Exercise extends lifespan in Atxn1^154Q mice by upregulating EGF and consequently downregulating of Cic. (A) Median survival was 272 days for non-exercised Atxn1^154Q mice vs 317 days for exercised Atxn1^154Q mice (45 day extension, N = 6 per group, p <0.01 by log-rank test). (B) Exercise caused a 50% increase in brainstem EGF levels in WT mice and a 74% increase in Atxn1^154Q mice measured by ELISA (mean ± SEM, N = 5 per group). (C) Exercise caused a 24% and a 22% decrease in brainstem Cic levels in WT and Atxn1^154Q mice, respectively measured by RT-qPCR (mean ± SEM, N = 5 per group). (D) Western blotting for Cic demonstrates an exercise-induced decrease in Cic levels in the brainstem while Atxn1 or Atxn1^154Q remain unaffected. (E) Primary brainstem neuronal cultures treated with recombinant EGF for 72 hours show a dose-dependent decrease in the level of Cic but not Atxn1.

Reduction of Cic rescues multiple phenotypes in SCA1 mice. A 50% reduction of Cic in Atxn1^154Q mice rescued Purkinje cell integrity (A and B, N = 4 per genotype), body weight (C), and premature lethality (D, p < 0.05 by log-rank test). The median lifespan of Atxn1^154Q mice in this cohort is in close agreement with the median lifespan from the unexercised Atxn1^154Q cohort (Figure 1), with some increased variability due to greater numbers of litters being used in this larger cohort. For body weight and lethality, N=11 for WT, N = 12 for Cic-L^+/- mice, N=14 for Atxn1^154Q mice, N= 16 for Atxn1^154Q; Cic-L^+/- mice.

Reduction of Cic rescues multiple behaviors in SCA1 mice. Reduction of Cic by 50% in Atxn1^154Q mice resulted in motor improvements in the open field assay (A), dowel test (B), rotarod (C), and conditioned fear assay of learning and memory (D). Legend for each panel indicated in Panel B. Values represent mean ± SEM with N= 9-10 per genotype. ANOVA with post-hoc t-tests performed where # or ## indicates p < 0.05 or p < 0.01 between WT and Atxn1^154Q mice, and * or ** indicates p < 0.05 or 0.01 between Atxn1^154Q and Atxn1^154Q; Cic-L^+/- mice.

Atxn1^154Q causes concomitant gain and loss of Cic transcriptional function. (A) Atxn1^154Q hyper-repressed cerebellar transcripts containing a Cic motif, but these were normalized to near WT levels in Atxn1^154Q; Cic-L^+/- mice measured by RT-qPCR (mean ± SEM, N = 6 per genotype) with more Cic was bound to their promoters in Atxn1^154Q vs WT littermates as measured by chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR, mean ± SEM, N = 4 per genotype). (B) Atxn1^154Q also de-repressed cerebellar transcripts containing a Cic motif, but these transcripts were further upregulated in Atxn1^154Q; Cic-L^+/- mice by RT-qPCR (mean ± SEM, N = 6 per genotype), with less Cic was bound to their promoters in Atxn1^154Q vs WT littermates measured by ChIP-qPCR (mean ± SEM, N = 4 per genotype). Data analyzed by ANOVA with post-hoc t-test where * indicates p < 0.05 between WT and Atxn1^154Q and # indicates p < 0.05 between Atxn1^154Q and Atxn1^154Q; Cic-L^+/- mice.