Student Theses and Dissertations

Date of Award

2026

Document Type

Thesis

Degree Name

Doctor of Philosophy (PhD)

Thesis Advisor

Nathaniel Heintz

Abstract

Huntington’ s disease (HD) is a late-onset, autosomal dominant neurodegenerative disorder caused by expansion of a CAG repeat in the HTT gene, yet the mechanisms that couple somatic CAG instability to selective neuronal vulnerability and widespread transcriptional toxicity in the human brain remain incompletely defined. In this thesis, we combine technical advances in fluorescence-activated nuclear sorting (FANS) with integrative molecular profiling of postmortem human brain tissue to delineate a unified pathogenic cascade. This cascade links cell-type-specific regulation of DNA repair pathways and somatic CAG expansion to region- and cell-type-specific epigenetic remodeling and transcriptional dysfunction in HD. By integrating and extending findings from human striatum, cortex, hippocampus, and cerebellum, we define both the cellular selectivity of somatic expansion and the downstream molecular mechanisms by which expanded HTT alleles drive neuronal toxicity. First, we establish and refine FANS-based methods that enable isolation of large numbers of nuclei from defined neuronal and glial populations in human caudate, putamen, cortex, hippocampus, and cerebellum. Antibody- and RNA probe–based nuclear labeling, coupled to deep sequencing methodologies, yields high-purity molecular profiles for striatal direct and indirect pathway medium spiny neurons (dMSNs and iMSNs), cholinergic and additional interneuron classes, multiple glial cell types, deep-layer cortical projection neurons, hippocampal pyramidal neurons and interneurons, cerebellar granule cells, and Purkinje cells. These optimized workflows support quantitative measurement of HTT and ATXN3 CAG tract lengths in defined cell types, alongside matched transcriptomic, genomic and chromatin accessibility datasets that can be analyzed at genome-wide scale. Using these approaches,we define the cell-type specificity of somatic CAG expansion in both HTT and ATXN3 across key brain regions. In the striatum, large somatic expansions of the mutant HTT exon 1 CAG tract occur selectively in dMSNs, iMSNs, and cholinergic interneurons, with mean somatic length gains exceeding 20 repeat units in MSNs and similarly robust expansions in cholinergic interneurons, whereas most interneuron classes and all major glial populations exhibit minimal or no expansion. Parallel analyses in the cerebellum and cortex reveal large expansions in deep layer cortical projection neurons and modest but significant CAG expansions in Purkinje cells compared with cerebellar granule cells and glia, consistent with emerging evidence for Purkinje cell vulnerability in HD. Importantly, in spinocerebellar ataxia type 3 donors, MSNs also show somatic expansion of the mutant ATXN3 CAG repeat, demonstrating that MSNs are intrinsically prone to expanding long CAG tracts at distinct genomic loci and implicating trans-acting mechanisms beyond local sequence context. Together, these data support a model in which extensive somatic expansion of the mutant HTT allele is necessary but not always sufficient for neuronal loss, as illustrated by expansion in cholinergic interneurons and Purkinje cells that are comparatively spared. To explain this cellular specificity, we interrogate cell-type-resolved expression and regulation of DNA mismatch repair (MMR) and other DNA repair pathways highlighted by human HD genome-wide association studies. In control human striatum, MSNs display markedly elevated nuclear levels of MSH2 and MSH3, components of the MutSβ complex, relative to other striatal neurons and glia, whereas FAN1 and additional protective repair factors do not show analogous MSN-specific upregulation. Biochemical assays demonstrate that MutSβ, but not MutSα, potently inhibits endo- and exonucleolytic cleavage of slipped CAG/CTG structures by FAN1, providing a mechanistic basis by which a high MutSβ:FAN1 ratio promotes repeat expansion rather than repair. Integrating FANS-seq with ATAC-seq and H3K27ac profiling, we identify cell-type-specific cis-regulatory elements in MSH2, MSH3, FAN1, POLE, and other HD GWAS candidate genes whose enhancer activity and DNA methylation status explain their divergent expression across neuron and glial populations. Notably, key MSH3 enhancers are present in human MSNs but absent in murine MSNs, underscoring the species-specific regulatory logic that shapes human somatic instability and highlighting human MMR enhancers as potential