Do you need stable expression for your experiments and do you expect a change in cell proliferation or survival as a function of aVb3?
Stable vs. transient expression is a huge issue here. Selection can be a tricky business, in that it can result in clonal artifacts, and can fail to yield the desired phenotype if there is a counter-selection process. For example if high integrin expression slows down proliferation or induces apoptosis in your system, then selection for "stable" cells will result in low expressors overtaking your population.
Due to the large size of integrin cDNAs, they do not express well from transgenes, and it is not possible to encode both an alpha and beta chain in the same viral construct, as this will exceed the packaging constraints of most viral vectors. In some cases, excess, unliganded integrin can result in unexpected and perhaps non-physiological responses. You would need to consider carefully how to control for such effects.
If endogenous aVb3 integrin is present in your cell line of interest, there are many good experimental tools to try and perturb the function of the endogenous integrin, rather than pursue the tricky (from both an interpretation and technical standpoint) experiment of forced expression. aVb3 can be blocking by adding cyclic RGD or the LM609 function-blocking antibody. Ligand-binding competent aVb3 levels can be checked by the WOW-1 antibody. These tools are relatively robust and reliable, in my experience.