Immunomodulatory Amnion-Derived Mesenchymal Stromal Cells Preserve Muscle Function in a Mouse Model of Duchenne Muscular Dystrophy

Major Points and Findings:

This is an animal study. No patients were treated. A Japanese team at the University of Tokyo, Nippon Medical School and Kaneka Corporation gave human amnion-derived mesenchymal stromal cells (hAMSCs) to mdx mice, the standard mouse model of Duchenne muscular dystrophy, by repeated tail-vein injection, and followed the mice for up to a year.

For a family considering treatment, the paper shows that these cells calm the immune system in the laboratory by pushing macrophages toward the anti-inflammatory M2 type, that repeated intravenous doses were tolerated by mice and reduced muscle inflammation, and that treated mice kept more grip strength, ran further and had slightly better heart function at one year. It does not show that any of this happens in children. The cells were amnion-derived, not the umbilical cord-derived cells used in our own programme, and they did not restore dystrophin, the missing protein that causes the disease. The cells were supplied by Kaneka Corporation, which part-funded the work and employs four of the authors.

Aim:

To test whether hAMSCs polarise human macrophages toward the M2 type, and whether repeated systemic hAMSC injection given early in the disease improves skeletal and cardiac muscle function in mdx mice.

Methods:

Cells. Amnion was taken, with consent, from fetal membranes of women having a caesarean delivery. Cells were isolated by enzymatic digestion, expanded, frozen and used between passages 1 and 5. They carried the usual MSC markers (CD73, CD90, CD105), lacked CD45, CD34, CD14, CD11b, CD19 and HLA-DR, and formed bone and cartilage in culture but not fat.

Laboratory work. Human blood mononuclear cells were co-cultured with hAMSCs at 20 to 1 and macrophage markers were measured by flow cytometry and qPCR, with indomethacin used to block prostaglandin E2 (PGE2).

Animals. Male mdx mice on a C57BL/10 background were randomly split into control and treated groups. Each dose was 8.0 × 10⁵ hAMSCs in 100 µL of saline into the tail vein, once a week, starting at 4 to 5 weeks of age. The short-term arm received four doses (n = 10, assessed at 12 and 18 weeks of age, against 14 control mdx and 7 wild-type mice). The long-term arm received four doses (n = 6) or six doses (n = 4) and was assessed at one year against 12 control mdx and 7 wild-type mice. Many outcomes were measured in subsets of 3 to 8 mice, given below. The prespecified primary outcome was long-term motor function (forelimb grip strength and running speed); histology was secondary. Investigators could not be blinded to mouse strain; running was recorded by an automated wheel over five days. Data were excluded when a mouse had lost weight, was injured or debilitated, or when a value was considered peculiar.

Results:

Immune effects in the dish. hAMSCs cut proliferation of activated T cells from 90.1% to 59.9% (p below 0.0001) and secreted PGE2. The share of CD206-positive (M2) cells rose modestly, from 11.3% to 14.2% (p = 0.035), while M1 cells did not change. Indomethacin abolished the CD206 rise, pointing to PGE2 as the mediator.

Tolerability. Body weight did not differ between treated and untreated mdx mice at any point. No other safety measures (organ examination, tumour screening, immune reaction to the human cells) are reported.

Short term (12 to 18 weeks of age, four doses):

• Serum creatine kinase, the muscle-leak marker, was lower four weeks after the last dose (435 ± 79 vs 583 ± 92 ng/mL, p = 0.038, n = 6 vs 4) but not at ten weeks (582 vs 700 ng/mL, p = 0.228). The authors call this transient.
• IL-6 in tibialis anterior muscle: 10.0 ± 3.8 vs 94.2 ± 64.1 pg/mL (p = 0.002, n = 4 vs 5).
• Grip strength at 12 weeks: 336.8 ± 21.0 g vs 280.0 ± 37.8 g (p = 0.0008; normalised to body weight 10.3 vs 8.8 g/g, p = 0.012; n = 10 vs 14).
• Macrophage (F4/80) area in the diaphragm: 14.0% vs 19.5% (p = 0.013, n = 4 vs 3). The proportion of macrophages carrying the M2 marker CD206 in tibialis anterior rose from 17.6% to 55.5% (p below 0.0001).
• Human cells were found in tibialis anterior four weeks after treatment, occupying only 0.0036% of the section area, and were gone by ten weeks. Dystrophin was not detected, so the cells did not repair the underlying defect.

Long term (one year, four or six doses, usually pooled):

• Mononuclear cell infiltration in tibialis anterior: 9.0% in controls, 7.3% after four doses, 4.7% after six; only the six-dose group reached significance (p = 0.038, n = 4 per treated group vs 14). No difference in the diaphragm.
• Centrally nucleated fibres, a sign of repeated degeneration and regeneration: 64.3% vs 82.0% (p = 0.014, n = 5 vs 7).
• Fibrosis was not reduced in tibialis anterior, diaphragm or heart.
• Grip strength: 296.5 ± 27.2 g vs 265.7 ± 26.3 g (p = 0.049; normalised 8.6 vs 7.3 g/g, p = 0.002; n = 8 vs 12). Wild-type mice scored 336.4 g, so treated mice sat between the two.
• Maximum running speed was not different (21.6 vs 20.5 m/min, p = 0.77). Daily running distance was greater (3,649 ± 1,025 vs 2,420 ± 788 m/day, p = 0.028, n = 8 vs 10), as was average distance per minute (18.3 vs 14.7 m/min, p = 0.047, n = 5 vs 5).
• Heart: no difference in gross morphology, fibrosis or macrophage polarisation, but less mononuclear infiltration (n = 6 vs 6) and higher left ventricular fractional shortening, 37.0 ± 2.3% vs 33.6% (p = 0.017; wild type 39.6%). Table 2’s header lists 8 treated mice while its footnote and the figure legend say 5, an inconsistency in the paper itself.

Conclusions:

The authors conclude that early, repeated systemic hAMSC treatment delayed inflammation and motor decline in mdx mice, probably through M2 macrophage polarisation rather than by the cells becoming muscle. They suggest it could be combined with steroids or with gene therapy.

Limitations the authors state:

• Heart disease in mdx mice is mild compared with patients, so the cardiac findings give limited information about humans. They propose confirming it in the dog model.
• The human cells persisted only briefly, and how secreted factors produce long-lasting effects is not established; the mechanism remains to be worked out.
• Long-term culture and expansion of MSCs carry risks of senescence and chromosomal change, and immunosuppressive capacity varies with tissue source.
• From the methods: investigators were not blinded to strain, some data were excluded by judgement, and long-term groups were small.

Background Information:

Duchenne affects about 1 in 3,500 male births. Steroids are the standard anti-inflammatory treatment but responses vary and side effects are common. The same team previously reported that repeated intravenous dental pulp stem cells maintained locomotor activity in mdx mice and dystrophic dogs. The authors note that intravenously given MSCs are largely trapped in the lungs and die there, and that their immune effect appears to come from how the recipient’s own phagocytes respond to the dying cells. The paper contains no umbilical cord-derived cell data and mentions no human trial of amnion-derived cells for Duchenne.