The overproduction of beta-amyloid (A beta) appears to be a primary cause of Alzheimer's disease (AD). A beta can be generated by proteolytic cleavage of precursor protein (beta APP) at beta- and gamma-secretase sites in both disease and normal cells. There is, however, no evidence that proteolytic processing of beta APP in sporadic AD-affected tissues differs qualitatively or quantitatively from that occurring in normal cells, and additional pathways for the enhanced production of A beta in sporadic AD which constitutes the majority of all AD cases should be considered. The major factor limiting the production of A beta in normal cells is cleavage at the alpha-secretase site within the A beta sequence. But, whereas the intact beta APP is a substrate for cleavage at the alpha-secretase site, the immediate precursor of A beta, 12-kDa C-terminal beta APP fragment, is not susceptible to the alpha-secretase cleavage but it can be cleaved by gamma-secretase thus generating A beta. Moreover, the gamma-secretase cleavage is not the rate-limiting step in the production of A beta. Therefore, the increase in production of the 12-kDa C-terminal beta APP fragment may be an efficient way to overproduce A beta. A mechanism for the generation of the 12-kDa fragment independently of beta APP is proposed. It postulates an additional step of amplification of mRNA, namely the antisense RNA-mediated generation of a truncated mRNA encoding 12-kDa C-terminal fragment. Initiation of translation at the first AUG in the truncated mRNA results in a polypeptide that is cleaved by gamma-secretase generating A beta. The proposed model makes several verifiable predictions and suggests new directions of experimentation that may lead to a better understanding of the mechanisms involved in AD.