Various data suggest that Alzheimer's disease results from the accumulation of amyloid beta (A beta) peptide fibrils and the consequent formation of senile plaques in the cognitive regions of the brain. One approach to lowering senile plaque burden in Alzheimer's disease brain is to identify compounds that will increase the degradation of existing amyloid fibrils. Previous studies have shown that proteoglycans and serum amyloid P (SAP), molecules that localize to senile plaques, bind to A beta fibrils and protect the amyloid peptide from proteolytic breakdown. Therefore, molecules that prevent the binding of SAP and/or proteoglycans to fibrillar A beta might increase plaque degradation and prove useful in the treatment of Alzheimer's disease. The nature of SAP and proteoglycan binding to A beta is defined further in the present study. SAP binds to both fibrillar and nonfibrillar forms of A beta. However, only the former is rendered resistant to proteolysis after SAP association. It is interesting that both SAP and proteoglycan binding to A beta fibrils can be inhibited by glycosaminoglycans and Congo red. Unexpectedly, Congo red protects fibrillar A beta from breakdown, suggesting that this compound and other structurally related molecules are unlikely to be suitable for use in the treatment of Alzheimer's disease.