It's late, and I'm tired, so I don't entirely trust my results here. However, if I'm right, this will explain these results.
The theoretical maximums, simply caused by diffraction, are 1555 lp/ph for f/5.6 and 2199 lp/ph for f/4.
So, the reason the 100-400L improved less is that there's less room to improve between the old one and the diffraction limit than there was for the old 400/4DO which has a higher diffraction limit due to its faster f-stop.
In fact, the result for the 100-400L II is just 1% less than what I'd calculate from a perfect f/5.6 lens behind an AA filter. And that 1% number is based on a rule-of-thumb for the AA filter that itself is less accurate than 1%.
We probably will see a bigger difference on high MP bodies where the lens will have a bigger impact. I'm not about to fool with calculations.