TL;DR
This study uses cryo-electron tomography (cryoET) to reveal previously unseen organelle abnormalities in neurons from Huntington's disease (HD) models, identifying enlarged mitochondrial granules and sheet aggregates in autophagic organelles as early disease markers. These findings, combined with proteomics and AI-driven quantification, offer new structural insights into HD pathogenesis and potential therapeutic targets.
What Problem Does This Paper Address?
Huntington's disease (HD) is a devastating neurodegenerative disorder caused by an expanded CAG repeat in the huntingtin gene, leading to a toxic huntingtin protein. While patient-derived induced pluripotent stem cells (iPSCs) offer a valuable model for studying HD, identifying early, reliable pathological biomarkers has been challenging. Existing models often require specific treatments to show overt disease phenotypes, making it difficult to study the earliest stages of the disease. This paper aims to address this by using advanced cryo-electron tomography (cryoET) to visualize the ultrastructure of neurons from both human iPSC-derived and mouse models of HD, looking for subtle, early organelle changes that might serve as biomarkers.
Methods
The researchers employed cryogenic electron tomography (cryoET), a high-resolution imaging technique that allows for the 3D reconstruction of cellular structures in a near-native state. This was applied to neurites of:
- Six human iPSC-derived neuronal cell lines with varying CAG repeat lengths (Q18 to Q109).
- Primary cortical neurons from BACHD, deltaN17-BACHD, and wild-type (WT) mouse models.
To analyze the vast amount of cryoET data, they developed a semi-automated, AI-based 3D segmentation method using convolutional neural networks (specifically, a 3D-UNet architecture) to detect and quantify mitochondrial granules.
Additionally, they performed mitochondrial proteomics on isolated mitochondria from HD patient iPSC-neurons and controls to identify differences in protein content.
Finally, they investigated the therapeutic potential of Protein Inhibitor of Activated STAT1 (PIAS1) knockdown, a protein previously linked to proteostasis and synaptic function in HD, by evaluating its effect on the observed organelle phenotypes.
Datasets
The study utilized:
- Human iPSC-derived neurons: Six lines with endogenous CAG repeat lengths ranging from normal (Q18, Q20) to pathogenic (Q53, Q66, Q77, Q109).
- Mouse primary neurons: From BACHD (full-length human mHTT with expanded polyQ), deltaN17-BACHD (lacking the first 17 amino acids of HTT), and wild-type (WT) control mice.
- Proteomics data: From isolated mitochondria of HD patient iPSC-neurons (Q109) and control iPSC-neurons (Q18).
No new benchmarks were introduced; rather, existing models were used for cryoET data collection and analysis.
Key Insights and Ideas
- Early structural hallmarks of HD: The study identified two novel, early structural abnormalities in HD neurons: enlarged granules within mitochondria and sheet-like aggregates within double membrane-bound autophagic organelles. These were consistently observed across both human and mouse HD models, suggesting they are fundamental to the disease process.
- Mitochondrial granules as potential RNA granules: The enlarged mitochondrial granules, particularly in higher polyQ lines, are hypothesized to be mitochondrial RNA granules (MRGs). This is supported by proteomics data showing altered RNA-binding proteins and the successful reduction of granule size upon knockdown of GRSF1, a key protein involved in MRG formation.
- Sheet aggregates in autophagic organelles: These aggregates, distinct from mitochondrial granules, appear to be a consequence of cellular stress and toxicity induced by mutant huntingtin. Their presence in autophagic organelles suggests a potential disruption in cellular waste clearance pathways.
- AI for quantitative structural analysis: The development of an AI-based segmentation tool for quantifying mitochondrial granules demonstrates a powerful approach to objectively characterize ultrastructural changes, overcoming the limitations of manual annotation.
- Therapeutic potential of PIAS1 modulation: Knockdown of PIAS1 (or its mouse homolog Pias1) ameliorated the enlarged mitochondrial granules in both human and mouse HD models, highlighting PIAS1 as a potential therapeutic target.
Key Results
- Mitochondrial abnormalities: HD neurons consistently showed enlarged, dense granules within mitochondria and distorted cristae. The Q109 line exhibited the most severe mitochondrial dysmorphology.
- Sheet aggregates: These were observed in the neurites of all HD patient and mouse model neurons, but were absent in control neurons. They were described as composed of densely interwoven slab-shaped and sheet-like structures.
- Proteomic differences: Mitochondrial proteomics revealed 177 differentially enriched peptides between HD (Q109) and control (Q18) iPSC-neurons, with 124 unique proteins depleted in HD. Notably, RNA-binding proteins were significantly altered.
- AI quantification: The AI algorithm successfully quantified mitochondrial granule volumes, showing a significant increase in HD neurons compared to controls. This enlargement was rescued by PIAS1 hetKO in iPSC-neurons and partially by Pias1 KD in mouse neurons.
- GRSF1 knockdown: Reducing GRSF1 in Q66 HD neurons led to a decrease in mitochondrial granule size and a significant increase in granule number per mitochondrion.
- PIAS1 intervention: PIAS1 hetKO in HD iPSC-neurons rescued both enlarged mitochondrial granules and sheet aggregates. Pias1 KD in mouse neurons reduced enlarged mitochondrial granules but not sheet aggregates.
Limitations and Future Work
- Nature of sheet aggregates: While the study provides strong evidence for their presence, the precise biochemical composition and origin of the sheet aggregates remain to be fully elucidated. Future biochemical analyses of purified aggregates are suggested.
- Cell-type specificity: The differential effects of Pias1 KD in human vs. mouse neurons warrant further investigation to understand cell-type specific responses.
- Mechanism of granule formation: While hypothesized to be MRGs, the exact mechanism of their aberrant accumulation and their precise chemical composition (e.g., RNA vs. calcium phosphate) require further study, potentially using elemental analysis with electron microscopy.
- Therapeutic translation: While PIAS1 modulation showed promise, further studies are needed to determine its efficacy and safety in more complex models and eventually in clinical settings.
- Broader applicability of cryoET: The study advocates for cryoET as a powerful tool for early disease phenotyping and preclinical evaluation of therapeutics, suggesting its application to other neurodegenerative disorders.
Summary updated