10 November 2004. ApoE, the major lipoprotein of the brain, is one of the strongest risk factors for late-onset Alzheimer disease (AD), a fact that no doubt spurred the ApoE minisymposium at the 34th annual meeting of the Society for Neuroscience in San Diego last month.
Studies on ApoE have exploded since it was discovered that those carrying one copy of ApoE4 allele are more likely to get the disease than those with only ApoE2 or ApoE3—having two copies of ApoE4 puts one at even greater risk. But exactly how the lipoprotein increases the likelihood of getting AD has not been nailed down. There are plenty of theories (see ARF related news story), one being that it interferes with the processing of amyloid precursor protein (AβPP) and the production of amyloid peptides (Aβ). In fact, Mike Irizarry, Massachusetts General Hospital, Boston, has recently published data showing that all variants of the cholesterol transporter inhibit γ-secretase cleavage of APP, leading to reduced production of Aβ and accumulation of APP C-terminal fragments. Irizarry and colleagues had also found that ApoE reduces transcriptional activation mediated by the partnership of APP intracellular domain (AICD) and the transcriptional coactivator Fe65 (see Irizarry et al., 2004). But curiously, when they had examined the effect of ApoE on processing of Notch, another substrate of γ-secretase, they found little effect. So by what mechanism might the lipoprotein affect γ-secretase cleavage of APP?
This was the question Irizarry addressed. One possibility is that binding of ApoE to its receptor—LDL receptor related protein (LRP)—triggers some event that specifically modulates cleavage of APP by γ-secretase. To test this, Irizarry incubated cells with ApoE in the presence or absence of the LRP blocker RAP (he used a variety of cell types including primary cultured neurons). He found that Aβ production was attenuated under both conditions. In fact, even in cells devoid of LRP, ApoE affected APP processing. Next, Irizarry tested if ApoE might be involved in redistribution of cholesterol in the cell, a process that has been implicated in modulation of γ-secretase (see ARF related news story). When he measured efflux of cholesterol from cells incubated with ApoE, he found only a very small effect; ApoE2, for example, caused about a twofold increase in cholesterol efflux. However, similar amounts of high-density lipoprotein caused a 15-fold increase in cholesterol efflux without affecting APP processing, so the slight cholesterol efflux evoked by ApoE2 is unlikely to explain the effect on APP.
Sticking with the cellular distribution theme, Irizarry then looked to see if ApoE might affect the localization of APP. To test this, he used an antibody to the C-terminal of APP to locate the protein in the cell, finding that addition of ApoE caused the precursor to shift from the cell surface to intracellular granules. The location of presenilin 1, one of the major proteins of the γ-secretase complex, was unaffected. The results suggest that ApoE does not have a direct effect on γ-secretase, but instead keeps the precursor protein and the protease separated.
Yadong Huang, University of California at San Francisco, has slightly different ideas on the link between ApoE and AD. Huang has hypothesized that ApoE cleaving enzyme or, AECE, may play a major role in neurotoxicity because it cleaves ApoE near the C-terminus, leaving a truncated protein that is neurotoxic (see ARF related news story from the Institute for the Study of Aging symposium on ApoE in July 2003).
AECE, it turns out, is a chymotrypsin-like serine protease which can cleave ApoE in various locations including at leucine 268 and methionine 272. To determine what impact these different cleavages may have, Huang’s lab have made several truncated ApoE constructs and expressed them in mice. Animals expressing the protein truncated at amino acid 271 show age-related neurodegeneration and loss of CA3 neurons in the hippocampus. The protein colocalizes with synaptophysin, suggesting that it is present in the presynapses, but it is not found near the dendritic marker MAP2, Huang reported. In contrast, mice expressing a shorter, 241-amino acid protein that lacks the lipid binding domain seem normal, indicating that lipid interaction may be essential for the truncated ApoE’s neurotoxicity.
The physiological relevance of this work has been questioned because of lack of evidence that neurons normally express ApoE, but Qin Xu in Huang’s group showed that ApoE can be expressed in CNS neurons under stress (SfN abstract 442.6). Using immunostaining and in situ hybridization, Xu showed that ApoE is expressed in CA1 hippocampal neurons that are subjected to excitotoxic injury using kainic acid. Xu has produced chimeric mice that have green fluorescent protein expressed under the endogenous ApoE promoter and hopes to use these animals to probe, in real-time, changes in expression of the lipoprotein.—Tom Fagan.