Ideally, internal control is a housekeeping gene that shows basal expression under all conditions. That is why B-actin is commonly used as a loading control. So, its expression should be consistent throughout, under all test conditions. However, minor variants can be understood due to technical errors while protein estimation, loading, and detection. That is very normal. But what you are reporting, is a very significant deviation. This should not have been the case. Maybe you can try again. Try to trace back the protein concentrations you loaded? Was it precise?
You should also check the available literature for any relation between AD and the B-actin levels. If it still does not help, then changing b-actin would become essential. You can also use GAPDH (ideal one), B-tubulin, Histone H2B, or cytochrome c oxidase. All these work really well and have been well documented in the literature. Good luck to you!