Psychopharmacology

, Volume 202, Issue 1, pp 37–51

Phospholipase A2 activation as a therapeutic approach for cognitive enhancement in early-stage Alzheimer disease

Authors

    • Laboratory of Neuroscience (LIM-27), Department and Institute of Psychiatry, Faculty of MedicineUniversity of São Paulo
  • Orestes V. Forlenza
    • Laboratory of Neuroscience (LIM-27), Department and Institute of Psychiatry, Faculty of MedicineUniversity of São Paulo
  • Wagner F. Gattaz
    • Laboratory of Neuroscience (LIM-27), Department and Institute of Psychiatry, Faculty of MedicineUniversity of São Paulo
Original Investigation

DOI: 10.1007/s00213-008-1351-0

Cite this article as:
Schaeffer, E.L., Forlenza, O.V. & Gattaz, W.F. Psychopharmacology (2009) 202: 37. doi:10.1007/s00213-008-1351-0

Abstract

Rationale

Alzheimer disease (AD) is the leading cause of dementia in the elderly and has no known cure. Evidence suggests that reduced activity of specific subtypes of intracellular phospholipases A2 (cPLA2 and iPLA2) is an early event in AD and may contribute to memory impairment and neuropathology in the disease.

Objective

The objective of this study was to review the literature focusing on the therapeutic role of PLA2 stimulation by cognitive training and positive modulators, or of supplementation with arachidonic acid (PLA2 product) in facilitating memory function and synaptic transmission and plasticity in either research animals or human subjects.

Methods

MEDLINE database was searched (no date restrictions) for published articles using the keywords Alzheimer disease (mild, moderate, severe), mild cognitive impairment, healthy elderly, rats, mice, phospholipase A2, phospholipid metabolism, phosphatidylcholine, arachidonic acid, cognitive training, learning, memory, long-term potentiation, protein kinases, dietary lipid compounds, cell proliferation, neurogenesis, and neuritogenesis. Reference lists of the identified articles were checked to select additional studies of interest.

Results

Overall, the data suggest that PLA2 activation is induced in the healthy brain during learning and memory. Furthermore, learning seems to regulate endogenous neurogenesis, which has been observed in AD brains. Finally, PLA2 appears to be implicated in homeostatic processes related to neurite outgrowth and differentiation in both neurodevelopmental processes and response to neuronal injury.

Conclusion

The use of positive modulators of PLA2 (especially of cPLA2 and iPLA2) or supplementation with dietary lipid compounds (e.g., arachidonic acid) in combination with cognitive training could be a valuable therapeutic strategy for cognitive enhancement in early-stage AD.

Keywords

Phospholipase A2ActivationLearningMemoryCognitive trainingTherapeuticAlzheimer diseaseMild cognitive impairment

Introduction

We have recently reviewed the alterations in both cholinergic and glutamatergic systems underlying cognitive impairment and neuropathology in Alzheimer disease (AD) and demonstrated the important participation of the enzyme phospholipase A2 (PLA2) in such alterations in different ways at different stages of AD (for review, see Schaeffer and Gattaz 2008). Briefly, results suggested the hypothesis that persistent inhibition of specific subtypes of intracellular Ca2+-dependent PLA2 (cPLA2) and Ca2+-independent PLA2 (iPLA2; perhaps iPLA2-VIB) at early stages of AD may play a central role in memory deficits and production of β-amyloid (Aβ) peptide through downregulation of cholinergic receptors (involving muscarinic acetylcholine receptor M1 and M2 subtypes) and glutamate receptors (involving NMDA, AMPA, and metabotropic glutamate receptor 1 subtype). As the disease progresses, Aβ-induced upregulation of cPLA2 (likely cPLA2-IVA) and secretory Ca2+-dependent PLA2 (sPLA2; probably sPLA2-IIA) results in increased extracellular glutamate levels, NMDA overstimulation, and excessive cPLA2 and sPLA2 activation, thus establishing a vicious cycle of neuronal injury and participating in the neurodegenerative process in AD.

In the present article, we focus on the therapeutic role of PLA2 stimulation by cognitive training and positive modulators, or of supplementation with arachidonic acid (AA; PLA2 product) in facilitating memory function and synaptic transmission and plasticity in either research animals or human subjects. Based on the evidences, we hypothesize that the use of positive modulators of PLA2 (especially of cPLA2 and iPLA2) or supplementation with dietary lipid compounds (e.g., arachidonic acid) to reverse the decline of PLA2 products involved in memory and neuritogenesis, in combination with a cognitive training protocol to induce neuronal activity underlying memory and neurogenesis, could be a valuable therapeutic approach to improve the cognitive function and perhaps delay the neurodegenerative progress of patients with early-stage AD.

Methods

MEDLINE database was searched with no date restrictions for all published articles written in English using the keywords Alzheimer disease (mild, moderate, severe), mild cognitive impairment, healthy elderly, rats, mice, phospholipase A2, phospholipid metabolism, phosphatidylcholine, arachidonic acid, cognitive training, learning, memory, long-term potentiation, protein kinases, dietary lipid compounds, cell proliferation, neurogenesis, and neuritogenesis. Articles related to the objective of this review that were identified through the MEDLINE search were included. Reference lists of the identified articles were checked to select additional studies of interest. Both original and review articles on PLA2 subtypes and neurogenic areas in the brain were selected for this review. Just original articles on the alterations of PLA2 and phospholipid metabolism in AD, role of PLA2 in memory and Aβ production, activity-dependent stimulation of PLA2 and related mechanisms, effects of lipid supplementation on memory, evidence of neurogenesis in AD, effects of learning on neurogenesis, and role of PLA2 in neuritogenesis were selected for this review.

