5-Lipoxygenase Activating Protein Reduction Ameliorates Cognitive Deficit, Synaptic Dysfunction, and Neuropathology in a Mouse Model of Alzheimer’s Disease

ARCHIVAL REPORT

5-Lipoxygenase Activating Protein Reduction Ameliorates Cognitive Deficit, Synaptic Dysfunction, and Neuropathology in a Mouse Model of Alzheimer’s Disease Phillip F. Giannopoulos, Jin Chu, Yash B. Joshi, Margaret Sperow, Jin-Guo Li, Lynn G. Kirby, and Domenico Praticò Background: 5-lipoxygenase activating protein (FLAP) is abundantly present in the central nervous system. Although its function has been extensively interrogated in the context of peripheral inflammation, novel roles for this protein are emerging in the central nervous system. The objective of our study was to investigate the functional role that FLAP plays in a mouse model of Alzheimer’s disease (AD) with plaques and tangles (i.e., 3xTg mice). Methods: By implementing a genetic knockout of FLAP and pharmacologic inhibition with a FLAP inhibitor (MK-591), we evaluated the effect on the AD-like neuropathology, cognition, and synaptic plasticity in the 3xTg mice. Results: We show that reduction of FLAP leads to amelioration of cognition and memory along with the rescuing of synaptic dysfunction at an early age before the development of overt neuropathology. Genetic knockout and pharmacologic inhibition of FLAP also yielded an improvement in AD pathology through a reduction in Aβ via the γ-secretase pathway and a decrease in tau phosphorylation through the cdk5 pathway. Conclusions: Our studies identify a novel functional role for FLAP in regulating memory and synaptic plasticity. They establish this protein at the crossroad of multiple pathways that ultimately contribute to the development of the entire AD-like phenotype, making it a viable therapeutic target with disease-modifying capacity for the treatment of this disease. Key Words: Amyloid beta, Alzheimer’s disease, behavior, FLAP protein, tau protein, transgenic mouse models

A

lzheimer’s disease (AD) is a chronic neurodegenerative disorder and the most common cause of dementia, affecting millions; because the number of cases is on the rise worldwide, it represents an enormous socioeconomic cost for the entire society (1). Before the development of the two hallmark brain pathologic lesions, neuritic plaques and neurofibrillary tangles, the prominent clinical manifestation of AD is an impairment of memory function, which is secondary to synaptic failure characterized by loss of synaptic density and function (2–6). The 5-lipoxygenase (5LO) is an enzyme abundantly present in the central nervous system, where its activity is dependent on the presence of another protein called 5-lipoxygenase activating protein (FLAP). From a biochemical point of view, they form a unique functional complex whose integrity is necessary for the full enzymatic activity, which then catalyzes the oxidation of arachidonic acid to produce bioactive lipid molecules such as leukotrienes (7). Although FLAP and 5LO function have been extensively interrogated in the context of peripheral inflammation, novel roles

From the Center for Translational Medicine (PFG, JC, YBJ, J-GL, DP), Department of Pharmacology (PFG, JC, YBJ, J-GL, DP), Center for Substance Abuse Research (MS, LGK), and Department of Anatomy and Cell Biology (LGK), Temple University School of Medicine, Philadelphia, Pennsylvania. Authors PFG and JC contributed equally to this work. Address correspondence to Domenico Praticò, M.D., 947 Medical Education and Research Building, 3500 North Broad Street, Philadelphia, PA 19140; E-mail: [email protected]. Received Feb 7, 2013; revised Mar 25, 2013; accepted Apr 11, 2013.

0006-3223/$36.00 http://dx.doi.org/10.1016/j.biopsych.2013.04.009

for these proteins are emerging in the central nervous system. Among them, new evidence indicates that 5LO may be involved in neurodegeneration because its genetic deficiency or overexpression in mouse models of AD-like amyloidosis significantly modulates their phenotype (8–10). However, because 5LO can also use other molecules aside from arachidonic acid as substrate without the presence of FLAP, it remains to be investigated whether the biological effects of 5LO genetic absence are due to the elimination of leukotriene metabolites or the disruption of other functions of 5LO independent from FLAP. To address this biological question, we used the triple transgenic mouse model of AD, 3xTg, which manifests the most salient features of the disease including synaptic dysfunction, memory impairments, amyloid beta, and tau pathology (11) and implemented a genetic and a pharmacologic approach by generating mice genetically deficient for FLAP (3xTgFLAPKO), and administering 3xTg mice with MK-591, a selective FLAP inhibitor (12). Compared with controls, we found that even before the development of overt neuropathology, 3xTg-FLAPKO and 3xTg mice receiving the inhibitor displayed a significant improvement in cognition and memory, which was secondary to a rescue of their synaptic dysfunctions and amendment of synaptic integrity. In addition, later in life the same mice showed a significant reduction of both Aβ and tau neuropathologies.

Methods and Materials All animal procedures were approved by the Institutional Animal Care and Usage Committee, in accordance with the U.S. National Institutes of Health guidelines. The 3xTg mice and mice genetically deficient for FLAPKO used in the study were reported previously (11,13). All the animals were backcrossed 10 times on the same genetic background C57BL6/SJL. The FLAPKO mice were crossbred several times with 3xTg mice to obtain founder BIOL PSYCHIATRY 2013;]:]]]–]]] & 2013 Society of Biological Psychiatry

