Methylome analysis of FTLD patients with TDP-43 pathology identifies epigenetic signatures specific to pathological subtypes

TL;DR

This study finds that different types of FTLD-TDP have unique DNA methylation patterns, suggesting they are epigenetically distinct.

Contribution

The study identifies subtype-specific epigenetic signatures in FTLD-TDP using the largest DNA methylation dataset to date.

Findings

Thousands of differentially methylated CpGs and hundreds of DMRs were identified in FTLD-TDP brains.

Most methylation changes are unique to specific FTLD-TDP subtypes, indicating distinct epigenetic profiles.

Hypomethylation of CAMTA1 was validated in TDP-A, with potential impacts on other genes in the locus.

Abstract

In the last decade, the importance of DNA methylation in the functioning of the central nervous system has been highlighted through associations between methylation changes and differential expression of key genes involved in aging and neurodegenerative diseases. In frontotemporal lobar degeneration (FTLD), aberrant methylation has been reported in causal disease genes including GRN and C9orf72; however, the genome-wide contribution of epigenetic changes to the development of FTLD remains largely unexplored.

We performed reduced representation bisulfite sequencing of matched pairs of post-mortem tissue from frontal cortex (FCX) and cerebellum (CER) from pathologically confirmed FTLD patients with TDP-43 pathology (FTLD-TDP) further divided into five subtypes and including both sporadic and genetic forms (N = 25 pairs per group), and neuropathologically normal controls (N = 42 pairs). Case-control differential methylation analyses were performed, both at the individual CpG level, and in regions of grouped CpGs (differentially methylated regions; DMRs), either including all genomic locations or only gene promoters. Gene Ontology (GO) analyses were then performed using all differentially methylated genes in each group of sporadic patients. Finally, additional datasets were queried to prioritize candidate genes for follow-up.

Using the largest FTLD-TDP DNA methylation dataset generated to date, we identified thousands of differentially methylated CpGs (FCX = 6,520; CER = 7,134) and several hundred DMRs in FTLD-TDP brains (FCX = 134; CER = 219). Of these, less than 10% are shared between pathological subgroups. Combining additional datasets, we identified, validated and replicated hypomethylation of CAMTA1 in TDP-A potentially also impacting additional genes in the locus. GO analysis further implicated DNA methylation in myelination and developmental processes, as well as important disease-relevant mechanisms with subtype specificity such as protein phosphorylation and DNA damage repair in TDP-A, cholesterol biosynthesis in TDP-B, and protein localization in TDP-C.

We identify methylation changes in all FTLD-TDP patient groups and show that most changes are unique to a specific pathological FTLD-TDP subtype, suggesting that these subtypes not only have distinct transcriptomic and genetic signatures, but are also epigenetically distinct. Our study constitutes an invaluable resource to the community and highlights the need for further studies to profile additional epigenetic layers within each FTLD-TDP pathological subtype.

The online version contains supplementary material available at 10.1186/s13024-025-00869-2.

Figures