The amyloid-β (Aβ) peptide is a destroyer of synapses, and its attack on neurotransmission is blamed for devastating memory loss experienced in Alzheimer disease. Now, two new studies suggest that Aβ’s initial assault is aimed squarely at synaptic glutamate receptors. The related papers, both collaborations between the labs of Paul Greengard of Rockefeller University and Gunnar Gouras from the Weill Medical College of Cornell University in New York, show that Aβ peptides, whether added to cultures or produced by neurons, decrease the number and activity of synaptic NMDA and AMPA glutamate receptors.

The findings, published June 17 in Nature Neuroscience, and in press in Neurobiology of Disease last April (with corrected proof available online), provide a mechanism for how Aβ could perturb synaptic function, plasticity and memory, and show that the deleterious effects of Aβ can start early in the life of neurons.

The NMDA receptor is well-known to regulate synapse density and memory formation, and the Nature Neuroscience report describes the fate of this complex in cultured cortical neurons dosed with Aβ1-42 peptide. First author Eric Snyder and colleagues show that one hour of Aβ exposure lowered the number of NMDA receptor subunits on the cell surface, due to increased endocytosis. The imbalance in NMDA trade resulted in loss of NMDA subunits from synapses, and the researchers showed that excess endocytosis required the α-7 nicotinic acid receptor, (known to bind Aβ), as well as the activation of the protein phosphatase 2B (PP2B). Downstream of PP2B, the phosphotyrosine phosphatase STEP was activated, which is known to dephosphorylate tyrosines on the NMDA NR2B subunit in a region that controls receptor endocytosis.

Lowering synaptic NMDA receptor levels had a functional impact, revealed when Snyder et al. used patch-clamping to measure NMDA-evoked currents—these were reduced in neurons treated with Aβ. Interestingly, the authors found that this reduction only occurred in those neurons in which Aβ alone induced an inward current. Associated with this Aβ-driven current, there was reduced phosphorylation of CREB, a transcription factor important for synapse survival and memory formation.

Several of the biochemical aberrations observed after treatment of neuronal cultures with Aβ, like lower STEP and CREB phosphorylation, are also seen in human Alzheimer brain or in mouse models of the disease. When the researchers checked NMDA receptor levels in neurons from Tg2576 mice that express mutated human APP, they found them to be half that in wild-type neurons. To prove the decrease was due to Aβ, they treated cultures with the γ-secretase inhibitor DAPT, and showed that the treated neurons recovered their full complement of NMDA receptors.

In the Neurobiology of Disease paper, first author Claudia Almeida of the Gouras lab and her collaborators draw a similar picture of Aβ’s injurious effects on the GluR1 subunit of the AMPA-selective glutamate receptor. Almeida et al. cultured embryonic neurons from the same Tg2576 transgenic Aβ-expressing mice to look for early synaptic changes. The cultured neurons produce and accumulate intracellular Aβ progressively, and the new work reveals early decreases in the presynaptic protein synaptophysin, and the postsynaptic proteins PSD95 and GluR1. Decreases in these proteins were associated with altered synaptic function, as glutamate-stimulated expression of the immediate early gene Zif268 was impaired, and the neurons displayed lower numbers of functional synapses. The remaining synapses were enlarged, reminiscent of the alterations observed in AD brains.

Synaptic changes were detected early, after less than 2 weeks in culture, and accumulation of Aβ was required for the changes, since, again, treatment of cultures with the γ-secretase inhibitor DAPT prevented the loss of surface expression of AMPA receptor subunits. The results parallel those seen with NMDA receptor, and in fact were obtained with the same cells. To complete the circle, Almeida et al. finally showed that addition of Aβ to the same cortical neurons used in the Nature Neuroscience paper caused a 32 percent reduction of cell-surface AMPA receptors.

How Aβ elicits these losses in glutamate receptors is unclear, but the earliest change detected in the neuronal cultures from AD mice was a decrease in the postsynaptic protein PSD95, implicated by Bill Klein and colleagues in Aβ toxicity (see Lacor et al., 2004). This protein may play a central role in synapse dysfunction, since cell surface expression of both the NMDA and AMPA receptors depends on it. Interestingly, the STEP-sensitive tyrosine phosphorylation of NR2B implicated in NMDA receptor trafficking by Snyder et al. was previously shown to regulate the binding of PSD95 to the NMDA receptor (Lavezzari et al., 2003).

