Post-Symptomatic Administration of hMSCs Exerts Therapeutic Effects in SCA2 Mice

Major Points and Findings:

This is a preclinical study in mice, and no patients were treated. It is included because it tests whether treatment still has an effect when it starts after symptoms have appeared, which is the situation of almost every family who contacts a clinic. Most animal studies treat before the disease shows itself. In this study, mice carrying the human spinocerebellar ataxia type 2 (SCA2) gene were treated only once they were already unsteady.

Human bone marrow mesenchymal stem cells (hMSCs) given into the spinal fluid three times, four weeks apart, slowed the loss of coordination and preserved Purkinje cell markers over the following six months. A single injection did not. The mice still declined, and results in six mice per group do not predict what happens in patients.

Aim:

To test whether hMSCs given after symptom onset protect the cerebellum and motor function in SCA2 mice, to compare single and repeated dosing at two dose levels, and to explore whether a secreted protein called FSTL1 is involved.

Methods:

The animals were transgenic mice expressing full-length human ataxin-2 in Purkinje cells (both sexes), with normal C57BL/6J mice as the comparison. Six animals per group were used for every behavioural and protein experiment. The work followed ARRIVE reporting guidelines and was approved by the Kyungpook National University animal committee.

Bone marrow MSCs came from eight healthy human donors and were collected with ethics approval and consent. Cells were not pooled. Each experiment used cells from a single donor, chosen for strong early growth. They were expanded to passage 3 or 4 under GMP conditions at CorestemChemon Inc. (Seoul) in medium containing 10% fetal bovine serum, and confirmed as CD29, CD44, CD73 and CD105 positive and CD34 and CD45 negative. Three of the authors are affiliated with that company. The paper declares no competing interests.

The team first showed that by 25 weeks of age the SCA2 mice had already lost neuronal markers and were failing motor tests, so treatment began at 26 weeks. Nine groups were randomly assigned: untreated, vehicle only (HypoThermosol storage solution), and SCA2 mice given 6 × 10⁴ or 1 × 10⁵ cells, either once or three times at four-week intervals.

For delivery, under ketamine and xylazine anaesthesia, the head was fixed in a frame, the membrane over the cisterna magna (the fluid space behind the cerebellum) was exposed, and cells in 20 microlitres were infused at 2 microlitres per minute through a 30-gauge needle, which was left in place for 10 minutes. The route is intrathecal, but it is a surgical one aimed directly at the cerebellum. It differs from a lumbar puncture in a patient.

Every four weeks from 26 to 50 weeks the mice were tested on an accelerating rotarod (4 to 40 rpm, three trials with an hour’s rest between) and given a 0 to 9 composite score from ledge walking, hindlimb clasping and gait. An open-field test measured distance travelled. At 50 weeks the cerebellum was analysed by western blot and immunofluorescence.

Results:

Motor function:

• At 26 weeks, before treatment, SCA2 mice stayed on the rotarod for 344.8 ± 11.2 seconds against 523.3 ± 10.7 for normal mice.
• At 50 weeks, untreated SCA2 mice managed 205.8 ± 14.6 seconds. Mice given three doses of 1 × 10⁵ cells managed 311.3 ± 13.0 seconds (p below 0.001). They were still below their own starting level and well below normal mice.
• Single injections did not prevent the behavioural decline at either dose.
• Open-field distance was 1,482.7 cm with repeated hMSCs against 1,072.3 cm untreated.
• The composite ataxia score was 2.0 ± 0.26 with repeated hMSCs against 3.83 ± 0.3 untreated (lower is better).

Cerebellar neurons: NeuN (granule cells) and calbindin (Purkinje cells) were significantly reduced in untreated SCA2 mice at 50 weeks and significantly preserved after three doses of 1 × 10⁵ cells. Some increase in NeuN was seen with a single higher dose, but the significant preservation of both markers was in the repeated-dose group.

Growth factors: BDNF, GDNF and CNTF were all low in SCA2 cerebellum from 25 weeks onward and were preserved by repeated treatment (p below 0.001). Staining located BDNF and GDNF in Purkinje cells and not in granule cells.

Inflammation: The microglial marker Iba1, the astrocyte marker GFAP, and TNF-alpha, IL-1 beta and iNOS were all raised in untreated SCA2 mice and reduced by repeated treatment.

Levels of FSTL1 and its partner TGF-beta 1 fell as the disease progressed and were preserved by repeated hMSC treatment. The authors propose that hMSCs act partly through the FSTL1 pathway.

No safety outcomes were reported. There are no data on deaths, surgical complications, immune reactions to human cells or tumour checks. The fate of the injected cells was not tracked in this study.

Conclusions:

The authors conclude that repeated hMSC administration after symptom onset, unlike single administration, preserved Purkinje cells, growth-factor support and motor performance in SCA2 mice, with benefit still visible 16 weeks after the last dose. They acknowledge limits. The FSTL1 link is an association, since they did not block or remove FSTL1 to prove it is required, and “another mechanism may also be involved”. They note that because MSCs are short-lived and SCA2 is progressive, treatment would probably work best at the pre-symptomatic or earliest stage, and that this may explain why some patients in an earlier clinical study (Tsai 2017) slipped back to their pre-treatment state within months. They suggest that clinical protocols need redesigning to give sustained benefit.

The methods show further limits. There were six animals per group, and the paper does not state that behavioural assessors were blinded. The transgenic model expresses the faulty gene only in Purkinje cells, whereas human SCA2 also affects the brainstem and nerves. Human cells were placed in mice without any reported immune monitoring.

Background Information:

SCA2 is the second most common dominantly inherited ataxia after SCA3. It is caused by an expanded CAG repeat in the ATXN2 gene. About 22 repeats is normal, up to 31 is harmless, and 33 or more causes disease. People with more than 45 repeats can develop symptoms in their teens. It affects the cerebellar cortex, pontine nuclei, inferior olives and peripheral nerves.

The paper touches on route of delivery. An earlier SCA2 mouse study found that intravenous MSCs delayed motor decline while cells injected directly into the brain did not. This group had previously shown that cells placed in the cisterna magna spread through the cerebrospinal fluid and cerebellum. Together with the single-versus-repeated finding here, this suggests that in trial design the dose schedule may be as relevant as dose size. Whether that holds in people has not been tested in a controlled trial.