Results

The role of phospholipase A2 in Alzheimer disease pathology

PLA2: functions, subtypes, and intracerebral distribution

Phospholipases A2 are a superfamily of hydrolytic enzymes that catalyze the cleavage of fatty acids from the sn-2 position of membrane glycerophospholipids to generate lysophospholipids and free fatty acids (Dennis 1994, 1997). PLA2 preferentially cleaves membrane phosphatidylcholine (PtdCho) over other glycerophospholipids (Nalefski and Falke 1996; Larsson et al. 1998; Nalefski et al. 1998; Larsson Forsell et al. 1999; Yang et al. 1999). PLA2-catalyzed hydrolysis of PtdCho originates lyso-PtdCho (Prokazova et al. 1998) and often AA as free fatty acid (Underwood et al. 1998; Mancuso et al. 2000). AA at high levels is converted to inflammatory mediators, such as prostaglandins and leukotrienes, by cyclooxygenases (COX) and lypoxygenases, respectively. AA at physiological levels can directly modulate neural cell function by various mechanisms, such as altering membrane fluidity, regulating ion channels, and facilitating neurotransmitter release and uptake. Some prostaglandins play an important role in normal neural activity by modulating the release of neurotransmitters. Lysophospholipids include 1-alkyl-2-lyso-sn-glycero-3-phosphocholine, the immediate precursor of platelet-activating factor (PAF), a potent inflammatory mediator. Lysophospholipids may also change membrane fluidity. Lyso-PtdCho can modulate ion channels and neurotransmitter release (Farooqui and Horrocks 2006). Finally, free choline generated by hydrolysis of membrane PtdCho is utilized to synthesize the neurotransmitter acetylcholine (Maire and Wurtman 1985; Blusztajn et al. 1987a, b).

So far, at least 24 enzymes that possess PLA2 activity have been identified in mammals, which are classified into four main groups: (a) secretory (extracellular) Ca2+-dependent PLA2 (sPLA2), subdivided into several groups: IB, II (IIA, IIC, IID, IIE, IIF), III, V, X, XII, XIII; (b) cytosolic Ca2+-dependent PLA2 (cPLA2), subdivided into six groups: IVA, IVB, IVC, IVD, IVE, IVF; (c) intracellular Ca2+-independent PLA2 (iPLA2), which consists of two major groups: VI (VIA, VIB) and PAF-acetylhydrolases (plasmatics), subdivided into groups VII (VIIA, VIIB) and VIII (VIIIA, VIIIB); and (d) plasmalogen-selective PLA2 (Six and Dennis 2000; Winstead et al. 2000; Kudo and Murakami 2002; Chakraborti 2003; Farooqui and Horrocks 2004; Ohto et al. 2005). In the human brain, sPLA2-IIE, V, X, and XII have been identified (Chen et al. 1994; Gelb et al. 2000; Suzuki et al. 2000). In addition, cPLA2-IVA, IVB, and IVC have been detected in the human brain: group IVA was detected in hippocampus, amygdala, substantia nigra, thalamus, subthalamic nucleus, and corpus callosum (Sharp and White 1993; Pickard et al. 1999); group IVB in temporal and frontal lobes, occipital pole, hippocampus, amygdala, caudate nucleus, putamen, substantia nigra, thalamus, subthalamic nucleus, cerebellum, and corpus callosum (Pickard et al. 1999; Song et al. 1999); and group IVC in hippocampus, amygdala, caudate nucleus, substantia nigra, thalamus, subthalamic nucleus, and corpus callosum (Underwood et al. 1998; Pickard et al. 1999). Finally, iPLA2 subtypes (VIA and VIB) have also been identified in the human brain (Larsson Forsell et al. 1999; Mancuso et al. 2000; Tanaka et al. 2000). In rat brains, sPLA2-IIA, IIC, and V, and iPLA2-VI have been detected in several regions, including the cerebral cortex, hippocampus, striatum, thalamus, hypothalamus, and cerebellum (Molloy et al. 1998). In addition, cPLA2-IVA has been detected within the rat brain, including the cerebral cortex, hippocampus (CA1, CA2, CA3, dentate gyrus), amygdala, striatum, thalamus, hypothalamus, cerebellum, and olfactory bulb (Owada et al. 1994; Molloy et al. 1998; Kishimoto et al. 1999). In mice brains, sPLA2-IIC and IIE, cPLA2-IVA and iPLA2-VIA have been identified (Lautens et al. 1998; Six and Dennis 2000; Suzuki et al. 2000; Bosetti and Weerasinghe 2003; Shinzawa et al. 2008). Data described in this subtitle are summarized in Table 1.
Table 1

Distribution of PLA2 subtypes in human and rodent brains

PLA2 subtypes

Human brain

Rat brain

Mouse brain

sPLA2-IIA

 

Cerebral cortex, hippocampus, striatum, thalamus, hypothalamus, cerebellum

 

sPLA2-IIC

 

Cerebral cortex, hippocampus, striatum, thalamus, hypothalamus, cerebellum

Brain (undefined region)

sPLA2-IIE

Brain (undefined region)

 

Brain (undefined region)

sPLA2-V

Brain (undefined region)

Cerebral cortex, hippocampus, striatum, thalamus, hypothalamus, cerebellum

 

sPLA2-X

Brain (undefined region)

  

sPLA2-XII

Brain (undefined region)

  

cPLA2-IVA

Hippocampus, amygdala, substantia nigra, thalamus, subthalamic nucleus, corpus callosum

Cerebral cortex, hippocampus (CA1, CA2, CA3, dentate gyrus), amygdala, striatum, thalamus, hypothalamus, cerebellum, olfactory bulb

Brain (undefined region)

cPLA2-IVB

Temporal lobe, frontal lobe, occipital pole, hippocampus, amygdala, caudate nucleus, putamen, substantia nigra, thalamus, subthalamic nucleus, cerebellum, corpus callosum

  

cPLA2-IVC

Hippocampus, amygdala, caudate nucleus, substantia nigra, thalamus, subthalamic nucleus, corpus callosum

  

iPLA2-VI

 

Cerebral cortex, hippocampus, striatum, thalamus, hypothalamus, cerebellum

 

iPLA2-VIA

Brain (undefined region)

 

Brain (undefined region)

iPLA2-VIB

Brain (undefined region)

  

