NR4A1 Methylation Associated Multimodal Neuroimaging Patterns Impaired in Temporal Lobe Epilepsy

Dongmei Zhi, Wenyue Wu, Bo Xiao, Shile Qi, Rongtao Jiang, Xingdong Yang, Jian Yang, Wenbiao Xiao, Chaorong Liu, Hongyu Long, Vince D. Calhoun, Lili Long, Jing SuiFrontiers in Neuroscience 2020 · DOI

Abstract

DNA hypermethylation has been widely observed in temporal lobe epilepsy (TLE). While NR4A1 knockdown has been reported to alleviate seizure severity in mouse models, the underlying methylation-imaging pathway modulated by aberrant NR4A1 methylation levels remains to be clarified in human patients. This study utilizes multi-site canonical correlation analysis with reference (MCCAR) to guide the fusion of three MRI features—functional connectivity (FC), fractional anisotropy (FA), and gray matter volume (GMV)—using blood-based NR4A1 methylation levels as a prior for 56 TLE patients and 65 healthy controls. Post-hoc correlations were evaluated between the identified NR4A1-associated brain components and clinical metrics.

Introduction

Temporal Lobe Epilepsy (TLE) is the most prevalent form of focal epilepsy, often characterized by resistance to standard pharmacological treatments. While genetic factors contribute to TLE risk, they account for only approximately 32% of the variance, highlighting the potential role of epigenetic mechanisms—how environmental factors change gene expression without altering the DNA sequence. DNA hypermethylation of the NR4A1 gene is a promising candidate, as its modulation has shown therapeutic potential in animal models. However, the link between peripheral molecular markers and macroscopic brain network disruptions in humans has remained elusive. This study aims to determine if NR4A1 methylation levels in the blood can serve as a biological "guide" to identify specific patterns of brain impairment in TLE patients.

Methodology: Multimodal Fusion via MCCAR

The study recruited 121 participants, including 56 TLE patients and 65 healthy controls. The TLE group was further divided into right-sided (n=27) and left-sided (n=26) subsets. Researchers employed the Multi-site Canonical Correlation Analysis with Reference (MCCAR) technique to integrate multimodal neuroimaging data guided by NR4A1 methylation levels. Unlike unsupervised methods, MCCAR uses a biological anchor to isolate specific brain networks most sensitive to the epigenetic marker across three modalities: Functional Connectivity (fMRI), Fractional Anisotropy (dMRI), and Gray Matter Volume (sMRI).

Figure 1: Overview of the MCCAR framework, illustrating the fusion of multimodal MRI data guided by NR4A1 methylation levels across all subjects and TLE subsets.

Key Results

The Blood-Brain Epigenetic Pathway

The analysis revealed significantly higher NR4A1 methylation levels in TLE patients compared to controls (0.34 vs 0.28, $p < 10^{-6}$). These elevated levels were quantitatively linked to a specific pattern of multimodal brain impairment:

  • Functional Connectivity (FC): Impaired strength in temporal-cerebellar and occipital-cerebellar circuits.
  • Structural Integrity (FA): Lower fractional anisotropy (decreased white matter integrity) in the cingulum and hippocampus.
  • Morphometry (GMV): Reduced gray matter volume in the thalamus, putamen, temporal pole, and cerebellum.
Figure 2: Detailed functional connectivity (FC) impairments identified through fMRI analysis, showing correlations between different brain regions.

Clinical and Cognitive Correlations

These neuroimaging "loadings" were significantly correlated with clinical outcomes. Higher methylation and associated brain damage predicted earlier disease onset and poorer cognitive performance, as measured by the Minimum Mental State Examination (MMSE) and executive function accuracy (Hayling Sentence Complete Test). Notably, patients with Right TLE exhibited more pronounced brain impairment and cognitive deficits compared to those with Left TLE, particularly showing more reduced GMV in the thalamus.

Discussion and Clinical Implications

These findings establish a potential imaging-methylation pathway where peripheral NR4A1 methylation levels reflect central nervous system disruptions. The involvement of the thalamus and putamen as critical nodes suggests that this epigenetic marker tracks the integrity of networks involved in seizure propagation. The lateralization effects observed underscore the complexity of TLE and the potential for blood-based epigenetic markers to serve as accessible biomarkers for brain health and disease progression.

Limitations and Future Work

The cross-sectional nature of the study precludes definitive causal conclusions regarding whether methylation causes brain changes or vice versa. Future research should focus on longitudinal assessments and broader epigenetic landscapes beyond a single gene to fully map the molecular drivers of TLE. Additionally, larger multi-center studies are needed to validate these imaging-methylation pathways for clinical diagnostics.

Summary updated

Abstract

DNA hypermethylation has been widely observed in temporal lobe epilepsy (TLE), in which NR4A1 knockdown has been reported to be able to alleviate seizure severity in mouse model, while the underlying methylation-imaging pathway modulated by aberrant methylation levels of NR4A1 remains to be clarified in patients with TLE. Here, using multi-site canonical correlation analysis with reference, methylation levels of NR4A1 in blood were used as priori to guide fusion of three MRI features: functional connectivity (FC), fractional anisotropy (FA), and gray matter volume (GMV) for 56 TLE patients and 65 healthy controls. Post-hoc correlations were further evaluated between the identified NR4A1-associated brain components and disease onset. Results suggested that higher NR4A1 methylation levels in TLE were related with impaired temporal-cerebellar and occipital-cerebellar FC strength, lower FA in cingulum (hippocampus), and reduced GMV in putamen, temporal pole, and cerebellum. Moreover, findings were also replicated well in both patient subsets with either right TLE or left TLE only. Particularly, right TLE patients showed poorer cognitive abilities and more severe brain impairment than left TLE patients, especially more reduced GMV in thalamus. In summary, this work revealed a potential imaging-methylation pathway modulated by higher NR4A1 methylation in TLE via data mining, which may impact the above-mentioned multimodal brain circuits and was also associated with earlier disease onset and more cognitive deficits.
BibTeX
@inproceedings{zhi2020nr,
  author = {Dongmei Zhi and Wenyue Wu and Bo Xiao and Shile Qi and Rongtao Jiang and Xingdong Yang and Jian Yang and Wenbiao Xiao and Chaorong Liu and Hongyu Long and Vince D. Calhoun and Lili Long and Jing Sui},
  title = {{NR4A1 Methylation Associated Multimodal Neuroimaging Patterns Impaired in Temporal Lobe Epilepsy}},
  booktitle = {Frontiers in Neuroscience},
  year = {2020},
  doi = {10.3389/fnins.2020.00727},
}
Figure