A tesseract isn't a "4d rotation". It's a 4d generalisation of a cube. The problem with a general shape is that there is no generalisation to 4d because there's no generalisable rule for generating it.
Think about a square. The rule for generating it is to write down all possible combinations of two zeros and ones - (0,0), (0,1), (1,0), and (1,1) - then connect points whose coordinates differ in only one place. That exact same rule generates a cube if you change "two" to "three". Now think of an arbitrary polygon drawn on a sheet of paper. How do you generalise the rule for drawing that particular arbitrary polygon to 3d?
The same is true generalising a 3d shape to a 4d one. If you can write a general rule for creating the shape that works in an arbitrary number of dimensions then you can generalise it to a 4d equivalent. But this is not possible for arbitrary polygons.
That said, you can always embed a 3d object in 4d space, just as you can imagine a 2d object in 3d space. You simply take your (x,y,z) coordinate triples and make them (x,y,z,0). You could then rotate this and view its 3d projection. The result would be to scale the object along one of its directions, and possibly to distort it slightly if perspective effects are simulated.