Reduced brain PLA2 activity in early-stage Alzheimer disease

Previous studies from our group showed reduced PLA2 activity in postmortem parietal and frontal cortices of AD patients, and this reduction was correlated with the severity of dementia (Gattaz et al. 1995a, 1996). These findings were supported by Ross and colleagues (1998), who reported reduced cPLA2 and iPLA2 activities in postmortem parietal and temporal cortices of AD patients as well as decreased cPLA2 activity in the hippocampus. Moreover, decreased iPLA2 activity was found in postmortem prefrontal cortex of frontal-variant AD patients (Talbot et al. 2000). These data are in line with in vivo spectroscopic measurements of brain membrane phospholipid turnover in early AD. PLA2-catalyzed hydrolysis of membrane glycerophospholipids leads to the production of metabolites observable by both proton (1H) and phosphorus (31P) magnetic resonance spectroscopy (MRS). The signal of choline-containing compounds observed by 1H-MRS is constituted of metabolites of PtdCho, especially glycerophosphocholine, phosphocholine, and free choline, and considered to reflect membrane turnover (Miller et al. 1996). With 31P-MRS, the phosphomonoester (PME, such as phosphocholine) and phosphodiester (PDE, such as glycerophosphocholine) signals observed correspond, respectively, to the precursors and breakdown products of phospholipids and the ratio between both provides in vivo information on membrane turnover (Pettegrew et al. 1987). Variations in the intensity of the choline, PME, and PDE signals have been reported in AD brains. In vivo 1H-MRS has provided evidence for a reduction of choline signal in medial temporal lobes (including the hippocampus, subiculum, and amygdala) of mild and moderate AD patients, and this reduction was correlated with loss of verbal memory (Jessen et al. 2000; Chantal et al. 2002). On the other hand, increased choline signal found in the posterior cingulate cortex of patients with mild cognitive impairment (MCI) and mild to moderate AD (Kantarci et al. 2000, 2003) may represent a compensatory mechanism to the deficiency of free choline for acetylcholine synthesis by increased degradation of membrane PtdCho (Blusztajn et al. 1986). In vivo 31P-MRS findings from our and other laboratories have shown elevations of PME levels and of the PME/PDE ratio in parietotemporal association and prefrontal cortices of patients with mild and moderate AD (Brown et al. 1989; Pettegrew et al. 1994, 1995; Cuénod et al. 1995; Forlenza et al. 2005). Additionally, we demonstrated that higher PME in the prefrontal cortex of mild and moderate AD patients was correlated with the severity of cognitive decline and particularly with functions such as memory, visual perception, orientation, and abstract thinking (Forlenza et al. 2005). Moreover, a significant increase in the PME/PDE ratio of approximately 50%, mostly due to a decrease in the PDE signal, was found in frontal and parietal cortices of patients with mild to moderate AD (Gonzalez et al. 1996). Collectively, these findings support the evidence for a decrement of PLA2 activity as an early manifestation in AD.

On the other hand, increased immunoreactivity of cPLA2-IVA was reported in reactive astrocytes associated with Aβ deposits in postmortem occipital, temporal, and frontal cortices of severe AD patients (Stephenson et al. 1996, 1999). Additionally, increased sPLA2-IIA-immunoreactive astrocytes were found in postmortem inferior temporal gyrus and hippocampal dentate gyrus and CA3 field of severe AD patients. In the inferior temporal gyrus, the majority of sPLA2-IIA-positive astrocytes were associated with Aβ-containing plaques (Moses et al. 2006). Finally, upregulation of cPLA2 mRNA was reported in the hippocampal CA1 field of moderate to severe AD patients (Colangelo et al. 2002), and upregulation of sPLA2-IIA mRNA was found in the hippocampus (confined mainly to dentate gyrus and CA3 field) of severe AD patients (Moses et al. 2006). In accordance with the upregulation of PLA2 in late stages of AD, Miatto and colleagues (1986) found decreased PME levels and increased PDE levels in autopsy specimens of frontal and parietal cortices from patients with severe AD. Additionally, Pettegrew and colleagues (1988) found increased PDE levels (reflecting excessive degradation of neural membranes) in cortical areas of patients with severe AD besides the finding of elevated PME levels (reflecting decreased degradation of neural membranes) in mild AD, hypothesizing that PME elevations occur early in the disease, whereas PDE increments occur later. Correlations with morphologic data further suggested that elevations of PME levels precede the appearance of neuritic plaques in early stages of AD, whereas elevations of PDE levels correlate with the density of neuritic plaques in late stages of the disease (Pettegrew et al. 1988). The findings in early AD are in accordance with in vitro studies suggesting that reduced PLA2 activity may contribute to Aβ production (Emmerling et al. 1993, 1996; Nitsch et al. 1997; Cho et al. 2006); the findings in late AD are in line with the studies of Stephenson et al. (1996, 1999) and Moses et al. (2006), showing a correlation between increased PLA2 immunoreactivity and Aβ-containing plaques. Finally, in autopsy specimens of parietal lobe gray matter from subjects with severe AD, greater PDE was correlated with the density of neurofibrillary tangles (Smith et al. 1993). Data described in this subtitle are summarized in Table 2.
Table 2

Alterations of PLA2 and phospholipid metabolism beginning at the early stages of Alzheimer disease

 

Brain region

PLA2 activity

PLA2 immunoreactivity

PLA2 mRNA

Choline signal

PME levels

PDE levels

PME/PDE ratio

MCI

Cingulate cortex

   

   

Mild AD

Frontal cortex

    

Parietal cortex

    

Temporal cortex

    

 

Cingulate cortex

   

   

Hippocampus

   

   

Moderate AD

Frontal cortex

    

Parietal cortex

    

Temporal cortex

    

 

Cingulate cortex

   

   

Hippocampus

  

↑ (c)

   

Severe AD

Frontal cortex

↓ (i)

↑ (c) Astrocytes

  

 

Parietal cortex

↓ (c + i)

   

 

Temporal cortex

↓ (c + i)

↑ (c + s) Astrocytes

     

Occipital cortex

 

↑ (c) Astrocytes

     

Hippocampus

↓ (c)

↑ (s) Astrocytes

↑ (c + s)

    

reduced, increased, c cPLA2, i iPLA2, s sPLA2

Taken together, the data suggest that decreased membrane phospholipid metabolism mediated by reduced activity of PLA2 (particularly cPLA2 and iPLA2) in the cerebral cortex and hippocampus is an early event in AD and may contribute to cognitive dysfunction and neuropathology at early stages of the disease. As the disease worsens, cPLA2 and sPLA2 levels (and accordingly the metabolism of phospholipids) become elevated in AD brains, supporting, thus, the hypothesis that there is an active inflammatory process occurring in AD.

