With the exception of familial cases, which can be traced to specific mutations, it is unclear why there’s a glut of amyloid-β (Aβ) in the brains of Alzheimer disease patients. The most simplistic explanations are that too much is formed or not enough of it is cleared out of the brain. Recent papers in Nature Medicine and the Journal of Biological Chemistry reveal factors that might affect Aβ supply and demand.

First, the supply-side economics. Todd Golde, Eddie Koo, and colleagues from the Mayo Clinic, Jacksonville, Florida, and the University of California at San Diego report in the April 17 online Nature Medicine that there are many small molecules, both natural and synthetic, that might lead to overproduction of Aβ. First author Thomas Kukar and colleagues screened a variety of compounds for their ability to modulate the activity of γ-secretase, the membrane protease that catalyzes the last step in the production of Aβ from Aβ precursor protein (AβPP). Their work was no doubt prompted by their previous findings that some nonsteroidal anti-inflammatories (NSAIDs), including indomethacin, can have an allosteric effect on γ-secretase, leading to a decrease in production of the most amyloidogenic form of the peptide, Aβ42, and a concomitant increase in shorter Aβs, such as Aβ38 (see ARF related news story). But this time, Kukar and colleagues found that some NSAIDs actually have the opposite effect, increasing the production of Aβ42.

In the screen, the authors used H4 cells expressing AβPP with the Swedish mutation, and Chinese hamster ovary (CHO) cells expressing wild-type precursor to measure Aβ production. Among the most potent compounds that turned up were the cyclooxygenase-2 (COX-2) inhibitor celecoxib, and fenofibrate, a peroxisome proliferator-activated receptor α (PPARα) agonist, which is prescribed to reduce plasma cholesterol. Celecoxib, at 10 micromolar, led to almost double the amount of Aβ42 being released from the cells (as measured by an ELISA assay), while having very little effect on production of Aβ38 and Aβ40. Fenofibrate caused up to a 3.5-fold increase in production of Aβ42 and a twofold decrease in production of Aβ38, though at admittedly very high concentrations (over 100 micromolar). These increases are similar to those resulting from familial AD mutations in AβPP or presenilin, the catalytic component of γ-secretase (see Scheuner et al., 1996).

The authors found that the compounds had similar effects in test tube assays, suggesting that the drugs act directly on γ-secretase, as opposed to on some activator, such as the ROCK/RhoA kinase that has been shown to contribute to the shedding of membrane protein ectodomains (see ARF related news story on statins, isoprenoids, and Aβ42 production). In fact, even more surprising is the finding that the RhoA activators farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP) also led to increased production of Aβ42 in the in vitro assays, suggesting that these isoprenoids can also modulate γ-secretase directly. In support of this idea, Kukar and colleagues found that including farnesyl and geranylgeranyl transferase inhibitors in these assays failed to prevent the isoprenoid activation of Aβ42 production.

The authors conclude that “the elevations in Aβ42 by these compounds are comparable to the increases in Aβ42 induced by Alzheimer disease-causing mutations in the genes encoding amyloid-β protein precursor and presenilins, raising the possibility that exogenous compounds or naturally occurring isoprenoids might increase Aβ42 production in humans.”

On the demand side of the equation, Rudy Tanzi and colleagues from Massachusetts General Hospital report in the April 25 Journal of Biological Chemistry that clearance of Aβ might be a significant problem in AD patients (this paper was first published online February 21).

First author Robert Moir and colleagues compared Aβ autoantibodies circulating in the plasma of AD patients with those in normal control volunteers. Such measurements have been made before, but only on antibodies to synthetic, unmodified, monomeric Aβ, which is only a fraction of the total Aβ in the brain. Instead, Moir and colleagues focused on autoantibodies to oligomeric, cross-linked species.

Tanzi and colleagues previously generated these CAPS, or cross-linked β-amyloid protein species, by using copper as an oxidizing reagent, and showed that CAPS are similar in size and stability to oligomers of Aβ extracted from AD brain tissue. Moir used the CAPS to capture autoantibodies from plasma of 59 AD patients and the same number of controls, and found a significant reduction, of about 25 percent, in the AD plasma. The authors also found that the levels of autoantibodies correlated with age at onset of disease: the lower the age of onset the lower the antibody titer.

