCA1 mice are significantly lighter than age- and sex-matched littermate controls at 32 weeks of age (Figure 4A). By 52 weeks of age, F1 mice, from cSCA1 × Sox2-Cre crosses, demonstrate a robust reduction in weight, as well (Figure 4A). The delay in weight loss can also be seen in the trend lines in Supplemental Figure 6, where SCA1 mice begin diverging in their weights from controls around 12 weeks of age, while the F1 mice from cSCA1 × Sox2-Cre do so around 28 weeks of age. We did not see a significant difference in total body weight in F1 mice of crosses between cSCA1 and either ChAT-Cre, Ckm-Cre, or Pcp2-Cre (Figure 4A and Supplemental Figure 6).

Figure 4

Cell-specific contributions of the polyglutamine-expanded ATXN1 protein on weight and lifespan. (A) Weights of SCA1 mice at 32 weeks of age and F1 offspring of cSCA1 mice crossed with various Cre lines: Sox2 (ubiquitous), ChAT (motor neuron), Ckm (skeletal muscle), and Pcp2 (Purkinje neuron) at 52 weeks of age. WT indicates either WT or Cre alone mice, SCA1 indicates Atxn1^154Q/+ mice, and cSCA1 indicates Cre × Atxn1^{154Q_flox_stop/+} mice. “M” signifies males, and “F” signifies females. SCA1 group: M WT, n = 13, and F WT, n = 9; M SCA1, n = 13, and F SCA1, n = 9. Sox2 group: M WT, n = 12 and F WT, n = 7; M cSCA1 n = 8, and F cSCA1, n = 6. ChAT group: M WT, n = 9, and F WT, n = 7; M cSCA1, n = 9, and F cSCA1, n = 6. Ckm group: M WT, n = 5, and F WT, n = 7; M cSCA1, n = 6, and F cSCA1, n = 6. Pcp2 group: M WT, n = 16, and F WT, n = 6; M cSCA1, n = 12, and F cSCA1, n = 7. Student’s t test. ****P < 0.0001. (B) Survival as measured either at the time of natural death or humane euthanasia in SCA1 and cSCA1 mice crossed with various Cre lines. SCA1 group: WT, n = 22; SCA1, n = 22. Sox2 group: WT, n = 57; cSCA1, n = 49. ChAT group: WT, n = 57; cSCA1, n = 48. Ckm group: WT, n = 12; cSCA1, n = 12. Pcp2 group: WT, n = 49; cSCA1, n = 31.

Individuals with SCA1 and SCA1 mice both demonstrate reduced lifespans (1, 19). We measured survival as either the point in which mice spontaneously die or are requested to be humanely euthanized by our veterinarian staff to avoid unnecessary pain and suffering. Our colony of SCA1 mice display a precipitous drop off in survival with a median survival of ~41 weeks of age (Figure 4B). Congruent with the total body weight data, we found a delay in premature death that was almost twice as long in the F1 mice of cSCA1 × Sox2-Cre mice (median survival ~75 weeks) compared with SCA1-alone mice (Figure 4B). F1 progeny from the cSCA1 mice crossed with the various Cre lines (ChAT-Cre, Ckm-Cre, or Pcp2-Cre) failed to demonstrate significant premature death compared with their control littermates.

We have previously linked decreased total body weight, skeletal muscle atrophy, and premature death with neuromuscular respiratory dysfunction around 24 weeks of age in SCA1 mice using whole body plethysmography (11). Given that only the F1 mice from cSCA1 × Sox2-Cre demonstrate weight loss and premature death, we next wanted to determine whether they also manifest neuromuscular respiratory dysfunction. Indeed, close to the time of premature death at 72 weeks of age, and not earlier at 24 or 48 weeks of age, we found that these mice develop a significant reduction in tidal volume with a compensatory increase in respiratory rate, hallmarks of neuromuscular respiratory weakness (Figure 5). At the earliest time point, 24 weeks of age, we paradoxically found that cSCA1 × Sox2-Cre demonstrated a slight but significant increase in tidal volume compared with their age- and sex-matched littermate controls. In addition to respiratory muscle strength, there are other contributors to tidal volume, including elastic recoil of the lung, airway patency, and chest wall anatomy, and perhaps one of these factors accounts for this unexpected finding. We did not assess breathing physiology in the cell-specific Cre crosses, given that they failed to have a reduction in body weight and lifespan discussed above.

