Post-Conference News: Oligomers, Protofibrils, and Fibrils—They Are All Bad
Extensive data supporting a central role for Aβ in the genesis of Alzheimer's disease notwithstanding, the amyloid hypothesis has its weaknesses. One is that a specific neurotoxic Aβ species and the nature of its effects on neuronal function in vivo have not been defined. In a symposium titled "Protein Aggregation and Neurodegenerative Disease-an Unfolding Story" (#674), Bruce Yankner reviewed studies with synthetic Aβ peptides that have shown fibrils like those present in amyloid plaques to be neurotoxic in vitro. Such studies have supported the concept that extracellular plaques of amyloid fibrils are principally responsible for neuronal dysfunction and loss in AD.
Changiz Geula (#584.6 and #584.7) reconfirmed the toxicity of fibrils in vivo in an elegant series of experiments, in which the authors injected small amounts of fibrillar Aβ (200 pg) into the cerebral cortex of aged rhesus monkeys. This produced a dense core similar to that of native Aβ plaques, causing altered tau phosphorylation, neuronal loss, microglial activation, and proliferation. The lesion size and extent of neuronal loss increased with survival time after injection, as did the number and activation state of microglia. Coinjecting microglial inhibitory factor reduced the number of activated microglia and the lesion size by half. In addition, Dr. Geula noted that activated microglia were associated with naturally occurring compact plaques present in the aged rhesus cortex. Together, these findings indicate that an important component of plaque toxicity results from microglial activation.
Another problem that dogs the amyloid theory is that the severity of dementia and the density of amyloid plaques correlate poorly. However, levels of soluble Aβ and the extent of synaptic loss correlate strongly with the severity of cognitive impairment, suggesting that an Aβ species preceding plaques causes early damage. Several groups (including Bruce Yankner's, # 674) have confirmed that protofibrils are neurotoxic. Dean Hartley (#584.11) presented compelling evidence, gained by whole cell patch-clamp analysis , that protofibrils alter the electrical activity of neurons independent of the fibrils' effect. The NMDA antagonist D-APV caused a 30% inhibition of fibril-induced activity, whereas it blocked protofibril-induced activity by 60%-70%. These effects were not simply a result of the "gumming up" of neuronal membranes but appeared to involve specific interactions, since removal of protofibrils allowed the electrical activity of neurons to return to normal.
These findings, together with behavioral and morphological alterations found in several strains of AβPP-transgenic mice prior to plaque formation, suggest that soluble Aβ oligomers are important effectors of neurotoxicity. But which Aβ species mediates these changes in vivo? Although discrete electrophysiological and morphological alterations have been detected in young AβPP transgenic mice prior to amyloid deposition (Mucke, #674), it is not possible to define whether monomers, soluble oligomers, protofibrils, or dispersed amyloid fibrils cause these early alterations in synaptic form and function. Our demonstration that cell-derived oligomers of human Aβ inhibit the late phase of LTP addresses this issue (#128.3 and #920.1). We used conditioned medium from CHO cells that overexpress human AβPP and secrete SDS-stable, low n-oligomers similar in size and concentration to those detected in both human brain and CSF. Upon microinjection of this medium into the lateral cerebral ventricle of a live anesthetized rat, recordings from the CA1 detected a dramatic decrease in the late phase of LTP. Biochemical and immunological manipulations revealed that the LTP block was not attributable to fibrillar or monomeric Aβ but was mediated by the oligomers. This approach overcomes limitations of using synthetic peptides or transgenic mice. It shows clearly that physiological levels of stable Aβ oligomers can alter a sensitive and validated measure of synaptic plasticity in the absence of effects by monomers or fibrils. We only examined the effects of secreted oligomers, yet we have detected oligomers in Golgi-like vesicles.
It seems likely-as Andrea Le Blanc suggested (#128.10)-that intracellular Aβ oligomers might also play a role in Aβ-mediated neurotoxicity. Le Blanc found that microinjection of Aβ1-42 into the cytoplasm of primary human neurons lead to a 50% loss of cells within two days, whereas extracellular application of the same peptide solution, or micro-injection of Aβ42-1, did not. She also reported that cell loss was similar for peptide preparations that had formed fibrils and for preparations containing monomers, dimers, and trimers.
Bruce Yankner (#674) also reported that Aβ42 accumulated in Down's syndrome astrocytes in a vesicular pattern similar to that reported for adult human brain (Gouras #584.1). Moreover, Aβ accumulation in these astrocytes was associated with altered mitochondrial membrane potential and an increase in tunnel-positive cells. However, not all astrocytes and neurons that showed evidence of Aβ accumulation were tunnel-positive, suggesting that Aβ accumulation induces apoptosis in a subset of neurons but is unlikely to be an early event in AD pathogenesis. After many years of painstaking work, it now seems clear that multiple species of Aβ are neurotoxic. Thus, rather than fixate on an individual species, one must consider the whole process of fibrillogenesis from the dimer on up.