Reduced PLA2 activity and memory impairment in research animals

Several studies in laboratory animals have shown that the inhibition of PLA2, as well as the blockade of AA production, impairs learning and memory, simulating deficits that are found since the earliest phases of AD and represent the most predominant cognitive changes in this disease. For instance, infusions of nonselective PLA2 inhibitors into the chick intermediate medial hyperstriatum ventrale impaired learning of a passive avoidance task, in which the aversive stimulus is an unpleasant tasting substance (Hölscher and Rose 1994). Additionally, intraperitoneal injection of a nonselective PLA2 inhibitor in rats impaired spatial learning tested in the Morris water maze (Hölscher et al. 1995). Furthermore, intracerebroventricular infusion in mice of a nonselective PLA2 inhibitor or a dual cPLA2 and iPLA2 inhibitor impaired memory formation of a step-through inhibitory avoidance task (in which a context, tone, and foot shock are presented together in an associative fashion; Sato et al. 2007), and a selective iPLA2 inhibitor impaired spatial learning tested in the Y-maze (Fujita et al. 2000). Recent studies from our group showed that infusions of dual cPLA2 and iPLA2 inhibitors or a selective iPLA2 inhibitor into rat hippocampal CA1 field impaired acquisition of short- and long-term memory (Schaeffer and Gattaz 2005) and retrieval of long-term memory of a step-down inhibitory avoidance task, in which fear conditioning is induced by a single exposure to a context followed by an electric foot shock (Schaeffer and Gattaz 2007).

Studies of the effect of inhibitors of COX pathway of AA metabolism have supported the involvement of AA (PLA2 product) in learning and memory. Thus, infusions of non-specific COX inhibitors, or specific COX-1 or COX-2 inhibitors into the chick intermediate medial hyperstriatum ventrale impaired both acquisition and consolidation of a passive avoidance task (Hölscher 1995a, b). Additionally, intraperitoneal injection of a nonspecific COX inhibitor or a specific COX-2 inhibitor impaired memory formation of a passive avoidance task in mice (Sato et al. 2007) and spatial memory retention in the Morris water maze in rats (Teather et al. 2002). Moreover, infusion of a selective COX-2 inhibitor into rat hippocampal CA1 field impaired both acquisition (Rall et al. 2003) and retention (Sharifzadeh et al. 2005) of spatial memory in the Morris water maze. Finally, the concentration of prostaglandins (COX products of AA metabolism) was enhanced in the chick intermediate medial hyperstriatum ventrale after passive avoidance training, and intracerebral infusions of specific COX-1 or COX-2 inhibitors prevented the training-related increase of prostaglandins release (Hölscher 1995b).

Reduced PLA2 activity and Aβ production in cultured cells

In AD, Aβ peptide plays a major role in neurodegeneration, which is the principal underlaying cause of the cognitive decline. The abnormal processing and/or increased expression of β-amyloid precursor protein (β-APP) is associated with Aβ formation. The cleavage of β-APP by α-secretase within the Aβ domain generates non-amyloidogenic C-terminal APP fragment and soluble APPα (sAPPα), which has neurotrophic and neuroprotective effects, whereas sequential processing of β-APP by β- and γ-secretases generates the Aβ peptide, which has neurotoxic effects (Caporaso et al. 1992; Fukushima et al. 1993; Fluhrer et al. 2002).

Although not yet fully understood, abnormal PLA2 metabolism may underlie the neurodegenerative process in AD. Several studies in cell lines support the involvement of PLA2 in neuropathology and, hence, neurodegeneration in AD. Such studies suggest that α-secretase processing of β-APP and, therefore, sAPPα release induced by cholinergic and glutamate receptor activation is regulated by PLA2-induced AA release. For instance, PLA2 activation in HEK 293 cells increased the release of sAPPα evoked by the stimulation of metabotropic glutamate receptor 1α subtype (Nitsch et al. 1997). In addition, PLA2 activation in CHO cells increased the secretion of sAPPα evoked by the stimulation of muscarinic acetylcholine receptor M1 subtype, whereas PLA2 inhibition decreased sAPPα release (Emmerling et al. 1993). Moreover, treatment of CHO cells with exogenous PLA2 or AA increased sAPPα release (Emmerling et al. 1996). Finally, activation of PLA2 or exposure to AA in SH-SY5Y cells increased the secretion of sAPPα evoked by the activation of muscarinic acetylcholine receptors, whereas PLA2 inhibition decreased sAPPα release (Cho et al. 2006). Therefore, abnormal PLA2 activity may favor Aβ formation. These data support the findings showing that reduced PLA2 activity correlates with the density of neuritic plaques in postmortem AD brains (Gattaz et al. 1995a, 1996), and elevations of PME levels precede the appearance of neuritic plaques in early AD (Pettegrew et al. 1988).

Stimulation of phospholipase A2 for the treatment of early-stage Alzheimer disease