Figure 5

Neuromuscular respiratory failure in cSCA1 mice crossed with Sox2-Cre. Whole-body plethysmography was performed on F1 progeny of cSCA1 × Sox2-Cre mice at 6, 12, or 18 months of age. Sox2 WT, n = 30; Sox2 cSCA1, n = 15. Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Selective reduction of polyglutamine-expanded ATXN1 rescues motor coordination but not impaired locomotion. To examine the effect of restricted pathogenic allele expression on motor function, we employed 2 behavioral assays, open field assay and rotarod assay. Open field assay entails placing a subject mouse in a chamber with sensors that measure how much and where the mouse moves within the chamber over a 30-minute time period (26). SCA1 mice have a reduction in the overall distance traveled when assessed by the open field assay (Figure 6A). As expected, F1 mice from cSCA1 × Sox2-Cre cross demonstrated a similar reduction in distance traveled. Interestingly, we also found a significant reduction in the distance traveled in F1 mice from the cSCA1 × ChAT-Cre cross. F1 mice from the cSCA1 × Ckm-Cre or cSCA1 × Pcp2-Cre crosses did not demonstrate any difference in distance traveled compared with their age- and sex-matched littermate controls. This implies that even the reduced expression of the pathogenic Atxn1 allele, in motor neurons specifically, may be sufficient to induce a motor impairment phenotype of decreased distance traveled. While these reduced levels are not sufficient to cause this impairment in Purkinje neurons or skeletal muscle.

Figure 6

Behavioral effects of cell type–specific restriction of polyglutamine-expanded ATXN1 protein. (A) SCA1 mice and F1 offspring of cSCA1 mice crossed with various Cre lines: Sox2 (ubiquitous), ChAT (motor neuron), Ckm (skeletal muscle), and Pcp2 (Purkinje neuron) were subjected to open field assay at 7 or 15 weeks of age, respectively. Total distance traveled in cm during the assay was plotted. W) indicates either WT or Cre alone mice, and cSCA1 indicates Cre × cSCA1 mice. SCA1 group: WT, n = 22; SCA1, n = 22. ***P < 0.001 with t test. Sox2 group: WT, n = 20; cSCA1, n = 15. **** P < 0.0001 with t test. ChAT group: WT, n = 16; cSCA1, n = 15. **P < 0.01 with t test. Ckm group: WT, n = 13; cSCA1, n = 13. P > 0.05 with t test. Pcp2 group: WT, n = 20; cSCA1, n = 19. P > 0.05 with t test. (B) Retention time on an accelerating rotating rod was measured in SCA1 and cSCA1 mice crossed with various Cre lines at 7 or 11 weeks of age, respectively. SCA1 group: WT, n = 22; SCA1, n = 22. Sox2 group: WT, n = 20; cSCA1, n = 16. ChAT group: WT, n = 17; cSCA1, n = 13. Ckm group: WT, n = 13; cSCA1, n = 13. Pcp2 group: WT, n = 16; cSCA1, n = 13. For each pair, a 2-way ANOVA was calculated with Sidak’s multiple comparisons test, *P < 0.05, **P < 0.01, and ****P < 0.0001.

Next, we assessed for motor coordination using the rotarod assay. In this assay, mice are placed on a rotating rod that is gradually accelerating in speed, and the time on the rod before they fall off (retention time) is measured as a proxy of overall motor coordination (26). SCA1 mice have been previously shown to have an impairment in their performance on the rotarod (19), and we independently confirmed this in our own cohort (Figure 6B). However, we did not find consistent or significant differences in rotarod performance in the F1 progeny of cSCA1 mice crossed with the various Cre lines, compared with their specific age- and sex-matched littermate controls. In addition, we aged a cohort of the progeny from the cSCA1 × Sox2-Cre cross to 36 weeks of age and did not find a significant difference in their rotarod performance compared with controls (Supplemental Figure 7). One caveat to point out is that none of the 4 Cre-alone lines demonstrated the typical improvement over trials days in rotarod performance as shown by WT mice. Previously published work has demonstrated a “Cre effect” on other behavioral assays, and perhaps the expression of Cre is playing a role in our rotarod results (27). The lack of significant rotarod impairment in the F1 mice from cSCA1 × Sox2-Cre and cSCA1 × Pcp2-Cre crosses demonstrates that the reduction in pathogenic allele expression we found in our cSCA1 mice is not adequate to trigger this cerebellar phenotype in SCA1.

