Feasibility and Safety of Cellular Therapy for In-Utero Repair of Myelomeningocele (CuRe Trial): A First-in-Human, Phase 1, Single-Arm Study

Major Points and Findings:

The CuRe trial is the first published human trial of stem cells given before birth for spina bifida. Six unborn babies with myelomeningocele had the standard open fetal repair operation at the University of California, Davis, with one addition. A patch carrying living placenta-derived mesenchymal stem cells (PMSCs) was laid directly on the exposed spinal cord before it was closed. The report covers safety up to the newborn hospital stay only, and follows more than a decade of sheep studies by the same group.

The patch could be applied in every case, and no baby had a wound problem, a cerebrospinal fluid leak, an infection or a tumour at birth. Walking, bladder and bowel function are not yet reported, because the authors deliberately held back all motor observations. The treatment was given during fetal surgery at around 25 weeks of pregnancy, so the results do not apply to cell infusions given to a child after birth.

Aim:

To test whether PMSCs seeded on an extracellular matrix sheet (PMSC-ECM) can be applied during in-utero myelomeningocele repair, and whether doing so harms the repair site or the newborn.

Methods:

The study was a phase 1, single-centre, single-dose, single-arm study, overseen by the US Food and Drug Administration (IND 24097), the California Institute for Regenerative Medicine and an independent data safety monitoring board (DSMB). It was registered as NCT04652908.

Eligibility criteria copied those of the MOMS trial, the 2011 study that established fetal surgery for this condition:

• Pregnancy between 19 and 26 weeks
• Myelomeningocele or myeloschisis with its upper edge between T1 and S1
• Hindbrain herniation confirmed on fetal MRI
• Normal karyotype
• Excluded: twin or multiple pregnancy, a fetal anomaly unrelated to the spinal defect, a maternal reason not to operate, conditions raising the risk of preterm birth, and psychosocial limitations

Cells: PMSCs were grown from the chorionic villus tissue of donated placentas in the university’s Good Manufacturing Practice facility, then frozen. Seventy-two hours before each operation a batch was thawed and tested for identity, sterility and viability. The cells were seeded onto a commercially available dural graft (Cook Biodesign) at 300,000 cells per square centimetre, incubated for 24 hours, and tested for sterility again before use. The cells are allogeneic (from an unrelated donor).

Operation: Mothers had general anaesthesia with an epidural, a low transverse cut in the abdomen, and a 5 to 8 cm opening in the uterus. The neurosurgical repair followed the MOMS technique. The only change was placing the patch on the neural placode with the cells touching the cord, then closing the dura over it (or using the patch itself as the dural layer), then closing the fetal skin. The fetus was scanned weekly until delivery.

Enrolment was staggered, and no new mother was enrolled until the previous baby had been delivered and checked. Stopping rules were set against MOMS event rates. A single maternal or perinatal death, two cerebrospinal fluid leaks, or three wound dehiscences would have triggered DSMB review. Each repair site was photographed and examined within 24 hours of birth. Contrast MRI was done within two weeks of birth, or as soon as the baby was stable, to look for abnormal tissue growth.

Results:

The six pregnancies were enrolled between 21 June 2021 and 5 December 2022. The mothers were 23 to 36 years old and all six were White and non-Hispanic. Lesion levels were L1, L2/3, L3, L4, S1 and S1. Five fetuses had myelomeningocele and one had myeloschisis. All had hindbrain herniation and two had clubfoot. Surgery took place between 24 weeks 5 days and 25 weeks 5 days.

The patch was applied in all six. There were no technical complications, and no mother or fetus needed a transfusion, resuscitation or emergency delivery.

Birth outcomes were as follows:

• All six were delivered by caesarean at a median of 34 weeks 5 days (range 33 weeks 2 days to 36 weeks 6 days). Birthweights ran from 2,065 g to 3,565 g.
• Preterm premature rupture of membranes occurred in 4 of the 6 pregnancies. The paper lists this in a table and does not discuss it in the text. With six patients and no control group, a link with the cell product can be neither shown nor excluded.
• Two babies were born before 34 weeks. One needed intubation for transient respiratory distress related to prematurity. None went home on oxygen.
• Hospital stay until discharge home ranged from 5 to 30 days. The median corrected gestational age at discharge was 37 weeks 4 days.

Primary safety endpoints: All six newborns had an intact, healed repair site with no cerebrospinal fluid leak, no infection, no skin separation and no abnormal tissue on examination. Postnatal MRI showed reversal of hindbrain herniation in all six and no sign of tumour. No baby needed a shunt or other fluid diversion procedure before discharge. There were no deaths. The authors recorded no adverse event related to the cell product, and state that as of 1 August 2025 no child had had a wound complication attributable to it.

Leg movement, walking, bladder and bowel outcomes are not reported. The authors consider early motor observations after birth to be of uncertain validity and left them out.

Conclusions:

The authors conclude that the patch did not interfere with the known benefits of fetal surgery and caused no cell-related adverse effects. On that basis the DSMB and the FDA allowed the trial to continue as a phase 1/2a study enrolling 35 patients without staggering.

The limitations they state are a sample of six and a report confined to birth outcomes. There is no control group, and MOMS serves as a historical comparison. Children will be followed at 3, 6, 9, 12, 18, 24 and 30 months and then yearly to age 6, using standard infant motor scales (including the Bayley and Peabody scales), anorectal manometry and urodynamics. Only that follow-up can show whether the cells improve function.

Two of the authors hold patents on the technology, assigned to the University of California.

Background Information:

Fetal repair lowers the need for a shunt, but the authors note that 58% of children operated on before birth in MOMS still could not walk independently at 30 months. Surgery prevents further damage in the womb but does not repair damage already done. This is the rationale for adding cells. They are meant to work through the substances they release, reducing inflammation and nerve cell death. They are not expected to become part of the spinal cord, and the authors say long-term survival of the cells is not the goal.

Two specific safety concerns were addressed. The team checked that the anti-inflammatory action of the cells did not stop the fetal wound from healing, because in earlier sheep work an amniotic membrane patch had had that effect. They also watched for tumour growth, either from the transplanted cells or from fetal tissue stimulated by them. Neither was seen, but the authors accept that six babies followed to birth cannot settle the question.