Intraneuronal Ab Accumulation-More Evidence, Less Controversy?

Live discussion with Gunnar Gouras held 26 November. See Transcript.

Background text by Gunnar Gouras.

See Comment by Mike D'Andrea and Robert Nagele

Some investigators have long speculated about an intraneuronal origin of amyloid plaques and about a role for intraneuronal Ab in Alzheimer's disease pathogenesis, but lacked convincing data for this idea (for background see previous Alzforum chat, 2000). We believe we now have enough evidence to conclude that, in AD, Ab accumulates within neurons, especially in their distal processes and synaptic compartments, and then leads to synaptic dysfunction from within. This would render plaques "extracellular" remnants of AbPP/ b-amyloid accumulating and degenerating processes.

Antibodies directed at the Ab domain of AbPP helped visualize intraneuronal Ab accumulation (i.e., Sparks et al., 1996; LaFerla et al., 1997). However, given that full-length AbPP predominates over Ab in cells and that these antibodies also recognize AbPP, they were insufficient to convince skeptics. The advent of specific antibodies against the C-terminus of Ab40 or Ab42 brought a new wave of reports demonstrating intraneuronal Ab accumulation (Gouras et al., 2000; D'Andrea et al., 2001; Gyure et al., 2001 , among others). Interestingly, these studies indicated that it was especially the less-secreted but more disease-linked Ab42 peptide that accumulated inside neurons. These immunohistochemical studies have increased interest in intraneuronal Ab in AD, but still were not enough to turn the field away from the prevailing view that extracellular Ab is what's toxic exclusively. This view remains entrenched in the field.

This year, however, several more articles have appeared that strengthen the case for intraneuronal Ab accumulation. Andrea Leblanc and colleagues ( Zhang et al., 2002) reported provocative data demonstrating that microinjection of Ab1-42, but not Ab1-40 or Ab42-1, into primary neurons was exquisitely toxic (see also Peter Lansbury's comments on this paper in his interview). Critics noted the artificial nature of exogenous addition of Ab that has not been processed physiologically. Then Jorge Busciglio, Bruce Yankner, and colleagues extended their elegant studies on cultured Down's syndrome astrocytes and neurons. While these studies focused on different questions, they also found that intraneuronal Ab42 accumulation is an early feature in Down's syndrome brain that can be accompanied by TUNEL staining (see news story). Our current paper (Takahashi et al., 2002) now provides further evidence for the reality of intraneuronal Ab, which will be hard to discount (though critics may, of course, choose to ignore it). This chat hopes to encourage critical input.

In brief, our paper provides an exciting new piece of the puzzle for the mechanism of synaptic dysfunction/loss and plaque formation in AD. We provide immuno-gold electron microscopy evidence that intraneuronal Ab accumulation causes structural pathology within neuronal processes, as well as in pre- and post-synaptic compartments. We had asked where, at an ultrastructural level, do Ab peptides accumulate with aging in the brains of Tg2576 mice expressing human AbPP containing the Swedish familial AD mutation? We found that Ab42 normally localizes to the outer membranes of multivesicular bodies (MVBs). Our most critical control is that this MVB Ab42 is not present in AbPP knockout mice. Moreover, as Tg2576 mice age, Ab42 accumulates especially in neuronal MVBs of distal processes/synaptic compartments, and eventually it is associated with synaptic pathology both in Tg2576 mouse and human AD brain. These data support the scenario that plaques are remnants of Ab-accumulating and degenerating neuronal processes. Our work provides a novel mechanism linking soluble Ab increases inside neuronal terminals with synaptic loss and plaque formation in AD.

I suggest we discuss these questions:

1) What is the source of MVB Ab42? Is some, all, or any of it derived from extracellular Ab42 and/or synaptically released Ab42?

2) Is there a normal function of MVB-associated Ab42?

3) What is the mechanism by which MVB Ab42 increases with aging?

4) What could we all do to persuade the field that AD, like a growing number of other neurodegenerative diseases, is characterized by intracellular protein accumulation of a genetically-linked peptide? For the sake of argument, let's turn the tables and state that we have ultrastructural evidence for intraneuronally accumulating Ab and now it is up to those who think otherwise to provide evidence to the contrary.

5) Why is intracellular Ab so difficult to detect biochemically, i.e. by immunoprecipitation/Western blot?

6) Can MVB Ab42 become neurotoxic? Does neurotoxicity derive from release of Ab42 from MVBs into "cytosol" (i.e., as per Zhang et al., 2002)? Does it derive from abnormal interactions of MVB-bound Ab with kinesins or tau, with neuronal dysfunction resulting as transport of vital cargo is disrupted (i.e., see Kamal et al., 2001; Sisodia, 2002)?