Our investigation begins with creating a potentially novel tool, a conditional SCA1 (cSCA1) mouse model, that allows for cell-specific expression of the pathogenic polyglutamine–expanded ATXN1. By driving expression of the toxic polyglutamine–expanded ATXN1 in a specific cell type, we can test what role pathology in this cell type plays on the overall phenotype of disease. In this manuscript, we discuss the generation of our cSCA1 mouse model and find that polyglutamine-expanded ATXN1 levels from the conditional mutant allele are reduced by more than half of those compared with the nonconditional polyglutamine–expanded ATXN1. While this finding was unexpected, we discovered, as the mice aged, that it allowed for a unique possibility: we could examine the effect of significantly reducing mutant ATXN1 levels on disease onset and progression.

cSCA1 mice with ubiquitous expression of polyglutamine-expanded ATXN1 develop declining weight, neuromuscular respiratory dysfunction, and premature death twice as late as the SCA1 mice. Furthermore, we found that this reduced level of toxic polyglutamine–expanded ATXN1 throughout the brain or specifically restricted in Purkinje neurons was not sufficient to produce a rotarod assay deficit. In contrast, this same reduced level of toxic protein in motor neurons was sufficient to drive a decreased distance traveled deficit in the open field assay. Importantly, when the cSCA1 mice were crossed with a motor neuron Cre driver (ChAT-Cre), the offspring failed to develop weight loss and premature death, as they do in the case of ubiquitous expression of the pathogenic allele (cSCA1 × Sox2-Cre). This could suggest that restricted expression of polyglutamine-expanded ATXN1 in motor neurons is not sufficient to drive these features of disease. An alternative explanation could be that the ChAT-Cre line has a lower recombination efficiency than the Sox2-Cre line; thereby, the former might manifest in even lower levels of polyglutamine-expanded ATXN1 in motor neurons than the latter, when crossed with cSCA1 mice. We conclude from these findings that reduction of polyglutamine-expanded ATXN1 levels rescue many disease features of SCA1. This could be due to the degree of reduction we observed, selective reduction of the pathogenic allele while preserving the WT allele level, the fact that the reduction is present early in life, or a combination of all 3.

Prior attempts aimed at reducing ATXN1 levels in mice, such as antisense oligonucleotides (ASO), microRNA (miRNA), RNA interference (RNAi), or altering modifiers of ATXN1 have failed to yield as robust a rescue as we found — in particular, the approximately 34-week extension in lifespan (28–36). In some cases, the reduction of polyglutamine-expanded ATXN1 levels was less than the 50% reduction we found. In other cases, the intervention reduced polyglutamine-expanded ATXN1 levels later in life, compared with our reduction, which was as early as P1. A common theme all these prior interventions share is that they reduced both WT and polyglutamine-expanded ATXN1. In contrast, we have a reduction specifically in the polyglutamine-expanded ATXN1, leaving the WT protein unaffected.

We are actively investigating the mechanism by which cSCA1 mice produce less polyglutamine-expanded ATXN1 compared with SCA1 mice. Results from the F1 offspring generated by crossing cSCA1 to Sox2-Cre mice refute the explanation that this is the result of decreased mRNA expression or incomplete recombination of the floxed stop cassette. While quantification of specific neuronal nuclei with ATXN1 aggregates in cSCA1 × Sox2-Cre mice ranges from 65% to 81%, we do not think this represents incomplete recombination; instead, it may represent a combination of the heterogenous formation of nuclear aggregates, together with the technical limitation of assessing images in 1 focal plane by microscopy. Sanger sequencing of the coding regions of Atxn1 did not demonstrate any novel mutations arising during the generation of the cSCA1 line. However, sequencing of intronic regions did reveal a single base deletion in the intron upstream of the first coding exon of Atxn1 (Supplemental Figure 8). This single nucleotide deletion is found in cis with the expanded repeat allele, unique to cSCA1 mice, and not present in SCA1 or WT mice. We are currently investigating a hypothesis that this mutation might reduce translation efficiency of polyQ ATXN1 via inclusion of a cryptic exon. Our findings suggest that the reduction of polyQ ATXN1 is at the level of protein regulation, either by affecting translation efficiency or stabilizing the nascent protein. By determining the precise mechanism for this reduction in toxic protein, we might learn important lessons that could be applied toward generating allele-specific therapies in SCA1.