2 BIOL PSYCHIATRY 2013;]:]]]–]]] animals (3xTg/FLAPKO), which were then crossed with each other and the animals from these crosses used for the studies. They were kept in a pathogen-free environment on a 12-hour light– dark cycle and fed a normal chow and water ad libitum. Male and female mice underwent behavioral testing at two age groups (6–8 and 12–14 months). Starting at 4 to 5 months of age, a separate group of 5-month-old male and female 3xTg mice were also randomized to receive MK-591 (generous gift of Merck Frost, Kirkland, Quebec, Canada; 40 mg/kg weight) in their chow diet; while some were sacrificed at 6 months of age for electrophysiology studies, others were kept on the same diet until they were 13 to 14 months old. Considering that each mouse eats approximately 5 g per day of the chow diet formulated for 320 mg MK-591 per kg diet (Harlan Teklad, Madison, Wisconsin), the final dose of the drug intake is approximately 40/mg/kg weight per day. Food consumption and body weight were measured during the study, and no difference was noted. Likewise, by the end of the study, no changes in feeding behavior and no differences in body weight between the two groups were recorded. At sacrifice, mice were perfused with ice-cold .9% phosphate-buffered saline containing ethylenediamine tetraacetate (2 mmol/L, pH 7.4), and brains were harvested. Behavioral Tests Y-Maze. The Y-maze apparatus consisted of three arms 32 cm (long)
10 cm (wide) with 26-cm walls (San Diego Instruments, San Diego, California). Each mouse was placed in the center of the Y-maze and allowed to explore freely through the maze over a 5-minute session to assess spontaneous alternating behavior. The sequence and total number of arms entered were video recorded. Any entry into an arm was considered legitimate if all four paws entered the arm. An alternation was defined as three consecutive entries in three different arms (1, 2, 3 or 2, 3, 1, etc.). The percentage alternation score was calculated using the following formula: (total alternation number/total number of entries −2)
100. Testing was always performed in the same room and at the same time to ensure environmental consistency as previously described (8). Fear Conditioning. Two weeks before sacrifice, fear-conditioning experiments were performed following methods as described previously (8,14,15). On the first day of testing, animals were placed into the conditioning chamber for 2 minutes before the onset of a sound, lasting 2 seconds and paired with a foot shock. Mice were removed from the chamber 60 seconds after the shock and on Day 2 were tested for contextual and cued fear conditioning, which is essentially a 24-hour memory retention test. Conditioning is measured by recording the freezing behavior as the complete absence of movement both during training as well as testing. The percentage time during which the mouse froze is calculated for analysis of contextual and cued fear memory assessments. Tests were conducted in a conditioning chamber equipped with black methacrylate walls, a transparent front door, a speaker, and grid floor (Start Fear System, Harvard Apparatus, Holliston, Massachusetts). Immunoblot Analyses Primary antibodies used in this article are summarized in Table 1. Proteins were extracted, sonicated, and centrifuged at 13,000 rpm for 45 minutes at 41C, and supernatants were used for immunoblot analysis, as previously described (8,14). Total protein concentration was determined by using BCA Protein Assay Kit (Pierce, Rockford, Illinois). Samples were electrophoretically separated using 10% Bis-Tris Gels or 3% to 8% Tris-Acetate Gel (Bio-Rad, Richmond, California), according to the molecular weight of the target molecule, and then transferred onto www.sobp.org/journal

P.F. Giannopoulos et al. nitrocellulose membranes (Bio-Rad). They were blocked with Odyssey blocking buffer for 1 hour and then incubated with primary antibodies overnight at 41C. After three washing cycles with Tris-buffered saline, membranes were incubated with IRDye 800CW or IRDye 680CW-labeled secondary antibodies (LI-COR Bioscience, Lincoln, Nebraska) at 221C for 1 hour. Signals were developed with Odyssey Infrared Imaging Systems (LI-COR Bioscience). Actin was always used as an internal loading control. Sarkosyl Insolubility Assay The assay for insoluble tau was performed as described previously (16). Briefly, ultracentrifugation and sarkosyl extraction (30 min in 1% sarkosyl) was used to obtain soluble and insoluble fractions of tau. Insoluble fractions were washed one time with 1% sarkosyl, then immunoblotted with total human tau (HT-7) antibody (Table 1). Biochemical Analyses Mouse brain homogenates were sequentially extracted initially in radioimmunoprecipitation assay for the Aβ soluble fractions, then in formic acid for the insoluble fractions; they were then assayed by a sensitive sandwich enzyme-linked immunosorbent assay kit (WAKO Chemicals, Richmond, Virginia), as previously described (9,10). Immunohistochemistry Primary antibodies used in this study are listed in Table 1. Immunostaining was performed as reported previously (9,10,14). First, serial 6-μm-thick coronal sections were mounted on 3-aminopropyl triethoxysilane–coated slides. Every eighth section from the habenular to the posterior commissure (8–10 sections per animal) was examined using unbiased stereologic principles. The sections for testing Aβ were deparaffinized, hydrated, and pretreated with formic acid (88%) and subsequently with 3% H2O2 in methanol. The sections for synaptophysin (SYP), postsynaptic density protein 95 (PSD95), and microtubule associated protein-2 (MAP-2) were deparaffinized, hydrated, subsequently pretreated with 3% H2O2 in methanol and then treated with citrate (10 mmol/L) or IHC-Tek Epitope Retrieval Solution (IHC World, Woodstock, Maryland) for antigen retrieval. Sections were blocked in 2% fetal bovine serum and then incubated with primary antibody overnight at 41C. The following day, sections were incubated with biotinylated antimouse immunoglobulin G (Vector Laboratories, Burlingame, California) and then developed by using the avidin-biotin complex method (Vector Laboratories) with 3,3′-diaminobenzidine as a chromogen. Light microscopic images were used to calculate the area occupied by Aβ immunoreactivity, and the cell densities of glial fibrillary protein and cluster of differentiation 45-immunopositive reactions by using the software Image-Pro Plus for Windows version 5.0 (Media Cybernetics, Bethesda, Maryland). The threshold optical density that discriminated staining from background was determined and held constant for all quantifications. The area occupied by Aβ immunoreactivity was measured by the software and divided by the total area of interest to obtain the percentage area of immunoreactivity. Electrophysiology Six-month-old mice (n ¼ number of slices/number of slices animals): wildtype (n ¼ 23/8); 3xTg (n ¼ 21/7); 3xTg-FLAPKO (n ¼ 21/6); 3xTg plus MK-591 (n ¼ 10/3) were sacrificed by rapid decapitation and brains were put into ice-cold artificial cerebral spinal fluid (ACSF) in which sucrose (248 mmol/L) was substituted for NaCl. Transverse hippocampal slices (400 μm thick) were cut using a Vibratome 3000 plus (Vibratome, Bannockburn, Illinois) and placed in ACSF (124 mM NaCl, 2.5 mmol/L KCl, 2 mmol/L

P.F. Giannopoulos et al.

