Simply put MU - is for Multi User; MIMO is Multi In Multi Out - is the technology that allows the radio to use multiple streams of data at once.
The 2x2,3x3,4x4 refers to how the antenna's are arranged and it simply equates to how many streams it can process, so 2x2 can handle 2 streams, a 4x4 can handle 4 streams.
So a 3x3 can handle 3 users?
Not really... it can handle 3 streams at once, those streams could all go to one user.
So a 4x4 is best then.
Yes that's true but it comes at a cost and depending on your situation it might be overkill for what you need. It then does relate to how many users you have per access point and what they are doing, Browsing the web has lower demands than watching streaming TV for example.