Learning and memory induce PLA2 activation in research animals

Animal research has elucidated some possible brain biochemical mechanisms related to experience-dependent stimulation, and PLA2 activation seems to be highly implicated here. Studies from our group showed that training of rats in the one-trial step-down inhibitory avoidance task (contextual fear task) increased the endogenous activity of PLA2 in the CA1 field of hippocampus (Schaeffer and Gattaz 2005). Additionally, our studies showed that re-exposure of rats to context after training (contextual memory retrieval) stimulated the activity of PLA2 in the CA1 field (Schaeffer and Gattaz 2007). In a recent study, we have extended our investigation in the hippocampus to the cerebral cortex of rats and found that PLA2 activity was increased in both frontal and parietal cortices around the time of training and during consolidation of the step-down inhibitory avoidance. Additionally, PLA2 activity was increased in the parietal cortex immediately after retrieval of the inhibitory avoidance (Schaeffer et al. submitted). Furthermore, training of chicks in a one-trial passive avoidance task was followed by enhanced concentration of AA (Clements and Rose 1996) and prostaglandins (Hölscher 1995b) in the intermediate medial hyperstriatum ventrale. Several mechanisms involving AA have been shown to contribute to the increase of glutamate in the synaptic cleft that occurs in LTP and, thereby, to modulate postsynaptic glutamate receptors. For instance, AA inhibited glutamate uptake into neurons and glial cells (astrocytes) from rat cerebral cortex and hippocampus (Yu et al. 1986, 1987; Barbour et al. 1989; Volterra et al. 1992). Additionally, prostaglandins (in particular PGE2) applied to the CA1 field of rat hippocampal slices induced increases in [Ca2+]i in astrocytes, thus, stimulating these cells to release glutamate; the release was drastically reduced by blocking PLA2. In turn, glutamate released from astrocytes upon PGE2 stimulation triggered [Ca2+]i elevations mediated by NMDA and AMPA receptors in a large number of neurons from the hippocampal CA1 field (Bezzi et al. 1998).

The other product of PLA2-catalyzed hydrolysis of PtdCho, lyso-PtdCho, is also implicated in learning and memory. In vivo induction of LTP increased the concentration of free AA and decreased the levels of PtdCho in postsynaptic densities prepared from rat hippocampal slices; such changes were due to enhanced PLA2 activity after LTP induction (Clements et al. 1991). Free choline generated by hydrolysis of membrane PtdCho in both cholinergic and non-cholinergic neurons is utilized by cholinergic neurons to synthesize the neurotransmitter acetylcholine (Maire and Wurtman 1985; Blusztajn et al. 1987a, b). Finally, lyso-PtdCho caused a long-lasting enhancement of acetylcholine-induced currents in Xenopus oocytes (Ikeuchi et al. 1997). These data suggest that PLA2-catalyzed hydrolysis of PtdCho into lyso-PtdCho, and subsequently, free choline is required for potentiation of cholinergic neurotransmission that accompanies LTP.

Several other biochemical mechanisms which are closely connected to PLA2 have been related to activity-dependent changes. Hence, it was demonstrated that memory consolidation of a passive avoidance task in chicks is specifically associated with elevations in the binding of ligands to the NMDA glutamate receptor in the intermediate medial hyperstriatum ventrale (Stewart et al. 1992). Additionally, training of rats in the step-down inhibitory avoidance resulted in an increase in NMDA NR1 subunit expression in the hippocampus (Cammarota et al. 2000). Exposure of rats to a discrimination task (spatial learning task) resulted in learning-induced activation of protein kinase C (PKC) in the hippocampus (Vázquez and Peña de Ortiz 2004). Exposure of rats to the step-down inhibitory avoidance resulted in increased activation of PKC in the frontal, parietal, and entorhinal cortices and hippocampus (Bernabeu et al. 1995; Cammarota et al. 1997; Young et al. 2002) and increased levels of PKCβ1 subunit in the hippocampus (Paratcha et al. 2000). In rats exposed to the one-trial step-through inhibitory avoidance task, Young and colleagues (2002) also observed increased activation of hippocampal PKC. Additionally, training of rats in the step-down inhibitory avoidance provoked an increase in the activity of Ca2+/calmodulin-dependent protein kinase II (CaMKII) and in the amount of CaMKIIα subunit in the hippocampus (Cammarota et al. 1998). Moreover, learning of the inhibitory avoidance task was accompanied by an increase in the phosphorylation of p38, p42, and p44 mitogen-activated protein kinase (MAPK) in the rat hippocampus (Alonso et al. 2002, 2003). In mice, re-exposure to context after training (contextual memory retrieval) stimulated the activity of p42 and p44MAPK in the hippocampus (Chen et al. 2005). Finally, step-down inhibitory avoidance training increased [3H]AMPA binding to the AMPA glutamate receptor (Cammarota et al. 1995, 1996; Bernabeu et al. 1997) and the levels of AMPA GluR1 receptor subunit in the hippocampus of rats (Bernabeu et al. 1997; Cammarota et al. 1998). Data described in this subtitle are summarized in Table 3.
Table 3

Biochemical mechanisms stimulated in rodent and chick brains during processing of new experience

Brain region

Phases of memory processing

Acquisition

Consolidation

Retrieval

Parietal cortex

PLA2, PKC

PLA2

PLA2

Frontal cortex

PLA2, PKC

PLA2

 

Entorhinal cortex

PKC

  

Hippocampus

PLA2, AA, NMDA, PKC, CaMKII, p38, p42 and p44 MAPK, AMPA

 

PLA2, p42 and p44 MAPK

IMHV

AA, prostaglandins

NMDA

 

IMHV intermediate medial hyperstriatum ventrale (chick brain)

As already mentioned, PLA2 activity is closely connected to all biochemical mechanisms described above. PLA2-dependent release of AA is a receptor-mediated process. In this way, activation of postsynaptic NMDA glutamate receptors raises postsynaptic [Ca2+]i and stimulates cPLA2, which generates AA, as found in mouse cortical neurons and rat hippocampal neurons and slices (Sanfeliu et al. 1990; Pellerin and Wolfe 1991; Lazarewicz et al. 1992; Stella et al. 1995). Many studies have demonstrated that PLA2 can be regulated by a variety of protein kinases. For example, activation of cPLA2 is regulated by p38MAPK (Zhou et al. 2003), p42MAPK (Lin et al. 1993; Nemenoff et al. 1993; Gordon et al. 1996), CaMK II (Muthalif et al. 2001), and PKC phosphorylation (Wijkander and Sundler 1991; Nemenoff et al. 1993), and iPLA2 activation is regulated by PKC phosphorylation (Underwood et al. 1998; Akiba et al. 1999). In turn, stimulation of PLA2 activity (with the PLA2 activator melittin) in the presence of Ca2+ in rodent cortical and hippocampal slices as well as membrane preparations increased [3H]AMPA and [3H]glutamate binding to the AMPA receptor (Massicotte and Baudry 1990; Baudry et al. 1991; Massicotte et al. 1991; Tocco et al. 1992; Catania et al. 1993; Bernard et al. 1995; Chabot et al. 1998; Gaudreault et al. 2004), whereas PLA2 inhibition and the Ca2+ chelator EGTA reduced agonist binding to AMPA (Bernard et al. 1993, 1995; Catania et al. 1993). Additionally, PLA2 inhibition in rat hippocampal slices prevented Ca2+-dependent formation of LTP in the CA1 field as well as the increase of [3H]AMPA binding to the AMPA receptor that characterizes LTP (Bernard et al. 1994). Moreover, selective inhibition of iPLA2-VIB prevented LTP induction in the CA1 field of rat hippocampal slices as well as the associated increase of [3H]AMPA binding to the AMPA receptor and the upregulation of AMPA GluR1 subunit levels in crude synaptic fractions (Martel et al. 2006). Collectively, the data suggest that PLA2 activation is induced in the brain of laboratory animals during processing of new experience.

