Reboxetine Treatment Reduces Neuroinflammation and Neurodegeneration in the 5xFAD Mouse Model of Alzheimer’s Disease: Role of CCL2 Irene L. Gutiérrez1,2 & Marta González-Prieto1,2 & Javier R. Caso1,2 & Borja García-Bueno1,2 & Juan C. Leza1,2 & José L. M. Madrigal1,2 Received: 31 January 2019 /Accepted: 3 July 2019 # Springer Science+Business Media, LLC, part of Springer Nature 2019 Abstract The reduction of brain noradrenaline levels is associated to the initiation of Alzheimer’s disease and contributes to its progression. This seems to be due mainly to the anti-neuroinflammatory actions of noradrenaline. The analysis of noradrenaline effects on brain cells demonstrates that it also regulates the production of the chemokine CCL2. In the present study, we analyzed the effect of the selective noradrenaline reuptake inhibitor, reboxetine, on the inflammatory and neurodegenerative alterations present in 5xFAD mice, and how the genetic removal of CCL2 affects reboxetine actions. We observed that the removal of CCL2 reduced the memory impairments in 5xFADmice as well as the neuroinflammatory response, the accumulation of amyloid beta plaques, and the degeneration of neurons in the brain cortex. The administration of reboxetine with osmotic pumps for 28 days also resulted in anti-inflammatory and neuroprotective changes in 5xFAD mice, even in the absence of CCL2. Yet, 6-month-old CCL2KO mice presented a significant degree of neuroinflammation and neuronal damage. These findings indicate that reboxetine treatment prevents the brain alterations caused by prolonged overproduction of amyloid beta, being these effects independent of CCL2, which is a mediator of the damage caused by amyloid beta in the brain cortex, but necessary for the prevention of the development of neurodegeneration in normal healthy conditions. Keywords Noradrenaline . Reboxetine . CCL2 .MCP-1 . 5xFAD . Neuroinflammation Introduction The loss of noradrenergic neurons and the subsequent reduc- tion of brain noradrenaline (NA) levels are two alterations known to occur before other processes characteristic of Alzheimer’s disease (AD) such as the accumulation of amy- loid β plaques or hyperphosphorylated tau [1]. In addition, there seems to be a cause-effect relationship between nor- adrenaline depletion and the progression of neuroinflamma- tion and neuronal damage. This has been demonstrated by removing locus coeruleus noradrenergic neurons with the neurotoxin DSP4 [2], by genetically depleting noradrenergic neurons [3], or by the administration of NA precursors [4]. In parallel, in vitro studies demonstrate that the protective actions Irene L. Gutiérrez and Marta González-Prieto contributed equally to this work. * José L. M. Madrigal jlmmadrigal@med.ucm.es Irene L. Gutiérrez irenlo02@ucm.es Marta González-Prieto martagp@ucm.es Javier R. Caso jrcaso@med.ucm.es Borja García-Bueno bgbueno@med.ucm.es Juan C. Leza jcleza@med.ucm.es 1 Department of Pharmacology and Toxicology, School of Medicine, Universidad Complutense de Madrid (UCM), Av. Complutense s/n, 28040 Madrid, Spain 2 Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Instituto de Investigación Neuroquímica (IUINQ-UCM) and Instituto de Investigación Sanitaria Hospital 12 de Octubre (Imas12), Madrid, Spain https://doi.org/10.1007/s12035-019-01695-6 Molecular Neurobiology (2019) 56:8628–8642 /Published online: 11 July 2019 http://crossmark.crossref.org/dialog/?doi=10.1007/s12035-019-01695-6&domain=pdf http://orcid.org/0000-0002-1409-9561 mailto:jlmmadrigal@med.ucm.es of NA are due, at least in part, to its ability to reduce the activation of cells such as microglia [5] and astrocytes [6] that indirectly damage neurons. This inhibition of neuroinflamma- tion seems to be the result of NA regulation of glial behavior, and therefore, it depends on the alteration of the intracellular functions of these cells. Based on this, we decided to study the mechanisms through which NA reduces the neuronal damage associated to AD. This way, we found that NA induces the expression and synthesis of the chemokine CCL2 (MCP-1), in astrocytes [7, 8]. CCL2 is best known for its ability to attract monocytes and other immune cells to the sites of injury [9]. And therefore, it is considered a mediator of inflammation. Consequently, the blockade of its actions either by genetic alterations or by other means has proven to reduce the neuronal damage produced by