17 September 2010. No, you haven’t accidentally hit the home page of the World Trade Organization. Amyloid precursor protein (APP) does export iron—from neurons—according to a paper in the September 2 Cell online. Researchers led by Ashley Bush at the University of Melbourne, Victoria, Australia, and Jack Rogers at Harvard Medical School report that APP oxidizes Fe2+ to Fe3+, loading it onto transferrin, a major iron transporter in the brain. (Incidentally, the transferrin gene is a genetic risk factor for Alzheimer disease.) By freeing neurons of Fe2+, which can generate reactive oxygen species (ROS), APP functions similarly to ceruloplasmin, a copper-based ferroxidase that prevents a rare form of dementia associated with iron accumulation in glia. According to the new research, zinc—which accumulates in amyloid plaques—poisons APP ferroxidase activity. This latter is drastically reduced in cortical tissue samples taken from AD patients, suggesting that excess zinc poisons APP in AD tissue. “The findings bring a lot of things together that people knew were implicated [in Alzheimer’s] but didn’t quite understand how they were associated. Now we can really see a big piece of the puzzle,” Bush said in an interview with ARF. The researchers propose a model, based on results from in vitro, cell, animal, and human tissue studies, whereby zinc, leaching from extracellular amyloid plaques, cripples APP ferroxidase activity, causing a buildup of iron that can, in turn, lead to oxidative damage and neurodegeneration. George Perry and colleagues from Case Western Reserve University, Cleveland, Ohio, noted in an e-mail to ARF that the work suggests that APP “represents a unique system, adapted to the brain, to cope with iron homeostasis,” (see full comment below).
“It had long been suspected that APP expression and function is regulated by iron and iron metabolism,” Rogers told ARF via e-mail. His group had previously uncovered an iron response element (IRE) in the 5’ untranslated region of APP mRNA (see Rogers et al., 2002). IRE sequences are found on transcripts for ferritin, the transferrin receptor, and other mRNAs that are suppressed by iron regulatory proteins (IRPs), which bind IREs when iron levels are low. When intracellular iron rises, it knocks IRPs off IREs, kick-starting protein synthesis. Recently, Rogers and colleagues reported that the same regulatory dance dictates APP translation (see Cho et al., 2010). Now, by outing APP as a ferroxidase and iron exporter, first author James Duce and colleagues show why that translational refinement makes sense.
In 2002, Rogers and colleagues noticed that APP contained an REXXE ferroxidase consensus motif and took out a patent on the discovery. Though only five amino acids long, this motif is very rare, said Bush, and is highly conserved among some ferroxidases in evolution. To see if the precursor protein functions as one of them, Duce and colleagues tested recombinant APP in an in vitro assay. Kinetic analysis revealed that it was better at oxidizing Fe2+ than ceruloplasmin. The authors then became suspicious that APP may act as the ferroxidase of cortical neurons, which do not express the copper ferroxidase. In follow-up experiments, Duce and colleagues found that APP bound to the ceruloplasmin partner ferroportin in cortical neurons (and in HEK293T cells, which also have no ceruloplasmin), and that primary cortical neurons from APP-negative mice accumulate Fe2+. The findings suggested that APP, itself a transmembrane protein, acts like the membrane-tethered ceruloplasmin, which takes iron from ferroportin, oxidizes it, and passes it out of the cell and onto transferrin.
How does this proposed role for APP fit with Alzheimer disease? For one, scientists know that sufficient iron accumulates in the brains of AD patients to be detected by magnetic resonance imaging (see Bartzokis et al., 1994; Pankhurst et al., 2008). In one study, iron magnetic resonance in the hippocampus correlated with rusty cognition as determined with the Mini-Mental State Exam (see Ding et al., 2009). Could waning APP ferroxidase activity be to blame for the iron buildup?
To take a first stab at this question, Duce and colleagues tested postmortem tissue samples from AD patients. Though total APP levels were normal (or even slightly higher than control tissue), ferroxidase activity was 75 percent lower than in cortical tissue from age-matched normal controls. Ferroxidase activity in AD cerebellar samples was no different from controls, however, suggesting something unique about cortical tissue. Because ferritin contains the same REXXE ferroxidase motif and is poisoned by zinc, which accumulates in amyloid-β plaques found in the cortex, Bush and colleagues wondered if the slightly larger metal might be preventing APP from oxidizing iron in the AD brain. Retesting the AD cortical tissue samples in the presence of a zinc chelator restored ferroxidase activity to normal levels. In addition, the researchers found a quantitative link between iron ferroxidase activity and amyloid plaques.
“We showed a beautiful inverse correlation between APP ferroxidase and Aβ levels in 23 cases,” said Bush. This association was statistically significant. Furthermore, the researchers found normal levels of APP ferroxidase in tissue samples from frontotemporal dementia and Parkinson disease patients. “So [ferroxidase inactivation] seems selective to tissue that has amyloid in it. Therefore, we could say that iron accumulation in the neocortex in Alzheimer disease is due to zinc, from amyloid, inhibiting APP ferroxidase,” said Bush.
One implication of this work is an added benefit of zinc-targeting drugs, such as clioquinol, which Bush co-developed as a therapy for AD. The rationale is that clioquinol, and the second-generation drug PBT2, block the interaction of zinc with Aβ (see ARF related news story). That could reduce the concentration of zinc in the extracellular space. “The implication is that drugs that we have been developing will rescue APP ferroxidase inhibition,” said Bush. He said Prana Biotechnology Ltd., plans to start a Phase 2b/3 trial of PBT2 at the end of this year (see related news).—Tom Fagan.
Duce JA, Tsatsanis A, Cater MA, James SA, Robb E, Wikhe K, Leong SL, Perez K, Johanssen T, Greenough MA, Cho H-H, Galatis D, Moir RD, Masters CL, McLean C, Tanzi RE, Cappai R, Barnham KJ, Ciccotosto GD, Rogers JT, Bush AI. Iron-export ferrodixase activity of beta-amyloid precursor protein is inhibited by zinc in Alzheimer’s disease. Cell. September 17, 2010; 142:1-11. Abstract