Brain Inflammation Emerges Years Before Symptoms in Spinocerebellar Ataxia Type 3, Landmark Imaging Study Finds

Spinocerebellar ataxia type 3, the most common hereditary ataxia in the world, has long been treated as a disease that announces itself with a stumble. A new study from a Chinese research consortium suggests that the real story begins far earlier, deep inside the brainstem, where an invisible wave of neuroinflammation appears to ignite years before a single symptom of unsteadiness emerges. The findings, published in Annals of Clinical and Translational Neurology, offer the most detailed multimodal map yet of how this fatal neurodegenerative disorder unfolds across the human lifespan, and they arrive at a moment when the first targeted therapies are finally entering clinical trials.

SCA3, also known as Machado-Joseph disease, is caused by an unstable expansion of CAG repeats in the ATXN3 gene, which produces a misfolded ataxin-3 protein that clumps inside the nuclei of neurons. The toxic aggregates gradually destroy neural circuits in the cerebellum, brainstem, and basal ganglia, producing progressive loss of coordination, pyramidal signs, parkinsonism, and a constellation of non-motor symptoms. There is currently no causal treatment, although antisense oligonucleotides designed to silence the ATXN3 gene are now being tested in humans. That therapeutic milestone has created an urgent demand for biomarkers capable of tracking the disease’s silent evolution, since preventive trials will need objective measures that can identify the right patients at the right time.

To build that map, a team led by researchers at the First Affiliated Hospital of Fujian Medical University recruited 229 genetically confirmed SCA3 patients and 104 age- and sex-matched healthy controls through the OSCCAR ataxia research network in south-east China between February 2019 and February 2023. Thirty-eight of the patients, roughly one in six, were still preataxic, meaning they carried the mutation but had not yet developed measurable gait disturbance. The cohort was remarkable in its breadth: 125 patients completed neuropsychological testing, 208 underwent structural MRI at 3 Tesla, and 50 received positron emission tomography with the radiotracer 18F-DPA-714, which binds to the translocator protein, or TSPO, a well-established molecular signature of activated microglia, the brain’s resident immune cells.

The neuropsychological results revealed a striking temporal choreography of non-motor symptoms. Cognitive performance on the Montreal Cognitive Assessment, anxiety scores on the Hamilton Anxiety Scale, and fatigue ratings on the Fatigue Scale-14 all diverged from healthy control values within roughly four to five years after ataxia onset. Depression scores crossed the clinical threshold around 6.3 years after onset, and sleep quality deteriorated last, at approximately 11.2 years. Excessive daytime sleepiness, measured by the Epworth Sleepiness Scale, never left the normal range across the entire disease course. The researchers used additive regression models with cubic P-splines to reconstruct these trajectories on a continuous time axis anchored to disease duration, allowing them to pinpoint precisely when each symptom domain crossed validated clinical cutoffs.

Structural MRI told an equally compelling story, but with a different timeline. Volume loss in the brainstem and ventral diencephalon fell below the normal range roughly 0.7 and 0.2 years before ataxia onset, respectively, while cerebellar white matter atrophy and pallidum shrinkage only became detectable after symptoms appeared. The brainstem volume alone distinguished SCA3 patients from controls with an area under the receiver operating curve of 0.9639, and the ventral diencephalon achieved 0.9284, although discrimination was weaker when only preataxic carriers were compared with controls. Volumes in the cerebellar gray matter, putamen, caudate, and thalamus declined gradually but remained within two standard deviations of the control mean throughout the observed disease course, suggesting that these regions are drawn into the degenerative process later and more slowly.

The PET findings upended conventional assumptions about the sequence of pathology. Uptake of 18F-DPA-714, normalized to the cerebellar cortex as a pseudo-reference region, was significantly elevated in the brainstem, ventral diencephalon, pallidum, and thalamus compared with controls, and reduced in the caudate. Crucially, the brainstem signal was already abnormal in preataxic carriers, before any measurable atrophy, and it remained elevated throughout the entire disease course. The brainstem SUVR distinguished patients from controls with an area under the curve of 0.97 overall and 0.90 in the preataxic subgroup, outperforming structural MRI in the earliest phase. The researchers verified that raw cerebellar cortical uptake was stable across groups after adjusting for age and sex, ruling out a reference-region artifact as the source of the signal.

