Allogenic Umbilical Cord-Derived Mesenchymal Stromal Cells Improve Motor Function in Prenatal Surgical Repair of Myelomeningocele: An Ovine Model Study

Major Points and Findings:

This preclinical study was carried out in fetal sheep. No children were treated. A French team in Paris, which is preparing to add umbilical cord-derived mesenchymal stromal cells (UC-MSCs) to the prenatal surgical repair of myelomeningocele, placed a fibrin patch carrying UC-MSCs from donor sheep on the exposed spinal cord of fetal lambs during the repair. Lambs repaired with the cell patch were compared with lambs repaired with the same patch without cells.

One day after birth, the lambs that had received cells moved better, none of them leaked urine, and under the microscope their spinal cords had more large nerve cells and much less scar tissue. These differences were statistically significant. Other measures, including nerve conduction tests, muscle size and sphincter thickness, showed no difference. Only nine lambs were born alive and they were followed for 24 hours, so the study says nothing about longer-term effects or safety.

Aim:

To investigate the effect of an allogenic (donor) UC-MSC patch applied during fetal surgery on motor and sphincter function at 24 hours of life in the sheep model of myelomeningocele.

Methods:

The study was approved by the French National Committee on Animal Research and followed the ARRIVE guidelines for animal research.

UC-MSCs were grown by the explant method from the umbilical cords of three lambs delivered by caesarean section near term. The cells were spindle-shaped, stuck to plastic and carried the sheep MSC markers CD29, CD90, CD73, CD44 and CD166, but not CD31 or CD45. For each cell patch, 17 million cells at passage 5 were mixed into fibrinogen from a commercial fibrin sealant kit (Evicel) and set with thrombin. Control patches were made the same way without cells. Both kinds were flexible and about 2 mm thick.

A spinal defect was surgically created in fetal lambs at day 75 of gestation (L1 to L5 laminectomy with removal of the dura) and repaired at day 89. The patch was laid on the spinal cord and the fetal skin was closed over it. The surgeons knew which patch was used. Lambs were delivered by caesarean section at day 140 to 142 and kept alive for 24 hours.

At 2 and 24 hours of life a paediatric rehabilitation physician and a veterinary surgeon, both blinded to the group, assessed movement with the Sheep Locomotor Rating scale (SLR, from 0 for complete paralysis to 15 for normal), as well as standing, reflexes, sensation and muscle wasting (amyotrophy). A lamb was counted as continent if it took a normal urination posture and did not leak when the bladder was pressed. At 24 hours, under anaesthesia, a blinded veterinary neurologist recorded muscle electrical activity, nerve conduction and spinal evoked potentials, and a spine scan measured the size of the defect. The lambs were then euthanised. A blinded neuropathologist measured spinal cord area, grey matter area, large neuron density and scar tissue (fibrosis) thickness, and all major organs were examined for tumours.

The study was planned with four lambs per group, assuming that 90% of treated and 10% of control lambs would score above 12 on the SLR. Groups were compared with the Mann-Whitney U test and the chi-square test.

Results:

Patches were placed in 13 fetuses. Four were lost: two miscarriages (one in each group) and two fetal deaths, one discovered at the time of repair and one treated fetus at caesarean delivery. Overall fetal loss was 31%. Nine lambs were born alive, 5 with a cell patch and 4 with a control patch.

Safety: All nine repairs had healed completely. No lamb had hindbrain herniation, and no tumour was found in any organ. The causes of the four fetal losses could not be determined because autopsy was not possible.

Movement:

• SLR at 2 hours: median 2 (range 1 to 7) with cells against 2 (1 to 6) without, p = 0.83.
• SLR at 24 hours: median 14 (range 13 to 15) with cells against 5 (4 to 14) without, p = 0.04.
• Able to stand unaided: 5 of 5 with cells, 2 of 4 without, p = 0.07 (not significant).
• Muscle wasting in the hind legs: 0 of 5 with cells, 3 of 4 without, p = 0.02.

