The human ApoE4 isoform is the biggest single risk factor for sporadic Alzheimer disease (AD). Through direct interaction with Aβ, ApoEs promote clearance of Aβ across the blood brain barrier (see, for example, Shibata et al., 2000), but they also drive the formation of toxic Aβ fibrils (see Ma et al., 1994). The ApoE4 allele correlates with increased Aβ deposition in AD (see Schmechel et al., 1993), so could inhibiting the ApoE/Aβ interaction be a useful therapeutic strategy? Recent work from Martin Sadowski and Thomas Wisniewski’s labs at New York University School of Medicine suggests that it may. Reporting in the December 5 PNAS, Sadowski and colleagues show that a modified fragment of Aβ that blocks the interaction of ApoE with endogenous Aβ reduces fibrillar and amyloid Aβ deposits in mouse models of AD, and also reduces cerebral amyloid angiopathy. In addition, a recent paper from Guojun Bu and colleagues at Washington University, St. Louis, Missouri, and Lilly Research Laboratories, Indianapolis, Indiana, shows that ApoE can drive endocytotic uptake and intracellular accumulation of Aβ raising the possibility that an inhibitor like the Aβ fragment might also protect against intraneuronal Aβ toxicity, which has become the focus of much recent research activity.

Sadowski and colleagues previously showed that an Aβ fragment (residues 12-28) containing solely the ApoE binding domain can complex with ApoE in vitro and prevent it from driving formation of toxic Aβ fibrils (see Sadowski et al., 2004). Now they have adapted this peptide for in vivo application, switching valine 18 for proline to make it non-fibrillogenic, using all D-stereoisomers to make it less biologically active, and protecting the N- and C-terminals by acetylation and amidation, respectively. The modifications generated a peptide, designated Aβ12-28P that has decreased immunogenicity and extended serum half-life.

Before testing the peptide in AD mouse models, Sadowski and colleagues made sure that it could still bind to ApoE. They found that though it does not bind as strongly as the native Aβ12-28 in vitro, the dissociation constant for the ApoE interaction is still very low (Kd is 32 +/- 5 nM), giving an inhibition constant of 13 nM, which is in the therapeutic range. Next, the researchers tried the inhibitor in single (APPSwe) and double (APPSwe/PS1) transgenic mice. They administered the drug for six months, three times a week by intraperitoneal injection, starting when Aβ deposits first appear in these models (age 12 months for single, and two months for double transgenics). After the treatment period the researchers found that total Aβ deposits, as determined from immunohistochemical analysis with two different antibodies (6E10 and 4G8), were reduced by about 50 and 45 percent in the neocortex and hippocampus, respectively, of both transgenic strains. Thioflavin S staining revealed that fibrillar Aβ was also reduced by 38 and 32 percent, respectively in the neocortex and hippocampus of single transgenics, and 22 and 32 percent in double transgenics. They found that ApoE deposition within Aβ plaques was also significantly reduced, again by about 50 percent in the two different mouse strains.

As the authors point out, one potential drawback to therapies that prevent aggregation of Aβ is that they might increase levels of soluble oligomeric forms of the peptide, which are now generally considered to be the most toxic (see ARF related news story). But the researchers found that levels of Aβ in formic acid extracts, representing soluble forms of the peptide, were also reduced. Soluble Aβ40 and 42 were lower by 25 percent in APPSwe mice and 44 and 32 percent, respectively, in APPSwe/PS1 transgenic animals. The authors also found that the reduction in Aβ load was accompanied by behavioral improvements. In a radial arm maze test of working memory, treated animals performed just as well as wild type, while untreated transgenic animals had significant reduction in working memory capability, making significantly greater numbers of errors.

The results indicate that Aβ12-28P does indeed block ApoE driven aggregation and deposition of Aβ in vivo and that this protects the animals against memory loss. But there may be other possible modes of action, including induction of an immune response or alteration of cholesterol levels (ApoE is a major cholesterol binding protein in the brain). That the researchers found a transient increase in total serum cholesterol following a single i.v. injection of Aβ12-28P suggests that alteration of lipid profiles might be linked to the effects of the treatment. However, Aβ antibody profiles were no different between treated and untreated animals, indicating that the peptide does not spur a strong immune reaction. Though active and passive immunotherapies are, of course, being actively pursued as potential therapies in their own right (see ARF related news story), there are some concerns that this strategy might clear parenchymal plaque burden but contribute to increased deposits of Aβ in the blood vessel deposits, potentially leading to increased cerebral amyloid angiopathy (see ARF related news story). But Sadowski and colleagues found that in APPSwe mice, Aβ12-28P reduced Aβ deposits in cortical blood vessels by 70 percent. “This observation demonstrates an additional therapeutic benefit of blocking the ApoE/Aβ interaction that has not been observed with immunization against Aβ,” write the authors.

