I agree, there is no way to compare cytoplasm to nucleus by WB, just differentially treated samples from the same compartment. If the protein changes are not extreme (like p65 after TNFa treatment) it will be hard to observe them.
Fractionation protocol is too long stepwise to assume that you have the same concentrations of protein at the end of it all. I quantify the protein sample using BCA or bradford prior to running it and load the same amount (from each compartment- the cyt and nuc are incomparable).
For quantification of protein shuttling I used an Imagestream analysis (Facs).
I also think you should modify your fractionation protocol. Usually you pellet the nuclei @ 1700-1500 G for 10 min, not max speed for 10 sec.
I agree with both Chris and Lee. Compare changes of target proteins pre/post treatment in each compartment separately, then compare changes with respect to each other. Internal control or housekeeping protein for each compartment also important to normalize your data.