This is a complex proposition since Aβ may begin to oligomerize intracellularly. Oligo- and polymerization reactions are highly concentration-dependent, therefore, limiting monomer production should target production of toxic assemblies. But as Claudio Soto's presentation (#584.9) made evident, anti-aggregation approaches may also prove viable.
Soto et al. are testing the usefulness of their "β-sheet breaker peptide iAβ5" in a mouse model of AD. The mice, which are transgenic for both human AβPP (London mutation V717I) and PS-1 (A246T mutation), develop plaques by six months of age and show signs of neuronal dystrophy and microgliosis. Chemically blocking the N- and C-termini with added acetyl and amide groups dramatically improved the stability of the iAβ5 peptide. Intracerebroventricular or intraperitoneal administration of the blocked peptide reduced amyloid burden by 67.3% and 46.5%, respectively, while increasing neuronal survival and decreasing astrogliosis and microglial activation. Further efforts to improve the bioavailability of iAβ5 and peptidomimetics are under way.
While the increased survival of neurons, the decreased inflammatory response, and the clearance of plaques are welcome news, the real test of this therapeutic approach will lie in whether it can destabilize neurotoxic species other than fibrils, as well. Will it destabilize fibrils only to stabilize protofibrils or oligomers? Only time will tell.—Dominic Walsh, Center for Neurologic Diseases, Harvard Institutes of Medicine, Boston
(Note: The author codiscovered amyloid protofibrils and is closely involved with some of the research discussed here.)
No Available References
No Available Further Reading
Post-Conference News: Fly-Fishing for γ-Secretase Function and Substrates
A number of presentations at this year's Society for Neuroscience meeting provided new insight into the biology of presenilin. First, about its function. Deletion of the notch receptor in Drosophila produces a characteristic, hypomorphic phenotype, in which the wings appear scalloped or notched and the eyes are too small or missing. In line with burgeoning literature suggesting that presenilin is intimately involved in notch processing, knocking out presenilin produces a phenotype in both Drosophila and mice that is indistinguishable from notch deletion. But how does removing presenilin cause this effect? Is it because presenilin plays an essential role in notch trafficking? Or is it, as Michael Wolfe et al. propose, because presenilin (which he believes to be an essential component of the γ-secretase complex) is the enzyme responsible for S3 cleavage and the release of the notch intracellular domain (NICD, #244.1)?
Using an in vitro γ-secretase assay system developed by Merck (Li et al. 2000), Wolfe found that a bacterially expressed C99-like substrate (C100-flag, consisting of AβPP's C99 plus an initiating methionine and a flag tag at the C-terminus) was cleaved in a presenilin-dependent manner and that the cleavage could be inhibited by the potent γ-secretase inhibitor DAPT (Dovey et al. 2001). Wolfe constructed an analogous notch-based substrate, N100-flag, and found that DAPT also prevented its γ-secretase processing. DAPT also blocked notch signaling in a reporter assay and prevented nuclear translocation of NICD in whole cells.
Having exhausted his repertoire of cell-free and cell-based assay systems, Wolfe turned to organism-based studies. He fed Drosophila larvae 1 mM DAPT and monitored their development. The larvae developed a phenotype indistinguishable from hypomorphic notch flies, whereas Drosophila fed PAPT, a structural analogue of DAPT with reduced inhibitory activity, developed normally. The effects of DAPT are time-specific, occurring during day four of development (mid- to late 3rd instar), and are similar to those seen with temperature-induced Notch mutants. Immunostaining revealed that DAPT affects the expression of proteins dependent on Notch signaling. In a beautiful demonstration of this effect, Wolfe showed that wingless is lost in the wing margins where its expression depends on notch, but not in other areas where its expression does not depend on Notch.
Using the zebrafish, Christian Haass reviewed data that appeared to corroborate and extend Wolfe's (#244.2). Haass also used the γ-secretase inhibitor DAPT, which in his system had two major effects: it led to disordered somite formation and it induced neurogenesis of motor neurons. Expression of recombinant NICD, the fragment released by γ-cleavage, completely reversed the effects of DAPT, demonstrating that the DAPT phenotype resulted from a blockade of Notch signaling at or before NICD production through γ-secretase. Together, these findings demonstrate that the major effects of presenilin loss or γ-secretase inhibition are on notch, at least developmentally. However, given the growing list of γ-secretase substrates, careful scrutiny of less dramatic phenotypes is warranted.
Now, on to those substrates. Two more came out of presentations by P. Marambaud and Nikolaos Robakis. These authors showed that both epithelial (E) and neuronal (N) cadherins bind to the C-terminal fragment of presenilin1 (244.4 and 464.6). This involves residues 760-771 of the cadherins, a domain that is also needed for binding the δ-catenin-like protein P120. Presenilin1 and P120 bind cadherins competitively, i.e. presenilin-1 destabilizes cadherin/P120 complexes. Robakis has previously shown that cell-cell contact and the formation of adhesion junctions cause presenilin1 to localize to the plasma membrane, where a fraction of it binds to E-cadherin (Baki et al. 2001).