BIOL PSYCHIATRY 2013;]:]]]–]]] 3

Table 1. Antibodies Used in This Study Antibody 4G8 APP PS-1 Nicastrin APH-1 Pen-2 CTFs

HT-7 AT-270 PHF-13 PHF-1 GFAP CD45 SYP (H-8) PSD95 MAP2 GSK3α/β p-GSK3α/β JNK SAPK/JNK P-SAPK/JNK Cdk5 P35/25 PP2A Actin

Immunogen

Host

Application

Source

aa 18–22 of human beta amyloid (VFFAE) aa 66–81 of APP {N-terminus} aa around valine 293 of human presenilin 1 aa carboxy-terminus of human nicastrin Synthetic peptide from hAPH-1a aa N-terminal of human and mouse Pen-2 a synthetic peptide [(C)KMQQNGYENPTYKFFEQMQN] corresponding to amino acids 751–770 of human precursor protein (APP), conjugated to KLH aa 159-163 of human tau Peptide containing phospho-T181 Peptide containing phospho-Ser396 Peptide containing phospho-Ser396/S404 aa spinal chord homogenate of bovine origin Mouse thymus or spleen aa 221-313 of SYP of human origin Purified recombinant rat PSD-95 Bovine brain microtubule protein aa 1-420 full length GSK-3β of Xenopus origin aa around Ser21 of human GSK-3a. aa of human JNK aa of recombinant human JNK2 fusion protein aa Thr183/Tyr185 of human SAPK/JNK aa C-terminus of Cdk5 of human origin aa C-terminus of p35 /25 of human origin aa 295-309 of catalytic subunit of human protein phosphatase 2A. Clone 1D6. aa C-terminus of Actin of human origin

Mouse Mouse Rabbit Rabbit Rabbit Rabbit Rabbit

IHC WB WB WB WB WB WB

Covance Millipore Cell Signaling Cell Signaling Millipore Invitrogen Santa Cruz

Mouse Mouse Mouse Mouse Mouse Rat Mouse Mouse Rabbit Mouse Rabbit Rabbit Rabbit Mouse Rabbit Rabbit Mouse

WB WB WB WB WB WB WB,IHC WB,IHC WB,IHC WB WB WB WB WB WB WB WB

Pierce Pierce Cell Signaling Dr. P. Davies Santa Cruz BD Pharmingen Santa Cruz Thermo Scientific Millipore Millipore Cell Signaling Cell Signaling Cell Signaling Cell Signaling Santa Cruz Santa Cruz Millipore

Goat

WB

Santa Cruz

APH-1, anterior pharynx defective-1; APP, amyloid precursor protein; CD45, custer domain 45; CDK, cylcin dependent kinase; CTF, carboxy terminal fragments; GFAP, glial fibrillary acidic protein; GSK, glycogen synthase kinase; hAPH, human anterior pharynx defective; HT-7, total human tau; IHC, immunohistochemistry; KLH, lysine, leucine, histidine; MAP-2, microtubule associated protein-2; PHF, paired helical filaments; PP, protein phosphatase; PSD95, postsynaptic density protein 95; SAPK/JNK, stress activated protein kinase/junamino terminal kinase; SYP, synaptophysin; WB, Western blot.

NaH2PO4, mmol/L CaCl2, 2 mmol/L MgSO4, 10 mmol/L dextrose, and 26 mmol/L NaHCO3) at room temperature to recover for 1 hour bubbled with 95% O2/5% CO2. Slices were transferred to a recording chamber (Warner Instruments, Hamden, Connecticut) and continuously perfused with ACSF at 1.5 to 2.0 mL/min flow, bubbled with 95% O2/5% CO2, and maintained by an inline solution heater (TC-324; Warner Instruments) at 321 to 341C. We recorded field excitatory postsynaptic potentials (fEPSPs) from the CA1 stratum radiatum by using an extracellular glass pipette (3–5 MΩ) filled with ACSF. Schaffer collateral/commissural fibers in the stratum radiatum were stimulated with a bipolar tungsten electrode placed 200 to 300 μm from the recording pipette. Stimulation intensities were chosen to produce a fEPSP that was one-third of the maximum amplitude, based on an input–output curve using stimulations of 0 to 300 μA, in increments of 20 mA. Paired-pulse facilitation experiments were performed using a pair of stimuli of the same intensity delivered 20, 50, 100, 200, and 1000 milliseconds apart. Baseline was recorded for 20 minutes before tetanization with pulses every 30 seconds. Long-term potentiation (LTP) at CA3 to CA1 synapses was induced by four trains of 100-Hz stimulation delivered in 20-second intervals. Recordings were made every 30 seconds for 3 hours following tetanization. The fEPSP rise/slope (mV/msec) between 30% and 90% was measured offline using Clampfit 10.3 (Molecular Devices, Sunnyvale, California) and normalized to the mean rise/slope of the baseline. Slices were eliminated if an unstable baseline was produced or if the normalized rise/slope dropped more than 20 to 50 mV/msec in an approximately 10-minute period. All the

tests were always performed by an experimenter who was unaware of the different genotypes and treatment. Data Analysis One-way analysis of variance, unpaired Student t test (twosided) and Bonferroni multiple comparison tests were performed using Prism 5.0 (GraphPad Software, La Jolla, California). All data are presented as mean ⫾ standard error of the mean. Significance was set at p ⬍ .05.

