Xiao-Dan Cai

Normal processing of the amyloid beta protein precursor (beta APP) results in secretion of a soluble 4-kilodalton protein essentially identical to the amyloid beta protein (A beta) that forms insoluble fibrillar deposits in Alzheimer's disease. Human neuroblastoma (M17) cells transfected with constructs expressing wild-type beta APP or the beta APP717 mutants linked to familial Alzheimer's disease were compared by (i) isolation of metabolically labeled 4-kilodalton A beta from conditioned medium, digestion with cyanogen bromide, and analysis of the carboxyl-terminal peptides released, or (ii) analysis of the A beta in conditioned medium with sandwich enzyme-linked immunosorbent assays that discriminate A beta 1-40 from the longer A beta 1-42. Both methods demonstrated that the 4-kilodalton A beta released from wild-type beta APP is primarily but not exclusively A beta 1-40. The beta APP717 mutations, which are located three residues carboxyl to A beta 43, consistently caused a 1.5- to 1.9-fold increase in the percentage of longer A beta generated. Long A beta (for example, A beta 1-42) forms insoluble amyloid fibrils more rapidly than A beta 1-40. Thus, the beta APP717 mutants may cause Alzheimer's disease because they secrete increased amounts of long A beta, thereby fostering amyloid deposition.

The 4-kilodalton (39 to 43 amino acids) amyloid beta protein (beta AP), which is deposited as amyloid in the brains of patients with Alzheimer's diseases, is derived from a large protein, the amyloid beta protein precursor (beta APP). Human mononuclear leukemic (K562) cells expressing a beta AP-bearing, carboxyl-terminal beta APP derivative released significant amounts of a soluble 4-kilodalton beta APP derivative essentially identical to the beta AP deposited in Alzheimer's disease. Human neuroblastoma (M17) cells transfected with constructs expressing full-length beta APP and M17 cells expressing only endogenous beta APP also released soluble 4-kilodalton beta AP, and a similar, if not identical, fragment was readily detected in cerebrospinal fluid from individuals with Alzheimer's disease and normal individuals. Thus cells normally produce and release soluble 4-kilodalton beta AP that is essentially identical to the 4-kilodalton beta AP deposited as insoluble amyloid fibrils in Alzheimer's disease.

The 4-kilodalton amyloid beta protein (A beta), which forms fibrillar deposits in Alzheimer's disease (AD), is derived from a large protein referred to as the amyloid beta protein precursor (beta APP). Human neuroblastoma (M17) cells transfected with constructs expressing wild-type beta APP or a mutant, beta APP delta NL, recently linked to familial AD were compared. After continuous metabolic labeling for 8 hours, cells expressing beta APP delta NL had five times more of an A beta-bearing, carboxyl terminal, beta APP derivative than cells expressing wild-type beta APP and they released six times more A beta into the medium. Thus this mutant beta APP may cause AD because its processing is altered in a way that releases increased amounts of A beta.

The approximately 4 kD (39-43 amino acid) polypeptide (amyloid beta protein, A beta) deposited as amyloid in Alzheimer's disease (AD) is derived from a set of 695-770 residue precursor proteins collectively referred to as the amyloid beta-protein precursor (beta APP). Using immunoblotting techniques, metabolic labeling, and sequencing we have analyzed beta APP derivatives in medium conditioned by: (1) human mononuclear leukemic (K562) cells expressing a model beta AP-bearing carboxyl-terminal beta APP derivative (2) human neuroblastoma (M17) cells transfected with constructs expressing full length beta APP and (3) M17 cells expressing only endogenous beta APP. In each case, we observed the release of a approximately 4 kD beta APP derivative essentially identical to the A beta found in AD amyloid. A similar, if not identical, beta APP fragment was readily detected in CSF from both Alzheimer's disease patients and controls. These observations indicate that the A beta is produced and released by normal processing of the beta APP. To determine if the production of A beta or A beta-tearing COOH-terminal beta APP derivatives is altered in cells expressing the mutant beta APPs linked to familial AD, we have compared M17 cells expressing wild type beta APP with those expressing mutant beta APPs (beta APP delta I or beta APP delta NL). After continuous metabolic labeling for 8 hours, cells expressing the beta APP delta NL mutant showed a 5-fold increase in the relative amount of an approximately 11.4 kD A beta-bearing carboxyl-terminal beta APP derivative, and they released 6-fold more 4 kD A beta into the medium. These observations provide strong evidence that: (1) the pathway producing A beta in cultured cells is highly relevant to AD and (2) the beta APP delta NL mutant causes AD because its processing is altered in a way that releases increased amounts of A beta.

The 39–43 amino acid (∼ kD) amyloid β protein (Aβ) deposited as amyloid Alzheimer's disease is an internal peptide beginning 99 residues from the COOH end of a much larger amyloid fiprotein precursor (βAPP). In cultured cells, normal processing of the PAPP produces ∼8.7, ∼9.6, ∼10.9, and ∼ 11.4 kD COOH-terminal derivatives that appear to contain all or part of the Aβ domain. In this study, we metabolically labeled transfected human neuroblastoma (Ml 7) cells with PH] phenylalanine plus (35SJmethionine and then radiosequenced the immunopre-cipitated COOH-terminal βAPP derivatives taking advantage of the fact that the Aβ has phenylalanines at positions 4, 19, and 20, and a single methionine at position 35. Our analysis of the derivatives produced by transfected Ml 7 cells expressing βAPP695 confirms that the ∼8.7 kD COOH-terminal derivative begins at Aβ17 and indicates that the ∼9.6 and ∼10.9 kD derivatives begin at Aβ10 and Aβ4 respectively. Significantly, we find that the 11.4 kD derivative begins at APr Thus normal PAPP processing produces a potentially amyloidogenic COOH-terminal derivative that has the Aβ domain intact at its amino terminus. We have previously shown that cells expressing βAPPWL, a mutant linked to familial Alzheimer's disease, produce an increased amount of the 11.4 kD COOH-terminal derivative and secrete more Aβ. Radiosequencing of these derivatives showed that the ANL mutant is cleaved at the same location as wild type βAPP producing an 11.4 kD COOH-terminal derivative and Aβ that both begin at Aβ Thus the ANL mutation appears to accelerate a cleavage that releases an 11.4 kD COOH-terminal derivative identical to that normally produced from wild type PAPP, and it appears that this 11.4 kD derivative is further processed to release excess Aβ.