As for intraneuronal Aβ, which may be the most toxic form to neurons (see ARF related news story), the report by Bu and colleagues in the November 24 Journal of Biological Chemistry, shows that ApoE and the low density lipoprotein receptor-related protein (LRP) have a hand in boosting that particular cache of Aβ.

The LRP receptor is expressed in the soma and dendrites of neurons and is involved in internalizing many different ligands, including ApoE. It also binds to Aβ. The role of LRP in AD is still somewhat debatable, with some studies suggesting that the receptor aids in Aβ clearance(see ARF related news story) and other suggesting that it promotes Aβ accumulation. On the latter score, Bu and colleagues have shown that overexpression of a mini human LRP receptor (mLRP2), which behaves by all counts like the full length, in the PDAPP mouse model of AD leads to increased levels of soluble Aβ and compromised memory performance (see ARF related news story). That study was done in aged, 22 month old mice. Now Bu and colleagues address what happens in young mice before plaque deposits are apparent, a fundamental question in the study of AD etiology and pathology.

First author Celina Zerbinatti and colleagues report that membrane-bound (Triton X-100 extractable) and insoluble (guanidine solubilizable) Aβ42 were significantly elevated (by 20 and 12 percent, respectively) in PDAPP/mLRP2 mice compared to PDAPP controls. “Although the effects are small, they demonstrate the involvement of LRP in regulating brain Aβ levels,” write the authors. Because they found no changes in APP levels or APP processing (CTF-β levels were unchanged) it is tempting to conclude that the increased Aβ levels resulted from reduced clearance. In support of this idea the authors found that levels of CSF Aβ were significantly decreased in PDAPP/mLRP2 vs PDAPP mice.

To determine exactly where Aβ was ending up, the authors used confocal microscopy to co-localize the peptide with cell markers. They found that in PDAPP animals Aβ42 localized with the neuronal marker NeuN and also with the lysosomal marker LAMP-1, suggesting that the peptide in located internally with lysosomes. Interestingly, the researchers also found significant loss of NeuN-co-localized Aβ42 in mice lacking ApoE compared to animals homozygous for the lipoprotein gene, suggesting that ApoE may contribute to the accumulation of intraneuronal Aβ. “Our results support the hypothesis that LRP binds and endocytosis Aβ42 both directly and via ApoE but that endocytosed Aβ42 is not completely degraded and accumulates in intraneuronal lysosomes,” write the authors.

The demonstration that PC12 cells clear Aβ from cell culture medium more rapidly if the cells also expressed mLRP2 adds weight to that argument. That ApoE3 or ApoE4 particles also increased that clearance suggests that blocking the ApoE/Aβ interaction, as with Sadowski’s Aβ peptide, may help reduce intraneuronal Aβ, too.—Tom Fagan

Comments

  1. Sadowski and colleagues examine the effects of a modified Aβ peptide 12-28P that has a proline in position 18, is acetylated at the N-terminus, amidated at the C-terminus, and is made of D-amino acids. In previous studies and in this paper, it is shown that this peptide can inhibit the ApoE4-stimulated fibrillogenesis of Aβ40 as well as ApoE4 binding to Aβ40. They then administered this peptide (1mg, 3x per week) over months, systemically, to two types of APP transgenic mouse models that develop plaques. The treatment resulted in improved performance in the radial arm maze, decreased plaques, decreased Aβ levels by ELISA, and decreased CAA. The effects of the peptide in vivo appear clear, and this approach, given its systemic nature of treatment, appears quite promising. This work should stimulate greater efforts in the important area of how to utilize ApoE as a therapeutic target, given its key role in AD pathogenesis.

