Amyloid deposits regress when the supply of fibril precursor proteins is sufficiently reduced, indicating that amyloid fibrils are degradable in vivo. Serum amyloid P component (SAP), a universal constituent of amyloid deposits, efficiently protects amyloid fibrils from proteolysis in vitro, and may contribute to persistence of amyloid in vivo. Drugs that prevent binding of SAP to amyloid fibrils in vivo should therefore promote regression of amyloid and we are actively seeking such agents. A complementary strategy is identification of critical molecular processes in fibrillogenesis as targets for pharmacological intervention. All amyloidogenic variants of apolipoprotein AI contain an additional positive charge in the N-terminal fibrillogenic region of the protein. This is unlikely to be a coincidence and should be informative about amyloidogenesis by this protein. The two amyloidogenic variants of human lysozyme, caused by the first natural mutations found in its gene, provide a particularly powerful model system because both the crystal structure and folding pathways of wild-type lysozyme are so well characterized. The amyloidogenic variant lysozymes have similar 3D crystal structures to the wild type, but are notably less thermostable. They unfold on heating, lose enzymic activity, and aggregate to form amyloid fibrils in vitro.