Cognitive training induces PLA2 activation in platelets of elderly subjects

In the search for peripheral correlates of cerebral enzymatic abnormalities, several studies have focused on PLA2 activity in platelet membranes (Gattaz et al. 1995b, 1996, 2004). Platelets are frequently used as peripheral markers for neurons because they share some common membrane and receptor properties, providing, thus, an interesting model for the investigation of metabolic abnormalities in AD (Zubenko et al. 1999). There is indirect evidence that brain and platelet PLA2 activity are under similar regulatory control. Forlenza and colleagues (2005) showed that reduced intracerebral phospholipid breakdown correlates with cognitive impairment in AD, and this finding is further correlated with lower platelet PLA2 activity, further supporting the hypothesis that platelet PLA2 activity may mirror the activity of the enzyme in the cerebral tissue. In a previous study, we found decreased PLA2 activity in platelets of patients with AD (Gattaz et al. 1996), and more recently, we found that lower activity of platelet iPLA2, present also in the brain, was correlated with the severity of cognitive decline in samples of patients with MCI and AD (Gattaz et al. 2004). It is worth noticing that MCI patients are at higher risk to convert to AD with an estimated rate of 10–30% per year (Petersen et al. 2001a, b). Such individuals display intermediate values for PLA2 activity as compared to normal controls and AD patients (Gattaz et al. 2004).

Procedures of nonpharmacological treatment such as cognitive training have been clinically performed and reported to be effective in improving memory function in elderly subjects with MCI (Rapp et al. 2002; Belleville et al. 2006; Wenisch et al. 2007) and early-stage AD (Clare et al. 2002; Abrisqueta-Gomez et al. 2004; Avila et al. 2004). In a 1H-MRS study, Valenzuela and colleagues (2003) reported an elevation of creatine and choline signals in the hippocampus of healthy older adults after 5 weeks of memory training; in this work, the authors concluded that such changes might be a marker of increased neuronal plasticity in these individuals. In a recent study conducted in our laboratory, we investigated the effects of cognitive training on platelet PLA2 activity in healthy elderly individuals. Twenty-three cognitively unimpaired older adults were randomly assigned to receive memory training or standard outpatient care only. Both groups were cognitively assessed by the same protocol, and the experimental group underwent a four-session memory training intervention. After 1 month, patients in the experimental group had significant increase in cPLA2, sPLA2, total platelet PLA2 activity, and significant decrease in platelet iPLA2 activity, suggesting that memory training may have a modulating effect in PLA2-mediated biological systems associated with cognitive functions and neurodegenerative diseases (Talib et al. 2008).

Dietary lipid compounds improve memory function in research animals and human subjects

Nonpharmacological interventions with dietary lipid compounds are recognized as important protective measures against the development of AD-related pathology in animal models. Animal studies have shown that long-term intake of dietary compounds enriched with AA preserves hippocampal synaptic plasticity and neuronal membrane fluidity and enhances spatial memory performance of aged rats. A significant correlation was observed between the behavioral performance and AA composition in hippocampal phospholipids (Kotani et al. 2003; Okaichi et al. 2005; Fukaya et al. 2007). Additionally, long-term dietary supplementation with PtdCho enhanced spatial learning ability of aged mice (Lim and Suzuki 2000a, b, 2002). Because PLA2 acting on membrane PtdCho generates lipid mediators such as AA (a facilitatory retrograde messenger in glutamatergic neurotransmission) and lyso-PtdCho (involved in the synthesis of acetylcholine), PLA2 activity may be one of the mediators of the beneficial effects on memory of dietary PtdCho.

In a recent study, Kotani and colleagues (2006) investigated for the first time the effects of supplementation with AA in human amnesic patients and demonstrated that subjects with MCI treated with the supplementation showed a significant improvement of the immediate memory and attention score, and subjects with organic brain lesions showed a significant improvement of immediate and delayed memories. Interestingly, in the cerebrospinal fluid of AD patients the levels of phospholipids and also fatty acids were found to be significantly reduced (Mulder et al. 1998). Additionally, blood plasma phospholipid and PtdCho levels were reported to be decreased in patients with AD and MCI (Conquer et al. 2000). It is noteworthy that the threshold levels of AA that promotes and facilitates neural cell injury and death still needs to be established, inasmuch as the downstream effects of the activation of certain PLA2 subtypes on the synthesis of proinflammatory cytokines.

Endogenous neurogenesis in Alzheimer disease

Mounting evidence has accumulated showing the existence of a phenomenon designated “adult neurogenesis” in the mammalian brain. De novo generation of neurons by neural progenitor cells in the adult brain occurs in specific brain areas, such as the subgranular zone of the dentate gyrus in the hippocampus and the rostral subventricular zone of the lateral ventricles. New neurons leaving the subgranular zone migrate into the adjacent granule cell layer of the dentate gyrus and neurons that arise in the subventricular zone migrate into the olfactory bulb (Yamashima et al. 2007). New neurons residing in the subventricular zone also enter the association neocortex (prefrontal, inferior temporal, posterior parietal; Gould et al. 1999b). Neurogenesis in these regions occurs in response to several physiological stimuli, including excitatory neurotransmission (Cameron et al. 1998) and learning (Gould et al. 1999a, c), as well as in response to various pathological situations, including epileptic seizures (Bengzon et al. 1997; Gray and Sundstrom 1998), excitotoxic and mechanical lesions (Gould and Tanapat 1997), and ischemia (Jiang et al. 2001; Jin et al. 2001; Kee et al. 2001; Zhang et al. 2001) in animal models.

