Why metformin could protect the brain by rewiring mitochondria

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Sep 1 2025

A new study shows metformin enhances myelin repair in human-based models by tuning mitochondrial metabolism, offering hope for multiple sclerosis treatment.

Study: Metformin alters mitochondria-related metabolism and enhances human oligodendrocyte function. Image Credit: Juan Gaertner / Shutterstock

In a recent study published in the journal Nature Communications, an international team of researchers investigated whether metformin enhances the differentiation and myelination of human oligodendrocyte progenitor cells (OPCs) across human-relevant models and defined mitochondria-related mechanisms that support neuroprotection.

Background

Every day, millions of brain messages rely on myelin to stay precise; when that insulation fails, movement, memory, and mood suffer. Multiple sclerosis (MS) removes myelin from axons, and aging reduces remyelination because OPCs become less responsive.

Metformin, a first-line therapy for type II diabetes mellitus, crosses the blood-brain barrier and alters the adenosine monophosphate (AMP): adenosine triphosphate (ATP) ratio by inhibiting mitochondrial Complex I, activating AMP-activated protein kinase (AMPK). 

Repurposing for neuroprotection faces a challenge: human oligodendroglia differ significantly from those of rodents. Better remyelination could slow the progression of disability and preserve independence.

Further research is needed to define mechanisms and benefits.

About the study

Researchers compared three human systems to assess the effects of metformin on oligodendroglia. They generated human embryonic stem cell (hESC)-derived OPCs in monolayer culture, produced cortical organoids containing oligodendroglia, and transplanted green fluorescent protein-labeled hESC-OPCs into the corpus callosum of Shiverer; recombination activating gene 2 (Rag2)-null mice to create human-mouse chimeras.

Metformin hydrochloride (100 μM) was applied to monolayers for 7 days and administered daily to organoids from day in vitro 60 to 70. Chimeras received oral metformin (300 mg/kg) for 21 days, beginning 42 days post-transplantation.

Differentiation and myelination were quantified by immunostaining for myelin basic protein (MBP), oligodendrocyte transcription factor 2 (OLIG2), and the mature marker adenomatous polyposis coli (APC; clone CC1), plus electron microscopy (EM) to compute myelinated-axon percentage and g-ratio (axon diameter divided by axon plus myelin diameter).

Single-cell ribonucleic acid sequencing (scRNA-seq) profiles were generated for cells, and single-nucleus RNA sequencing (snRNA-seq) datasets from the brain and spinal cord were integrated using canonical correlation analysis and an artificial neural network (ANN) to compare identities. 

Differential expression and Gene Ontology (GO) analyses were used to test pathway changes. Mechanistic readouts included in situ hybridization for NDUFA11 and EIF1, and Western blotting for TOMM20 and CHCHD2.

Study results

In monolayer cultures, metformin increased human oligodendrocyte differentiation within seven days. Intermediate oligodendrocytes showed a mean increase of 0.70 ± 0.2 SEM (fold change) and more mature OLIG2+ MBP+ cells rose by a mean increase of 0.52 ± 0.23 SEM (fold change) versus vehicle, comparable to clemastine fumarate. scRNA-seq showed these cells resembled fetal rather than adult oligodendroglia: OPC clusters aligned with adult OPCs, whereas oligodendrocytes mapped to immature, committed oligodendrocyte progenitor cell (COP)-like states, with persistent SRY-box transcription factor 2 (SOX2) expression marking immaturity. Notably, this contrasts with rat models where metformin only aids aged OPCs, highlighting species-specific responses.

In cortical organoids, a metformin pulse from day in vitro 60–70 did not change counts of CC1+ or MBP+ cells but significantly expanded MBP area (mean increase of 0.45 ± 0.18 SEM), indicating more myelin protein per area without altering cell number. Integrated analyses against adult snRNA-seq again placed most organoid oligodendroglia in COP or immature oligodendrocyte compartments.

The strongest effects appeared in human-mouse chimeras. After transplantation of hESC–derived OPCs into the Shiverer; Rag2-null corpus callosum, 46.77% ± 4.39 SEM of regional cells were human and 70.51% ± 2.28 SEM of those were OLIG2+. By EM, a mean of 16.15% ± 1.88 SEM of axons were myelinated at baseline.

A 21-day oral course of metformin increased myelinated axons from 21.44 ± 2.3 SEM% to 28.21 ± 1.9 SEM% and reduced the g-ratio from 0.84 ± 0.004 SEM to 0.81 ± 0.009 SEM, independent of axonal diameters and consistent with thicker myelin. Although mature oligodendrocyte counts (CC1+) did not change, myelin output per axon improved, implying enhanced function per cell.

Mitochondrial structure and gene programs shifted with treatment. Metformin increased mitochondrial profile area in axons and glia, consistent with changes in mitochondrial content or dynamics. Transcriptomics in human chimera oligodendrocytes revealed the upregulation of NDUFA11, COX8A, and EIF1, which supports the translation of mitochondrial messages.

In situ hybridization confirmed higher NDUFA11 and EIF1 signals in human OLIG2+ cells, and Western blots in hESC oligodendroglial monocultures showed increases in TOMM20 and CHCHD2, consistent with heightened mitochondrial activity and dynamics.

Importantly, effects were not limited to transplanted human cells. In mouse corpus callosal oligodendrocytes, astrocytes, microglia, and neurons, metformin elevated EIF1 and COX8A, indicating broader metabolic tuning rather than a strictly cell-autonomous action.

Finally, in a single-nucleus dataset of MS donor brains, oligodendrocytes from two individuals known to have taken metformin before death expressed more EIF1 than two untreated MS donors, echoing the chimera signal despite small numbers. 

Together, metformin increased myelin proteins and sheaths across models and rewired mitochondria-related metabolism in ways that support oligodendrocyte function.

Conclusions

To summarize, this study shows that metformin enhances myelin proteins in vitro and myelin sheaths in vivo, with adult-like transcriptional similarity most evident in the chimera model.

In chimeras, myelinated axons rose and g-ratio fell without expanding mature oligodendrocyte counts, implying more myelin per cell. Transcriptional and protein signatures, including NDUFA11, COX8A, EIF1, TOMM20, and CHCHD2, are consistent with altered mitochondrial function and metabolism. 

Limitations include fetal-like cells with persistent SOX2 expression, absence of demyelination or inflammation, lack of direct mitochondrial respiration measurements, and few MS donors. 

The findings align with ongoing clinical testing of metformin’s neuroprotective potential in MS. Overall, the evidence supports testing metformin as a neuroprotective, remyelination-enhancing therapy in MS.