Results Genetic Absence and Pharmacologic Blockade of FLAP Ameliorates Cognition To evaluate the behavioral effect of FLAP genetic absence and pharmacologic inhibition with MK-591, which by binding specifically to FLAP completely prevents its activity (12), mice were initially tested in the Y-maze at two ages: 6 to 8 and 12 to 14 months old. Initially, no differences were noted among the control and treated groups when considered in regard to their general activity as assessed by the total number of arm entries at both ages (Figure 1A). However, when we counted the number of alternations in the same test, we observed that 3xTg mice had a much lower number of alternations resulting in a significant lower percentage in comparison with the wildtype and FLAPKO mice. In contrast, compared with the 3xTg mice, the 3xTg-FLAPKO and 3xTg receiving MK-591 had a greater number of alternations resulting in a significantly higher percentage in both age groups, www.sobp.org/journal

4 BIOL PSYCHIATRY 2013;]:]]]–]]]

P.F. Giannopoulos et al. Figure 1. Genetic absence and pharmacologic inhibition of 5-lipoxygenase activating protein (FLAP) ameliorates behavioral deficits of triple transgenic (3xTg) mice. (A) Total number of arm entries for wild-type (WT), genetically deficient for FLAP (FLAPKO), 3xTg, 3xTgFLAPKO, and 3xTg-MK-591 mice at 6 to 8 and 12 to 14 months of age. (B) Percentage of alternations between the same groups of mice (*p ⬍ .001). (C) Contextual fear memory response in WT, FLAPKO, 3xTg, 3xTg-FLAP, and 3xTg-MK-591 mice. (D) Cued recall fear memory response in the same groups of mice (6–8 months: n ¼ 8 for WT, n ¼ 7 for FLAPKO, n ¼ 12 for 3xTg, n ¼ 12 for 3xTg-FLAPKO, n ¼ 7 for 3xTg-MK-591; 12–14 months: n ¼ 7 for WT, n ¼ 7 for FLAPKO, n ¼ 11 for 3xTg, n ¼ 12 for 3xTg-FLAPKO, n ¼ 10 for 3xTg-MK-591; *p ⬍ .001). Values represent mean ⫾ SEM.

Figure 2. 5-lipoxygenase activating protein (FLAP) reduction modulates Aβ peptide levels of deposition in brains of triple transgenic (3xTg) mice. (A) Radioimmunoprecipitation assay (RIPA)-soluble extractable Aβ1-40 levels in cortex of 3xTg, 3xTg genetically deficient for FLAP (FLAPKO), and 3xTgMK-591 mice at 14 months of age (n ¼ 9 for 3xTg and n ¼ 11 for 3xTg-FLAPKO; *p ⬍ .001). (B) Formic acid (FA) extractable Aβ1-40 levels in cortex of 3xTg, 3xTg-FLAPKO, and 3xTg-MK-591 mice at 14 months of age (n ¼ 9 for 3xTg and n ¼ 11 for 3xTg-FLAPKO; *p ⬍ .001). (C) RIPA-soluble extractable Aβ1-42 levels in cortex of 3xTg, 3xTg-FLAPKO, and 3xTg-MK-591 mice at 14 months of age (n ¼ 9 for 3xTg and n ¼ 11 for 3xTg-FLAPKO; *p ⬍ .001) (D) FA-extractable Aβ1-42 levels in cortex of 3xTg, 3xTg-FLAPKO, and 3xTg-MK-591 mice at 14 months of age (n ¼ 9 for 3xTg and n ¼ 11 for 3xTg-FLAPKO; *p ⬍ .001). (E) Representative brain sections from 14-month-old 3xTg and 3xTg/FLAPKO mice immunostained with 4G8 antibody. (F) Quantification of the area occupied with Aβ immunoreactivity in the brains of the same group of mice (*p ⬍ .001). (G) Representative brain sections from 14-month-old 3xTg and 3xTg MK-591 mice immunostained with 4G8 antibody. (H) Quantification of the area occupied with Aβ immunoreactivity in the brains of the same group of mice (*p ⬍ .001). (I) Representative Western blots of amyloid precursor protein (APP), PS-1, nicastrin, APH-1, Pen-2, carboxyl terminal fragments (CTF)-β, and CTF-α in brain homogenates from 14-month-old 3xTg and 3xTg-FLAPKO mice. (J) Densitometric analyses of the immunoreactivities to the antibodies shown in panel F (*p ⬍ .01; n ¼ 4 for 3xTg; n ¼ 4 for 3xTg-FLAPKO). (K) Representative western blots of APP, PS-1, nicastrin, APH-1, Pen-2, CTF-β, and CTF-α in brain homogenates from 14-month-old 3xTg and 3xTg-MK-591 mice. (L) Densitometric analyses of the immunoreactivities to the antibodies shown in H (*p ⬍ .01) (n ¼ 4 for 3xTg; n ¼ 4 for 3xTg-MK-591). Values represent mean ⫾ SEM.

www.sobp.org/journal

P.F. Giannopoulos et al.

BIOL PSYCHIATRY 2013;]:]]]–]]] 5

Figure 3. 5-lipoxygenase activating protein (FLAP) modulates tau phosphorylation and metabolism in the brains of triple transgenic (3xTg) mice. (A) Representative Western blot analyses for total tau (HT-7) and phosphorylated tau at residues T181(AT270), S396(PHF-13), and S396/S404(PHF-1) in homogenates of brain cortices from 3xTg mice and 3xTg mice genetically deficient for FLAP (3xTg-FLAPKO) at 14 months of age. (B) Densitometric analyses of the immunoreactivities to the antibodies shown in panel A (*p ⬍ .01). (C) Representative Western blot analyses for total tau (HT-7) and phosphorylated tau at residues T181(AT270), S396(PHF-13), and S396/S404(PHF-1) in homogenates of brain cortices from 3xTg mice and 3xTg mice treated with MK-591. (D) Densitometric analyses of the immunoreactivities to the antibodies shown in panel A (*p ⬍ .001). (E) Representative Western blot analyses for insoluble tau in brain homogenates from cortices of 3xTg and 3xTg-FLAPKO mice at 14 months of age. (F) Densitometric analyses of the immunoreactivity to the antibody shown in panel E (*p ⬍ .001). (G) Representative Western blot analyses for insoluble tau in brain homogenates from cortices of 3xTg and 3xTg mice treated with MK-591. (H) Densitometric analyses of the immunoreactivity to the antibody shown in G (*p ⬍ .001).