All told, these two papers, by tracing a direct route from Aβ peptide production to synaptic alterations via the downregulation of glutamate receptors, suggest that prolonged depression of NMDA- and AMPA-mediated neurotransmission could kick off the pathological changes seen in AD brain. An important question that remains is: Where does the process start? Is it intracellular or extracellular Aβ that matters, and what are the earliest events? The cultured neurons described in these papers should provide a useful system to answer these questions.—Pat McCaffrey


  1. These are very interesting studies. There has been growing evidence that some of the primary targets of AD and APP are synapses. These studies support this view. Furthermore, there has been recent interest in the relation between APP processing and synaptic transmission and in the trafficking of postsynaptic receptors. These studies provide important molecular evidence that such processes are key targets of Aβ. The molecular details of how APP derivatives affect synapses will be an important area of research, since judicious modulation of these processes may open therapeutic avenues to the treatment of AD.

  2. These papers provide intriguing evidence for a link among β-amyloid, excitatory synaptic transmission, and altered membrane trafficking. It has been known for some time that early stages of Alzheimer disease (AD) are associated with learning impairments and cognitive decline before the prototypical pathological hallmarks of plaques and tangles. These learning impairments have been linked to altered transmission at excitatory synapses, and in particular learning-related forms of plasticity such as long-term potentiation and long-term depression in the hippocampus. Now, Greengard, Gouras, and colleagues reveal that β-amyloid—the toxic peptide which accumulates in AD—exerts an unexpected influence over the abundance of both primary types of neurotransmitter at excitatory synapses: the AMPA- and NMDA-type glutamate receptors. These findings emphasize the critical need for an understanding of the cell biology of postsynaptic receptor trafficking under healthy physiological conditions and how such cellular processes go awry in the early stages of AD.

  3. I would be interested in how the community thinks this study ties in with the prescription of memantine (an NMDA-receptor antagonist) for moderate to severe Alzheimer disease. If NMDA receptor deficits contribute significantly to Alzheimer disease, would not this treatment be expected to have a detrimental rather than a beneficial effect?

  4. These two papers from the Greengard and Gouras labs identify specific pre- and postsynaptic defects in cultured neurons induced by exposure to the β-amyloid peptide (Aβ). It is well-established that synaptic loss likely occurs early in the Alzheimer pathological cascade, and rodent studies have demonstrated a specific depression in long-term potentiation associated with Aβ1-42. These two new studies, therefore, provide mechanistic insights into neuronal alterations potentially associated with AD memory loss. The Snyder et al. study demonstrates a specific loss of surface NMDA glutamate receptors in cortical neurons exposed to Aβ. Importantly, they go far beyond this observation and provide evidence for an Aβ-dependent molecular cascade that involves the α7 nicotinic receptor, protein phosphatase 2B, and tyrosine phosphatase, ultimately culminating in enhanced endocytosis of the NMDA receptor. In the Almeida et al. study, cultured primary neurons from the well-studied Tg2576 AD mouse model were used to demonstrate both presynaptic (reduced synaptophysin protein levels) and postsynaptic (reduced AMPA glutamate receptor subunit and PSD-95 protein levels, general reduction in dendritic spines) deficits. A perhaps important difference in these two studies is that Almeida et al. relied on endogenous production of Aβ in the primary neuronal culture, which they have previously shown to result in substantial intracellular pools of Aβ.

    As it becomes clear that Aβ can have fairly specific effects on synaptic function, an important question arises: Does Aβ have a natural role regulating synaptic function? If so, treatments that eliminate Aβ might have negative effects on synaptic function. However, Aβ is produced by many cells throughout the body, which probably would not be predicted for a protein with specific or important roles in synaptic function. If the Aβ peptide does not normally have a direct role in synaptic function, why does it have the observed effects on synaptic components? One possibility is that Aβ has a more general biological role (e.g., regulating cholesterol uptake), and a byproduct of an excess or abnormal form of this activity is the observed effects on synapses. Alternatively, Aβ1-42 (or some oligomer thereof) may simply be a toxic protein, and the apparently specific synaptic deficits result from low-grade toxicity affecting particularly vulnerable components of synaptic function. The case for highly Aβ-specific toxicity would be stronger if other peptides (e.g., random amphipathic peptides, or other proteins claimed to form toxic oligomers) could be shown not to have similar effects in the primary neuronal culture models.