    There are some issues regarding the mechanism of the effects that are seen that can be clarified in future studies. In regard to the effects of 12-28P, whether its actions in the brain on Aβ deposition are via ApoE could be further clarified. Is the effect of 12-28P on Aβ deposition and behavior (CNS effects) via 12-28P binding ApoE in the brain, in the blood, both, or via another effect? Does the treatment reverse existing deposits or only prevent new ones? While mouse ApoE is fibrillogenic in vivo, human ApoE isoforms, including ApoE4, delay the onset of Aβ deposition relative to the absence of ApoE. It will be useful to determine the effects of this peptide in APP/E4 mice. Finally, many treatments that can decrease Aβ in the brain affect processes which not only include ApoE/Aβ interactions but also Aβ degrading enzymes, secretases, and synaptic activity (to name but a few). It will also be important to assess 12-28P effects in regard to these types of activities.

    View all comments by David Holtzman
  2. We have shown that blocking the interaction of Aβ and ApoE with a synthetic peptide-Aβ12-28P that mimics the ApoE binding site on Aβ constitutes a highly promising approach for reducing Aβ accumulation in the CNS and preventing memory impairment in AD transgenic (Tg) mice (Sadowski et al., 2006). Aβ12-28P was designed to be non-toxic and non-fibrillogenic, as well as having improved BBB permeability and increased serum half-life compared to residues 12-28 of Aβ (please also see Sadowski et al., 2004). In our studies, we have demonstrated that the therapeutic effect of Aβ12-28P can be only achieved in the presence of ApoE. We have shown that Aβ12-28P does not function as a β-sheet breaker; hence, its effect on Aβ deposition cannot be mediated by direct disaggregation of Aβ deposits but by neutralizing of ApoE's effect on Aβ aggregation and its sequestration within the CNS. We have also demonstrated that Aβ12-28P does not stimulate production of anti-Aβ antibodies (since it is a very weak immunogen and is given without an adjuvant). Therefore, its therapeutic action cannot be mediated through an anti-Aβ vaccination effect.

    One significant advantage of this novel approach is that blocking the ApoE/Aβ interaction not only reduces deposition of Aβ in the brain parenchyma, but also reduces the burden of cerebral amyloid angiopathy, without inducing perivascular hemorrhages. This contrasts with vaccination approaches which do not reduce cerebral amyloid angiopathy and may increase the risk of intracerebral bleeding. Observations of Pattson et al. (2006) made on brains of subjects vaccinated with AN-1792 indicated that although the burden of parenchymal deposits is effectively decreased, vascular deposits remain unchanged or may even be increased. This study also demonstrated that vaccinated subjects showed an increase in the pool of soluble Aβ. Therefore, approaches focusing on ApoE as a therapeutic target offer a unique opportunity to address both parenchymal and vascular deposition of Aβ.

    Although our studies have been designed to study mainly the effect of blocking the ApoE/Aβ interaction on Aβ deposition in the CNS, there are numerous indications that this approach may offer additional therapeutic benefits. For example, several lines of evidence suggest the importance of ApoE in mediating neuronal re-uptake of Aβ via LRP receptors. This leads to neuronal Aβ accumulation and degeneration, as recently demonstrated in the elegant study by Zerbinatti et al. (2006, with Drs. Bu and Holtzman as senior authors). Therefore, blocking the ApoE/Aβ binding may potentially reduce accumulation of Aβ inside neurons and the associated toxicity. Another potential mechanism by which blocking the Aβ/ApoE interaction may have a beneficial role is in the dynamics of Aβ oligomer and Aβ fibril formation. ApoE was demonstrated to interact both with Aβ oligomers (Naslund et al., 1995) and Aβ fibrils. Future experiments have been planned to investigate how blocking ApoE/Aβ binding affects formation of oligomers and their equilibrium with fibrillar Aβ. Furthermore, studies of Bell et al. (2006, with Drs. Zlokovic and Holtzman as senior authors) indicated that binding of Aβ to ApoE reduces its efflux across the BBB. Therefore, another potential benefit of ApoE/Aβ binding antagonists would be improved clearance of Aβ across the BBB.

