At first glance, the pair of proteins that form the key pathological features of Alzheimer disease seem quite an odd couple. One is a peptide snipped from a membrane-associated precursor. The other is a microtubule-binding protein normally found in nerve axons. Yet a new study by Gunnar Gouras and colleagues at Weill Cornell Medical College, New York, suggests that toxic forms of amyloid-β (Aβ) and tau get cozy where it may matter most in AD—at synapses. Reporting 3 September in the Neurobiology of Aging online, the team used dual-labeling immuno-electron microscopy (dual EM) to reveal colocalization of Aβ42 and hyperphosphorylated tau in postsynaptic terminals of hippocampal neurons in two well-characterized AD mouse models (Tg2576 and 3xTg-AD) and in postmortem human AD brain. The findings provide “new insight into the anatomical divergence of the two pathologies,” Gouras told ARF. “That has been a mystery—why plaques are in one place and tau in another. This links them at the pathological level at synapses.”

Earlier work from Gouras’s and other labs suggested that buildup of intracellular Aβ42 accumulation precedes neuronal dysfunction (see, e.g., Gouras et al., 2000; Wirths et al., 2001). Gouras’s group had also shown that in Tg2576 mice, Aβ42 localizes primarily to late endosomal vesicles, with marked buildup in distal processes, especially postsynaptic compartments (Takahashi et al., 2002). Nothing of the sort was said of tau at the time. But other groups had spotted hyperphosphorylated tau lurking around plaques in dystrophic neurites from human subjects and AD transgenic mice (Moechars et al., 1999; Otth et al., 2002; Sturchler-Pierrat et al., 1997), so that’s where first author Reisuke Takahashi and colleagues looked when they decided to re-analyze the Tg2576 mice. Sure enough, by immunofluorescence they found hyperphosphorylated tau in abnormal neurites surrounding Aβ42-containing plaques. They saw similar tau patterns in human AD biopsy brain tissue (see figure).


Together at last: Aβ42 (green) and phospho-tau (red) cozy up in dystrophic neurites around an amyloid plaque in human AD brain tissue. Areas of Aβ-tau colocalization appear in yellow. Image credit: Estibaliz Capetillo-Zarate

However, these experiments didn’t have the resolution to distinguish intracellular from extracellular expression. “You need EM if you want to say anything about what's inside a process or synapse,” Gouras said. With help from EM expert and Cornell colleague Teresa Milner, the researchers cranked up the magnification several hundredfold to get a closer look at the dystrophic neurites around plaques. They saw hyperphosphorylated tau on filamentous structures and in microtubule-associated clusters near Aβ42 in Tg2576 dendritic profiles, including postsynaptic compartments. It is widely believed that synapses fizzle long before neurons die, and synaptic loss is the best correlate of fading cognition in AD (Coleman and Yao, 2003).

Because Tg2576 mice have scant tau pathology, Gouras and colleagues turned to a newer AD mouse model, 3xTg-AD (Oddo et al., 2003 and ARF related news story) for a better look at Aβ-tau interactions. With mutations in APP, tau, and presenilin-1, the triple transgenics not only accumulate intraneuronal Aβ and subsequent plaques but also develop tau tangles (LaFerla et al., 2007). “Here we had a powerful model to study the potential relationship between Aβ and tau,” recalled Gouras. “They're the two hallmarks, but is there any relation?”

Evidence for an Aβ-tau connection in AD has been trickling in. For example, Lennart Mucke and colleagues at the Gladstone Institute and the University of California, San Francisco, showed recently that APP transgenic mice with halved levels of tau protein are protected from Aβ-induced cognitive decline and premature lethality (see ARF related news story). Aβ and tau seem to be linked, “but we really don't know how,” Gouras said. “The triple transgenic allows us to look at an isolated system.”

CA1 neurons in the hippocampus of these mice were well suited for this study because of their abundant intraneuronal Aβ and clearly stratified anatomy. Their distal dendritic processes end in the same place—the stratum lacunosum-moleculare (SLM). Curiously, hyperphosphorylated tau appeared most prominently in the SLM (by three phospho-tau-specific antibodies), and this was the 3xTg-AD brain region showing the strongest colocalization with Aβ42. No such colocalization was seen in axons or presynaptic terminals.

Was the Aβ-tau relationship one-to-one in the SLM? Zooming in with dual EM, the researchers were able to look at individual SLM dendritic terminals in 13-month-old 3xTg-AD mouse brain. Of 78 identifiable terminals, 29.2 percent were positive for both Aβ42 and phospho-tau, 48.6 percent were negative for both, and far fewer were only positive for Aβ42 (13.9 percent) or phospho-tau (8.3 percent). For the most part, Gouras said, “it's not like one postsynaptic compartment has Aβ accumulation and another has phospho-tau. Phospho-tau develops in those postsynaptic compartments that have Aβ accumulation.” Gouras noted that the Aβ-tau colocalization reported in the SLM could also be occurring in vulnerable neurons elsewhere in the brain, but those areas are harder to study because the anatomy is not so well defined.

