Memories come with a hefty price tag. When energy is plentiful, neurons expend a lot of it to make new memory proteins at synapses; alas, under stress, neurons cut back on that protein synthesis to conserve resources. Could that impair memory and cognition? In the August 11 Nature Neuroscience, scientists led by Eric Klann, New York University, report that boosting protein synthesis improves memory in a mouse model of Alzheimer's disease (AD). They deleted kinases, including one called PERK, that deactivate eukaryotic initiation factor 2α, a key regulator of RNA translation. Their results suggest that phosphorylation of eIF2α may play a role in the synaptic dysfunction that characterizes AD and that suppressing the relevant kinases might offer a novel therapeutic approach.

The work "highlights the importance of translational control at the synapse in the pathophysiology of Amyloidβ (Aβ), and perhaps, neurodegeneration generally," wrote Benjamin Wolozin, Boston University School of Medicine, Massachusetts, to Alzforum in an email. He noted that protein synthesis pathways represent potentially important targets for therapeutic development, that have remained untapped in AD drug discovery (see full comment below).

When cells encounter stressful situations, they phosphorylate eIF2α. This shuts down general protein synthesis, and boosts the expression of select stress response genes, some of which help reconstitute misfolded proteins. Previously, researchers led by Robert Vassar at Northwestern University Feinberg School of Medicine in Chicago, Illinois, reported that eIF2α phosphorylation enhances expression of β-secretase (BACE-1), which initiates amyloidogenic processing of amyloid β precursor protein (see ARF related news story). In addition, previous studies reported that phosphorylated eIF2α abounds in the human AD brain compared to age-matched postmortem samples, and that the modified protein tags degenerating neurons (see Chang et al., 2002). However, whether the modification plays a role in causing the synaptic deficits and memory impairment in Alzheimer's, and whether it could be altered to modulate these deficiencies, remained unclear.

To answer these questions, lead author Tao Ma and colleagues first used Western blots to look for changes in phosphorylated eIF2α levels in the brains of mice that model AD. They found more of the modified protein in the hippocampi and prefrontal cortices of 10-12 month-old APP/PS1 animals compared to controls. Immunohistochemistry and Western blots in postmortem human brains similarly supported earlier studies, revealing almost three times as much phosphorylated eIF2α in hippocampi from AD patients than controls.

Does Aβ affect eIF2α phosphorylation? It is known that Aβ blocks long-term potentiation (LTP), a form of synaptic plasticity that requires new protein synthesis (for a review, see Klann et al., 2004). Normally, high-frequency electrical stimulation of hippocampal slices causes dephosphorylation of eIF2α, promotes protein synthesis, and induces LTP. However, when the researchers applied synthetic Aβ oligomers to the slices, eIF2α remained phosphorylated, no new protein was made, and electrical stimulation elicited no LTP.

Could preventing phosphorylation of eIF2α rescue these deficits? The researchers generated conditional knockout mice lacking PKR-like ER kinase (PERK) in the excitatory neurons of the forebrain and hippocampus. Applying Aβ to tissue slices from those brain regions had no effect on LTP and eIF2α remained dephosphorylated.

To find out if deleting PERK also protected against Aβ in vivo, Ma and colleagues crossed the PERK knockouts with APP/PS1 mice. At 10-12 months, the PERK-negative offspring phosphorylated eIF2α in the hippocampus and prefrontal cortex less than did control mice. LTP appeared normal, and the mice performed as well as controls in tests of cognition and spatial memory, including the Morris water maze, Y water maze, and the object location test. The animal made normal amounts of PSD-95, synaptophysin, and other proteins involved in synaptic plasticity and memory. Interestingly, though BACE-1 levels were unchanged in these mice, they had less Aβ in their hippocampi than APP/PS1 animals. The PERK-deficient mice had wild-type levels of the Aβ-degrading enzyme neprilysin; this may explain their normal Aβ load, since APP/PS1 mice are known to have less neprilysin.

Crossing knockouts of another eIF2α kinase, GCN2, with APP/PS1 mice also yielded offspring with normal LTP, spatial learning, and memory. Altogether, these findings suggest that lowering eIF2α phosphorylation rescues AD phenotypes, claim the authors.

How might eIF2α contribute to Alzheimer's and other neurodegenerative diseases? Klann hypothesizes that in the later stages of disease, reactive oxygen species keep cells chronically stressed and eIF2α constantly phosphorylated, which would suppress genes necessary for synaptic plasticity and memory. Since eIF2α is hyperphosphorylated in prion disease models, as well (see Moreno et al., 2012), this mechanism could be at play in multiple neurodegenerative diseases and represent a common therapeutic target, suggested the authors.

In practice, reducing eIF2α phosphorylation could prove tricky, since it is a routine protective response, Klann told Alzforum. Scientists would likely have to target a treatment to specific areas in which eIF2α regulation had gone out of balance. Klann’s lab previously showed that targeted disruption of PERK in the forebrain made it difficult for mice to learn and unlearn behaviors while adapting to new environments and tasks (see Trinh et al., 2012). Interestingly, scientists recently reported that intraperitoneal injection of a small molecule called ISRIB, which blocks PERK's phosphorylation of eIF2α, enhanced spatial and fear-associated learning in wild-type mice (see Sidrauski et al., 2013 ).

