The Therapeutic Potential of Bone Marrow Mesenchymal Stem Cells-Derived Exosomes for Retinal and Optic Nerve Diseases

Major Points and Findings:

Exosomes are tiny packets (30 to 150 nanometres across) that cells release to signal to one another. Much of the effect of mesenchymal stem cells in damaged tissue appears to be mediated by these packets. This 2026 review from the Second Hospital of Jilin University gathers the laboratory evidence on exosomes from bone marrow MSCs in diseases of the retina and optic nerve, including optic nerve injury and glaucoma.

All of the evidence is preclinical and comes from cell cultures, rats and mice. The animal results are consistent. In their closing section the authors state that there are no active interventional clinical trials of MSC exosomes for retinal or optic nerve disease, that safety is “far from established”, and that the rodent models used “often offer an overly optimistic forecast of clinical success”. This is the current state of the evidence for exosome treatment of optic nerve conditions.

Aim:

To summarise how bone marrow MSC-derived exosomes (BMSC-exos) form, what they carry and how they act, to evaluate the evidence in retinal and optic nerve diseases, and to set out the barriers to clinical use.

Methods:

A narrative review. The authors do not describe a search strategy, inclusion criteria or quality assessment, so study selection cannot be checked. They chose bone marrow as the cell source because it has the largest body of neuroprotection research, while acknowledging that exosomes from adipose, umbilical cord and gingival MSCs are easier to obtain.

Results:

Exosomes form inside the cell and carry the membrane markers CD9, CD63 and CD81 plus ALIX and TSG101. The usual isolation method, ultracentrifugation, also pulls down some larger vesicles, which is why many recent papers use the term “small extracellular vesicles”. Exosomes are not necessarily benign. The review notes that in other settings they carry disease-causing proteins in Alzheimer’s disease, contribute to the destruction of insulin-producing cells in diabetes, and can trigger death of T lymphocytes.

Their cargo includes microRNAs (including miR-21, miR-146a and the miR-17-92 cluster) and neurotrophic factors (BDNF, NGF, CNTF, GDNF and others). Mead and Tomarev showed that the microRNAs contribute to the effect. When they knocked down Argonaute-2, a protein needed for microRNAs to work, the exosomes lost much of their protective effect on retinal ganglion cells. After uptake the vesicles stayed active in retinal cells for more than 28 days.

Optic nerve crush (rats):

• Mead and Tomarev (2017) gave human BMSC-exos, 3 × 10⁹ particles once a week for three weeks. The retinal nerve fibre layer thinned from 48.4 µm to 33.8 µm in treated eyes, against 48.2 µm to 18.0 µm in controls. Function was preserved in more than 50% of ganglion cells, and regrowing axons reached 114.2 µm against 43.7 µm in controls.
• Cui and colleagues (2021) gave a single dose of 3 × 10⁹ particles and observed the animals for 30 days. They found less ganglion cell death through the PI3K/Akt pathway and lower levels of TNF-alpha, IL-1 beta, IL-6, IL-8 and MCP-1.
• You and colleagues (2025) gave a single intravitreal dose and observed for 21 days. Treatment dampened the IL-17, TNF, NF-kB and HIF-1 pathways and stopped microglia shifting into an injury-responsive state.

Glaucoma models: Two studies by Mead and colleagues (2018) used genetically glaucoma-prone DBA/2J mice and rats with raised eye pressure, the latter injected monthly for nine months. The review reports severely degenerating axons at 40% with treatment against 66% in controls, better-preserved ganglion cell electrical responses (24.9 µV against 18.5 µV) and slower nerve fibre layer loss.

Other models reviewed: In retinal ischaemia-reperfusion, dying cells fell by more than 50% and ERG recovery improved. Exosomes from hypoxia-preconditioned cells protected better. Benefits are also reported in models of retinal degeneration, diabetic retinopathy, optic neuritis, oxygen-induced retinopathy and retinal detachment. The review makes one comparison between cell sources. Bone marrow exosomes lack the hypoxia-driven pro-angiogenic microRNAs miR-210 and miR-378 found in exosomes from some other MSC sources, which may make them less likely to promote unwanted new blood vessels in the eye.

Doses and routes: In the 13 tabulated animal studies, doses ranged from 1 × 10⁶ to 1 × 10¹¹ particles, or 1 to 100 µg of protein, mostly in volumes of 1 to 5 µL. Eleven gave a single dose, and observation ran from 7 days to 9 months, mostly 3 to 4 weeks. Delivery was into the eye in essentially all of them. None of the reviewed optic nerve studies gave bone marrow exosomes into a vein or into the spinal fluid.

Conclusions:

The authors conclude that BMSC-exos are “a promising preclinical strategy” and that their value in people “remains to be proven”. They name four barriers:

• The animal models are a poor match. Nearly all data come from acute injury in rodents, which does not resemble slow human diseases such as glaucoma. There are almost no data from large animals or primates. Without these, they write, “the therapeutic benefits remain theoretical”.
• Safety is unknown. Injecting a mixed population of vesicles into the eye carries risks of inflammation, fibrosis or autoimmune reaction from uncharacterised cargo. Existing safety observations are short, typically 28 days. No maximum tolerated dose, formal toxicology or long-term biodistribution studies exist.
• Clinical trials are lacking. ClinicalTrials.gov holds five relevant entries. Two, for macular hole (NCT03437759) and retinitis pigmentosa (NCT05413148), have “unknown” status. The other three concern surface eye disease treated with drops. None targets the optic nerve.
• The product is inconsistent. Exosome content varies with the tissue source, the donor (one study found exosomes from female donors protected ganglion cells better than those from male donors), the culture conditions and the batch. This makes standardisation and consistent manufacturing difficult. Delivery that reaches the retina and optic nerve and lasts is also unsolved.

Their recommended next steps are chronic disease models in large animals, regulatory-grade toxicology, potency assays and more uniform vesicle preparations.

Background Information:

The review lists the drawbacks of transplanting whole bone marrow MSCs as low efficiency of targeted differentiation, possible immune rejection and possible tumour formation. By comparison, exosomes cannot divide, provoke less immune response, are more stable and are easier to store.

For MSC therapy, the work supports the idea that the cells act mainly by releasing protective and anti-inflammatory signals and do not become new nerve cells. It does not show that exosomes, or the cells that make them, restore an optic nerve that has already atrophied. In the animal experiments treatment began at or shortly after the injury, when ganglion cells were still alive to be protected. The authors describe moving from acute injury models to “complex human disease contexts” as an unmet challenge.