Seems like it'd be ideal to just work with a single strain; you'd rule out many confounding factors. I don't see why having multiple different strains of mouse would make the results more generalisable to humans.

It's still a controlled trial with some compelling results and analysis. This is an excellent basis for further research.