Given the findings of this study, another point to consider is the role of misfolded polyglutamine-expanded proteins propagating through the brain. The prion model for neurodegeneration stipulates that misfolded protein aggregates are able to seed and self-propagate transcellularly to different regions of the brain as the disease course advances (37). Aged F1 offspring from our cSCA1 mice crossed with cell-specific Cre lines demonstrated exquisite selectivity in the cell types that formed intranuclear ATXN1 aggregates. For example, when crossed with a Purkinje neuron Cre driver, we found typical intranuclear ATXN1 aggregates in Purkinje neurons, but no aggregates were found elsewhere in the brain, including within deep cerebellar nuclei to which Purkinje neurons project. This would argue against the ability of these protein aggregates to self-propagate throughout the nervous system. A caveat of this may be that the expression of polyglutamine-expanded ATXN1 in our model is low and that perhaps it is not sufficient to propagate aggregates to other neurons. However, we did find that the main driver of pathology was soluble protein levels of the polyglutamine-expanded ATXN1 and not the presence of intranuclear ATXN1 aggregates, in agreement with previous work (38).

Insight from this study strengthens our conceptualization of how pathology arises in SCA1. We propose that the extension of lifespan in the cSCA1 mice is derived not only from expressing less than half the levels of polyglutamine-expanded ATXN1, but also expressing normal WT ATXN1 levels in cSCA1 mice. Additional support for this proposition comes from a prior study in which a serine residue (S776) critical for promoting ATXN1 protein stability after phosphorylation was mutated to an alanine. When this mutation was introduced specifically on the polyglutamine-expanded ATXN1, there was a partial rescue of SCA1 pathology (39). However, when both the WT and polyglutamine-expanded alleles contained the phospho-dead mutation at S776, the mice continued to develop characteristic SCA1 pathology at nearly the same pace as unmodified SCA1 mice (39). Another point of discussion our data raises, which is particularly relevant regarding therapeutic approaches, is the timing of reduced polyglutamine-expanded ATXN1 levels. Our cSCA1 mice presumably exhibit reduced levels of the toxic protein from conception, although the earliest we assessed levels in this manuscript was at P1. It’s not clear whether reduction of polyglutamine-expanded ATXN1 to a similar magnitude later in life would have the same protective effect. Determining the precise amount of reduction in polyQ ATXN1 needed to rescue each cell-specific phenotype, determining when this reduction is needed, and determining whether it needs to be mutant allele specific will be critical for the design of novel and effective gene-based therapies in SCA1 and — potentially — for the broader category of neurodegenerative proteinopathy-based diseases.

Mouse husbandry. SCA1 mice, whose generation and characterization were previously described (19), were backcrossed to C57BL/6J background for a minimum of 10 generations. All mouse lines were kept on a 12-hour light/dark cycle. The following commercial mouse lines were purchased from the Jackson Laboratory: B6.Cg-Edil3Tg(Sox2-cre)1Amc/J, B6;129S6-Chattm2(cre)Lowl/J, B6.FVB(129S4)-Tg(Ckmm-cre)5Khn/J, and B6.129-Tg(Pcp2-cre)2Mpin/J.

Generation of cSCA1 mouse model. SCA1 mice were used to target the pathogenic CAG repeat allele with knock-in of a floxed stop cassette. Specifically, this cassette was inserted upstream of the first coding exon in Atxn1. We targeted sgRNAs to a small segment of unique sequence in this area only found on the CAG-expanded repeat allele and not on the WT allele. Using the “Guide Design Resources” website created by the Zhang lab (https://zlab.bio/guide-design-resources), we selected 4 of the top sgRNAs for the region of interest: regions 2, 8, 11, and 13 (Supplemental File 1). Fertilized embryos from SCA1 mice were collected and microinjected with the mixture of sgRNAs, Cas9 protein, and our donor DNA using established techniques (40). The donor DNA was synthesized from GeneArts (Thermo Fisher Scientific) and contained 3 SV40 polyadenylation transcriptional termination signals concatemerized and floxed by loxP sites (Supplemental File 1). The design of the flox-stop-flox cassette was chosen from previous studies in the literature (41). cSCA1 mice were backcrossed to C57BL/6J background for a minimum of 7 generations.

