That would be awesome. Perhaps if it were possible dump the state of memory B cells (I'm wondering if flow cytometry can do this), classify and inventory immunoglobulins, filter them out via perfusion, and custom program naive B cells and add them back to the body. Memory T cell types seem way messier but some should be tunable with epigenetic levers. We will eventually learn how to hack the adaptive immune system, but it's going to take a lot of research time and effort to produce to a specific autoimmune therapy for a specific clinical problem likely customized to a single individual.