Preclinical Stem Cell Therapy in Fetuses with Myelomeningocele: A Systematic Review and Meta-Analysis

Major Points and Findings:

This is a systematic review of animal research. No children were treated. A team led from University College London collected every published study up to May 2020 in which stem cells were given to animal fetuses with myelomeningocele, the most severe form of spina bifida, and pooled the results where the studies were similar enough.

They found 26 studies, mostly in rats and sheep, and most of them used mesenchymal stem cells. Giving stem cells before birth did not lower the number of animals that survived to term. In rats, an injection of stem cells into the fluid around the fetus made it much more likely that skin grew over the defect. In sheep, adding human placenta-derived mesenchymal stem cells to fetal surgery improved leg movement after birth. The authors call the findings encouraging, but note that most studies had a high risk of bias, that all animals were assessed only up to 24 hours after birth, and that no human trial had yet been done.

Aim:

To summarise the safety and efficacy of stem cells given in the womb to animal models of myelomeningocele, and to judge whether the evidence justifies moving to a clinical study.

Methods:

The review followed the PRISMA guidelines and was registered in advance with PROSPERO (CRD42019160399). Five databases (MEDLINE, Embase, Web of Science, Scopus and the Cochrane Library) were searched from their start until May 2020, and reference lists of earlier reviews were checked by hand.

Studies were included if animals with myelomeningocele received stem cells of any type in the womb and were compared with animals given fetal surgery alone, a saline injection or no treatment. Studies giving cells only after birth, and studies not in English, were excluded.

Two reviewers selected the studies and extracted the data independently, with a third settling disagreements. Risk of bias was rated with the SYRCLE tool for animal studies. Results were pooled in meta-analyses only where the animal model, cell type and outcome were the same. The outcomes were survival, coverage of the defect, the spinal cord under the microscope, and nerve function.

Results:

Of 648 records found, 54 were read in full and 26 studies were included.

What was studied:

• Cell type: mesenchymal stem cells in 20 of 26 studies (77%), from placenta, amniotic fluid or bone marrow. The others used embryonic stem cells, neural stem cells, neural crest stem cells or skin grown from induced pluripotent stem cells. None used umbilical cord cells.
• Human cells were given to the animals in 12 studies (46%).
• Animals: rats in 15 studies (58%), in which the defect was caused by giving the mother retinoic acid; sheep in 7 (27%), in which the defect was created surgically; chick embryos in 3 and rabbits in 1.
• Every study assessed the animals at birth or up to 24 hours after birth.

Safety (survival to term):

• Rats given an injection of rat amniotic fluid stem cells into the amniotic fluid: no difference in survival against saline (relative risk 1.03, 95% CI 0.92 to 1.16).
• Sheep given human placenta-derived stem cells on the spinal cord during fetal surgery: no difference in survival against surgery alone (relative risk 0.94, 95% CI 0.78 to 1.13).
• Overall the review found no significant effect of stem cells on survival in any model.

Coverage of the defect:

• Thirteen studies, all in rats, chick embryos or one rabbit study, looked at whether skin grew over the defect.
• In four rat studies pooled together, a single injection of amniotic fluid stem cells into the amniotic fluid made complete coverage far more likely than saline (relative risk 16.35, 95% CI 3.27 to 81.79). The coverage was mostly a thin, rudimentary skin.
• In the rabbit study, half of the treated animals showed some coverage but none had complete coverage.

Spinal cord and movement:

• Eleven studies reported on the spinal cord or on nerve function, and 10 of them used mesenchymal stem cells.
• In five sheep studies pooled together, adding human placenta-derived stem cells to fetal surgery improved the Sheep Locomotor Rating score, a 0 to 15 scale of hind leg function, by 5.18 points on average (95% CI 3.36 to 6.99) compared with surgery alone.
• Some of these sheep studies also found more large nerve cells (motor neurons) in the spinal cord, although the differences were not significant in every study.
• In rats, bone marrow stem cells injected into the spinal cord were linked to less cell death in the cord, more sensory nerve cells, and stronger nerve signals to a leg muscle.
• One sheep study using mouse neural stem cells found no improvement in walking (2 of 6 against 2 of 8). The cells did not turn into nerve cells but did produce the nerve growth factors GDNF and BDNF.

Quality of the studies: most studies carried a high risk of bias, mainly from selective reporting of outcomes and from poor description of how animals were allocated and whether researchers were blinded. None fully described animal housing or blinded outcome assessment.

The abstract gives seven studies in the meta-analyses, while the discussion gives nine.

Conclusions:

The authors conclude that injecting amniotic fluid stem cells into the amniotic fluid is safe and effective in helping the defect close in small animals, with no data yet in large animals, and that human placenta-derived stem cells placed on the spinal cord during fetal surgery are safe and improve leg movement in newborn lambs. They describe the findings as “encouraging for clinical translation”, but say the best cell source, dose and delivery device still need to be established, with proper toxicology studies, before a first human trial.

Limitations the authors state:

• Only studies published in English were included.
• Most included studies had a high risk of bias, with little detail on randomisation, allocation and blinding.
• Publication bias may lead to the benefit of stem cells being overestimated.
• No study followed the animals beyond 24 hours after birth, so longer-term nerve function is unknown.
• The response of the fetal immune system to donor cells has not been studied.

Background Information:

The authors explain that fetal surgery for myelomeningocele, usually done between 23 and 26 weeks of pregnancy, reduces the need for a brain shunt and improves walking, but is not a cure. At 30 months of age about half of children treated before birth still need a catheter to pass urine and more than half cannot walk without orthoses. Fetal surgery is also offered in only a limited number of centres.

They suggest that mesenchymal stem cells mainly help through the substances they release rather than by replacing nerve tissue. In the animal studies the cells calmed inflammation in the nervous system, supported nerve cells and encouraged new blood vessels, through growth factors, signalling proteins and small vesicles.

They also point out that amniotic fluid is already collected from most women considering fetal surgery, to check the baby’s chromosomes, so a baby’s own amniotic fluid stem cells could in principle be grown for treatment.