Fibroblast culture, electroporation, and RT-PCR. Primary fibroblasts derived from ear clippings from a sequence confirmed cSCA1 founder mouse and an unmodified SCA1 mouse. In brief, a clipping of ear tissue was washed in sterile phosphate buffer solution (PBS). Then, the tissue was transferred to a 1 mL solution of 0.25% Trypsin and 4 mg/mL collagenase type 1 (Worthington-Biochem, LS004196) and incubated for 1 hour at 37°C. During this incubation, the sample received 5 pulses of max speed vortex. At the conclusion of the incubation the ear tissue was triturated with a p1000 5 times, and then the entire solution was added to a 10 mm tissue culture plate with 9 mL of DMEM (containing 10% FBS and penicillin and streptomycin antibiotics). The next day, the ear tissue was removed from the plate and the media was exchanged. After 1 week of growth, cells were lifted off the plate with 0.25% Trypsin and replated in a 6-well tray after being electroporated (Nucleofector, Lonza) with either empty vector provided with Lonza kit or a vector expressing Cre recombinase. Forty-eight hours after electroporation, RNA was harvested from cells and DNase treated (TURBO DNA-free Kit Invitrogen, AM1907), and RT-PCR was performed using a primer pair flanking the CAG-expanded repeats 4291-FW PolyQ (5′-ACCTTCCAGTTCATTGGGTC-3′) and 4292-RV PolyQ (5′-GCTCTGTGGAGAGCTGGA-3′).

Western blotting. Homogenates of whole brain in P1 pups or cerebellar, brainstem, spinal cord, or tibialis anterior skeletal muscle tissue from 3-week-old mice (n = 3–5 per genotype) were prepared by Dounce homogenization in NETN buffer (100 mM NaCl, 20 mM Tris-HCl [pH 8.0], 0.5 mM EDTA, 1.5% NP-40, 1× Xpert Protease Inhibitor Cocktail from GenDEPOT and 1× Xpert Phosphatase Inhibitor Cocktail from GenDEPOT). Samples were sonicated for a total of 10 times, incubated for 30 minutes at 4°C, and spun at 11,000g at 4°C for 15 minutes. Protein concentrations of the supernatant were measured using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). Samples were diluted and prepared in NuPAGE sample reducing agent (Invitrogen) and NuPAGE LDS Sample Buffer (Invitrogen). The samples were then boiled for 10 minutes and then run on NuPAGE 4%–12% Bis-Tris gel (1.5 mm, 15-well gels; Invitrogen) or NuPAGE 3%–8% Tris-Acetate Protein Gels (1.5 mm, 15-well gels; Invitrogen) in MES running buffer (50 mM MES, 50 mM Tris base, 0.1 % SDS, 1 mM EDTA). The proteins were subsequently transferred to Immobilon-FL PVDF membranes (Invitrogen, 0.45 μm). After blocking for 1 hour at room temperature with 1:1 Odyssey Blocking Buffer TBS (LI-COR Biosciences) in TBS, membranes were probed overnight at 4°C with anti-ATXN1 (11750VII in house, 1:2000; ref. 42), anti-GAPDH (MilliporeSigma, 6C5, 1:10,000), anti-Vinculin (Sigma-Aldrich, V9131, 1:5,000) or anti-Cic (1:1,000 in house; ref. 5) in 1:1 Odyssey Blocking Buffer and TBS (LI-COR Biosciences) in Tris-buffered saline (5 mM Tris [pH 7.5], 120 mM NaCl) with 0.1% Tween-20 (TBST). The secondary antibodies used were anti–rabbit IgG (H&L) DyLight 680–conjugated (Rockland Immunochemicals, 611-144-002, 1:10,000) and anti–mouse IgG (H&L) DyLight 800–conjugated (Rockland Immunochemicals, 610-145-002, 1:10,000) in 1:1 Odyssey Blocking Buffer and TBS (LI-COR Biosciences) in TBST. The membranes were washed 3 times with TBST and then imaged using the Odyssey CLx Imaging System (LI-COR Biosciences).

qPCR. RNA from cerebellum tissue was isolated from 3-week-old mice (n = 3–5 per genotype). Total RNA was isolated using TRIzol from Invitrogen following the manufacturer’s instructions. RNA was quantified using NanoDrop 1000 (Thermo Fisher Scientific), and random-primed cDNA was prepared from 1 μg total RNA using M-MLV Reverse Transcriptase (Invitrogen). qPCR was then performed with PowerUp SYBR Green Master Mix (Applied Biosystems), and samples were run on a real-time PCR detection system (Bio-Rad, CFX96). All samples were analyzed in triplicate, and Atxn1 expression levels were normalized to the expression of the housekeeping gene Gapdh. Primers for Gapdh and Atxn1 were obtained from MilliporeSigma. Primers used for Gapdh were Gapdh_FW (5′-AGGTCGGTGTGAACGGATTTG-3′) and Gapdh_RV (5′-TGTAGACCATGTAGTTGAGGTCA-3′), and primers used for Atxn1 were Atxn1_FW (5′-GAGAATCGAGGAGAGCCAC-3′) and Atxn1_RV (5′- AGACTTCGACACTGACCTG-3′).

