Can targeting APP processing upstream of amyloid generation improve efficacy over downstream approaches?