    Another general issue is the relationship between synaptic dysfunction in AD and the more extreme neuronal cell loss that occurs later in the disease. One possibility is that synaptic dysfunction itself leads to neuronal death. Alternatively, synaptic dysfunction might be the "tip of the iceberg" of a more general pathological process. However, a third possibility, that Aβ-dependent synaptic deficits and neuronal loss are actually independent processes, cannot yet be ruled out. If the studies discussed here can be extended by developing treatments that block Aβ-dependent synaptic deficits, it may be possible to determine the relationship between synaptic dysfunction and neuronal cell death using recent transgenic mouse models that appear to capture much of human AD pathology.

  5. These papers, dealing with an effect of Aβ on postsynaptic mechanisms, when considered in combination with the recent paper from the Ferreira lab which shows an effect of Aβ on presynaptic mechanisms, show that there is more than one target through which Aβ can have a deleterious effect on synaptic function. Furthermore, other data indicating loss of dynamin 1 transcript in AD brain without loss of PSD 95 transcript (Yao et al., 2003) suggest that these effects on synaptic function may occur prior to the loss of synapses by still living neurons. [On the other hand, there are the Scheff data (reviewed in Scheff and Price, 2003) showing increased size of remaining synapses as other synapses are lost.] That still living neurons lose synapses in AD is suggested by data showing loss of synaptophysin message in selected affected single neurons in AD brain (e.g., Callahan et al., 2002) and by decreased synapse/neuron ratio in AD brain (Bertoni-Freddari et al., 1996). These data together suggest a progression of 1) decreased synaptic function, 2) loss of synapses by still living neurons and, finally, 3) loss of synapses by neuron death.


    . Deterioration threshold of synaptic morphology in aging and senile dementia of Alzheimer's type. Anal Quant Cytol Histol. 1996 Jun;18(3):209-13. PubMed.

    . Progressive reduction of synaptophysin message in single neurons in Alzheimer disease. J Neuropathol Exp Neurol. 2002 May;61(5):384-95. PubMed.

    . Beta-amyloid-induced dynamin 1 depletion in hippocampal neurons. A potential mechanism for early cognitive decline in Alzheimer disease. J Biol Chem. 2005 Sep 9;280(36):31746-53. PubMed.

    . Synaptic pathology in Alzheimer's disease: a review of ultrastructural studies. Neurobiol Aging. 2003 Dec;24(8):1029-46. PubMed.

    . Defects in expression of genes related to synaptic vesicle trafficking in frontal cortex of Alzheimer's disease. Neurobiol Dis. 2003 Mar;12(2):97-109. PubMed.

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Synaptic targeting by Alzheimer's-related amyloid beta oligomers. J Neurosci. 2004 Nov 10;24(45):10191-200. PubMed.
  2. . Differential binding of the AP-2 adaptor complex and PSD-95 to the C-terminus of the NMDA receptor subunit NR2B regulates surface expression. Neuropharmacology. 2003 Nov;45(6):729-37. PubMed.

Further Reading


  1. . Beta-amyloid-induced dynamin 1 depletion in hippocampal neurons. A potential mechanism for early cognitive decline in Alzheimer disease. J Biol Chem. 2005 Sep 9;280(36):31746-53. PubMed.
  2. . Age-related loss of synaptophysin immunoreactive presynaptic boutons within the hippocampus of APP751SL, PS1M146L, and APP751SL/PS1M146L transgenic mice. Am J Pathol. 2005 Jul;167(1):161-73. PubMed.
  3. . A focus on the synapse for neuroprotection in Alzheimer disease and other dementias. Neurology. 2004 Oct 12;63(7):1155-62. PubMed.

Primary Papers

  1. . Regulation of NMDA receptor trafficking by amyloid-beta. Nat Neurosci. 2005 Aug;8(8):1051-8. PubMed.
  2. . Beta-amyloid accumulation in APP mutant neurons reduces PSD-95 and GluR1 in synapses. Neurobiol Dis. 2005 Nov;20(2):187-98. PubMed.