Safety and Efficacy of Human Umbilical Cord-Derived Mesenchymal Stromal Cells in Fetal Ovine Myelomeningocele Repair

Major Points and Findings:

This preclinical study was carried out in fetal sheep. No children were treated. A French team in Paris is preparing a fetal surgery programme that adds umbilical cord-derived mesenchymal stromal cells (UC-MSCs) to the prenatal repair of myelomeningocele. The paper reports their first test of human UC-MSCs, a cell type widely used in clinical cell therapy, placed on the spinal cord of fetal lambs.

Lambs that received cells had a better-preserved spinal cord on histology, and the difference was statistically significant. They also tended to move better, although the movement results did not reach statistical significance. Only nine lambs were born alive and they were studied for 24 hours. Nearly half of all operated fetuses were lost. The authors state that translation to humans must wait for longer-term animal data.

Aim:

To assess the safety and efficacy of a fibrin patch containing human UC-MSCs, used as an add-on to fetal surgical repair of myelomeningocele in the sheep model.

Methods:

UC-MSCs were isolated by the explant method from three human umbilical cords, donated with consent by women after vaginal delivery. The cells met the International Society for Cellular Therapy definition. They were plastic-adherent and spindle-shaped, positive for CD73, CD90 and CD105, and negative for CD45, CD11b, CD14, CD34 and CD79a.

The patch was made the day before surgery from a commercial fibrin sealant kit (Evicel). For the cell patch, 17 million UC-MSCs at passage 5 were mixed into the fibrinogen solution before it was set with thrombin in a dish. Control patches were identical but contained no cells. On time-lapse microscopy the cells moved freely through the patch and took on their normal spindle shape over 72 hours.

A spinal defect was surgically created in fetal lambs at day 75 of gestation (L1 to L5 laminectomy with removal of the dura). It was repaired 14 days later, at day 89, when the patch was laid on the spinal cord and the fetal skin closed over it. Ewes were checked daily for general condition, temperature and wound healing. Lambs were delivered by caesarean at day 142.

A veterinarian blinded to the type of patch scored movement at 2 and 24 hours of life using the Sheep Locomotor Rating scale (0 means complete paraplegia, 15 means normal). Lambs were then euthanised. All major organs were examined by eye and under the microscope for tumours. At the L3-L4 level of the repair the team measured spinal cord area, grey matter area, the number of large neurons (motor neurons) per square millimetre of grey matter, and the thickness of scar tissue. Groups were compared with the Mann-Whitney U test.

Results:

A defect was created in 17 fetuses. Two were stillborn before repair, so 15 were repaired, 8 with a cell patch and 7 with a control patch. After repair, 2 of 8 were lost in the cell group and 4 of 7 in the control group. Overall fetal loss was 47% (8 of 17). In all, 9 lambs were born alive, 6 with cells and 3 controls.

Safety: No ewe developed fever or any other complication, and there was no sign of rejection of the human cells. The repair had healed completely in all nine lambs, with no cerebrospinal fluid leak and no infection. No tumour was found in any lamb on clinical examination or in any organ, 54 days after the cells were placed. No lamb had hindbrain herniation.

Movement:

• Locomotor score at 2 hours: median 4.0 (range 0 to 13) with cells against 2.0 (0 to 5) without, p = 0.32.
• Locomotor score at 24 hours: median 15.0 (range 2 to 15) with cells against 2.0 (2 to 5) without, p = 0.07.
• Able to stand unaided: 3 of 5 with cells, 0 of 3 without, p = 0.08.
• Muscle wasting: 1 of 6 with cells, 2 of 3 without, p = 0.13.

None of these differences reaches the conventional threshold of p below 0.05. At least one treated lamb scored only 2, which is no better than the controls.

Spinal cord under the microscope:

• Large neuron density: 9.9 per mm² of grey matter with cells against 6.3 without (p = 0.04).
• Grey matter area: 6.2 against 1.9 mm² (p = 0.01).
• Spinal cord area: 16.1 against 7.5 mm² (p = 0.02).
• Scar thickness around the cord: 70.9 against 253.7 micrometres (p = 0.10). Scar thickness at the repair site: 691.3 against 1,684.4 micrometres (p = 0.18). Both were lower with cells, without reaching significance.

The abstract describes the locomotor scores as means. The results table in the body of the paper labels them as medians with ranges, which also fits the statistical test used. This summary follows the table.

Conclusions:

The authors conclude that human UC-MSCs used alongside fetal myelomeningocele surgery were safe in sheep, that the cells appear to act through the substances they secrete rather than by turning into nerve tissue or spreading through the body, and that use in humans “will be considered following confirmation of these safety results over a longer period in the animal model”.

Limitations the authors state:

• Small numbers in each group, with probable lack of statistical power.
• A high fetal loss rate, although similar to other large-animal fetal surgery. In a review of 232 lambs from the leading US group, survival was only 43%.
• Lambs could not be kept alive beyond 24 hours for financial reasons, so nothing is known about medium or long-term safety or function. A follow-up study at 1 and 6 months of life is under way.
• The benefit was smaller than in the team’s earlier experiments with sheep UC-MSCs. They suspect donor-to-donor variability, since the three cords used may not have been strong producers of protective factors. They are testing two remedies, pooling cells from several donors and a potency test to select high-secreting donors.

Background Information:

The authors note that even after fetal surgery 71% of children cannot walk independently at age eight and 75% need daily urinary catheterisation at age ten. Several groups are therefore adding cells to the operation. A Californian group using placenta-derived cells has begun a human trial. The French group says it is the only one using umbilical cord cells.

In the team’s laboratory work, human UC-MSCs secrete brain-derived neurotrophic factor and hepatocyte growth factor, which protect nerve cells, and anti-scarring factors. In an earlier tracking study using fluorescent cells, the transplanted cells did not travel to distant organs.

For safety in people, the authors refer to two large meta-analyses of randomised trials of UC-MSCs for other conditions, mostly given intravenously to adults and children, in which transient fever was the only adverse event linked to the cells and no tumours were reported. None of those data come from use in a fetus, which the authors acknowledge has never been studied for UC-MSCs.