1 December 2005. The story of intraneuronal Aβ, for years that of a few researchers calling into the wind, has generated a steady buzz at the 35th Annual Meeting of the Society for Neuroscience, held November 12-16 in Washington, D.C. Until recently, most researchers have tended to argue that the Aβ peptide can’t be of much consequence inside the neuron, chiefly because neurons secrete it soon after making it. But now, a growing number of laboratories are studying intraneuronal Aβ, and at the conference the topic drew frequent questions after talks and at posters. Active research these days includes the study of intraneuronal Aβ in mouse models and the mystery of how intraneuronal and extracellular pools are related. The cutting edge of this field has advanced past merely trying to detect it inside neurons and toward the issue of oligomeric species (see upcoming SfN news summary). Exactly how its accumulation might damage the neuron has become a burning question. Here, synaptic biology, the role of Aβ in cellular signaling pathways and proteasome degradation, as well as interactions between intraneuronal Aβ and tau, are emerging areas that engage scientists in lively speculation. Selected highlights from the conference follow below in a four-part news series on intraneuronal Aβ. As always, your Alzforum reporters invite kvetches, kudos, and comments.
See also Part 2, Part 3, and Part 4.
Mice Strains Multiply
Research using mouse models to study intraneuronal Aβ goes back a decade. In 1995, Frank LaFerla first reported that a strain of transgenic mice that expressed Aβ inside neurons—but not another strain that expressed it extracellularly—developed pathology and lost neurons (LaFerla et al., 1995). LaFerla followed up with a study on intraneuronal Aβ and p53 activation in mice, and another one showing intraneuronal Aβ in human AD brain. In 2000, Gunnar Gouras at Weill Medical College of Cornell University in New York City and colleagues demonstrated intraneuronal Aβ accumulation in human brain (Gouras et al., 2000), but still, the topic was slow to catch on. Two years later Gouras’s group detected intraneuronal Aβ in young Tg2576 mice (Takahashi et al., 2002), and soon after that the triple transgenic mice produced by LaFerla’s group, then at University of California, Irvine, drew widespread attention (Oddo et al., 2003). At SfN, the knowledge base for this topic broadened when additional investigators introduced their models and showed similar findings.
For example, Robert Vassar, of Northwestern University in Chicago, presented an analysis of his 5XFAD transgenic mice, which coexpress three different AD-causing APP mutations plus two different PS1 mutations driven by the Thy-1 promoter. This genetic combination serves to crank up Aβ42 production so that researchers need not wait 6 months before they can seriously study their mice. In this accelerated model, cerebral Aβ levels rise rapidly beginning at 6 weeks of age, and amyloid deposits begin to appear at 2 months, Vassar reported. At 4 months, learning and memory deficits emerge, and neurons begin to degenerate. On neurodegeneration, Vassar noted that his lab has not yet done definitive stereologic counts, but he has seen synaptophysin levels drop and massive neuronal loss in relevant areas, for example, cortical layer 5 and subiculum. Gliosis accompanies these changes. By 9 months, the mice’s brains are packed with plaques, Vassar reported. Also at this time, the female mice (but not the males) suffer a sudden spike in levels of the proinflammatory cytokine Il-1b and start dying.
More to the point of the topic at hand, however, the Vassar group sees a punctate pattern of intraneuronal Aβ starting at 6 weeks and accumulating with age. (Researchers generally interpret punctate staining as indicating accumulation.) Thioflavin S staining indicates that a portion of this intraneuronal Aβ is aggregated, Vassar said, and some of it appears to occur in neurites. Neurons containing ample Aβ appear to form plaques, Vassar added, leading him to support the hypothesis that aggregating intraneuronal Aβ might cause neurodegeneration and then form the nidus of beginning plaques. It’s worth noting that preliminary studies with tau antibodies have not so far indicated a link between the aggressive Aβ42 production and tau pathology (see ARF SfN tau series), but this question needs more work.