7) Therapeutic implications: If plaques are remnants of degenerated neuronal processes, can treatment strategies that reduce plaques but not intraneuronal Ab succeed? Or is the resultant extracellular Ab also toxic to surrounding cells and processes, in line with the current hypothesis of extracellular Ab toxicity? It is important to keep in mind the possibility that something may damage synapses or axonal transport and reduce plaques. It was not widely understood that for plaques to occur one needs functioning neurons with adequate axonal/dendritic transport. In evaluating plaque-reducing drugs for AD therapy, one therefore needs to look carefully at possibly more subtle toxicity. Recent articles showing that lesioning the perforant pathway reduces plaque burden (see story), as well as our work, both imply that drugs toxic to synapses and/or axonal transport may be great at reducing plaques yet would most likely be harmful as treatment strategies for AD.

Additional Reading:

LaFerla FM, Troncoso JC, Strickland DK, Kawas CH, Jay G. Neuronal cell death in Alzheimer's disease correlates with apoE uptake and intracellular Ab stabilization. J Clin Invest 1997, 100(2): 310-320 Sparks, D.L. Intraneuronal b-amyloid immunoreactivity in the CNS. Neurobiol Aging 1996, 17(2): 291-299

Recent relevant reviews:

Bayer TA, Wirths O, Majtenyi K, Hartmann T, Multhaup G, Beyreuther K, Czech C. Key factors in Alzheimer's disease: beta-amyloid precursor protein processing, metabolism and intraneuronal transport. Brain Pathol 2001, 11(1): 1-11 Echeverria V, Cuello AC. Intracellular Ab amyloid, a sign for worse things to come? Mol Neurobiol 2002, 26(2-3): 299-316

Glabe C. Intracellular mechanisms of amyloid accumulation and pathogenesis in Alzheimer's disease. J Mol Neurosci 2001, 17(2): 137-45. Nixon RA, Mathews PM, Cataldo AM. The neuronal endosomal-lysosomal system in Alzheimer's disease. J Alzheimers Dis 2001, 3(1): 97-107

Tabira T, Chui DH, Kuroda S. Significance of intracellular Ab42 accumulation in Alzheimer's disease. Front Biosci 2002, 7:a44-9

Comment by Mike D'Andrea and Robert Nagele - Posted 4 December 2002

We have observed intraneuronal Ab accumulation consistently in Alzheimer brains (D'Andrea et al., 2002). In our opinion, the relationship between the local extent of this accumulation and the appearance of amyloid plaques is clear-cut. Since intraneuronal accumulation of Ab appears to be an early event in AD pathogenesis, the door is now open for new therapeutic strategies aimed at blocking this accumulation.

Our recent work has shown that Ab42 binds with exceptionally high affinity to the alpha7 nicotinic acetylcholine receptor (a7), raising the possibility that this interaction on the surfaces of a7-expressing neurons, like cortical pyramidal neurons, might somehow drive internalization of the Ab42-a7 complex (Nagele et al., 2002). In support of this idea, Ab42 and a7 are co-localized in the Ab42-positive accumulations in neurons of AD brains. This not only provides a plausible mechanism for Ab internalization and accumulation, but also may explain why cholinergic and cholinoceptive neurons are particularly vulnerable.

In addition, given the fact that the neuronal perikaryon is responsible for maintenance of the rather extensive dendritic trees and synaptic contacts, it is reasonable to suspect that these massive Ab42 accumulations exist at the expense of vital cellular machinery and, thus, eventually impair the ability of these neurons to maintain their dendrites and synaptic contacts. In view of this, we would not be surprised at all to find that intraneuronal accumulation of Ab peptides is related directly to the observed synaptic loss that is observed early in AD pathogenesis.

We have also provided additional evidence to suggest that there is a direct connection between eventual neuronal cell death and amyloid plaque formation; in fact, we have provided evidence that it is the death and lysis of amyloid-burdened neurons that appears to give rise to many amyloid plaques, at least the sub-population of plaques that is referred to as "dense core" plaques. If this proves to be the case, then the highly publicized "clearing" of plaques using the vaccine approach may have little benefit, since the neurons are already lost. We hope that more attention is given to the mechanisms and consequences of intracellular Ab accumulation, so that more promising therapies aimed at blocking intraneuronal accumulation (and therefore plaque formation) can be realized. —Michael D'Andrea and Robert Nagele, J&J Pharmaceutical Research and Development, Spring House, Pennsylvania.