A loading control is usually a 'house keeping" type product that is commonly and widely expressed and not affected by the treatment/manipulation/phenomenon you are studying
It depends on the cell line/primary cells you are using and what you are studying. For example this paper http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3204251/ shows that GADPH is not suitable for some cell types, but other markers are. What cells are you using and what treatment will they get?
Daniel is right, there is generally no one perfect control. GAPDH expression is also known to change in response to changes in nutrient levels. I've also found at least one example of significant beta actin expression change:
In general, I would always recommend to use at least two, rather even three housekeeping genes, at least as long as you are not sure which ones are stable. We routinely also use HPRT. When using all three in parallel, you can get a quite good feeling which ones are stable in your cell system wrt to your perturbations. Then you can stick to the two most stable ones to do your analyses.
Thank you so much guys for quite satisfying answers. FYI Daniel, actually i am working with different brain regions of rats and checking expression level of some neurotransmitters' receptor after long term administration of psychostimulants.
Ah, so you are comparing kind of different cell types? Or mainly comparing with and without treatment? Because when comparing different cell types the whole thing gets much more complicated and unstable. In that case I would really try to make sure to use exactly the same amount of RNA/cDNA, test different housekeepers in some replicates and hope that one or two of them appear to be stable. Otherwise the choice of the normalizer can change your results quite dramatically.