Mauro Costa-Mattioli, Baylor College of Medicine, Houston Texas, agreed that this could be an exciting avenue to explore for treatment of AD. He previously reported that a drop in eIF2α phosphorylation enhances memory in mice (see ARF related news story on Costa-Mattioli et al., 2007). However, he echoed the concern about PERK inhibition. "We need to be cautious—PERK regulates stress in the endoplasmic reticulum that is induced by misfolded proteins," he said. "If a cell cannot cope with such stress, it may end up with a large amount of misfolded protein, which can ultimately be toxic," he added.—Gwyneth Dickey Zakaib

Comments

  1. The paper by the Klann group presents an elegant study that highlights the importance of translational control at the synapse in the pathophysiology of Amyloidβ, and perhaps, neurodegeneration generally. When Aβ causes synaptic toxicity, it necessarily causes prolonged changes at the synapse, and a large part of this occurs by modifying protein synthesis. The pathways regulating this biology represent potentially important targets for therapeutic development, yet these pathways have been largely untapped in drug discovery efforts. Klann's study demonstrated the importance of protein translation using multiple independent approaches, utilizing knockouts and inhibitors of two enzymes that act at the same hub in translational control, eIF2α.

    Interestingly, Stephen Strittmatter presented a study at AD/PD 2013, which showed that Aβ increases phosphorylation of eIF2α, and does so in a manner dependent on PrP. So we now have multiple approaches coming from multiple laboratories all focusing on the same system. I'm sure that we will be hearing much more about this in the upcoming months and years.

    View all comments by Benjamin Wolozin
  2. The exciting paper by Ma et al. contributes to the accumulating evidence that PERK is involved in neurodegeneration and is potentially a therapeutic target (reviewed in (1)). Our group reported previously the activation of the unfolded protein response (UPR) in AD brain, including specific evidence for the activation of the translational PERK pathway. We detected the active form of the kinase and its phosphorylated substrate peIF2α (2,3)). This fits with the study by Ma et al.

    Essentially, these authors show that interference in basic homeostatic pathways can be employed to rescue deficits in synaptic plasticity in an APP/PS1 model. The authors are rightfully cautious to translate their findings to potential treatment for human disease, as interference in homeostatic stress pathways may also have a downside.

    It will be interesting to find out more mechanistic detail about the effects of PERK deletion on the progression of AD pathology, in particular tau pathology, as this is closely connected with UPR activation in human brain (3,4). Future studies will have to address how PERK deletion affects cellular stress resilience and whether and how neurons adapt to this type of interference.

    References:

    . Endoplasmic reticulum dysfunction in neurological disease. Lancet Neurol. 2013 Jan;12(1):105-18. PubMed.

    . The unfolded protein response is activated in Alzheimer's disease. Acta Neuropathol. 2005 Aug;110(2):165-72. PubMed.

    . The unfolded protein response is activated in pretangle neurons in Alzheimer's disease hippocampus. Am J Pathol. 2009 Apr;174(4):1241-51. PubMed.

    . The unfolded protein response is associated with early tau pathology in the hippocampus of tauopathies. J Pathol. 2011 Nov 21; PubMed.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Paper Alert: Energy Deprivation Drives up BACE Translation
  2. Memories—The Long, the Short, and the Schemas

Paper Citations

  1. . Phosphorylation of eukaryotic initiation factor-2alpha (eIF2alpha) is associated with neuronal degeneration in Alzheimer's disease. Neuroreport. 2002 Dec 20;13(18):2429-32. PubMed.
  2. . Biochemical mechanisms for translational regulation in synaptic plasticity. Nat Rev Neurosci. 2004 Dec;5(12):931-42. PubMed.
  3. . Sustained translational repression by eIF2α-P mediates prion neurodegeneration. Nature. 2012 May 24;485(7399):507-11. PubMed.
  4. . Brain-specific disruption of the eIF2α kinase PERK decreases ATF4 expression and impairs behavioral flexibility. Cell Rep. 2012 Jun 28;1(6):676-88. PubMed.
  5. . Pharmacological brake-release of mRNA translation enhances cognitive memory. Elife. 2013;2:e00498. PubMed.
  6. . eIF2alpha phosphorylation bidirectionally regulates the switch from short- to long-term synaptic plasticity and memory. Cell. 2007 Apr 6;129(1):195-206. PubMed.

Further Reading

Papers

  1. . Biochemical mechanisms for translational regulation in synaptic plasticity. Nat Rev Neurosci. 2004 Dec;5(12):931-42. PubMed.
  2. . Brain-specific disruption of the eIF2α kinase PERK decreases ATF4 expression and impairs behavioral flexibility. Cell Rep. 2012 Jun 28;1(6):676-88. PubMed.
  3. . Pharmacological brake-release of mRNA translation enhances cognitive memory. Elife. 2013;2:e00498. PubMed.
  4. . Sustained translational repression by eIF2α-P mediates prion neurodegeneration. Nature. 2012 May 24;485(7399):507-11. PubMed.
  5. . Phosphorylation of eukaryotic initiation factor-2alpha (eIF2alpha) is associated with neuronal degeneration in Alzheimer's disease. Neuroreport. 2002 Dec 20;13(18):2429-32. PubMed.
  6. . ApoE ε4 is associated with eIF2α phosphorylation and impaired learning in young mice. Neurobiol Aging. 2012 Aug 7; PubMed.
  7. . eIF2alpha phosphorylation bidirectionally regulates the switch from short- to long-term synaptic plasticity and memory. Cell. 2007 Apr 6;129(1):195-206. PubMed.

Primary Papers

  1. . Suppression of eIF2α kinases alleviates Alzheimer's disease-related plasticity and memory deficits. Nat Neurosci. 2013 Sep;16(9):1299-305. PubMed.