Autoantibodies are secreted by plasma cells and have an essential role in driving the renal manifestations of autoimmune diseases such as systemic lupus erythematosus and antineutrophil cytoplasmic autoantibody-associated vasculitis. Effective depletion of autoreactive plasma cells might be the key to curative treatment of these diseases. Two major plasma-cell compartments exist: short-lived plasmablasts or plasma cells, which result from differentiation of activated B cells, and long-lived plasma cells, which result from secondary immune responses. Long-lived plasma cells reside in survival niches in bone marrow and inflamed tissue and provide the basis of humoral memory and refractory autoimmune disease activity. Unlike short-lived plasmablasts, long-lived plasma cells do not respond to conventional immunosuppression or to therapies that target B cells. Existing therapies that target long-lived plasma cells, such as proteasome inhibitors and antithymocyte globulin, as well as promising approaches that target survival factors, cell homing or surface molecules, deplete the whole memory plasma cell pool, including cells that secrete protective antibodies. By contrast, we have developed a novel strategy that uses an affinity matrix to deplete pathogenic long-lived plasma cells in an autoantigen-specific manner without removing protective plasma cells. Targeting B-cell precursors to prevent replenishment of autoreactive long-lived plasma cells should also be considered.