Wost

No, you can't make the assumption that actin will be equally expressed across cell types, as you can't with any other gene that I am aware of, but it can be used internally within one cell type

The problem is that even within the same cell type expression of so-called HKP might significantly vary depending on your experimental condiitions. So as soon as you change a single parameter e.g. by treating your cells with a compund of interest the expression level of HKP might change. The problem is very similar to validation od RT-PCR where it is now well acdcepted that normalization/validation should be based on multiple housekeeping genes rather than a single one. The more data points (proteins) you use for normalization the better. So in consequence, as indicated in my other post, a total protein detection approach can be more robust and less prone to errors. You might use reversible stains like Ponceau or downstream compatible detections like stain-free to visulize total protein content for each lane on your blot. Another advantage of this approach is that differences in protein load can be identified before you start wasting a lot of time on blocking, antibody incubation, washing etc.