By using a novel, robust, fully-automated approach, we have shown:
- The density of activated microglia is a very sensitive measure of disease state in the α-synuclein PFF mouse model
- The phenotypic state of the microglia can be more finely quantified using:
- The distribution of activation scores, which can disentangle effects from the increase in the cell density and the changes in their morphology
- The distribution of activated microglia across different subclasses with distinct morphology
This approach may provide sensitive measures for the preclinical assessment of the efficacy of putative disease-modifying therapeutic agents in this mouse model of Parkinson's disease.
In summary, we have shown that our approach can precisely identify cells and characterize the morphology of microglia. Using our well-established mouse alpha-synuclein seeding & spreading model, we have shown that the density of activated microglia, as classified by our machine learning model, can be a very sensitive metric of the disease state. Compared to the Iba-1 stain density, it shows increased significant differences between groups.
In addition, we have shown how the phenotypic state of the microglia can be more finely quantified. First, using the distribution of activation scores from our machine learning model, we have observed how we could disentangle effects from the increase in cell density and the changes in the morphology. Second, the distribution of activated microglia across different subclasses with distinct morphologies could also be quantified.
The quantification of microglial morphology on Iba-1 stained IHC sections may provide a sensitive measure in preclinical therapeutic efficacy studies in this model.