    Unlike mouse ApoE, human ApoE exists in three different isoforms, E2, E3, and E4, which have differential effects on the risk of AD occurrence, the age of onset, and the magnitude of Aβ deposition. In vitro, all three isoforms were shown to promote Aβ fibril assembly with the ApoE4 isoform having the strongest effect. Transgenic mice studies designed to study Aβ deposition in the setting of various human ApoE have demonstrated that expression of human ApoE markedly delays Aβ deposition relative to mouse ApoE, but nevertheless, amyloid deposition occurs with an isotype gradation similar to that seen in AD patients (E4>E3>E2; Fagan et al., 2002). Therefore, although our studies have shown a benefit of blocking the ApoE/Aβ interaction, they have to be expanded to include analysis of the relative benefit in various transgenic lines expressing different human ApoE isoforms. Such studies will address the effectiveness of blocking ApoE/Aβ binding in the setting of different human ApoE isoforms, which is a prerequisite for advancing this form of treatment into clinical studies. Since it has been shown that development of fibrillar Aβ deposits is dependent on ApoE3 and E4 expression, it is likely that blocking either the Aβ/ApoE3 or E4 interaction will be beneficial. However, it is possible that the magnitude of the therapeutic response to Aβ/ApoE binding antagonists will be dependent on ApoE isotype expression.

    Although ApoE has been somewhat neglected in the past decade of AD research, it constitutes a valid and attractive therapeutic target. We hope that our study will spur more extensive research toward evaluating this alternative therapeutic approach for AD, which may result in the generation and testing of novel compounds applicable to humans. The use of compounds blocking the ApoE/Aβ interaction would not preclude the additional use of other emerging treatment strategies such as secretase inhibitors or passive immunization. These Aβ targeting therapies would likely have a synergistic effect on the overall therapeutic outcome.

    References:

    . Blocking the apolipoprotein E/amyloid-beta interaction as a potential therapeutic approach for Alzheimer's disease. Proc Natl Acad Sci U S A. 2006 Dec 5;103(49):18787-92. PubMed.

    . A synthetic peptide blocking the apolipoprotein E/beta-amyloid binding mitigates beta-amyloid toxicity and fibril formation in vitro and reduces beta-amyloid plaques in transgenic mice. Am J Pathol. 2004 Sep;165(3):937-48. PubMed.

    . Apolipoprotein E and low density lipoprotein receptor-related protein facilitate intraneuronal Abeta42 accumulation in amyloid model mice. J Biol Chem. 2006 Nov 24;281(47):36180-6. PubMed.

    . Human and murine ApoE markedly alters A beta metabolism before and after plaque formation in a mouse model of Alzheimer's disease. Neurobiol Dis. 2002 Apr;9(3):305-18. PubMed.

    View all comments by Martin Sadowski

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Aβ Star is Born? Memory Loss in APP Mice Blamed on Oligomer
  2. Madrid: News from the Vaccine Front—Phase 1 Hopefuls
  3. Sorrento: Immunotherapy Update Hot Off Lectern of AD/PD Conference
  4. SfN: Where, How Does Intraneuronal Aβ Pack Its Punch? Part 1
  5. It’s a RAP—Loss of LRP Increases Amyloid Deposition in Mice
  6. Lipoproteins and Amyloid-β—A Fat Connection

Paper Citations

  1. . Clearance of Alzheimer's amyloid-ss(1-40) peptide from brain by LDL receptor-related protein-1 at the blood-brain barrier. J Clin Invest. 2000 Dec;106(12):1489-99. PubMed.
  2. . Amyloid-associated proteins alpha 1-antichymotrypsin and apolipoprotein E promote assembly of Alzheimer beta-protein into filaments. Nature. 1994 Nov 3;372(6501):92-4. PubMed.
  3. . Increased amyloid beta-peptide deposition in cerebral cortex as a consequence of apolipoprotein E genotype in late-onset Alzheimer disease. Proc Natl Acad Sci U S A. 1993 Oct 15;90(20):9649-53. PubMed.
  4. . A synthetic peptide blocking the apolipoprotein E/beta-amyloid binding mitigates beta-amyloid toxicity and fibril formation in vitro and reduces beta-amyloid plaques in transgenic mice. Am J Pathol. 2004 Sep;165(3):937-48. PubMed.

Further Reading

Primary Papers

  1. . Blocking the apolipoprotein E/amyloid-beta interaction as a potential therapeutic approach for Alzheimer's disease. Proc Natl Acad Sci U S A. 2006 Dec 5;103(49):18787-92. PubMed.
  2. . Apolipoprotein E and low density lipoprotein receptor-related protein facilitate intraneuronal Abeta42 accumulation in amyloid model mice. J Biol Chem. 2006 Nov 24;281(47):36180-6. PubMed.