In a report in Brain Research earlier this month, Pat McGeer and coworkers address this question by comparing mRNA levels of the major plaque-related enzymes in postmortem samples of seven AD cases and eight controls. The scientists extracted RNA from areas of high plaque pathology (hippocampus, midtemporal gyrus), areas of moderate plaque pathology (midfrontal, occipital, and motor cortices), and areas of light plaque pathology (caudate, medial thalamus, cerebellum, substantia nigra), as well as from the heart, liver, spleen, and kidney. Then they amplified, and quantified, cDNAs of the three known AβPP isoforms (695, 751, 770), β-secretase (BACE), presenilin-1 (PS1, thought to be γ-secretase), and the A-β-degrading enzyme neprilysin.

All mRNAs were expressed in all tissues. At the same time, pronounced and tissue-specific differences clearly indicated that in Alzheimer's, plaque-prone brain areas generate more A-β and remove less of it than other brain areas, than control brain, and than peripheral organs.

APP695 was by far the dominant isoform in brain, confirming previous work in the field. It was expressed at much higher levels in hippocampus and midtemporal gyrus from AD than control brain, but there was no significant difference between AD and controls in brain areas of low plaque formation. BACE levels showed no significant difference between Alzheimer's and control samples in either the brain or the periphery. PS1, however, was much more highly expressed in brain than in the periphery. Every tested brain area except one had significantly higher PS1 levels in AD samples versus controls, with hippocampus and midtemporal gyrus showing the greatest difference. Neprilysin showed the opposite trend. It was much more highly expressed in peripheral organs than in brain; the lowest levels occurred in Alzheimer's hippocampus and midtemporal gyrus.

Together, these data explain why plaques grow abundantly in certain brain areas but not others, and never in peripheral organs, says McGeer. They also explain the previous paradox that PS1 and PS2 levels did not correlate with where pathology appears in AD when these enzymes were initially measured in human and animal brain. PS expression patterns make sense when viewed in the context of all Aβ-related proteins examined in this study, McGeer says.

The work does not speak to whether the AD-specific expression pattern is central to, or a consequence of, the fundamental disease process, the authors write. McGeer did not include insulin-degrading enzyme or other potential Aβ-degrading enzymes (see related news item) because he considers the evidence implicating them to be weaker than the data for neprilysin.-Gabrielle Strobel.

Reference:Yasojima K et al. Relationship between b amyloid peptide-generating molecules and neprilysin in Alzheimer's disease and normal brain. Brain Res 2001 Nov 16;919(1):115-21 Abstract