Continence: None of the 5 treated lambs and all 4 control lambs were incontinent (p = 0.002). The authors note this was a simple clinical judgement, not confirmed by bladder testing or histology.

Nerve tests and spine scans: Spontaneous muscle electrical activity, nerve conduction amplitudes and speeds, and spinal evoked potentials did not differ significantly between groups. The spine scans showed defects of similar size and spinal angulation in both groups.

Spinal cord under the microscope:

• Large neuron density: 14.5 per mm² of grey matter with cells against 5.6 without (p = 0.001).
• Scar tissue between the spinal cord and the skin: 453 against 3,921 micrometres thick (p = 0.03).
• Scar tissue around the spinal cord: 47 against 158 micrometres thick (p = 0.0001).
• Grey matter area (8.2 against 7.1 mm², p = 0.74) and spinal cord area (21.8 against 16.7 mm², p = 0.44) did not differ significantly.

Muscles and sphincters: The weight and fibre size of two hind-leg muscles, and the thickness of the urinary and anal sphincters, were similar in both groups.

The paper has some reporting inconsistencies. The text calls several results “mean” values while the tables and methods give medians. A few ranges in the extraspinal findings table cannot be correct (for example a control median of 13.7 with a range of 45.3 to 49.4), and the spine scan table repeats the numbers of the muscle weight rows. These do not affect the main findings above.

Conclusions:

The authors conclude that UC-MSC patches used alongside prenatal myelomeningocele repair in sheep “improve motor and sphincter functions”, together with better spinal cord preservation and less fibrosis, and that more studies are needed on long-term safety and on whether the approach can be translated to humans.

Limitations the authors state:

• Small sample size, partly due to fetal loss, which is common in this animal model.
• Follow-up of only 24 hours, because caring for possibly disabled lambs is costly and complex, which limits conclusions on effectiveness, tolerability and safety.
• A subjective measure of continence that was not confirmed histologically.
• Nerve conduction and evoked potential tests showed no difference, possibly because the study was too small or because nerve circuits differ between sheep and humans.

Other points also call for caution. The surgeons were not blinded. Standing ability was not significantly different. The 24-hour movement result rests on 5 against 4 animals, and one control lamb scored 14, the same as the treated median.

How this study differs from the team’s human-cell study: The same group later repeated the experiment with UC-MSCs from human umbilical cords (report summary). In this earlier study the cells came from sheep, movement and continence differed significantly at 24 hours and scar tissue was significantly reduced, while spinal cord and grey matter area did not differ. In the human-cell study the pattern was nearly reversed. Spinal cord area, grey matter area and neuron density were significantly better, but the movement results and the reduction in scar tissue did not reach significance. The authors of the later study suggested that the three human donor cords may have been weaker producers of protective factors. Both studies were small and followed the lambs for only 24 hours.

Background Information:

The authors note that even after fetal surgery, 71% of children with myelomeningocele cannot walk independently at age eight and 75% need daily urinary catheterisation by age ten. The team’s earlier sheep experiment found only a small, non-significant effect of UC-MSCs on movement at 2 hours of life, which they think was too soon after birth to judge.

A Californian group using early-gestation placental MSCs on a porcine matrix in the same model reported a median SLR of 15 with cells against 6 without, and more preserved neurons. That group found no difference in bladder function, so the continence finding here is new and unconfirmed.

The authors think UC-MSCs act mainly by releasing growth and protective factors rather than by becoming nerve cells, but state that the exact mechanisms are not yet known. They suggest the reduction in scar tissue may be relevant to tethered cord, a common problem after myelomeningocele repair. In an earlier tracking study by the group, the transplanted cells were not found in distant organs.

The study was funded by two French medical research charities, the Fondation de l’Avenir and the Fondation pour la Recherche Médicale. The authors declared no conflicts of interest. The journal published an accompanying commentary by Anna L. David.