Systematic review and meta-analysis of preclinical studies testing mesenchymal stromal cells for traumatic brain injury

Abstract

Mesenchymal stromal cells (MSCs) are widely used in preclinical models of traumatic brain injury (TBI). Results are promising in terms of neurological improvement but are hampered by wide variability in treatment responses. We made a systematic review and meta-analysis: (1) to assess the quality of evidence for MSC treatment in TBI rodent models; (2) to determine the effect size of MSCs on sensorimotor function, cognitive function, and anatomical damage; (3) to identify MSC-related and protocol-related variables associated with greater efficacy; (4) to understand whether MSC manipulations boost therapeutic efficacy. The meta-analysis included 80 studies. After TBI, MSCs improved sensorimotor and cognitive deficits and reduced anatomical damage. Stratified meta-analysis on sensorimotor outcome showed similar efficacy for different MSC sources and for syngeneic or xenogenic transplants. Efficacy was greater when MSCs were delivered in the first-week post-injury, and when implanted directly into the lesion cavity. The greatest effect size was for cells embedded in matrices or for MSC-derivatives. MSC therapy is effective in preclinical TBI models, improving sensorimotor, cognitive, and anatomical outcomes, with large effect sizes. These findings support clinical studies in TBI.

Conflict of interest statement

The authors declare that the study was conducted with no commercial or financial relationships that could be construed as creating potential competing interests.

© 2021. The Author(s).

Figures

Fig. 1. PRISMA flow diagram of the studies.

Flow chart represents the selection process. A total of 80 studies were included in the meta-analysis.

Fig. 2. Characteristics of the 88 studies in the qualitative synthesis, quality score, and risk of bias.

a Number of publications per year. b World map with a color scale indicating the number of papers published in each country (image adapted from Freepik.com). c–f Pie charts of features of publications related to MSC source (c), time of administration (d), delivery route (e), and modifications (f). g Distribution of quality scores. h Percentages of studies meeting each quality score criterion. i Risk of bias: percentages of low risk (green), unclear risk (yellow), and high risk (red) for each category.

Fig. 3. Sensorimotor outcome assessment up to 5 weeks post-treatment.

Forest plots show mean effect size and 95% CI of naïve and labeled MSCs for neurologic assessments (a), hindlimb function (b), and coordination (c) up to 5 weeks post-treatment. The graph on the right shows the last time point (up to 5 weeks) for all studies. Dots represent the single studies, and diamonds indicate pooled data. Vertical gray bars represent the mean and 95% CI of the pooled estimated effect size.

Fig. 4. Summary of sensorimotor, cognitive, and anatomical outcomes.

Forest plots show mean effect size and 95% CI of naïve+labeled MSCs for all sensorimotor tests (a), cognitive assessment (b), and contusion volume (c). Dots represent the single studies, and the diamonds represent pooled data. Vertical gray bars indicate the mean and 95% CI of the pooled estimated effect size.

Fig. 5. Subgroup meta-analysis on MSC-related variables.

Forest plots of neurologic assessments at last time points of naïve+labeled MSCs stratified by source (a), type of transplant (b), time (c), and route of administration (d) and dose (e). Dots represent the single studies, and the diamonds represent pooled data. Vertical gray bars indicate the mean and 95% CI of the pooled estimated effect size.

Fig. 6. Effect of MSC modifications.

Forest plots show mean effect size and 95% CI of genetically modified MSCs, matrix-embedded MSCs, and MSC derivatives for neurologic assessment (a), cognitive function at 4–5 weeks (b), and contusion volume at 4–5 weeks (c). Dots represent the single studies, and diamonds pooled data. Vertical pale gray bars indicate the mean and 95% CI of the pooled estimated effect size of the single modifications and vertical dotted bars represent 95% CI of the naïve + labeled MSC group.

Fig. 7. Begg’s funnel plots.

Funnel plots representing publication bias for sensorimotor outcomes of naïve+labeled MSCs at the last time points: a neurological, b hindlimb function, c and coordination assessments. Publication bias for cognitive function of naïve+labeled MSCs assessed at early (d) or late (e) post-treatment time points. Publication bias for contusion volume at early (f) or late (g) post-treatment time points.

Fig. 8. Graphical summary of the results, illustrating the main categories of variables in the meta-analysis.

The top panel refers to the stratified meta-analysis on neurologic assessment using naïve + labeled MSCs. The background colors of the circles are indicative of the effect size (as absolute SMD) as in the color scale at the bottom. Image created with BioRender.com.