When adhesion junctions undergo remodeling, for instance during cell differentiation, cadherins must dissociate from the cytoskeleton in a process Robakis found to be presenilin- and γ-secretase-dependent. First, E-cadherin is cleaved extracellularly between residues 700 and 701, generating a C-terminal fragment (CTF1) analogous to the C83 of AβPP. Next, CTF1 cleavage between L731 and R732 generates a shorter stub, CTF2. Merck's γ-secretase inhibitor L685,458 blocks production of CTF2. E-cadherin CTF2 is not generated in fibroblasts that lack presenilin1 nor in cells expressing dominant negative presenilin1 double aspartate mutations. Further, a point mutation within the E-cadherin/presenilin1/P120 binding site at residue 761 not only abolishes the E-cadherin/presenilin1 interaction, but also blocks generation of E-cadherin CTF2. These data suggest that presenilin1/γ-secretase regulate E-cadherin disassembly in a manner similar to AβPP and notch processing. However, N-terminal sequencing of E-cadherin CFT2 indicates that this may not be another example of intramembraneous proteolysis, as CTF2 begins at a site C-terminal of the transmembrane domain.
In a presentation largely focused on the effects of nicastrin on AβPP processing and Aβ production, Paul Murphy alluded to his group's recent report that the ErbB4 receptor tyrosine kinase also appears to be processed by γ-secretase (464.1). As with E-cadherin, ErbB4 processing is blocked by γ-secretase inhibitors and the expression of the dominant negative double aspartate mutations (see related news item).
Together with the elegant demonstration by Cao and Sudhof (see related news item) and Kimberly et al., 2001, that AβPP-CTFγ can translocate to the nucleus, the data described here strongly suggest that γ-secretase represents a unique proteolytic activity, which mediates a common processing event for disparate receptors. If this is true, therapeutic targeting of γ-secretase for Alzheimer's treatment may be fraught with danger. However, if γ-secretase cleavage of different receptors is mediated by different ligands, then development of specific agonists or antagonist of AβPP may offer an alternate route for therapeutic intervention.—Dominic Walsh, Center for Neurologic Diseases, Harvard Institutes of Medicine, Boston
- γ-Secretase Found to Cleave Receptor Tyrosine Kinase ErbB-4
- Long-elusive Function for APP Cleavage Product Comes into View: It's Gene Transcription
- Li YM, Xu M, Lai MT, Huang Q, Castro JL, DiMuzio-Mower J, Harrison T, Lellis C, Nadin A, Neduvelil JG, Register RB, Sardana MK, Shearman MS, Smith AL, Shi XP, Yin KC, Shafer JA, Gardell SJ. Photoactivated gamma-secretase inhibitors directed to the active site covalently label presenilin 1. Nature. 2000 Jun 8;405(6787):689-94. PubMed.
- Dovey HF, John V, Anderson JP, Chen LZ, de Saint Andrieu P, Fang LY, Freedman SB, Folmer B, Goldbach E, Holsztynska EJ, Hu KL, Johnson-Wood KL, Kennedy SL, Kholodenko D, Knops JE, Latimer LH, Lee M, Liao Z, Lieberburg IM, Motter RN, Mutter LC, Nietz J, Quinn KP, Sacchi KL, Seubert PA, Shopp GM, Thorsett ED, Tung JS, Wu J, Yang S, Yin CT, Schenk DB, May PC, Altstiel LD, Bender MH, Boggs LN, Britton TC, Clemens JC, Czilli DL, Dieckman-McGinty DK, Droste JJ, Fuson KS, Gitter BD, Hyslop PA, Johnstone EM, Li WY, Little SP, Mabry TE, Miller FD, Audia JE. Functional gamma-secretase inhibitors reduce beta-amyloid peptide levels in brain. J Neurochem. 2001 Jan;76(1):173-81. PubMed.
- Baki L, Marambaud P, Efthimiopoulos S, Georgakopoulos A, Wen P, Cui W, Shioi J, Koo E, Ozawa M, Friedrich VL Jr, Robakis NK. Presenilin-1 binds cytoplasmic epithelial cadherin, inhibits cadherin/p120 association, and regulates stability and function of the cadherin/catenin adhesion complex. Proc Natl Acad Sci U S A. 2001 Feb 27;98(5):2381-6. PubMed.
- Kimberly WT, Zheng JB, Guénette SY, Selkoe DJ. The intracellular domain of the beta-amyloid precursor protein is stabilized by Fe65 and translocates to the nucleus in a notch-like manner. J Biol Chem. 2001 Oct 26;276(43):40288-92. PubMed.