Recent studies have provided initial evidence that endogenous neurogenesis may be active in the adult human brain in response to neurodegenerative processes such as AD. For example, increased expression of immature neuronal marker proteins (doublecortin, polysialylated nerve cell adhesion molecule, neurogenic differentiation factor, and TUC-4) was reported in postmortem hippocampus of AD patients, with immuhistochemical localization to neurons in the subgranular zone of the dentate gyrus (neuroproliferative), granule cell layer (the destination of neurons of the subgranular zone), and CA1 region, which is the principal site of hippocampal pathology in AD (Jin et al. 2004). Moreover, increased expression of Ki-67, a cell-cycle marker protein, was found in the nuclei of neurons and glial cells in all subregions of postmortem hippocampus of AD patients, with the highest levels in the dentate gyrus (Nagy et al. 1997). In further studies, increased expression of Ki-67 and proliferating cell nuclear antigen, cell-cycle marker proteins, and mini-chromosome maintenance protein 2, a marker for chromosomal replication, was observed in the nuclei of neurons and glial cells in the hippocampus and entorhinal cortex of elderly human brains with different extents of AD pathology (Wharton et al. 2005). These data suggest that increased neurogenesis in AD brains may give rise to cells that replace neurons lost in the disease. In AD-like mice models, it has been demonstrated that neurogenesis in the hippocampal subregions is significantly increased at early stages of neurodegeneration; however, the survival of newly generated neurons is impaired at late stages of neurodegeneration, suggesting that strategies to enhance endogenous neurogenesis could have greater therapeutic potential when given to AD patients at early stages of the disease (López-Toledano and Shelanski 2007; Chen et al. 2008; Gan et al. 2008).

Learning and memory promote adult neurogenesis in research animals

Although its functional significance is not completely understood, several lines of evidence suggest the role of dentate gyrus neurogenesis in learning and memory. For instance, exposure to an enriched environment with opportunities for social interaction, exploration, and physical activity increased the survival of adult-generated neurons, cell proliferation, and neurogenesis in the dentate gyrus of adult mice (Kempermann et al. 1997, 1998a). Environmental stimulation also increased the survival of adult-generated neurons in the dentate gyrus of adult rats (Nilsson et al. 1999) and aged mice (Kempermann et al. 1998b). Additionally, adult rats housed in an enriched environment showed improved performance in a spatial learning task (Nilsson et al. 1999). Furthermore, training on a hippocampus-dependent spatial learning task (Morris water maze) enhanced the survival of new neurons in the dentate gyrus of adult rats (Ambrogini et al. 2000; Dupret et al. 2007; Epp et al. 2007). Finally, training on associative learning tasks that require the hippocampus increased the survival of adult-generated neurons in the rat dentate gyrus (Gould et al. 1999a; Dalla et al. 2007). These data indicate that learning and memory provide a cellular or molecular environment that favors both proliferation and survival of adult-generated neurons in the dentate gyrus. Additionally, the data suggest that experimental approaches to promote de novo neurogenesis may potentially improve neurocognitive function and provide an effective therapy for AD.

PLA2, neuronal homeostasis, and neuroplasticity

The relationship between PLA2 as well as their metabolites (AA and lyso-PtdCho) and neurite outgrowth has been evaluated in several tissue culture models. PLA2 and their products have been implicated in neuritogenesis, in both neurodevelopmental processes and response to neuronal injury. For instance, in cultures of cerebellar neurons treated with exogenous AA, or the PLA2 activator melittin, or a nonselective PLA2 inhibitor, AA was shown to be involved in an intracellular signaling pathway leading to neurite outgrowth, in a process dependent on the activation of neuronal Ca2+ channels (Williams et al. 1994a, b). Accordingly, neurite outgrowth was stimulated in mouse neuroblastoma × rat glioma hybrid NG108-15 cells overexpressing human COX-2 and showing neurotoxic levels of PGE2 (COX product of AA metabolism), β-APP, and Aβ peptide (Kadoyama et al. 2001). It has been recently shown that secretory phospholipases A2 exhibit neurotrophic effect. Hence, sPLA2-X was shown to protect cultured cerebellar granule neurons from apoptosis; this survival-promoting effect was correlated with the extent of sPLA2-induced AA release and was dependent on the activation of neuronal Ca2+ channels (Arioka et al. 2005). Accordingly, sPLA2-X was shown to exhibit neurotrophin-like activity in the rat pheochromocytoma cell line PC12; this neuritogenic effect of sPLA2 was mediated by generation of lyso-PtdCho and subsequent activation of G2A, a G-protein-coupled receptor involved in membrane phospholipid metabolism (Ikeno et al. 2005). Additionally, expression of sPLA2-X in PC12 cells was demonstrated to facilitate neurite outgrowth, particularly when combined with a suboptimal concentration of nerve growth factor (NGF). The neurite-extending ability of sPLA2-X depended on the production of lyso-PtdCho. Moreover, NGF-induced neurite extension of PC12 cells was modestly but significantly attenuated by an anti-sPLA2-X antibody or by a small interfering RNA for sPLA2-X (Masuda et al. 2005). In a very recent study, Masuda and colleagues (2008) demonstrated that expression of sPLA2-III in PC12 cells facilitated neurite outgrowth, whereas expression of a catalytically inactive sPLA2-III mutant or use of sPLA2-III-directed small interfering RNA reduced NGF-induced neuritogenesis. sPLA2-III also suppressed neuronal death induced by NGF deprivation; these neuritogenic and neurotropic effects of sPLA2 were mediated by generation of lyso-PtdCho.

