The genome, the proteome, and the transcriptome have all been successfully trawled for links to AD. How about the lipidome? In today’s Nature Neuroscience online, researchers led by Lennart Mucke at the University of California, San Francisco report a lipidomics approach that landed group IV phospholipase A2, and the arachidonic acid it releases from phospholipids, as potentially important players in Alzheimer disease pathology. The researchers found that both are elevated in transgenic mouse models of the disease and that the enzyme is activated in AD postmortem tissue. “There have been other studies on lipids but we used broad-scale profiling that was not really biased to any particular set of metabolites,” Mucke told ARF. The researchers showed that both the lipase and arachidonic acid mediate amyloid-β toxicity in cells, and contribute to memory, behavior, and lifespan losses in transgenic mice. “The work is truly spectacular,” said Tobias Hartmann, University of Saarland, Homburg, Germany, in an interview with ARF. “I really like how broadly they tried to prove the issue by going up to the point of mortality and synaptic function. This is very well done.” Hartmann, who studies links between lipid metabolism and amyloid-β precursor protein processing, was not involved in the work.

Lipids have been linked to AD pathology for some time. Apolipoprotein E, for example, is the strongest genetic risk factor for sporadic forms of the disease, while epidemiological data (see ARF related news story) and work in animals suggest that certain fatty acids, such as docosahexaenoic acid, may protect against Alzheimer’s (see ARF related news story). Phospholipid metabolites, such as arachidonic acid and prostaglandins, also play important roles in regulating neural activity and inflammation, processes that are altered in AD. And Hartmann’s own work has shown a dynamic interplay between cell membrane lipid flux and APP processing (see ARF related news story). But despite the extensive links between lipids and Alzheimer disease, the relationship between the two has not been comprehensively examined. To address this, first author Rene Sanchez-Mejia and colleagues used a mass spectrometry approach, comparing levels of 44 lipid metabolites in the hippocampus and cortex of control and transgenic APP mice.

Sanchez-Mejia and colleagues found that arachidonic acid and its metabolites, including leukotriene B4 (LTB4) and prostaglandins E2 and B2, were significantly higher in the hippocampus of hAPP mice (J20 line). The increase was most pronounced for LTB4, which was around twice as high in the transgenic mice hippocampus than in control tissue. LTB4 was also significantly higher in the cortex of APP animals. Epoxyeicosatrienoic acids (EETs) and other arachidonic acid metabolites also trended higher in APP mouse cortex, some of them reaching significant elevations. Levels of other lipids were unaffected. “Because different isoforms of PLA2 have specificities for particular fatty acids, this fatty acid profile implicates a specific isoform of PLA2,” write the authors.

Of the three isoforms of phospholipase A2 (PLA2) found in the mammalian brain—group II, IVA and VIA—the researchers focused on the group IVA isoform (GIVA-PLA2) because the GII-PLA2 is absent from the J20 line due to an inbred deletion and because unlike GVIA-PLA2, GIVA-PLA2 is activated by kinases linked to AD.

Sanchez-Mejia and colleagues found that, as in the rat brain, GIVA-PLA2 is robustly expressed in the mouse brain, something that has been unclear until now. But more importantly, they found that levels of phosphorylated GIVA-PLA2 are 1.5-fold higher in the hippocampus of hAPP mice relative to controls and about fivefold higher in postmortem tissue isolated from mild, moderate, or severe AD patients. Because of technical difficulties related to preservation and storage, the researchers were not able to measure lipid profiles in AD postmortem tissue.

What is the relationship between GIVA-PLA2, its metabolites, and AD pathology? The researchers found no change in the enzyme or its metabolites in the I5 transgenic mouse line, which expresses normal human APP and has much lower levels of Aβ42. This suggests that Aβ might be the driving force leading to altered lipid profiles. The authors confirmed this by treating primary cultured neurons with isolated Aβ42, which caused a rapid phosphorylation of the enzyme and the release of arachidonic acid. Over the same time frame (10 minutes), cell surface GluR1 AMPA receptor subunits increased twofold. This was transient, however, and over the course of an hour the AMPAR levels fell back to normal and then subsequently fell below normal levels. Over a period of days, cells treated with Aβ42 lost viability. Interestingly, these effects could be blocked by the GIVA-PLA2 inhibitor AA-COCF3 (arachidonyl trifluoromethyl ketone), suggesting that Aβ toxicity depends on the lipase. In support of this, the researchers found that they could mimic Aβ42-driven AMPAR changes and cell loss by adding arachidonic acid instead. These findings gel with previous studies, for example from Roberto Malinow’s lab at Cold Spring Harbor Laboratory, New York, linking Aβ to an initial excitotoxic response followed by loss of AMPA receptors and synaptic suppression (see ARF related news story).