Another new mouse model that features intraneuronal Aβ is that of Lars Lannfelt’s group at Uppsala University, Sweden. Data presented by Lars Nilsson demonstrate that a line that expresses the Swedish APP mutations, which enhance β-secretase cleavage of APP, in addition also expresses the Arctic mutation in Aβ that Lannfelt’s group had identified earlier (Nilsberth et al., 2001). This mutation yields a form of Aβ that is particularly aggressive in forming soluble protofibrils. In these ArcSweTg mice, too, staining for aggregated intraneuronal Aβ predated extracellular Aβ aggregates. It also appeared qualitatively different from the extracellular Aβ, suggesting that aggregates on either side of the cell membrane have structural differences, Lannfelt noted. Aβ deposition in these ArcSwe mutants occurs so rapidly that the mice have compact senile plaques by 5 to 6 months of age. However, there was little evidence of neuronal loss. This data is published online (Lord et al., 2006).
The Lannfelt mice are similar to a strain made in the laboratory of Lennart Mucke at the University of California, San Francisco. These mice also express the human Arctic Aβ mutation plus the Swedish and Indiana mutations in APP, though technical details differ. A brief communication about these mice appeared last year (Cheng et al., 2004), but the paper did not mention intraneuronal Aβ. (Incidentally, Irene Cheng presented further data on these mice at the SfN conference; see upcoming Alzforum news summary on Aβ oligomers.)
Lannfelt’s and Vassar’s mice featuring intraneuronal Aβ follow a model published last year by Laurent Pradier and Thomas Bayer (Casas et al., 2004), but these scientists did not present further data in Washington, D.C.
Do Oligomers Start Forming Inside Neurons?
The LaFerla lab, from the University of California, Irvine, did present a range of follow-up studies of their mice. One of them focuses on intraneuronal oligomerization of Aβ; it appeared online this month (Oddo et al., 2005). The scientists collaborated with William Klein, also at Northwestern, and Charles Glabe, also at UC Irvine, to characterize the time course of Aβ oligomerization in the triple transgenic mice with Klein’s M71/3 (anti-ADDL) antibody, which recognizes 12-24mers of Aβ, and Glabe’s A11, which detects larger species. In brief, the researchers first detect significant amounts of soluble, non-oligomeric Aβ at 4 months of age and Aβ oligomers by 6 months of age, both intraneuronally, in the CA/subiculum region of hippocampus. The researchers suggest that the oligomers begin forming inside neurons between the ages of 2 and 6 months and then accumulate progressively. By 1 year, the presence of Aβ oligomers in this brain region had changed from being intraneuronal to being primarily extracellular and occurring near plaques.
The intracellular oligomers were near the cell body and also in neurites near synaptic sites, Salvatore Oddo reported in Washington. Further experiments indicate that they colocalize with tau in the somatodendritic compartment of 6-month-old triple transgenic mice, but not with hyperphosphorylated forms of tau in old mice. While it is tempting to speculate about an interaction between Aβ oligomers and tau at synapses, this colocalization does not prove it by any means, the scientists caution in their paper.
Using a broader panel of Aβ-antibodies, Oddo et al. found that their staining patterns overlapped only partly, suggesting that Aβ occurs in different conformations inside neurons. Together with other data, this supports the notion held by some scientists that separate aggregation pathways for Aβ might exist side-by-side in vivo. This idea further holds that a portion of the Aβ oligomers may be physically stable, rather than fleeting fibrillization intermediates on the way to plaques. This would make them an attractive new target for future therapies.
Finally in this study, the researchers extended previous experiments using anti-Aβ antibodies, which had reduced both Aβ and early tau pathology (Oddo et al., 2004). Trying to target oligomeric Aβ this time, they injected the A11 antibody into the hippocampus of 12-month-old triple transgenic mice and found that it, too, reduced not only intraneuronal Aβ but also tau pathology. This suggests that intraneuronal Aβ oligomers represent a link between Aβ generation and tau, bolstering their status as an attractive target, the scientists write.—Gabrielle Strobel.
See also Part 2, Part 3, and Part 4 of this series.