Both cPLA2 and iPLA2 may play a role in neuronal growth and differentiation, as initially demonstrated by Smalheiser and colleagues (1996) that exogenous AA enhanced neuritogenesis in cultured NG108-15 cells, while the administration of a dual cPLA2 and iPLA2 inhibitor delayed the initial outgrowth of neurites. In a recent study conducted in our laboratory, we addressed the effects of the chronic treatment of mature primary cultures of rat neurons with a dual cPLA2 and iPLA2 inhibitor or a selective iPLA2 inhibitor and demonstrated that the early and sustained inhibition of PLA2 precludes the in vitro development of primary cortical and hippocampal neurons. Furthermore, the inhibition of PLA2 disrupted the viability of both immature and mature neurons, affecting their neurite-bearing morphology and cellular viability (Forlenza et al. 2007). These data suggest the potential contribution of PLA2 enzymes to neuronal growth and differentiation in both neurodevelopmental processes and response to neuronal injury.

Discussion

The data summarized above have established a crucial involvement of PLA2 in the mechanisms of learning and memory and suggest a substantial role of reduced PLA2 activity in the impairments of learning and memory in AD. There is increasing evidence showing that reduced in vivo membrane phospholipid turnover in the cerebral cortex and hippocampus is an early event in AD and correlates with the severity of cognitive decline. PLA2 is a key enzyme in membrane phospholipid metabolism. Thus, there is a suggestion that the reduced activity of PLA2 (particularly cPLA2 and iPLA2) reported in AD cortex and hippocampus may be implicated in the reduction of phospholipid turnover early in the disease. Interestingly, convergent evidence on behavioral and biochemical levels points to the involvement of both cPLA2 and iPLA2 in learning and memory. In this way, nonselective PLA2 inhibition (including cPLA2 + iPLA2 inhibition) and selective iPLA2 inhibition in brains of laboratory animals impaired spatial learning and contextual learning and retrieval. PLA2 acting on membrane phospholipids generates lipid mediators such as lyso-PtdCho and AA. Considering these actions of PLA2, the biochemical mechanisms by which intracerebral blockade of AA production impaired spatial learning and retrieval and contextual learning in laboratory animals are likely to involve reduced glutamate levels in the synaptic cleft, thereby decreasing glutamate binding to postsynaptic NMDA and AMPA receptors that occurs in LTP. Additionally, decreased levels of lyso-PtdCho mediated by PLA2 inhibition may result in reduced levels of free choline (precursor of acetylcholine), thus, contributing to learning and memory deficits. Together, the data provide evidence for a disordered brain phospholipid metabolism mediated by reduced activity of PLA2 (cPLA2 and iPLA2) early in AD pathology and suggest that such changes may account for memory impairment in the disease.

There is increasing evidence from studies in laboratory animals suggesting that PLA2 activation is induced in the healthy brain during processing of new experience. PLA2 activity is closely connected to several brain biochemical mechanisms (protein kinases, neurotransmitter receptors) related to experience-dependent stimulation. Therefore, PLA2 activation observed in the rat hippocampus and parietal cortex during contextual learning and retrieval and frontal cortex during contextual learning might facilitate memory through upregulation of AMPA glutamate receptors at the synapse via a PKC- and/or MAPK-dependent cascade. This positive modulation of AMPA receptors and memory by PLA2 has been shown to involve the production of AA. Additionally, PLA2-mediated production of lyso-PtdCho and, subsequently, free choline (precursor of acetylcholine) is required for mechanisms of synaptic plasticity underlying learning and memory. Studies with healthy elderly subjects further suggest that memory training may have a facilitating effect in PLA2-mediated biological systems associated with cognitive functions and neurodegenerative diseases. These data support the use of cognitive training as a promising nonpharmacological approach to stimulate PLA2 at least in healthy elderly subjects for the prevention of neurocognitive deficits.

The findings showing that interventions with dietary AA and PtdCho improve memory function in aged rodents and human amnesic patients indicate that the adverse effects on memory of AA and PtdCho deficiencies seen in AD and MCI may be reversible with a dietary supplementation. However, it is worth noticing that the threshold levels of AA that promote and facilitate neural cell injury and death still needs to be established, inasmuch as the downstream effects of the activation of certain PLA2 subtypes on the synthesis of proinflammatory cytokines.

Several lines of evidence suggest that learning and memory provide a cellular or molecular environment that favors both proliferation and survival of adult-generated neurons in the rodent hippocampus. Recent studies have provided initial evidence that endogenous neurogenesis may be active in the hippocampus of AD patients. Therefore, cognitive training might contribute to promote neurogenesis in AD, thus, delaying the neurodegenerative progress and improving the cognitive function. Finally, there is reasonable evidence that PLA2 enzymes (cPLA2, iPLA2, and sPLA2-III and X) are implicated in homeostatic processes related to neurite outgrowth and differentiation, in both neurodevelopmental processes and response to neuronal injury. Collectively, the data suggest that stimulation of PLA2 by either cognitive training or positive modulators might contribute to the growth and survival of newly generated neurons in the brain of AD patients.

A recent study conducted by Rozzini and colleagues (2007) demonstrated that subjects with MCI treated with neuropsychological training in combination with acetylcholinesterase inhibitors showed higher improvements in memory function than subjects treated only with acetylcholinesterase inhibitors. However, a combination of cholinergic and glutamatergic dysfunction appears to underlie the symptomatology of AD; thus, treatment strategies should address impairments in both systems. Because reduced PLA2 activity appears to have a crucial participation in both cholinergic and glutamatergic alterations underlying cognitive decline and neuropathology in AD (for review, see Schaeffer and Gattaz 2008), we believe that the stimulation of PLA2 may offer new strategies for the treatment of AD. Specifically, we believe that the use of positive modulators of PLA2 (especially of cPLA2 and iPLA2) or supplementation with dietary lipid compounds (e.g., AA) to reverse the decline of PLA2 products involved in memory and neuritogenesis, in combination with a cognitive training protocol to induce neuronal activity underlying memory and neurogenesis, could be a valuable therapeutic approach to improve the cognitive function and perhaps delay the neurodegenerative progress of patients with early-stage AD.

Copyright information

© Springer-Verlag 2008