suggesting an improvement in their working memory (Figure 1B). Next, mice underwent fear-conditioning testing. No differences were observed during the training session among the different groups (not shown), and a similar result was observed when they were assessed in the contextual fear conditioning recall paradigm (Figure 1C). In contrast, in the cued phase of the fear-conditioning testing, we observed that FLAPKO and wildtype mice exhibited similar levels of freezing at both ages, but 3xTg mice had significantly lower freezing percentages that were normalized in the 3xTg-FLAPKO and 3xTg mice treated with MK-591 (Figure 1D). Unavailability of FLAP Reduces Brain Aβ Level and Deposition A week after completion of the behavior tests (age 14 months), mice were sacrificed, and brains were harvested and assayed for levels and deposition of Aβ. As shown in Figure 2A–D, we observed that 3xTg-FLAPKO and 3xTg mice receiving MK-591 displayed a significant decrease in the amount of radioimmunoprecipitation assay–soluble and formic-acid soluble Aβ 1-40 and 1-42. Confirming the enzyme-linked immunosorbent assay data, we found that the same conditions led to a significant decrease in Aβ deposition in their brains (Figure 2E–H). To investigate the mechanism responsible for this change, we assessed the steadystate levels of amyloid precursor protein along with its cleavage products. As shown in Figure 2I through 2L, no differences were noticed for total amyloid precursor protein. However, compared with controls, the 3xTg-FLAPKO and 3xTg treated with MK-591 had a significant decrease in the steady-state levels of the four components of the γ-secretase complex, PS-1, nicastrin, Pen-2, and anterior pharynx defective-1, but not significant changes in carboxyl terminal fragments-α and carboxyl terminal fragments-β

were observed for both groups (Figure 2I–L). No differences were observed in the α-secretase and β-secretase pathways between the two groups (not shown). Also no changes were detected in the steady-state levels of neprilysin and insulin-degrading enzyme, which are involved in Aβ degradation, nor in apolipoprotein E, an Aβ chaperon, that could justify the decrease in Aβ (not shown). Effect of FLAP Modulation on Tau Phosphorylation Next we explored the effect of FLAP knockout and pharmacologic blockade on tau metabolism. As shown in Figure 3, although we did not observe any changes in the levels of total soluble tau among the groups, compared with the 3xTg, the 3xTg-FLAPKO and 3xTg mice treated with MK-591 had a significant decrease in its phosphorylated forms at epitopes S396, as recognized by the antibody paired helical filaments (PHF)-13, and epitopes S396/S404, as recognized by the antibody PHF-1 (ratios: PHF-13/ tau ¼ .52; PHF-1/tau ¼ .62; Figure 3A–D). In contrast, no changes were detected for the phosphorylation site as recognized by the antibody AT270 (T181). Additionally, compared with 3xTg we observed that 3xTg-FLAPKO and 3xTg receiving MK-591 had a significant reduction in the levels of insoluble tau (Figure 3E–H). To investigate the molecular mechanism responsible for the hypophosphorylation of tau, we next assayed some of the kinases that are considered major regulators of posttranslational modification of this protein. First, we did not observe any significant differences among the different groups in the steady-state levels of total and phosphorylated forms of glycogen synthase kinase 3-α and glycogen synthase kinase 3-β, stress activated protein kinase/junamino terminal kinase and total and www.sobp.org/journal

6 BIOL PSYCHIATRY 2013;]:]]]–]]]

P.F. Giannopoulos et al.

Figure 4. Tau hypophosphorylation in the genetic absence and pharmacologic inhibition of 5-lipoxygenase activating protein (FLAP) is due to changes in specific kinases. (A) Representative Western blot analyses for Cdk5, p35, p25, glycogen synthase kinase (GSK)3α, and GSK3β protein levels in brain homogenates from triple transgenic (3xTg) mice and 3xTg mice genetically deficient for FLAP (3xTg-FLAPKO) mice at 14 months of age. (B) Densitometric analyses of the immunoreactivities to the antibodies from the previous panel (3xTg, n ¼ 4; 3xTg-FLAPKO, n ¼ 4; *p ⬍ .001). (C) Representative Western blot analyses for Cdk5, p35, p25, GSk3α, GSK3β, p-GSK3 α, p-GSK3 β, stress activated protein kinase/junamino terminal kinase (SAPK/JNK)1, SAPK/JNK2, p-SA PK/ JNK1, p-SAPK/JNK2, and phosphatase protein-2 (PP2)A protein levels in brain homogenates from 3xTg and 3xTg-MK-591 mice at 14 months of age. (D) Densitometric analyses of the immunoreactivities from the previous panel (3xTg, n ¼ 4; 3xTg-MK-591, n ¼ 4; *p ⬍ .001). Values represent mean ⫾ SEM.

phosphorylated stress activated protein kinase/junamino terminal kinase (Figure 4A–D). However, we found that compared with controls, 3xTg-FLAPKO and 3xTg receiving MK-591 had a statistically significant decrease in levels of Cdk5 kinase and its coactivators p35 and p25 (Figure 4A–D). No changes in the levels of phosphatase-A2 were detected among the different groups of mice (Figure 4A–D). Genetic Absence and Pharmacologic Blockade of FLAP Ameliorates Synaptic Integrity Because changes in tau phosphorylation state and solubility have been correlated with modifications of synaptic integrity in AD, we evaluated this aspect of the 3xTg mice phenotype. Compared with the wildtype group, 3xTg mice displayed a significant decrease in the steady state levels of two main synaptic proteins, PSD-95 and SYP, whereas the levels of the proteins were restored to wildtype levels in both the 3xTg/FLAPKO mice and the 3xTg mice treated with MK-591 (Figure 5A–E). A similar result was obtained also when the dendritic protein MAP2 was assayed www.sobp.org/journal