“There seems to be an aberrant excitatory activity in the brain in models of Alzheimer’s disease as well as in the brains of patients with Alzheimer’s disease,” said Mucke. “We’ve been particularly excited by the interesting effect that arachidonic acid has in that it actually changes the neurotransmitter receptors on the neuronal surface in a way that would make these neurons at least temporarily more excitable.”

How Aβ causes activation of GIVA-PLA2 is not clear. Mucke suggests that it may be through calcium release. In support of that, the researchers found that calcium chelators, such as EGTA, could block Aβ-driven arachidonic acid release from primary neuronal cultures. “The molecular link is not clear, but it is astonishing to see that it [Aβ] works in 10 minutes,” said Hartmann. This tells a lot about the molecular mechanism, he suggested, excluding, for example, gene regulation and trafficking mechanism. “But there are so many different mechanisms that can regulate arachidonic acid that this is hard to speculate,” he said. Activation of GIVA-PLA2 by Aβ has also been linked to the NMDA receptor, which allows calcium influx (see Shelat et al., 2008 and related comment below).

One thing that Hartmann did note was the effect of Aβ dose on GIVA-PLA2 phosphorylation. Fifty μM Aβ had little effect, whereas at a fivefold lower dose lipase phosphorylation jumped 2.5-fold. “We saw something similar in our work and related it to Aβ aggregation,” he said. “That’s the only obvious thing about Aβ that would explain how you could increase the effect by decreasing the concentration.”

How does this work relate to the human condition and does it lead to any new potential treatments? By ablating or reducing the GIVA-PLA2 gene the researchers were able to dramatically improve learning and memory in hAPP mice. The animals also showed significant reductions in anxiety and premature mortality. The work suggests that blocking GIVA-PLA2 could offer a new potential treatment for AD. There are some inhibitors of this enzyme in use, but Mucke said that it is not known how good their bioavailability would be in the brain. AA-COCF3, for example, would have to be infused into the brain because it cannot pass the blood-brain barrier. “The medicinal chemistry may have to be tweaked some more to turn them into viable therapeutics,” said Mucke. “Hopefully, our study might encourage that.”

In the meantime, one alternative to the pharmaceutical approach is a dietary one. Omega-3 fatty acids have received a lot of attention recently as potential protective factors for AD. Docosahexanoic acid, for example, can reduce amyloid burden in mouse models of the disease (see Lim et al., 2005). “DHA and arachidonic acid are counterplayers,” said Hartmann. DHA is incorporated into the brain and can displace arachidonic acid. “A very efficient way to reduce arachidonic acid in the brain is by eating DHA. Of course, that’s a slow process, occurring over weeks, months, and even years,” said Hartmann. One clinical trial of DHA suggested that it may benefit people with very mild AD (see Freund-Levy et al., 2006 and ARF related news story), and there are other trials ongoing, including one by the Alzheimer’s Disease Cooperative Study (see ARF related news story). While we await definitive news on DHA, there’s probably no harm in eating more fish.—Tom Fagan.

Sanchez-Mejia RO, Newman J, Toh S, Yu G-Q, Zhou Y, Halabisky B, Cisse M, Scearce-Levie K, Cheng IH, Gan L, Palop JJ, Bonventre JV, Mucke L. Phospholipase A2 reduction ameliorates cognitive deficits in a mouse model of Alzheimer’s disease. Nature Neuroscience 2008 October 19; advanced online publication.


Make a Comment

To make a comment you must login or register.