IHC. Mouse tissue was collected and fixed for 24 hours in 10% formalin. Standard techniques were used to embed tissue in paraffin and to take 10 μm sections in serial section every 100 μm. Sections were deparaffinized and rehydrated. Slides were treated with antigen retrieval by boiling for 9 minutes in 10 mM sodium citrate, 0.05% Tween 20, pH 6.0. The ATXN1 11NQ antibody, generated in the Zoghbi laboratory (38), was incubated at a concentration of 1:2000 for approximately 12 hours at 4°C. Primary antibody was detected with biotinylated goat anti–rabbit IgG (1:1000) (Jackson ImmunoResearch Laboratories) and visualized using an ABC reagent kit (Vector Laboratories) according to the manufacturer’s recommendations. All IHC experiments were performed in triplicate, and figures show representative results. Images were collected with an Echo Revolve light microscope. For quantification analysis, a minimum of 25 nuclei per mouse were counted within a region of interested determined by the anatomical location and morphology of the cells. The percentage of those cells with positive ATNX1 intranuclear aggregates was averaged from 3 mice per genotype, and then the mean ± SEM was plotted.

Behavioral assays. All behavioral tests were performed during the light phase of the 12-hour light/dark cycle by an experimenter blinded to the genotype of the mice. The mice had access to food and water ad libitum, except during the assay. All mice were age- and sex-matched within experiments, and littermate controls were used when possible. For each test, the mice were habituated for 30 minutes in the test room before testing. Testing was done at a room brightness of 200 lux with white noise playing at 60 dB.

Open field assay. Seven- or 15-week-old mice were habituated for 30 minutes in the test room at 200 lux with white noise playing at 60 dB. Each mouse was placed singly in the clear Plexiglas box of the open field apparatus (OmniTech Electronics) and allowed to move freely for 30 minutes. Total distance traveled among other parameters were recorded using Fusion activity monitoring software.

Rotarod test. The rotarod test was performed at 7 or 11 weeks of age to evaluate coordination and motor skill acquisition (type 7650; Ugo Basile). The mice were placed on the rotating rod (3 cm diameter, 30 cm long) for 4 trials every day for a period of 4 days. Each trial lasted a maximum of 10 minutes. The rod accelerated from 4 to 40 rpm in 5 minutes and remained at 40 rpm for the remaining 5 minutes. The amount of time it took for the mice to fall was recorded. Two subsequent rotations around the rod were also counted as a fall.

Whole body plethysmography. We measured the breathing phenotypes as previously described (11, 43). In short, mice were placed in Buxco whole-body plethysmography chambers and received fresh room at a rate of 0.5 L/min. Breath waveforms were recorded using Phonemah 3 software. We derived breathing parameters (tidal volume, frequency, and minute volume) using custom written MATLAB (Mathworks) code. We only included breaths recorded after the mice were able to habituate to the recording chambers for 1 hour. We reduced the inclusion of artifacts from sniffing or moving by excluding breaths that had an inspiratory time under 0.025 seconds, expiratory time over 10 seconds, or a 2-fold disparity between expiratory and inspiratory tidal volume. We further filtered out periods at which more than 10% breaths were taken at a frequency higher than 600 breaths/min. Only mice from which we recorded more than 150 reliable breaths were included in our statistical analysis. Statistical significance was determined using a 2-tailed t test at each time point.

Statistics. All data are represented as the mean ± SEM. When possible, experimental analysis was performed in a manner that blinded investigators to the genotype of the animal. Statistical tests were carried out in accordance with the experimental design. Two-group comparisons used Student’s 2-tailed t test, whereas multigroup comparisons used 1- or 2-way ANOVA. Survival analysis used log-rank test. In each case, P < 0.05 was considered statistically significant.

Study approval. Animal care and experimental procedures were approved by the IACUC of Baylor College of Medicine, according to NIH guidelines (Guide for the Care and Use of Laboratory Animals, National Academies Press, 2011).