(Figure 5A–E). These results were further confirmed in brain sections of the same mice when they were assessed by immunohistochemical analyses (Figure 5C–F). Finally, we observed that, compared with brain homogenates from 3xTg, both 3xTg-FLAPKO and MK-591-treated mice had a significant decrease in glial fibrillary protein and cluster of differentiation 45 immunoreactivities, markers of astrocytes and microglia cells activation, respectively (Figure 5G–J). Involvement of FLAP in Synaptic Dysfunction Because the genetic absence and pharmacologic inhibition of FLAP in 3xTg mice yielded significant improvements in memory at an early age (6 months) before plaque and tangle pathology, we explored its effect on synaptic function at this young age. Initially, we investigated basal synaptic transmission by generating input/output (I/O) curves and measuring fEPSPs elicited in the CA1 region of the hippocampus by stimulation of the Schaffer collaterals at increasing strength of stimulations and intensities. As shown in Figure 6A, there were no differences observed in the

P.F. Giannopoulos et al.

BIOL PSYCHIATRY 2013;]:]]]–]]] 7

Figure 5. Genetic absence and pharmacologic inhibition of 5-lipoxygenase activating protein (FLAP) ameliorates synaptic biomarkers and decreases neuroinflammation in triple transgenic (3xTg) mice. (A) Representative Western blot analyses for synaptophysin (SYP), postsynaptic density protein 95 (PSD95), and microtubule associated protein-2 (MAP-2) in brain homogenates from cortices of wildtype and 3xTg mice and mice genetically deficient for FLAP (3xTg-FLAPKO). (B) Densitometric analyses of the immunoreactivities from panel A (*p ⬍ .001). (C) Representative immunohistochemical stainings in hippocampus for SYP-, PSD95-, and MAP-2-positive areas in brain sections of 3xTg and 3xTg-FLAPKO mice. (D) Representative Western blot analyses for SYP, PSD95, and MAP-2 in brain homogenates from cortices of wildtype, 3xTg, and 3xTg-MK-591 mice. (E) Densitometric analyses of the immunoreactivities from panel D (*p ⬍ .001). (F) Representative immunohistochemical stainings in hippocampus for SYP-, PSD95-, and MAP-2-positive areas in brain sections of 3xTg and 3xTg-MK-591 mice. (G) Representative Western blot analyses for glial fibrillary acidic protein (GFAP) and custer domain (CD)45 in brain homogenates from 3xTg and 3xTg-FLAPKO mice. (H) Densitometric analyses of the immunoreactivities to the antibodies from the previous panel (3xTg, n ¼ 4; 3xTg-FLAPKO, n ¼ 4; *p ⬍ .001). (I) Representative Western blot analyses for GFAP and CD45 in brain homogenates from 3xTg and 3xTg-MK-591 mice. (J) Densitometric analyses of the immunoreactivities to the antibodies from the previous panel (3xTg, n ¼ 4; 3xTg-MK-591, n ¼ 4; *p ⬍ .01). Values represent mean ⫾ SEM.

I/O curves among the groups considered (wildtype, 3xTg, 3xTgFLAPKO, 3xTg MK-591). We then measured short-term plasticity by examining paired-pulse facilitation, which is due to an activitydependent presynaptic modulation of transmitter release (17). Similar to what was observed in the I/O curves, no differences were noted in paired-pulse facilitation among any of the groups analyzed (Figure 6B). Finally, we investigated LTP in the CA1 region of the hippocampus, which is thought to measure neuronal plasticity along with being a key player in memory and cognition (18). In this test, we found that, compared with WT, 3xTg mice had a significant reduction in LTP responses. However, the genetic absence and pharmacologic inhibition of FLAP in the 3xTg completely restored the LTP responses to a level comparable to that of the WT mice (Figure 6C–E).

Discussion Our study shows that genetic absence or pharmacologic inhibition of FLAP can prevent important functional impairments such as memory deficits and synaptic dysfunction in a transgenic

mouse model of AD with plaques and tangles together with an amelioration of their neuropathology. Emerging evidence supports a role for the 5LO pathway in AD pathogenesis (9,19); however, because this enzyme can have biological activities unrelated to its ability of oxidizing fatty acids, it is still unclear whether the elimination of these derivatives (i.e., leukotrienes) or some other unrelated 5LO function is responsible for this effect. This fact is especially true because most of the data obtained so far have used animal genetically deficient for 5LO, which cannot rule out both scenarios. For this reason, we focused our attention on the protein FLAP, the only known function of which is to transfer arachidonic acid selectively to 5LO and enhance its sequential oxidative metabolism (7). Our results support the first hypothesis and, together with the previous knowledge on this pathway, establish it as a key player in the development of the entire AD phenotype and a valid target with disease-modifying therapeutic potential for AD treatment. Over the years, there has been increasing evidence suggesting that changes in synaptic integrity and function are the earliest phenotypic manifestations in the pathogenesis of AD (2–4). Thus, www.sobp.org/journal

8 BIOL PSYCHIATRY 2013;]:]]]–]]]

P.F. Giannopoulos et al.

