Abstract
Microglia morphological studies have been limited to the process of reviewing the most common characteristics of a group of cells to conclude the likelihood of a "pathological" milieu. We have developed an Imaris-software-based analytical pipeline to address selection and operator biases, enabling use of highly reproducible machine-learning algorithms to quantify at single-cell resolution differences between groups. We hypothesized that this analytical pipeline improved our ability to detect subtle yet important differences between groups. Thus, we studied the temporal changes in Iba1+ microglia-like cell (MCL) populations in the CA1 between P10-P11 and P18-P19 in response to intrauterine growth restriction (IUGR) at E12.5 in mice, chorioamnionitis (chorio) at E18 in rats and neonatal hypoxia-ischemia (HI) at P10 in mice. Sholl and convex hull analyses differentiate stages of maturation of Iba1+ MLCs. At P10-P11, IUGR or HI MLCs were more prominently 'ameboid', while chorio MLCs were hyper-ramified compared to sham. At P18-P19, HI MLCs remained persistently 'ameboid' to 'transitional'. Thus, we conclude that this unbiased analytical pipeline, which can be adjusted to other brain cells (i.e., astrocytes), improves sensitivity to detect previously elusive morphological changes known to promote specific inflammatory milieu and lead to worse outcomes and therapeutic responses.