different causes [10]. However, as for many other bioactive agents, the complete suppression of CCL2 activity may have deleterious consequences. This is due to its relevant role as a mediator of a critical process such as the immune response and also to its involvement in many other processes through- out the body [11]. In fact, studies from our group and others demonstrate that the production of CCL2 by astrocytes pro- tects neurons against injuries such as excitotoxicity [7], HIV- tat [12], methylmercury [13], or endotoxemia [14] among others. Therefore, according to the existing information, the maintenance of CCL2 concentrations seems to be necessary to mount an appropriate inflammatory response, but an excess of this chemokine could be as detrimental as a deficiency of it. Interestingly, NA regulation of CCL2 production by astro- cytes has proven to be affected by the activation state of these cells. This way, the massive release of CCL2 resulting from astroglial exposure to a pro-inflammatory stimulus is prevented by NA, being this inhibitory effect of NA opposite to the one observed in resting conditions [15]. Therefore, in this study, we have analyzed if NA exerts its neuroprotective effects also in CCL2KO mice in order to as- certain the potential involvement of CCL2 in NA actions. For this, we used the antidepressant and NA reuptake inhibitor reboxetine to elevate brain concentrations of NA. In addition, given the dual protective/deleterious roles we found for CCL2 in the CNS, we considered relevant to further explore the impact of long-term suppression of CCL2 activity in the brain cortex. For this purpose, we have analyzed the degree of neuroinflammation and neurodegeneration in aged CCL2KO mice. Material and Methods Mouse Models All experimental protocols adhered to the guidelines of the Animal Welfare Committee of the Universidad Complutense of Madrid, Spain (PROEX 052/17), and according to European Union laws (2010/63/EU). Wild-type (WT) C57BL/6, 5xFAD (strain B6.Cg- Tg(APPSwFlLon, PSEN1*M146L*L286V)6799Vas/ Mmjax), and CCL2KO (B6.129S4-Ccl2tm1Rol/J) mice were obtained from The Jackson Laboratory. These mice were maintained for over 10 generations on a C57Bl6 background. 5xFAD mice express 3 mutant forms of human APP (the Swedish mutation: K670N, M671L; the Florida mutation: I716V; the London mutation: V717I) under the control of the murine Thy-1 promoter, and the murine presenilin-1 (PSEN1) with the M233T and L235P familial AD (FAD)– linked mutations expressed under the control of the mouse PSEN1 promoter. Hemizygous 5xFAD mice were crossed with homozygous CCL2KO mice. The resultant 5xFAD+/−/CCL2+/− mice were backcrossed with CCL2KO mice to generate 5xFAD+/ −/CCL2KO mice. For these studies, 5-month-old male WT, heterozygous 5xFAD, homozygous CCL2KO, and 5xFAD/CCL2KO (het- erozygous 5xFAD and homozygous CCL2KO) were used. Five or more mice were included in each experimental group. Mice were housed up to 4 per cage in a controlled 12:12 h light/dark cycle, with food provided ad libitum. All efforts were made to minimize animal suffering and to reduce the number of animals used. Reboxetine Treatment Alzet osmotic minipumps (model 2004, delivering 0.25 μl/h for at least 28 days) were loaded with saline serum (vehicle) or with filter-sterilized reboxetine mesylate (50 mg/ml in saline serum) and kept submerged in isoosmolar saline solution at 37 °C for approximately 2 h to initiate a steady flow delivering an approximate dose of 10 mg/kg of body weight each day. At 5 months of age, all the mice were anesthetized by inhalation of isoflurane (induced in a chamber at 4% and maintained at 2% using a nose cone), a small inci- sion was made behind the neck, the pumps were im- planted subcutaneously between the scapulae, and the incision was closed with sutures. Twenty-nine days after the implantation of the pumps, the animals were killed by terminal anesthesia using sodium pentobarbital (250 mg/kg i.p. Vetoquinol®, Madrid, Spain) and sub- jected to transcardial perfusion with saline. The brains were removed, one hemibrain was dissected, and corti- cal areas were excised from the brain, snap-frozen, and kept at − 80 °C. The other half was post-fixed in 4% paraformaldehyde overnight and cryoprotected in 15% sucrose during the following 24 h. After this, regularly spaced series of 25-μm-thick coronal sections were col- lected in cryoprotectant solution