Figure 6. Genetic absence or pharmacologic inhibition of 5-lipoxygenase activating protein (FLAP) rescues synaptic dysfunction in triple transgenic (3xTg) mice. (A) Input/output (I/O) curves and representative field excitatory postsynaptic potentials (fEPSPs) at increasing stimulus strengths (0–300 mA) are shown for wildtype and 3xTg mice, 3xTg mice genetically deficient for FLAP (3xTg-FLAPKO), and 3xTg mice treated with MK-591 at 6 months of age. (B) Mean fEPSP slopes as a function of interpulse interval between the first and second fEPSPs evoked at CA3–CA1 synapses in slices from the same mice at 20, 50, 100, 200, and 1000 milliseconds in the same animals. (C) fEPSP slopes were recorded for 3 hours and expressed as the percentage of pretetanus baseline in the same mice. (D) Long-term potentiation (LTP) magnitudes expressed as the percentages of baseline for 0 to 10 minutes posttetanus (274.6% ⫾ 8.5% for wildtype, n ¼ 23 slices; 159.9% ⫾ 13.8%, n ¼ 19 slices for 3xTg; 269.7% ⫾ 10.3%, n ¼ 21 slices for 3xTg-FLAPKO; 272.5% ⫾ 13.2%, n ¼ 10 slices for 3xTg-MK-591). (E) For the same groups of mice, LTP magnitudes expressed as the percentages of baseline for 170 to 180 minutes posttetanus (202.6% ⫾ 6.5% for wildtype; 121.0% ⫾ 3.4% for 3xTg; 197.1% ⫾ 11.5% for 3xTg-FLAPKO; 156.9% ⫾ 6.9%; 202.5% ⫾ 9.4% for 3xTg-MK-591; *p ⬍ .0001). Values represent mean ⫾ SEM.

synaptic loss, as reflected by alterations in specific synaptic markers, is a characteristic of early stage AD pathology that associates better with clinical cognitive impairments rather than typical evaluation of the hallmark brain lesions (20,21). However, the mechanisms underlying this phenomenon are not fully elucidated. The recent availability of better mouse models of the disease recapitulating more faithfully the diverse aspects of the AD phenotype, including synaptic dysfunction and pathology, represent an invaluable tool in this direction. Notably, the 3xTg mice exhibit age-dependent deficits in synaptic plasticity, such as LTP, learning, and memory impairments along with the classical Aβ deposition and tau pathology (11). In our study, we first showed that memory impairments as well as synaptic dysfunction are early-age (6 months) events that take place before the accumulation of overt pathologic brain lesions. Next, we demonstrated that genetic absence of FLAP per se did not influence any of the memory tests completed in our set of behavioral experiments and that, compared with their controls, the 3xTg-FLAPKO and 3xTg MK-591 mice did not display any difference in general motor activities. However, these two groups of mice showed a significant improvement in their memory performance as early as 6 months of age, as demonstrated in the Y-maze paradigm, which assesses working memory in the mice through the recording of spontaneous alternation behavior (8). Additionally, the same groups of mice showed improvements in their learning memory ability, as evaluated by the fear conditioning test paradigm. Consequently, in this setting, we observed that 3xTg mice performed www.sobp.org/journal

significantly worse than the two groups of mice genetically deficient and pharmacologically inhibited for FLAP in the cued recall, but not in the contextual paradigm, indicating a possible role of the amygdala. Data from transgenic mouse models of the disease as well as human AD samples have convincingly demonstrated that the biochemical substrate for cognitive impairments and memory deficits are changes in synaptic density as reflected by changes in presynaptic markers, such as synaptophysin, and postsynaptic markers (20–23). In support of this knowledge, in our study we observed that compared with 3xTg, the 3xTg-FLAPKO and 3xTg MK-591 mice had a significant increase in three distinct protein markers of synaptic integrity (i.e., synaptophysin, PSD95, and MAP2). This finding provides a biochemical and molecular mechanism for the behavioral improvements in the 3xTg mice secondary to the genetic absence and inhibition of FLAP. In association to these changes, we found a significant reduction in the amyloidotic phenotype, which was secondary to an effect on the γ-secretase pathway (8). Interestingly, we did not observe any difference between wildtype and FLAPKO mice in the levels of this secretase, suggesting that this effect is transgene-dependent (PFG, DP; personal communication; March 2013). In addition to the effect on Aβ pathology, we also observed an influence on tau metabolism. Accordingly, we observed that both FLAP genetic absence and pharmacologic inhibition resulted in a significant decrease in tau phosphorylation at specific epitopes and that this biological

P.F. Giannopoulos et al. effect was mediated by the cdk-5 kinase pathway but not other kinases or phosphatases (8). To further support the involvement of the FLAP in the improvement of the behavioral and memory deficits, we next explored the effects that both conditions had directly on synaptic function by performing electrophysiologic studies. First, we observed that among the different groups of mice, there were no differences in basal synaptic transmission, indicating that this function is not altered in any of the implemented experimental conditions. Next, no significant differences were noted also when we measured short-term plasticity by examining paired-pulse facilitation, which is secondary to an activity-dependent presynaptic modulation of transmitter release (17). This observation suggests that the transient increase in the concentration of intraterminal calcium produced by an invading action potential is similar among all groups. Taken together, these data tell us that there is no change in the probability of transmitter release in any of these groups of mice (11,24). Lastly, we assessed the LTP response, which is a type of plasticity that is thought to have a major role in learning and memory functions (25). As reported previously (11), there was a significant difference in LTP responses between the wildtype and 3xTg mice, with the latter showing deficits. However, the genetic absence of FLAP in the 3xTg mice was sufficient to restore their LTP responses to a level comparable to the ones measured in the wildtype mice. The biological importance of this finding was corroborated by the demonstration that FLAP pharmacologic blockade was also sufficient to rescue the abnormal synaptic phenotype of the 3xTg mice. Altogether, these results unravel a new aspect of the neurobiology of FLAP in the context of AD pathogenesis by demonstrating its indispensable role for the 5LO-dependent biological effects on synaptic function and plasticity, along with learning and memory. Bearing in mind the additional anti-Aβ and anti-tau effect secondary to FLAP modulation, our findings have critical translational implications for AD because they establish this pathway as a major player in the development of the full spectrum of the disease’s phenotype. This new information, by demonstrating the pleiotropic role of this protein in AD pathogenesis, makes it not only a valid pharmacologic target but most important represents a unique therapeutic opportunity with a true disease-modifying potential for the treatment of AD. This work was supported by U.S. National Institutes of Health Grant Nos. AG033568 and NS071096 to DP; T32DA07237, which supports MS; and P30DA13429 to the Center for Substance Abuse Research. Additional funding was provided by a grant from the Alzheimer Association to DP (Grant No. NPSPAD-10-170775). The authors report no biomedical financial interests or potential conflicts of interest. 1. Citron M (2010): Alzheimer’s disease: Strategies for disease modification. Nat Rev Drug Discov 9:387–398. 2. DeKosky ST, Scheff SW (1990): Synapse loss in frontal cortex biopsies in Alzheimer’s disease: Correlation with cognitive severity. Ann Neurol 27:457–464. 3. Scheff SW, Scott SA, DeKosky ST (1991): Quantitation of synaptic density in the septal nuclei of young and aged Fischer 344 rats. Neurobiol Aging 12:3–12.

