Human Umbilical Cord Blood Mesenchymal Stem Cells Ameliorate Autism-Like Behaviors in a Valproic Acid-Induced Mouse Model via the IGF-1/Akt Signaling Pathway:

Major Points and Findings:

This laboratory study was carried out in mice and demonstrates no effect in children. It is included here because it addresses a question the clinical trials cannot answer, namely the mechanism by which cord-derived cells might change anything at all. The clinical literature repeatedly notes that the assumed mechanism has never been confirmed in humans. Animal work of this kind is where a testable answer is likely to come from.

Aim:

To test whether human umbilical cord blood mesenchymal stem cells (hUC-MSCs) improve autism-like behaviour in a mouse model, to check for toxicity, and to identify the molecular pathway responsible.

Methods:

Pregnant Kunming mice were given valproic acid (300 mg/kg intraperitoneally) on gestation days 12 and 13, a long-established way of producing autism-like behaviour in rodent offspring. The offspring then received 5 × 10⁵ hUC-MSCs injected into the lateral ventricles of the brain.

Three behavioural tests were used: a three-chamber social test (n = 14 control, 13 valproic acid, 22 treated), a marble-burying test for repetitive behaviour and an open field test for anxiety (n = 14, 11, 15 respectively). The authors also used Western blot for protein phosphorylation, quantitative PCR for gene expression and Golgi staining to count dendritic spines, and they grew primary neuron cultures in medium conditioned by the cells.

Results:

Behaviour: Some measures improved and others did not. In the social novelty test, control and treated mice preferred an unfamiliar mouse over a familiar one, whereas untreated valproic acid mice showed no such preference. Treated mice buried significantly fewer marbles, indicating less repetitive behaviour. Anxiety-like behaviour did not meaningfully change. Distance travelled and mean velocity in the open field were comparable to untreated animals, and although treated mice tended to spend more time in the centre, that difference was not statistically significant. The effect was therefore limited to sociability and repetitive behaviour.

Cortical neurons in untreated valproic acid mice had reduced dendritic spine density, and treatment restored it. The authors state this is the first time the effect of hUC-MSCs on cortical dendritic development in autism has been visualised and confirmed at a microscopic level.

Conditioned medium, which contains what the cells secrete but no cells, promoted dendritic sprouting and branching in cultured neurons on its own. The authors take this as support for a paracrine mechanism, in which the benefit comes from secreted factors and does not depend on cells replacing tissue. The same reasoning lies behind the exosome and nasal-spray trials now starting in humans.

Pathway: Treated mice showed restored phosphorylation of IGF-1R and Akt. Gene expression shifted in a consistent direction. GAP-43 and synaptophysin (both markers of synaptic plasticity) and the anti-inflammatory IL-10 rose, while the pro-apoptotic Bax and Caspase-3 and the pro-inflammatory IL-6 and IL-1β fell. The IGF-1/Akt pathway is a plausible target because it drives neuronal survival, suppresses apoptosis and supports axon and dendrite extension.

Toxicity: Over two weeks there was no effect on body weight or general health. No deaths occurred, and H&E staining showed no organ abnormalities.

Conclusions:

hUC-MSCs relieved core social dysfunction in this model through neurodevelopmental, anti-inflammatory and anti-apoptotic effects mediated by IGF-1/Akt signalling.

Background Information:

The authors list the following limitations of their work:

• No multi-omics evidence (RNA-Seq or proteomics) to confirm the molecular picture.
• No protein-level validation of the key inflammatory mediators, owing to limited sample.
• No cell residual counts, no long-term survival tracking, no biodistribution analysis, so where the cells went and how long they persisted is unknown.
• The valproic acid model produces autism-like behaviour but does not reproduce the genetic heterogeneity of autism in people, which limits how far these findings can be translated.

The last limitation is one that applies to all work of this type. A mouse model generates hypotheses about patients and provides no evidence in them. A mechanism demonstrated in an inbred strain with a chemically induced phenotype may or may not be the mechanism operating in any individual child.