Some authors use B-actin and others GAPDH, but I haven't found any explanation. Even within the same article, gapdh is used to normalize proBDNF, and B-actin to BDNF. Could someone give me an explanation?
there shouldn't be a better one, unless there is a specific modulation of the housekeeping in your experiment. You have a 16 and a 28 kDa protein to detect and they both run far enough from b-actin (42kDa) and GAPDH (36kDa) to blot them without stripping. If you have to check, on the same membrane, the level of some other proteins choose the housekeeping that is not on the way.
It is basically based on the experiment as well as on the tissue you are performing the experiment in. As an example: For studies of tumor tissue in comparison to nonmalignant tissue GAPDH should not be used since GAPDH is upregulated depending on the malignant state (the worse the more) - meaning it is not a great laoding control for this kind of experiment. It can also be regulated by influencing glycolysis (e.g. by inhibitors etc). But it certainly depends on the experiment.
For your problem it might be because BDNF can occur as a primer with 27 kDa (if it is a non-denaturing gel), too close to GAPDH. Maybe that is the fact why they use Actin instead of GAPDH.
I'm doubting in my election not because their molecular weight, but rather because BDNF is involved in cell morphological and metabolic changes, that may vary actin and GAPDH when BDNF vary.