Comments on this content

  1. The article by Sanchez-Mejia et al. (2008) in Nature Neuroscience reveals for the first time an important link between group IV phospholipase A2 (GIVA-PLA2) and cognitive deficits in the mouse model of Alzheimer’s disease. Using an unbiased lipidomics analysis protocol, this study demonstrated an increase in arachidonic acid (AA) and some of its metabolites in the brain tissues of transgenic mice expressing familial AD-mutant (hAPP) as compared with non-transgenic controls. The increase in AA was attributed to activation of GIVA-PLA2, a Ca2+-dependent enzyme with multiple phosphorylation sites (Sun et al., 2004). This phospholipase A2 prefers releasing AA and is stimulated by signaling cascades induced by G protein-coupled receptor agonists (Xu et al., 2002; Sun et al., 2004). The study by Sanchez-Mejia demonstrated the ability for oligomeric Aβ to stimulate phosphorylation of GIVA-PLA2 in mouse cortical neurons. This observation is similar to that reported by us earlier using rat cortical neurons (Shelat et al., 2008). However, while the study by Sanchez-Mejia et al. linked the increase in phospho-GIVA-PLA2 to activation of the AMPA receptor, Shelat et al. related Aβ activation to the NMDA receptor. Besides G protein-coupled receptors, oxidant compounds can also stimulate phospho-GIVA-PLA2 (Xu et al., 2003). In the study with rat cortical neurons, neuronal excitation by NMDA as well as by Aβ is linked to production of reactive oxygen species (ROS) by NADPH oxidase (Shelat et al., 2008). Consequently, these studies demonstrating a link between Aβ-mediated increase in ROS through NADPH oxidase and activation of GIVA-PLA2 provide important support for the oxidative hypothesis for AD and the observed increase in lipid peroxidation products during early phase of AD (Sun et al., 2007).

    Activation of GIVA-PLA2 not only produces AA for synthesis of a large number of eicosanoid metabolites, but AA is a known retrograde messenger and may directly modulate synaptic activity (see discussion in Shelat et al., 2008). Furthermore, GIVA-PLA2 activation produces lyso-phospholipids, compounds with detergent-like properties that may perturb membrane properties. There is also evidence that activation of GIVA-PLA2 can cause changes in neuronal membranes, including membrane physical properties (Hicks et al., 2008) and mitochondrial dysfunction (Zhu et al., 2006). In addition, increases in oxidative stress and MAPK pathways, signaling pathways associated with GIVA-PLA2, can alter changes in cytoskeleton properties and intercellular connections in astrocytes (Zhu et al., 2005). Based on these data, it is enticing to propose that in addition to targeting GIVA-PLA2, antioxidants may be included as potential therapy to ameliorate neurotoxicity of Aβ and progression of AD.

    A novelty in the study by Sanchez-Mejia et al. (2008) is the use of transgenic mice produced by crossing PLA2g4a-deficient mice with hAPP mice to provide data demonstrating that reduction of GIVA-PLA2 is associated with improved learning and memory. Indeed, this study together with other recent studies on GIVA-PLA2 provides strong support for the role of oxidative-membrane phospholipid degradation in neurodegenerative diseases and for future studies targeting NADPH oxidase and GIVA-PLA2 as potential therapeutics for treatment of these diseases in general and AD in particular.


    . Amyloid-beta peptide induces temporal membrane biphasic changes in astrocytes through cytosolic phospholipase A2. Biochim Biophys Acta. 2008 Nov;1778(11):2512-9. PubMed.

    . Amyloid beta peptide and NMDA induce ROS from NADPH oxidase and AA release from cytosolic phospholipase A2 in cortical neurons. J Neurochem. 2008 Jul;106(1):45-55. PubMed.

    . The roles of NADPH oxidase and phospholipases A2 in oxidative and inflammatory responses in neurodegenerative diseases. J Neurochem. 2007 Oct;103(1):1-16. PubMed.

    . Phospholipase A2 in the central nervous system: implications for neurodegenerative diseases. J Lipid Res. 2004 Feb;45(2):205-13. PubMed.

    . Role of PKC and MAPK in cytosolic PLA2 phosphorylation and arachadonic acid release in primary murine astrocytes. J Neurochem. 2002 Oct;83(2):259-70. PubMed.

    . Oxidant-mediated AA release from astrocytes involves cPLA(2) and iPLA(2). Free Radic Biol Med. 2003 Jun 15;34(12):1531-43. PubMed.

    . Phospholipases A2 mediate amyloid-beta peptide-induced mitochondrial dysfunction. J Neurosci. 2006 Oct 25;26(43):11111-9. PubMed.

    . Hydrogen peroxide alters membrane and cytoskeleton properties and increases intercellular connections in astrocytes. J Cell Sci. 2005 Aug 15;118(Pt 16):3695-703. PubMed.