BIOL PSYCHIATRY 2013;]:]]]–]]] 9 4. Selkoe DJ (2002): Alzheimer’s disease is a synaptic failure. Science 298: 789–791. 5. Palop JJ, Chin J, Mucke L (2006): A network dysfunction perspective on neurodegenerative diseases. Nature 443:768–773. 6. Scheff SW, Price DA, Schmitt FA, Dekosky ST, Mufson EJ (2007): Synaptic alterations in CA1 in mild Alzheimer disease and mild cognitive impairment. Neurology 68:1501–1508. 7. Radmark O, Werz O, Steinhilber D, Samuelsson B (2007): 5-Lipoxygenase: Regulation of expression and enzyme activity. Trends Biochem Sci 32:332–341. 8. Chu J, Praticò D (2012): Involvement of 5-lipoxygenase activating protein in the amyloidotic phenotype of an Alzheimer’s disease mouse model. J Neuroinflammation 9:2094–2099. 9. Chu J, Praticò D (2011): Pharmacological blockade of 5-lipoxygenase improves the amyloidotic phenotype of an Alzheimer’s disease transgenic mouse model. Am J Pathol 178:1762–1769. 10. Firuzi O, Zhuo J, Chinnici CM, Wisniewski T, Praticò D (2008): 5-Lipoxygenase gene disruption reduces amyloid-β pathology in a mouse model of Alzheimer’s disease. FASEB J 22:1169–1178. 11. Oddo S, Caccamo A, Shepherd JD, Murphy MP, Golde TE, Kayed R, et al. (2003): Triple transgenic model of Alzheimer’s disease with plaques and tangles: Intracellular Abeta and synaptic dysfunction. Neuron 39:409–421. 12. Diamant X, Timmersa MC, van der Venna H, Friedman BS, De Smet M, Depre M, et al. (1995): The effect of MK-0591, a novel 5-lipoxygenase activating protein inhibitor, on leukotriene biosynthesis and allergeninduced airway responses in asthmatic subjects in vivo. J Aller Clin Immunol 95:42–51. 13. Byrum RS, Goulet JL, Griffiths RJ, Koller BH (1997): Role of the 5lipoxygenase-activating protein (FLAP) in murine acute inflammatory responses. J Exp Med 185:1065–1075. 14. Yang H, Zhuo J, Chu J, Chinnici C, Pratico D (2010): Amelioration of the Alzheimer’s disease phenotype by absence of 12/15-lipoxygenase. Biol Psychiatry 68:922–929. 15. Zhuo JM, Portugal GS, Kruger WD, Wang H, Gould TJ (2010): Dietinduced hyperhomocysteinemia increases amyloid-beta formation and deposition in a mouse model of Alzheimer’s disease. Curr Alzheimer Res 7:140–149. 16. Andorfer C, Kress Y, Espinoza M, De Silva R, Tucker KL, Barde YA, et al. (2003): Hyperphosphorylation and aggregation of tau in mice expressing normal human tau isoforms. J Neurochem 86:582–590. 17. Zucker RS, Regehr WG (2002): Short-term synaptic plasticity. Annu Rev Physiol 64:355–405. 18. Bliss TVP, Collingridge GL (1993): A synaptic model of memory: Longterm potentiation in the hippocampus. Nature 361:31–39. 19. Chu J, Giannopoulos PF, Ceballos-Diaz C, Golde TE, Pratico D (2012): 5-Lipoxygenase gene transfer worsens memory, amyloid, and tau brain pathologies in a mouse model of Alzheimer disease. Ann Neurol 72:1762–1769. 20. Dickson DW, Crystal HA, Bevona C, Honer W, Vincent I, Davies P (1995): Correlations of synaptic and pathological markers with cognition of the elderly. Neurobiol Aging 16:285–304. 21. Masliah E, Mallory M, Alford M, DeTeresa R, Hansen R, Mckeel LA, et al. (2001): Altered expression of synaptic proteins occurs early during progression of Alzheimer’s disease. Neurology 56:127–129. 22. Sze CI, Troncoso JC, Kawas C, Mouton P, Price DL, Martin LJ (1997): Loss of the presynaptic vesicle protein synaptophysin in hippocampus correlates with cognitive decline in Alzheimer disease. J Neuropathol Exp Neurol 56:933–944. 23. Terry RD, Masliah E, Salmon E, Butters DP, DeTeresa N, Hill R, et al. (1991): Physical basis of cognitive alterations in Alzheimer’s disease synapse loss is the major correlate of cognitive impairment. Ann Neurol 30:572–580. 24. Hessler NA, Shirke AM, Malinow R (1993): The probability of transmitter release at a mammalian central synapse. Nature 366:569–572. 25. Kandel ER, Schwartz JH (1982): Molecular biology of learning: Modulation of transmitter release. Science 218:433–443.

www.sobp.org/journal