EM is a slow and laborious technique, dual EM even more so, Gouras told this reporter. “It is not like immunofluorescence that you can do overnight. This paper was four and a half to five years of work.” Other researchers reporting Aβ-tau colocalization in synapses earlier this year chose against microscopy and instead used flow cytometry to quantitate the two proteins in postmortem human AD brain synaptosomes—one-micron sac-like structures formed by nerve endings at synapses. In their analysis, nearly three-fourths of the synaptosomes were Aβ-positive. About a quarter of these also stained with phospho-tau antibody, most prominently those in the entorhinal cortex, a brain area with early AD vulnerability (see ARF related news story). In mice, this region could be dissected from brain tissue for EM analysis of presynaptic areas, Gouras said. Unfortunately, though, the entorhinal cortex in 3xTg-AD mice doesn’t have much AD pathology, so a mouse that more closely models human disease would be needed.

In the meantime, the new data have stirred up further intrigue about Aβ and phospho-tau, and what they might be doing together at synapses. Tau is mostly spread throughout nerve cells, Gouras said, “but there are instances where we see beginnings of phospho-tau accumulation at sites of Aβ. We want to understand the biology there.”—Esther Landhuis


  1. My colleague and I would also like to echo the importance of the connection between amyloid, tau, and neuronal dysfunction. The concept that tau levels within the neuron dictate the toxic response to Aβ clearly works in both directions. Our lab, in conjunction with the Ferreira and Binder labs, showed that primary cultures (Rapoport et al., 2002) of tau knockout neurons were resistant to Aβ-induced cell death. These same tau knockout mice were mated to APP transgenics by Mucke’s lab and they also showed that loss of tau impairs amyloid mediated damage. It stands to reason, then, that increased intraneuronal levels of hyperphosphorylated tau would promote amyloid mediated neuronal damage. Our unique bigenic mouse models (APPSw/NOS2-/- and APPSwDI/NOS2-/-) clearly demonstrate that non-mutated mouse tau becomes hyperphosphorylated at AD-like sites in the presence of amyloid deposition. Furthermore, the increased levels of amyloid and hyperphosphorylated tau are associated with profound neuronal loss in multiple brain regions (Colton et al.; Wilcock et al.). In addition to this neuronal loss, the work by Gouras and colleagues suggest that colocalization of phospho-tau and amyloid may also affect synapses in a way that could further impair neuronal function. These intimate interconnections between tau, amyloid, synapses, and neuronal loss may be a critical starting point for the downward spiral observed in AD brains.


    . Progression of amyloid pathology to Alzheimer's disease pathology in an amyloid precursor protein transgenic mouse model by removal of nitric oxide synthase 2. J Neurosci. 2008 Feb 13;28(7):1537-45. PubMed.

    . NO synthase 2 (NOS2) deletion promotes multiple pathologies in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A. 2006 Aug 22;103(34):12867-72. PubMed.

    . Tau is essential to beta -amyloid-induced neurotoxicity. Proc Natl Acad Sci U S A. 2002 Apr 30;99(9):6364-9. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Synapses Sizzle in Limelight of Symposium Preceding Neuroscience Conference, Orlando: Day 2
  2. APP Mice: Losing Tau Solves Their Memory Problems
  3. Synapses: Aβ and Tau Reside There, Cdk5 Shows Up Again

Paper Citations

  1. . Intraneuronal Abeta42 accumulation in human brain. Am J Pathol. 2000 Jan;156(1):15-20. PubMed.
  2. . Intraneuronal Abeta accumulation precedes plaque formation in beta-amyloid precursor protein and presenilin-1 double-transgenic mice. Neurosci Lett. 2001 Jun 22;306(1-2):116-20. PubMed.
  3. . Intraneuronal Alzheimer abeta42 accumulates in multivesicular bodies and is associated with synaptic pathology. Am J Pathol. 2002 Nov;161(5):1869-79. PubMed.
  4. . Early phenotypic changes in transgenic mice that overexpress different mutants of amyloid precursor protein in brain. J Biol Chem. 1999 Mar 5;274(10):6483-92. PubMed.
  5. . AbetaPP induces cdk5-dependent tau hyperphosphorylation in transgenic mice Tg2576. J Alzheimers Dis. 2002 Oct;4(5):417-30. PubMed.
  6. . Two amyloid precursor protein transgenic mouse models with Alzheimer disease-like pathology. Proc Natl Acad Sci U S A. 1997 Nov 25;94(24):13287-92. PubMed.
  7. . Synaptic slaughter in Alzheimer's disease. Neurobiol Aging. 2003 Dec;24(8):1023-7. PubMed.
  8. . Triple-transgenic model of Alzheimer's disease with plaques and tangles: intracellular Abeta and synaptic dysfunction. Neuron. 2003 Jul 31;39(3):409-21. PubMed.
  9. . Intracellular amyloid-beta in Alzheimer's disease. Nat Rev Neurosci. 2007 Jul;8(7):499-509. PubMed.

Other Citations

  1. Tg2576

Further Reading


  1. . Intraneuronal Alzheimer abeta42 accumulates in multivesicular bodies and is associated with synaptic pathology. Am J Pathol. 2002 Nov;161(5):1869-79. PubMed.
  2. . Intraneuronal Abeta42 accumulation in human brain. Am J Pathol. 2000 Jan;156(1):15-20. PubMed.
  3. . Intracellular amyloid-beta in Alzheimer's disease. Nat Rev Neurosci. 2007 Jul;8(7):499-509. PubMed.

Primary Papers

  1. . Co-occurrence of Alzheimer's disease ß-amyloid and τ pathologies at synapses. Neurobiol Aging. 2010 Jul;31(7):1145-52. PubMed.