The authors point out that the nature of the autoantibodies is unclear. The chemical modifications seen in CAPS are common to many proteins, for example, so the higher titer in normal plasma could reflect antibodies to non-Aβ epitopes. But plasma antibodies to Aβ have been shown to act as a peripheral sink, drawing Aβ out of the brain, so the results could reflect lower clearance of Aβ from the brain in AD patients. In fact, passive immunotherapy using intravenously administered immunoglobulin has recently shown promise as a treatment for AD (see ARF related news story). The authors write that their findings may be useful for diagnosis and facilitating future designs of reagents for Aβ vaccination and antibody perfusion therapies aimed at treating and preventing AD.—Tom Fagan


  1. In this interesting paper, Kukar and colleagues report that many NSAIDs and their derivatives selectively raise Aβ42 and lower Aβ38 secretion. These results are surprising, given the group’s original finding that another subset of NSAIDs do just the opposite, namely, reduce Aβ42 and increase Aβ38 production (Weggen et al., 2001). What makes their work even more intriguing is that endogenous isoprenoids, GGPP and FPP, also increase Aβ42 and decrease Aβ38, suggesting that this effect may have physiological relevance. The compounds do not activate RhoA and ROCK, thus excluding protein isoprenylation as a mechanism, but instead appear to act directly on the γ-secretase complex. Importantly, treatment of wild-type and Tg2576 mice with Celecoxib and a novel compound, FT-1, increased cerebral levels of Aβ42 levels but left those of Aβ40 unchanged. Their results raise the potential concern that exposure to similar compounds in the environment may raise Aβ42 in humans and cause AD.

    An intriguing aspect of this work is that rather large and diverse sets of compounds appear to directly bind to the γ-secretase complex and modulate the cleavage site on APP. This suggests that γ-secretase may be quite sensitive to conformational changes. Indeed, the authors propose that their compounds mimic FAD mutations, and it is likely that FAD mutations have an effect, although probably small, on γ-secretase conformation. Slight conformational changes could subtly shift the relative positions of APP and the γ-secretase active site, and thus alter the spectrum of Aβ species that are generated. While the physiological relevance of Aβ remains unknown, one could speculate that the different Aβ isoforms may have specific biological functions. If so, the isoprenoids may be the endogenous modulators of Aβ isoforms by acting directly on γ-secretase to adjust the Aβ population to suit particular physiological requirements.

    Accumulating evidence implicates a role for isoprenoids in APP and Aβ metabolism. Indeed, our recent results (Cole et al., in press) demonstrate that low isoprenoid levels (induced by statin treatment) cause the intracellular accumulation of Aβ. We attribute this effect to reduced isoprenylation of small GTPases (e.g., rhos and rabs) that are required for vesicular transport. In contrast, Kukar et al. raised isoprenoid levels in the cell and observed the reciprocal increase and decrease in Aβ42 and Aβ38, respectively. We propose that isoprenoids may be functioning in two separate pathways, depending on their concentrations in the cell: At low concentrations, isoprenoids may primarily affect the activity of small GTPases that in turn influence Aβ trafficking, while at high concentrations, they may directly modulate the γ-secretase complex to alter Aβ isoform balance. The cellular effects of the isoprenoids are clearly complex, and further studies are required to investigate these hypotheses and determine the multifaceted actions of the isoprenoids in APP and Aβ metabolism.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Anti-inflammatory Drugs Side-Step COX Cascade to Target Aβ
  2. Statins and AD—What Role Isoprenoids?
  3. Pilot Study Shows Promise of Passive Immunotherapy

Paper Citations

  1. . Secreted amyloid beta-protein similar to that in the senile plaques of Alzheimer's disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked to familial Alzheimer's disease. Nat Med. 1996 Aug;2(8):864-70. PubMed.

Further Reading

Primary Papers

  1. . Diverse compounds mimic Alzheimer disease-causing mutations by augmenting Abeta42 production. Nat Med. 2005 May;11(5):545-50. PubMed.
  2. . Autoantibodies to redox-modified oligomeric Abeta are attenuated in the plasma of Alzheimer's disease patients. J Biol Chem. 2005 Apr 29;280(17):17458-63. PubMed.