Circadian rhythms are disrupted in patients and preclinical models of Machado-Joseph disease

Abstract

Machado-Joseph disease (MJD) is caused by an abnormal CAG repeat expansion in the ATXN3 gene, leading to the expression of a mutant ataxin-3 (mutATXN3) protein. Patients with MJD exhibit a wide range of clinical symptoms, including motor incoordination. Emerging evidence highlights circadian rhythm disruptions as early indicators and potential risk factors for the progression of neurodegenerative conditions. Circadian rhythms are regulated by internal clocks, with the suprachiasmatic nucleus (SCN) acting as the master pacemaker to synchronize timing across the body's behavioural and physiological functions. While sleep disturbances have been observed in MJD, the role of clock regulation in its pathophysiology remains largely unexplored in spinocerebellar ataxias. This study aimed to investigate circadian rhythms, characterize associated disruptions and uncover the mechanisms underlying clock dysregulation in patients and preclinical models of MJD. Circadian activity in MJD patients was assessed over 2 weeks using actigraphy, while in a YAC-MJD transgenic mouse model, circadian rhythms were examined through: (i) wheel-running experiments; (ii) telemetry-based monitoring of core body temperature; (iii) immunohistochemical analysis of the neuropeptides arginine vasopressin (AVP) and vasoactive intestinal polypeptide (VIP) in the SCN and paraventricular nucleus (PVN); and (iv) quantitative real-time PCR evaluation of clock gene expression in the cerebellum. The impact of mutATXN3 on clock mechanisms was further investigated using Bmal1/Per2-luciferase reporters. MJD patients exhibited a progressive decline in robustness of behavioural rhythms, demonstrated by negative correlations between the circadian function index, rest-activity fragmentation and sleep efficiency with MJD clinical scales. YAC-MJD mice exhibited reduced activity levels and increased behavioural fragmentation, and they required three additional days to re-entrain after a jet lag protocol compared to controls. Disrupted core body temperature rhythms were observed, including a phase advance and elevated temperature (∼1°C) at the onset of the active period. Furthermore, transgenic mice showed reduced levels of VIP and AVP in the SCN and PVN and decreased clock gene expression in the cerebellum. Lastly, we found new mechanistic evidence that wild-type ATXN3 activates the promoters of Bmal1 and Per2, whereas mutATXN3 loses the capacity to drive Per2 upon polyglutamine expansion. Overall, our findings indicate that central clock dysfunction in MJD is associated with impaired clock gene expression and disruptions in activity and temperature rhythms. This study provides the first robust evidence of circadian rhythm dysregulation and underlying mechanisms in MJD, paving the way for identifying new biomarkers and developing novel circadian-based interventions to tackle MJD and possibly other spinocerebellar ataxias.

Keywords: Machado-Joseph disease; circadian rest-activity rhythms; core body temperature; core clock genes; spinocerebellar ataxia type 3; suprachiasmatic nucleus.

Figure 1

Disease progression in patients with MJD is associated with a circadian rhythm decline. (A) Schematic representation of MJD pathological features development and actigraphy study design. (B) Representative actograms of three MJD patients in different disease states followed for 2 weeks. Participants were classified by SARA score as mildly ataxic (3.0 ≤ SARA ≤ 7.5; dark blue), moderately ataxic (8.0 ≤ SARA ≤ 23.5; purple) or severely ataxic (SARA ≥ 24.0; red). Mild ataxia is represented by Patient MJD #1; moderate ataxia by Patient MJD #6; and severe ataxia by Patient MJD #7. Blue peaks represent activity (PIM) normalized for each patient, light blue sleeping states, light green resting states, light yellow light phases, light grey dark phases and purple off-wrist periods. Black (night) and white (day) bars are defined according to sunrise and sunset times of the study. (C) Representative average daily graphs depicting the hours of higher and lower activity across the different disease states. Mean PIM value calculated for each minute based on the raw data. (D) Correlation of SARA score with circadian function index (CFI), reflecting the overall decline in circadian robustness along the disease course. Data distribution assessed using the Shapiro–Wilk test. Pearson correlation was used to determine the correlation coefficient (r), goodness-of-fit (R²) and P-value. 95% confidence intervals are shown. MJD = Machado-Joseph disease; PIM= Proportional Integration Mode; SARA = Scale for the Assessment and Rating of Ataxia. Created in BioRender. Almeida, L. (2025) https://BioRender.com/oakgzd4.

Figure 2

YAC-MJD transgenic mouse model shows impaired circadian activity rhythms and decreased re-entrainment capacity. (A) Schematic representation of the protocol for wheel-running experiments performed with the YAC-MJD transgenic mouse model under different light-dark conditions. (B) Representative double-plotted actograms of wheel-running activity of 5-month-old male MJD YAC84.2/84.2 (homozygous, n = 3) and wild-type control mice (WT, n = 3) in constant darkness, after 8 days of light-dark (LD) entrainment. Total wheel revolutions determined (C) during the complete study, (D) in 7 h of the inactive period (E) and for two parts of the day. (F) Number of bouts, intervals of 18 consecutive minutes, where wheel-running revolutions were <90. Fragmentation was determined in six representative hours of the active phase during the days of LD entrainment. Homozygous mice showed (F) increased fragmentation and (C–E) significantly reduced levels of activity in specific parts of the day. During the free-running period, under the dark-dark (DD) light scheme, the regression line of the activity on-set was used to determine (G) imprecision, R², representing variability in the start of activity and (H) innate circadian periods of the animals. (I and J) Actograms from mice subjected to a 4-h phase advance jet lag experiment. Representative results of 7-month-old male homozygous (n = 6) compared to WT control mice (n = 5) are shown on the left (I), while 13-month-old male MJD YAC84.2 (hemizygous, n = 3) and WT control mice (n = 3) are shown on the right (J). (K) Homozygous mice required 3.42 additional days to re-entrain to the new LD cycle and (L) hemizygous mice required 3.00 more days for the same re-entrainment, compared to their respective WT control mice. Purple arrows represent the day of re-entrainment. Arrows between two days represent the average quantification by two independent researchers. Data in C–H, K and L are presented as mean ± standard error of the mean. Grey rectangles represent the dark phase. Unpaired t-tests were performed for comparison between groups. *P < 0.05, **P < 0.01, ns = not significant. MJD = Machado-Joseph disease. Created in BioRender. Almeida, L. (2025) https://BioRender.com/k8okexd.