and stored at − 20 °C in 24-well plates until processing. Mol Neurobiol (2019) 56:8628–8642 8629 Y Maze One day before the implantation of the pumps and again 29 days later, mice were tested in a Y maze. The test was carried out to obtain results for spontaneous alternation per- formance which is an index of memory and spatial learning. The maze is composed of three equally spaced arms (120°, 35 cm long, 5 cm wide, and 10 cm high) with gray, non- reflective base plate and walls. Each mouse was placed in the centerof themazeandallowedtomovefreelyfor8min.Themaze was cleaned with 75% alcohol between each trial to prevent ol- factory clues. An arm entrywas scoredwhen 85%ormore of the mouse body was in the arm. All test trials were video-recorded, tracked, and analyzed with ANY-maze™ tracking software. Percentage of alternation was determined as follows: number of triads containing entries into all three arms/maximum possible alternations (total number of arms entered/3) × 100. mRNA Analysis Total cytoplasmic RNA from tissues was prepared using TRIzol reagent (Thermo Fisher Scientific), aliquots converted to cDNA using random hexamer primers and SuperScript® Reverse transcriptase (Thermo Fisher Scientific), and mRNA levels estimated by quantitative real-time PCR (QPCR). PCR conditions were 35 cycles at 95 °C for 10 s, annealing at 60 °C for 15 s, and extension at 72 °C for 30 s followed by 5 min at 72 °C in a Corbett Rotorgene. Reactions were carried out in the presence of Sybr Green (Biotools). Relative mRNA levels were calculated by comparison of take-off cycles and normal- ized to values for GAPDHmeasured in the same samples. The primers used (all listed 5′ to 3′) were: IL-1β f: TGGAGAGTGTGGATCCCAAGCAAT IL-1β r: TGCTTGTGAGGTGCTGATGTACCA MIP1a f: ACTGACCTGGAACTGAATGCCTGA MIP1a r: ATGTGGCTACTTGGCAGCAAACAG GFAP f: AAGGTCTATTCCTGGCTGCACAGT GFAP r: AGCTTGGAGAGCAACAGCTAGTCA COX2 f: ATGAGTGGTAGCCAGCAAAGCCTA COX2 r: TACTGAGTACCAGGCCAGCACAAA mPGES-1 f: CCTAGGCTTCAGCCTCACAC mPGES-1 r: CAGCCTAATGTTCAGCGACA GAPDH f: TGCACCACCAACTGCTTAGC GAPDH r: GGCATGGACTGTGGTCATGAG CCL2 f: AGCAGGTGTCCCAAAGAAGCTGTA CCL2 r: AAAGGTGCTGAAGACCTTAGGGCA Western Blot Samples were homogenized by sonication in 500 μl of PBS mixed with a protease inhibitor cocktail (Complete, Roche Farma, Madrid, Spain) followed by a centrifuga- tion at 12.000g for 15 min at 4 °C. After adjusting protein levels in the resultant supernatants, homogenates mixed with Laemmli sample buffer (Bio-Rad, Hercules, CA, USA) and 20 μl (1 mg/ml) were loaded and the proteins size separated in 10% SDS-polyacrylamide gel electrophoresis (90 V). The proteins were then trans- ferred to polyvinyl difluoride membranes, which were blocked with 5% milk in Tris-buffered saline containing 0.1% Tween-20 for 1 h and incubated overnight at 4 °C with primary antibodies against GFAP (ab4678, Abcam), COX2 (SC1747, Santa Cruz), amyloid β (Aβ) (ab2539, Abcam), SMI32 (NE1023, Merck), annexin V (8555, Cell Signaling), or β-actin (A5441, Sigma). This was followed by incubation with anti-IgG-horseradish peroxidase-labeled secondary antibodies for 1 h at room temperature and subsequent detection with an enhanced chemiluminescence detection kit (ECL, Amersham Life Science). Blots were imaged using an Odyssey® Fc System (Li-COR Biosciences). Several exposition times were analyzed to ensure the linearity of the band inten- sities. The resulting bands were quantified by densitom- etry (ImageJ). All densitometries are expressed in arbi- trary units (AU). Immunohistochemistry Brain sections were washed with PBS for 5 min and blocked with 10% normal serum and 0.2% Triton X- 100 in PBS at 25 °C for 1 h. Next, they were incubated with primary antibodies (Table 1) diluted in 1% normal serum and 0.2% Triton X-100 in PBS at 4 °C for 18 h. After this, the sections were washed three times with PBS and incubated for 1 h at 25 °C with secondary antibodies diluted 1:500 in PBS with 1% normal serum. Then, they were washed three times for 5 min with PBS and post-fixed in 3.7% paraformaldehyde in PBS for 20 min. Autofluorescence was quenched with 50 mM NH4Cl in PBS for 15 min. Finally, after washing the sections with PBS for 3 min, they were mounted onto the slides from PBS and then coverslipped with Fluoroshield™ (Sigma). Images were obtained on a Nikon Eclipse Ti-S microscope equipped with a Digital Sight digital camera and NIS- Elements imaging software