How to classify mouse tissue cells across time?

[buddingyeast]

I have mouse datasets from the same tissue across multiple developmental timepoints performed in biological duplicate. It seems straightforward to integrate cells between two duplicates at a given timepoint, but should I also integrate across time?
It seems quite clear looking at the data that some marker genes and presumably cell types are maintained across time, but others are not. Biologically this makes sense as many cells are going to differentiate over time. Are there any strategies to link cells across 2 or more timepoints beyond integration?

Thanks.