Figure 3

YAC-MJD homozygous mice show a disruption in core body temperature rhythms. (A) Representative scheme of a surgical abdominal implementation of a temperature data logger for core body temperature (CBT) monitoring. (B) Graphical representation of CBT recorded every 5 min for 16 days (n = 4). Lines represent the mean of 7-month-old male MJD YAC84.2/84.2 (homozygous) and wild-type (WT) control mice. (C) Plot of average CBT values along the day. Homozygous mice showed a very evident increase in temperature at the beginning of the active phase. (D) Violin plot of median CBT showing the overall increase of CBT in the homozygous mice compared to WT. Grey rectangles indicate the dark phases. Multiple t-tests were performed for comparison of CBT values between groups at each time point (C), corrected for multiple comparisons using the Holm-Sidak method. CBT median values (D) were compared between groups by an unpaired t-test. *P < 0.05, **P < 0.01; (* to ****) dashed blue lines = *P < 0.05 to ****P < 0.0001. MJD = Machado-Joseph disease. Created in BioRender. Almeida, L. (2025) https://BioRender.com/kl57wrl.

Figure 4

YAC-MJD homozygous mice show reduced levels of the synchronizing neuropeptides vasoactive intestinal polypeptide (VIP) and arginine vasopressin (AVP) in the hypothalamic suprachiasmatic nucleus (SCN) and paraventricular nucleus (PVN). (A) Fluorescent immunohistochemical analysis of AVP and VIP illustrating the levels of these neuropeptides in the SCN, PVN and the supraoptic nucleus (SON) of the hypothalamus. (B) SCN AVP and VIP neuronal projections to the PVN, essential to regulate circadian rhythms. Representative co-immunofluorescence images of AVP (in red) and VIP (in green) in the (E) SCN and (G) PVN of MJD YAC84.2/84.2 (homozygous) and wild-type (WT) mice upon constant darkness. (C) Quantification of VIP and (D) AVP immunoreactivity showing decreased levels of both neuropeptides in the SCN and (H) of AVP in the PVN of homozygous mice compared to control animals. (F and I) A 65% and 78% reduction of AVP-positive cells was observed in the SCN and PVN, respectively, of homozygous mice compared to WT control mice. Data in (C, D, F, H and I) are presented as mean ± standard error of the mean. Values between groups were compared by unpaired t-tests (n = 3 animals per group; >12 PVN and >14 SCN unilateral regions, analysed for each animal). *P < 0.05, **P < 0.01, ***P < 0.001. Scale bars = 500 μm (A), 100 μm (B), 50 μm (E and G). MJD = Machado-Joseph disease.

Figure 5

YAC-MJD homozygous mice show a decline in the expression of core clock genes in the cerebellum. Expression levels of four core clock genes (A) Bmal1, (B) Per2, (C) Clock and (D) Cry1 at three time points during the day (every 8 h at zeitgeber time, ZT: 3, 11, and 19) analysed in the cerebellum of 7-month-old MJD YAC84.2/84.2 (homozygous) and wild-type (WT) control mice (n = 4–5 mice per time point/sex/genotype). Dramatic decreases observed in specific time points of all the analysed genes in the homozygous mice when compared to WT controls. (A) Bmal1 exhibited the most significant dysregulation, with decreases observed at ZT 11 h, ZT 19 h and in average expression levels for the three time points. Expression levels of target genes are shown relative to the reference gene Hprt. Data are presented as mean ΔΔCT values ± standard error of the mean. Grey rectangles indicate the dark phases. Unpaired t-tests were performed to assess differences in clock gene expression between homozygous and WT control mice at each time point of the day*,** and in average expression levels^##. Statistical significance was set as: *P < 0.05, **P < 0.01, ^##P < 0.01. MJD = Machado-Joseph disease. Created in BioRender. Almeida, L. (2025) https://BioRender.com/pdjbfnk.

Figure 6

Wilt-type ATXN3 activates the promoter of Bmal1 and Per2, while mutATXN3 loses the capacity to drive Per2 upon polyglutamine expansion. (A) Schematic representation of the experimental conditions. N2a cells were co-transfected with Bmal1-Luc or Per2-Luc reporter plasmids and human constructs of either mutATXN3 (72 CAG repeats) or wild-type (WT) ATXN3 (27 CAG repeats). Cells were harvested for luciferase assays 24 h after synchronization with 50% horse serum shock. Analysis of interaction with (B) Bmal1 and (C) Per2 promoters by measurement of luciferase bioluminescence in N2a cells. Both WT and mutant forms of ATXN3 drive the promoter of Bmal1; however, only the WT ATXN3 form can drive the transcription of Per2, a capacity that is lost upon polyglutamine expansion in mutATXN3. (B and C) Values are expressed as mean ± standard error of the mean of n = 4. Values between groups are compared with one-way ANOVA followed by Dunnett’s multiple comparisons test. **P < 0.01, ****P < 0.0001, ns = not significant. Created in BioRender. Almeida, L. (2025) https://BioRender.com/cc8fpet.