Mesenchymal Stromal/Stem Cells in Chronic Incomplete Traumatic Spinal Cord Injury: A Phase I/II Double-Blind Placebo-Controlled Multicentre Trial

Major Points and Findings:

This Spanish trial is the most rigorous test so far of a treatment approach close to the one used at our clinic, in which donor umbilical cord (Wharton’s jelly) mesenchymal cells are given by lumbar puncture. The design was randomised, double-blind and placebo-controlled. The treatment was safe but did not work better than placebo on any outcome. People in both groups improved over the year. In chronic spinal cord injury, improvement after an injection therefore does not show that the injection had an effect.

The trial was small (18 people) and used only two modest doses, so it does not rule out a benefit from other regimens. Results from uncontrolled studies should nevertheless be read with caution.

Aim:

To assess the safety of two intrathecal doses of allogeneic Wharton’s jelly MSCs (WJ-MSCs) in chronic incomplete cervical spinal cord injury, and to look for effects on neurological scores, function, pain, spasticity, nerve conduction tests and quality of life over 12 months.

Methods:

Phase I/II, randomised 1:1, double-blind, placebo-controlled, at Institut Guttmann (Badalona) and the University Hospital Complex of A Coruña. Registered as NCT05054803. Recruitment ran from October 2021 to June 2022.

Participants had to meet the following criteria:

• Traumatic injury between C1 and T1, confirmed on MRI
• Incomplete injury, AIS grade B, C or D
• 1 to 5 years since injury
• Age 18 to 70, living near a study centre

Exclusions included ventilator dependence, pregnancy, spinal surgery planned within 12 months, neurodegenerative disease, cancer within five years, any previous cell therapy, and any contraindication to lumbar puncture or rehabilitation.

Fifty-four people were screened and 36 excluded: 17 because their injury was complete (AIS A), 15 because it was not traumatic, 2 with tumours and 2 who declined. Eighteen were randomised, nine to each arm. The placebo arm had four AIS B, three C and two D. The cell arm had two, four and three. Mean age was 41, and mean time since injury was 2.3 years (placebo) and 2.8 years (cells).

The treatment: Cells came from the Barcelona blood and tissue bank, made under GMP from donated cords, with a mean viability of 91%. The dose was 0.7 to 1.3 million viable cells per kg, given twice, at baseline and at 3 months. Placebo was a saline solution with 2% human albumin in an identical coded syringe. At each session about 4 mL of spinal fluid was withdrawn at L3-L4, about 4 mL of product was injected over 2 to 5 minutes, and the participant lay flat under observation for 2 to 4 hours. Everyone followed an individualised rehabilitation programme for roughly 6 to 8 months.

Assessments: ISNCSCI motor, light touch and pin prick scores were recorded at 1, 3, 4, 6 and 12 months. Independence (SCIM III), walking (WISCI II), grip strength, spasticity, neuropathic pain and quality of life were measured at 6 and 12 months. Evoked potentials, reflex studies, electrical perception thresholds, mood and community integration were assessed at baseline and 12 months. Spinal fluid was checked for anti-HLA antibodies and donor DNA. Seventeen participants completed follow-up. One man in the cell arm, found to be 77 and so over the age limit, left after the first injection because he could not tolerate the nerve tests. He was kept in the safety analysis only.

Results:

Safety: There were 80 adverse events, 26 in the placebo group and 54 in the cell group. Fifty-two were mild, 24 moderate and 4 severe. The severe events (urinary tract infections, a pressure ulcer and a respiratory infection, in one placebo and two cell recipients) were judged unrelated to treatment. No life-threatening or unexpected events occurred.

• Post-puncture headache occurred in 6 of 9 on placebo and 7 of 9 on cells. One cell recipient needed hospital admission for it.
• The table of events linked to the procedure or possibly to treatment lists urinary retention with autonomic dysreflexia in three cell recipients and autonomic dysreflexia in a fourth. No placebo recipient had these. One cell recipient had two days of low-grade fever.
• No new anti-HLA antibodies and no persisting donor DNA were found in spinal fluid.

Motor scores: Both groups improved significantly. Placebo rose from a mean of 44.0 to 52.4 (p = 0.009), cells from 47.8 to 54.6 (p = 0.048). The difference between groups was not significant (p above 0.6).

The remaining outcomes were as follows:

• Light touch and pin prick scores showed no significant change within either group and no difference between groups.
• For AIS grade, one participant improved (C to D), and that person was in the placebo group. No cell recipient changed grade.
• SCIM III independence scores rose slightly in both groups. The estimated difference at 12 months was 2.16 points (95% CI -0.80 to 6.08, p = 0.130).
• Grip strength improved a little in both groups, which the authors put down to rehabilitation, natural recovery or test familiarity.
• Evoked potentials, reflex studies and sensory thresholds showed no treatment effect.
• Spasticity was stable. Neuropathic pain was low to begin with. Mean scores went from 2.9 to 1.9 on placebo and 2.1 to 2.2 on cells.
• Quality of life, mood and community integration did not differ.

Conclusions:

The authors conclude that intrathecal WJ-MSCs were feasible and well tolerated, and that their findings “do not provide evidence supporting the superiority of intrathecal MSC monotherapy delivered in a two-dose regimen over a placebo”. They stress that spontaneous fluctuation, rehabilitation and practice effects can all produce apparent improvement, and that without a control group such changes “could easily be misinterpreted as evidence of efficacy”.

The authors state several limitations. The sample was too small to detect modest effects, especially across such varied injuries, and no power calculation was made for efficacy. Follow-up of 12 months may be too short. Passive spread through spinal fluid may deliver few cells to a scarred lesion, and two doses may be too few. Rehabilitation differed between participants, multiple outcomes were tested without correction, and blinding success was not checked.

Background Information:

To explain why the cells may not have helped, the discussion points to the biology of a long-standing injury, with dense glial scar, cystic cavities and an environment that blocks regrowth. MSCs act mainly by releasing short-lived signalling molecules and do not replace nerve cells. That signal may be too brief to change a stable scar.

The authors contrast their result with open-label work. They cite a Jordanian study comparing autologous bone marrow MSCs with umbilical cord MSCs, the Mayo Clinic adipose MSC study in which 7 of 10 improved an AIS grade, and a single-patient report of repeated high doses. They describe this evidence as “anecdotal rather than definitive”.

For future work they suggest more intensive dosing, such as three intrathecal doses of about 100 million cells, tested under controlled conditions. They also propose combining cells with lesion-targeted delivery or scaffolds, and selecting people with more preserved tissue or a shorter time since injury.