Infusion of Human Umbilical Cord Tissue Mesenchymal Stromal Cells in Children With Autism Spectrum Disorder:

Major Points and Findings:

This is a small phase I safety study from Duke University. Twelve children with autism spectrum disorder received one, two or three intravenous infusions of mesenchymal stromal cells grown from donated umbilical cord tissue (hCT-MSCs). The authors count only two other published reports of cord tissue MSCs in autism.

The study answers a narrow question. It shows that the product can be made to GMP standards and that infusions in young children with autism are medically tolerable. It was not designed to show whether the cells work. Half of the children improved on two of three autism measures, but there was no control group and the authors say plainly that the placebo effect in autism trials means the improvements cannot be attributed to the treatment. Five of the nine children without anti-HLA antibodies at baseline developed them after treatment, without symptoms.

Aim:

To test the safety and feasibility of intravenous infusions of allogeneic, third-party hCT-MSCs in children with autism, and to collect descriptive data on autism symptoms for planning a later randomized trial.

Methods:

The study was an open-label, phase I dose-escalation trial at Duke University, conducted under FDA IND 17313 and registered as NCT03099239.

Children aged 2 to 11 with a DSM-5 diagnosis of autism, informed by the ADOS-2, were eligible. Children were ineligible if they had a coexisting psychiatric condition (other than ADHD), a genetic syndrome, a history of autoimmune disease or immunosuppressive therapy, an available qualified autologous cord blood unit, or any previous cellular therapy. They also needed genetic testing that did not indicate a cause for the autism, normal Fragile X testing, normal organ function and a normal immune screen by history and absolute lymphocyte count.

Twelve children were enrolled in 2017, nine boys and three girls, median age 6.4 years (range 4 to 9). Median nonverbal IQ was 38.5 (range 22 to 91).

The cells came from cord tissue donated to the Carolinas Cord Blood Bank, expanded to passage 2 under GMP and cryopreserved in 10% DMSO. Three lots from three donors were used, each child receiving cells from a single lot. Each dose was 2 × 10⁶ cells per kilogram, given through a peripheral IV over 30 to 60 minutes after premedication with diphenhydramine and methylprednisolone. Cohort 1 (three children) had one infusion, cohort 2 (three) had two, and cohort 3 (six) had three, at two-month intervals, 27 doses in total.

Safety was assessed by examination, blood tests, immune profiles and anti-HLA antibodies at baseline and six months, with questionnaires up to 12 months after the last dose. Autism measures were the Vineland-3 Socialization subscale (improvement defined as a rise of 3 points or more), the PDDBI Autism Composite (a fall of at least 5 points) and the clinician-rated CGI-Improvement at six months, all analysed descriptively.

Results:

All 12 children received their planned doses. Two product-related adverse events occurred. One child in cohort 2 had a hypersensitivity reaction with mild hypotension during the second infusion, on an occasion when the parent had asked for diphenhydramine to be omitted; the infusion was stopped, fluids and extra methylprednisolone were given, and the child recovered fully. One child in cohort 3 had moderate hypotension (lowest 78/30) after the third infusion, which resolved with IV fluids. The authors think both were most likely reactions to DMSO.

A further 66 nonserious adverse events, all mild and none attributed to the product, were recorded in 11 of 12 children. The most common was agitation during IV placement and infusion, which resolved the same day in every case. Twenty-two psychiatric or behavioural events were reported in seven children: aggression (2), agitation (5), anxiety (3), defiant behaviour (2), depression (1), emotional lability (1), insomnia (3), intentional self-injury (1) and stereotypies (4). Three children accounted for 17 of these 22. Events per child rose with the number of doses (medians of 2, 5 and 7.5 in cohorts 1, 2 and 3) but the trend did not reach significance (p = 0.05). There were no serious adverse events, no graft-versus-host disease and no concerning changes in blood tests.

Three children already had class I anti-HLA antibodies at baseline. Of the nine who did not, five developed low-titre antibodies by six months, persisting beyond 12 months. All three with broad-spectrum antibodies had received the same lot, and all four children at least haploidentical to their donor developed antibodies, against one of eight less well matched. None had any clinical effect.

On the autism measures, 6 of 12 children improved on at least two of the three (four on all three, two on two). Two of the six had received one dose, two had two and two had three, so dose number was not linked to improvement. Ten of 12 improved on at least one measure. On the CGI-I, one child was rated “much improved”, eight “minimally improved” and three “no change”.

Conclusions:

The authors report “favorable safety outcomes of a small, phase I trial” and state that “additional trials are needed to fully assess the safety and efficacy of this approach”. The infusions were “safe and well tolerated” and “suggestions of improvement in core symptoms of ASD were described in 50% of participants”. On this basis the centre opened a phase II randomized, double-blind trial (NCT04089579).

The limitations are the authors’ own. “Because this was an open-label trial and in light of the well-known placebo effect in ASD clinical trials, it is uncertain whether these improvements are related to the treatment.” The sample “may be too small to detect” any relationship between antibodies, dose number and outcome. Outcome was variable, with some children improving substantially and others not at all. The two infusion reactions led the authors to insist that MSC therapy in children be given “in a controlled setting with a secure IV, physiological monitoring, and medical personnel in attendance”.

Background Information:

The authors’ rationale is that neuroinflammation, microglial activation or immune dysregulation may drive symptoms in a subset of children with autism, and that MSCs modulate those processes through paracrine signalling rather than engraftment. Their product was shown in the laboratory to suppress T-cell responses and reduce microglial activation before clinical use.

Of the two earlier reports they cite, one gave cord blood mononuclear cells with or without intrathecal cord tissue MSCs, with transient fever in five children as the only treatment-related side effect. The other gave about 36 million cells intravenously every 12 weeks to 20 children, with no premedication and multiple donors per child; mild to moderate adverse events followed 20% of doses. The agitation seen during infusions, and the lack of any link between dose number and improvement, led the authors to plan a higher single dose rather than repeated doses for their next trial.