therapeutic targets. Having established how enhancer-driven control of DNA repair genes sets the propensity for somatic CAG expansion, we next dissect how expansion of the mutant HTT allele triggers downstream epigenetic and transcriptional toxicity. Genome-wide chromatin accessibility analyses reveal that cell types with unstable HTT alleles, most prominently striatal MSNs, undergo extensive, highly cell-type-specific alterations in chromatin structure, with thousands of accessible regions, particularly H3K27ac-positive active enhancers, losing accessibility in HD. These enhancer accessibility losses strongly correlate with reduced expression of their cognate genes, whereas regions gaining accessibility show limited association with active enhancer marks and gene induction. High-resolution mapping of DNA methylation and hydroxymethylation shows that repressed enhancers in HD MSNs accumulate both 5-methylcytosine and 5-hydroxymethylcytosine across CpG and non-CpG contexts, with 5hmC accumulation concentrated in nucleosome-depleted, transcription factor–binding regions. At the same time, transcripts encoding core components of the active DNA demethylation pathway, including TET1 and TDG, are significantly downregulated in HD MSNs, indicating a failure to complete active demethylation and implicating inhibitory cytosine modifications as drivers of enhancer dysfunction. Extending beyond enhancers, integrative analysis of gene body chromatin, H3K27ac, and cytosine modification states in MSNs identifies three mechanistically distinct classes of dysregulated genes. Super-enhancer-regulated genes that define MSN identity and function (for example, PDE10A, PCP4, PHACTR1) show broad H3K27ac domains, near-complete gene body demethylation in controls, and robust expression, yet gain mCG and hmCG in HD alongside enhancer repression, consistent with impaired demethylation across both enhancer and gene body domains. A second set of moderately expressed, non–super-enhancer-regulated genes exhibits coordinated increases in mCG, mCH, and hmCH with loss of hmCG, again correlating with transcriptional repression. A third class of genes is aberrantly derepressed in HD MSNs, characterized by loss of gene-body DNA methylation, implicating reduced binding of neuronal repressors such as MeCP2, and increased expression without corresponding enhancer activation. Many strongly dysregulated genes across these classes are directly linked to haploinsufficient neurological disorders and striatal degeneration in humans, including PDE10A, PDE8B, ANO3, and TAF1, reinforcing the notion that their altered expression has substantial functional consequences for MSN circuitry. Finally, quantitative DNA methylation metrics and single-molecule RNA in situ hybridization demonstrate that these epigenetic and transcriptional abnormalities occur in the majority of surviving MSNs in HD brains, rather than in a rare subset of severely affected cells. FANS-seq–based transcriptomics further reveal MSN-specific repression of nuclear and RNA processing genes, distinct regulation of DNA repair pathways, and induction of autophagy and lysosomal programs that likely reflect adaptive and maladaptive responses to mutant HTT toxicity. Together, these data support a two-phase model of HD pathogenesis in human neurons: in phase 1, cell-type-specific enhancers in MMR and related genes establish a high MutSβ:FAN1 environment in selected neuronal populations, driving somatic expansion of the mutant HTT CAG tract; in phase 2, somatically expanded HTT engages transcriptional machinery) and epigenetic regulators (i.e. TCERG1 and MED15) to impair active DNA demethylation, silence key enhancer- and super-enhancer–regulated identity genes, and aberrantly derepress other gene sets, resulting in widespread transcriptional toxicity across the majority of vulnerable neurons. This integrated framework reconciles somatic repeat instability with selective neuronal vulnerability and highlights convergent therapeutic strategies that combine inhibition of somatic CAG expansion with restoration of neuronal DNA demethylation and enhancer function in the human brain, while also laying the foundation for cell type informed DNA biomarkers for disease progression.

Comments

A Thesis Presented to the Faculty of The Rockefeller University in Partial Fulfillment of the Requirements for the degree of Doctor of Philosophy

License and Reuse Information

Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.

Download

Available for download on Tuesday, April 20, 2027

Included in

Life Sciences Commons

Share

COinS