Synthesizing these trajectories, the team proposed a four-stage model of SCA3 progression. The first stage, the TSPO-PET biomarker stage, encompasses preataxic carriers with elevated brainstem and ventral diencephalon inflammation but preserved brain volumes. The second, the MRI biomarker stage, adds measurable regional atrophy while patients remain clinically preataxic. The third stage is defined by manifest ataxia with a SARA score of 3 or higher and preserved global cognition. The fourth and final stage marks the emergence of clinically significant non-motor symptoms, indicated by a MoCA score below 26, alongside progressive basal ganglia and thalamic volume loss and sustained inflammatory signaling. The model parallels biomarker-driven staging frameworks developed for Alzheimer’s disease and Huntington’s disease, positioning SCA3 within a growing family of neurodegenerative disorders whose natural histories can now be quantified before symptoms appear.

The biological interpretation carries significant implications for drug development. The authors argue that neuroinflammation in SCA3 is not merely a byproduct of dying neurons but an active, early participant in the pathogenic cascade, consistent with neuropathological evidence of glial activation in the pons of both patients and mouse models. Preclinical studies have shown that reducing neuroinflammation ameliorates neuropathology and improves motor coordination in Machado-Joseph disease mice, and animal experiments demonstrate that 18F-DPA-714 uptake is reversible following anti-inflammatory or immunomodulatory interventions. This reversibility makes TSPO-PET a candidate pharmacodynamic endpoint for future disease-modifying trials, allowing investigators to confirm that an experimental therapy is actually engaging its target in the brain.

The study also situates TSPO-PET alongside fluid biomarkers such as neurofilament light chain, which likewise shows pronounced abnormalities in the preataxic phase before plateauing in manifest disease. The authors interpret this pattern as reflecting the biology of dynamic rate markers: before phenoconversion, a large pool of still-viable but stressed axons sheds NfL rapidly, while mutant ataxin-3 aggregation triggers an equally rapid microglial response. As the disease advances and the axonal pool shrinks, both signals settle into a chronic plateau. This shared trajectory underscores why early-stage stratification, rather than late-stage monitoring, may be where these biomarkers deliver their greatest clinical value.

The authors acknowledge important limitations. The design was cross-sectional, so temporal trajectories were reconstructed statistically rather than observed within individuals over time. TSPO is expressed not only by microglia but also by reactive astrocytes and endothelial cells, limiting cellular specificity, and tracer binding depends on the rs6971 polymorphism, although nearly 96 percent of this East Asian cohort were high-affinity binders. The staging cutoffs, while clinically grounded, may miss subtler changes, and the cohort was ethnically and geographically homogeneous. Replication in independent, multicenter cohorts will be essential. Nevertheless, the framework represents a concrete step toward precision medicine in SCA3, offering trial designers a molecularly grounded timeline for deciding when to intervene, which patients to enroll, and which endpoints to trust.

Subject of Research: Stage-specific patterns of neuroinflammation, brain atrophy, and non-motor symptoms in spinocerebellar ataxia type 3

Article Title: Stage‐Specific Patterns of Imaging and Non‐Motor Biomarkers in Spinocerebellar Ataxia Type 3

Article References: Li, C., Ye, Z.-X., Chen, X.-Y., Qiu, L.-L., Chen, Z.-L., Qiu, Y.-S., Lin, W., Yuan, R.-Y., Yu, X.-T., Huang, C.-Y., Chen, X.-Y., Lin, M.-T., Hu, J.-P., Chen, W.-J., Wang, N., Yao, S.-B., Gan, S.-R., & Fu, Y. (2026). Stage‐Specific Patterns of Imaging and Non‐Motor Biomarkers in Spinocerebellar Ataxia Type 3. Annals of Clinical and Translational Neurology, Article acn3.70555. https://doi.org/10.1002/acn3.70555

Image Credits: AI Generated

DOI: 10.1002/acn3.70555

Keywords: spinocerebellar ataxia type 3, Machado-Joseph disease, ATXN3, neuroinflammation, TSPO PET, 18F-DPA-714, biomarkers, MRI atrophy, microglial activation, non-motor symptoms, disease staging, clinical trials