Mitochondrial protein alterations in vascular dementia: evidence from Mendelian randomization, transcriptomics, and a chronic hypoperfusion model - Summary - MDSpire

Mitochondrial protein alterations in vascular dementia: evidence from Mendelian randomization, transcriptomics, and a chronic hypoperfusion model

  • By

  • Qian Liu

  • Huizhong Tan

  • Keke Tong

  • Ruhai Luo

  • Hanquan Li

  • Feng Qiu

  • Shiliang Wang

  • Le Xie

  • Xiuli Zhang

  • Dahua Wu

  • July 9, 2026

  • 0 min

Share

Objective:

To identify systemic circulating biomarkers associated with vascular dementia (VaD) risk and investigate corresponding neuropathological mitochondrial changes in the brain.

Approach:
  • Mendelian Randomization Analysis: Two-sample MR analysis of plasma proteomics and GWAS data to identify candidate proteins associated with VaD risk.
  • Transcriptomic Analysis: Assessment of differential expression of candidate genes using the GEO dataset GSE122063.
  • In Vivo Validation: Validation of proteins in a Bilateral Common Carotid Artery Occlusion (2-VO) rat model through evaluation of pathological features and protein expression levels.
Key Findings:
  • FourproteinswerenominallylinkedtoVaD:protectivefactorsAIFM1andCOX5B,andriskfactorsNDUFV2andNUDT5.COX5Bwasidentifiedasthemostrobusttargetwithconsistentevidenceacrossallanalyticallayers.NUDT5showedpartialconsistency,whileAIFM1andNDUFV2displayedcontradictionsbetweengenetic/transcriptomicdataandinvivoexpression.DonepezilimprovedVaDpathologybutdidnotaltertheexpressionofthestudiedproteins.
Interpretation:

COX5B is highlighted as a high-confidence driver of mitochondrial impairment in VaD, while inconsistencies with other proteins are acknowledged.

Limitations:
  • Thestudy'sfindingsarebasedonaspecificsetofproteinsandmaynotencompassallmitochondrialproteinsinvolvedinVaD.Therelianceonexistingdatasetsmaylimitthediscoveryofnovelproteins.
Conclusion:

The study identifies COX5B as a potential target for further investigation into its role in VaD, while also noting the complexities surrounding AIFM1, NDUFV2, and NUDT5.

Original Source(s)

Related Content