Safety and Efficacy of Mesenchymal Stem Cell Therapy in Multiple System Atrophy: Systematic Review

Major Points and Findings:

Multiple system atrophy (MSA) is a fast-moving neurodegenerative disease with no treatment that slows it. Its cerebellar form, MSA-C, presents mainly as ataxia. This systematic review covers the clinical trials of cell therapy in MSA published up to September 2025. It found seven studies with 123 participants, most of them with MSA-C.

The procedures were tolerated and side effects were mostly mild. Several studies saw slower worsening for some months after treatment. The authors do not regard this as proof of efficacy, because only two of the seven studies were randomised and the data were too inconsistent to pool in a meta-analysis.

Aim:

To assess the safety of mesenchymal stem cell (MSC) therapy in adults with probable or confirmed MSA, and secondarily its effect on disease progression measured with the Unified MSA Rating Scale (UMSARS: total score, Part I for daily activities, Part II for the motor examination).

Methods:

The review was registered on PROSPERO (CRD420251137238) and followed PRISMA 2020. PubMed, Scopus, Web of Science and Cochrane CENTRAL were searched to 16 September 2025 with no start date. Two reviewers worked independently. Bias was assessed with Cochrane RoB 2 (randomised trials) and ROBINS-I (other studies). Patients with other neurodegenerative disease, or a Mini-Mental State score of 24 or below, were excluded.

Of 951 records, 21 were read in full and seven were included:

• Lee 2008, South Korea. Open-label randomised trial, 11 treated and 18 controls, autologous bone marrow MSCs, intra-arterial then intravenous.
• Lee 2012, South Korea. Randomised, double-blind, placebo-controlled, 16 treated and 17 placebo, all MSA-C, autologous bone marrow MSCs (one intra-arterial dose of 4 × 10⁷ cells, then three intravenous doses at months 1, 2 and 3).
• Tsai 2017, Taiwan. Open-label, allogeneic adipose MSCs intravenously, in six SCA3 patients and one MSA-C patient (only the MSA patient counts here).
• Singer 2019, USA. Phase I/II dose escalation, 24 patients (19 MSA-C), autologous adipose MSCs intrathecally, from a single dose of 1 × 10⁷ cells up to two doses of 1 × 10⁸ a month apart.
• Chung 2021, South Korea. Phase I, 9 MSA-C patients, autologous bone marrow MSCs intra-arterially at three dose levels.
• Yu 2021 and Gong 2022, China. Studies of 20 and 7 patients using umbilical cord blood mononuclear cells, which are not cultured MSCs, injected through a lateral atlanto-occipital space puncture at the top of the neck.

Mean ages were 55 to 58. Counting from the review’s Table 1, the 123 participants include 35 control or placebo patients, so about 88 people with MSA received cells.

Results:

Risk of bias: Lee 2012 was rated “some concerns” because the intra-arterial procedure could not be fully blinded and some secondary outcome data were missing. Lee 2008 was rated high risk. All five single-arm studies were rated serious or critical. Without a control group, a treatment effect cannot be separated from natural variation, regression to the mean, other treatments or placebo response.

Safety:

• No serious toxicity was attributed to the cells in any study.
• With intra-arterial delivery, small ischaemic spots on MRI were the main concern. In Lee 2008, 7 treated patients had such lesions, without symptoms, and 6 had fever. In Lee 2012 there were 4 acute ischaemic lesions in the cell group and 6 in the placebo group, which suggests the catheter procedure was responsible and the cells were not.
• With intrathecal delivery (Singer), 50 adverse events were logged, 26 of them attributed to the MSC product. MRI showed thickened or mildly enhancing nerve roots at the base of the spine in all patients in the medium and high dose tiers. About half of them had mild back discomfort and none had neurological deficits.
• With cord blood mononuclear cells, Yu reported 3 mild headaches and 1 slight fever, and Gong reported 1 fever and 1 case of puncture-site pain. All resolved within hours to days.
• The one event the reviewers considered possibly cell-related was leptomeningeal enhancement (a sign of blood-brain barrier disturbance) in one low-dose patient in Chung’s trial.

Efficacy:

• Lee 2012 was the only blinded trial. UMSARS total rose from 40.1 to 49.6 with cells and from 39.5 to 55.8 with placebo over a year (p = 0.047). For Part II the p-value was 0.008. The difference became apparent around day 240 and then narrowed.
• Singer 2019 reported progression of 0.40 ± 0.59 UMSARS points per month against 1.44 ± 1.42 in matched historical controls (p = 0.004), with the higher dose tiers doing better. Treated patients still worsened, by about 15% over the year, against 53.5% in the historical comparison.
• In Chung 2021 the medium and high dose groups worsened more slowly than the low dose group, but not significantly (p = 0.096).
• In Gong 2022, UMSARS total improved from 42.6 to 25.7 at the best point, 3 to 6 months after treatment, then drifted back to 35.1 by last follow-up at 3 to 5 years.
• In Yu 2021, Part II improved from 30.2 to 24.3 at six months.

Two patterns recur. Benefit peaked at roughly 3 to 8 months and then faded. Patients treated earlier, with lower starting scores, tended to respond better.

Conclusions:

The authors describe MSC therapy as a “promising investigational approach” that may temporarily slow deterioration, especially early in the disease, and then state that “the current evidence is insufficient to establish definitive clinical efficacy.” They call for larger randomised trials with standardised cell products, at least 24 months of follow-up, testing of repeat dosing at around 6 to 8 months, objective biomarkers such as neurofilament light chain, and a direct comparison of intrathecal and intra-arterial delivery.

The review has some internal inconsistencies. The abstract says adverse events occurred in 65 to 70% of participants, but this figure is not derived anywhere in the body. The stated eligibility criteria exclude non-comparative studies, yet five of the seven included studies are single-arm. One sentence attributes a placebo comparison to Singer’s trial, which had no placebo arm, although the discussion correctly assigns it to Lee 2012. The screening numbers do not add up exactly. The review is most useful as an overview of the existing trials.

Background Information:

In MSA, misfolded alpha-synuclein builds up in oligodendrocytes, the cells that insulate nerve fibres. Median survival is 6 to 10 years from first symptoms. Diagnosis in life is imperfect. The review cites a figure of 62% of clinically diagnosed cases confirmed at autopsy, so any small trial may include people who do not have MSA.

The reviewers point out that the two Chinese studies used cord blood mononuclear cells, a mixed population of blood stem cells, immune cells and endothelial progenitors. These differ biologically from culture-expanded MSCs and should not be assumed to give the same results. The mechanism proposed for all of these products is trophic support and immune modulation (release of BDNF, GDNF and VEGF, less glial activation). Replacement of lost neurons is not expected.