  2. Comment by Rene Sanchez-Mejia and Lennart Mucke
    We thank Tom Fagan, Tobias Hartmann, James Lee, and Grace Sun for their kind words about our paper. A few comments may help clarify some of the issues raised. The excellent work by Roberto Malinow (Hsieh et al., 2006) has identified mechanisms that may underlie the Aβ-dependent downregulation of AMPA receptors (AMPAR). However, to our knowledge, his work has not linked Aβ to an initial excitotoxic response or addressed the fact that, at the network level, Aβ elicits epileptiform activity (Palop et al., 2007). Our most recent study revealed for the first time that exposure of neurons to Aβ leads to an immediate increase in surface AMPAR levels. This increase in AMPAR was associated with an immediate increase in neuronal activity in response to Aβ and arachidonic acid (AA). These early Aβ-induced alterations may help explain the aberrant excitatory network activity observed in hAPP mice (Palop et al., 2007). In our culture model, the increase in AMPAR eventually subsided, and continued exposure to Aβ led to a decrease in surface AMPAR levels, consistent with the results previously reported by Roberto Malinow (Hsieh et al., 2006) and others (Shankar et al., 2007). As outlined in our paper, it is tempting to speculate that the delayed decrease in surface AMPAR levels may be triggered by the earlier increase in AMPAR and the resulting excitotoxicity.

    Our study also revealed that exposure of neurons to Aβ oligomers increases the activation/phosphorylation of GIVA PLA2 at relatively low concentrations of Aβ. Tobias Hartmann suggested that the decrease in phosphorylated GIVA PLA2 that we observed at higher concentrations of Aβ might be related to changes in the aggregation state of Aβ. Although we did exclude this possibility in our study by monitoring the aggregation state of Aβ oligomers after adding them to the culture medium, we would like to offer an alternate explanation. It is likely that at higher concentrations Aβ caused more profound cellular toxicity, resulting in impairments in the function of kinases that phosphorylate GIVA PLA2. These possibilities are not mutually exclusive.

    The comments by James Lee and Grace Sun further underline the potential significance of our work and provide important additional discussion points.


    . AMPAR removal underlies Abeta-induced synaptic depression and dendritic spine loss. Neuron. 2006 Dec 7;52(5):831-43. PubMed.

    . Aberrant excitatory neuronal activity and compensatory remodeling of inhibitory hippocampal circuits in mouse models of Alzheimer's disease. Neuron. 2007 Sep 6;55(5):697-711. PubMed.

    . Natural oligomers of the Alzheimer amyloid-beta protein induce reversible synapse loss by modulating an NMDA-type glutamate receptor-dependent signaling pathway. J Neurosci. 2007 Mar 14;27(11):2866-75. PubMed.

Comments on Primary Papers for this Article

No Available Comments on Primary Papers for this Article


News Citations

  1. Talking Fish Fat And Cholesterol
  2. Fish Oil Swims Ahead in Dietary Brain Protection Race
  3. A Better GRIP on the Aβ-Lipid Connection
  4. AMPA Receptors: Going, Going, Gone in Aβ-exposed Synapses, PSD95 Knockouts
  5. First Trial of Fish Fats Shows Promise for Early AD
  6. Chicago: Phase 2 News—Therapeutic Breakfast Food?

Paper Citations

  1. . Amyloid beta peptide and NMDA induce ROS from NADPH oxidase and AA release from cytosolic phospholipase A2 in cortical neurons. J Neurochem. 2008 Jul;106(1):45-55. PubMed.
  2. . A diet enriched with the omega-3 fatty acid docosahexaenoic acid reduces amyloid burden in an aged Alzheimer mouse model. J Neurosci. 2005 Mar 23;25(12):3032-40. PubMed.
  3. . Omega-3 fatty acid treatment in 174 patients with mild to moderate Alzheimer disease: OmegAD study: a randomized double-blind trial. Arch Neurol. 2006 Oct;63(10):1402-8. PubMed.

External Citations

  1. Apolipoprotein E
  2. Alzheimer’s Disease Cooperative Study

Further Reading


  1. Chicago: Phase 2 News—Therapeutic Breakfast Food?
  2. First Trial of Fish Fats Shows Promise for Early AD
  3. Fish Oil Swims Ahead in Dietary Brain Protection Race
  4. Talking Fish Fat And Cholesterol
  5. A Better GRIP on the Aβ-Lipid Connection
  6. AMPA Receptors: Going, Going, Gone in Aβ-exposed Synapses, PSD95 Knockouts