White Matter Imaging Phenotypes Mediate the Negative Causality of Mitochondrial DNA Copy Number on Sleep Apnea: A Bidirectional Mendelian Randomization Study and Mediation Analysis

Qiaohui Ying,1,2,* Mingwei Wang,3,* Zichen Zhao,1,2 Yongwei Wu,1,2 Changyun Sun,1,2 Xinyi Huang,1,2 Xin Zhang,1,2 Jie Guo1,2 1Department of Orthodontics, School and Hospital of Stomatology, Cheeloo College of Medicine, Shandong University, Jinan, Shandong, People’s Republic of China; 2Shandong Provincial Clinical Research Center for Oral Diseases, Shandong University, Jinan, Shandong, People’s Republic of China; 3Institute of Blood Transfusion, Chinese Academy of Medical Sciences and Peking Union Medical College, Chengdu, Sichuan, People’s Republic of China*These authors contributed equally to this workCorrespondence: Jie Guo, Department of Orthodontics, School and Hospital of Stomatology, Shandong University, No. 44-1 Wenhua Road West, Jinan, Shandong, 250012, People’s Republic of China, Tel +86 13583104999, Fax +86 53188382923, Email [email protected]: Sleep apnea (SA), associated with absent neural output, is characterised by recurrent episodes of hypoxemia and repeated arousals during sleep, resulting in decreased sleep quality and various health complications. Mitochondrial DNA copy number (mtDNA-CN), an easily accessible biomarker in blood, reflects mitochondrial function. However, the causal relationship between mtDNA-CN and SA remains unclear. This study aimed to investigate the causality between mtDNA-CN and SA while identifying potential mediating brain imaging phenotypes (BIPs).Methods: Two-sample bidirectional Mendelian randomisation (MR) analysis was performed to estimate the causal relationship between mtDNA-CN and SA, with further validation using Bayesian framework-based MR analysis. A two-step approach was employed to evaluate causal relationships between BIPs, mtDNA-CN and SA, utilising the “product of coefficients” method to assess the mediating effects of BIPs. Multiple testing errors were corrected using the Benjamini–Hochberg method.Results: Genetically predicted mtDNA-CN had a negative causal effect on SA (OR = 0.859, 95% CI = 0.785– 0.939, P = 3.20× 10− 4), whereas SA did not have a causal effect on mtDNA-CN (OR = 1.0056, 95% CI = 0.9954– 1.0159, P = 0.2825). Among 3935 BIPs, two features related to white matter microstructure served as partial mediators: the second eigenvalue from diffusion MRI data analysed by tract-based spatial statistics in the right posterior thalamic radiation, with a mediation proportion of 11.37% (P = 0.0450), and fractional anisotropy in the right sagittal stratum, with a mediation proportion of 12.79% (P = 0.0323).Conclusion: This study demonstrated a causal relationship between mtDNA-CN and SA, with specific brain white matter microstructure phenotypes potentially acting as mediators. These findings highlight the potential of mtDNA-CN as a biomarker for SA and underscore its relevance in guiding future therapeutic strategies targeting mitochondrial health and brain white matter microstructure.Keywords: mitochondrion, sleep disorder, brain structure, magnetic resonance imaging, causal relationship