and with a Leica SP8 confocal microscope equipped with a Leica DFC350FX digital camera. Fluorescence intensity was quantified using NIH ImageJ soft- ware. At least five sections were included in each experimen- tal group. The frontal cortex (from + 2 to − 1 mm to the breg- ma) was analyzed, and images of the primary somatosensory area were obtained. The detection threshold was set manually and the same value was applied for the quantification of all groups. Mol Neurobiol (2019) 56:8628–86428630 Fluorescent Labeling of Aβ Plaques Detection of Aβ plaques with Fluoro-Jade C was per- formed according to the method described [16]. Briefly, the brain sections, prepared as described above, were transferred to separate wells containing 500 μl of PBS in a 24-well plate. After 5 min, the sections were trans- ferred to different wells containing PBS. This process was repeated one more time for a total of three PBS washes. After this, the sections were incubated in a 0.001% solution of Fluoro-Jade C (AG325, Merck) made in PBS for 10 min. The sections were then washed three times in PBS (5 min each), and autofluo- rescence was quenched with 50 mM NH4Cl in PBS for 15 min. Finally, after three additional PBS washes (3 min each), they were mounted onto slides from PBS and then coverslipped with Fluoroshield™ (Sigma). Images were obtained on a Nikon Eclipse Ti- S microscope equipped with a Digital Sight digital cam- era and NIS-Elements imaging software. Statistical Analysis All experiments were undertaken at least in triplicate. Data were analyzed by one-way analysis of variance (ANOVA), followed by Newman-Keuls multiple comparison tests. For those experiments that included all four genotypes and two treatments, two-way ANOVA analyses followed by Tukey’s post hoc comparisons were performed. p values < 0.05 were considered significant. Table 1 Primary antibodies used in immunohistochemistry studies Antigen Dilution Manufacturer Amyloid β 1:200 Abcam (ab2539) Annexin V 1:1000 Cell Signaling (8555) GFAP 1:1000 Abcam (ab4674) NeuN 1:400 Merck (ABN78A4) Anti-neurofilament H non-phosphorylated (SMI32) 1:1000 Merck (NE1023) CX3CR1 1:100 R&D Systems (AF5825) Fig. 1 Memory deficits in 5xFAD are reduced by CCL2 depletion. a Spatial working memory of WT, 5xFAD, 5xFAD/CCL2KO, and CCL2KO mice was assessed by spontaneous alternation in a Y maze. b Total number of arm entries performed by the mice during the test. Data are means ± SE of n ≥ 18 replicates per group. **p < 0.01 vs. WT. φp < 0.05 vs. 5xFAD. c Spatial working memory was also assessed at the end of the treatment inWT, 5xFAD, 5xFAD/CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx). d Total number of arm entries performed by the mice during the test performed at the end of the treatment. Data are means ± SE of n ≥ 9 replicates per group Mol Neurobiol (2019) 56:8628–8642 8631 Results CCL2 Removal Reduces Memory Loss in 5xFAD Mice The evaluation of spatial working memory with the Y maze task allowed us to observe a significant reduction of about 20% in spontaneous alternation in the 5xFAD mice compared with the WT ones (Fig. 1a). This is in agreement with the memory impairments characteristic of this mouse strain [17]. The total number of arm entries recorded during the 8 min the mice were in the maze did not differ significantly between the four groups (Fig. 1b) indicating that all the mice had a similar explorative behavior. Interestingly, the 5xFAD mice with the suppressed expression of CCL2 (5xFAD/CCL2KO) achieved higher percentages of spontaneous alternation than the 5xFAD mice expressing CCL2. In fact, there were no significant dif- ferences of spontaneous alternation between 5xFAD/ CCL2KO and WT mice. However, the analysis of the same factors performed at the end of the pharmacological treatment did not allow us to de- tect significant differences (Fig. 1c, d). Therefore, we cannot conclude that the reboxetine treatment applied affects the Fig. 2 Regulation of IL1β and MIP1α expression by reboxetine administration and CCL2 depletion. mRNA levels of IL1β (a), MIP1α (b), and CCL2 (c) were analyzed in brain cortex samples from WT, 5xFAD, 5xFAD/CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx). Data are means ± SE of n ≥ 6 replicates per group. ***p < 0.001, **p < 0.01, *p < 0.05 vs. WT. φφφp < 0.001, φφp < 0.01, φp < 0.05 vs. 5xFAD. δδδp < 0.001 vs. CCL2KO. θp < 0.05 vs. 5xFAD/ CCL2KO Fig. 3 Distribution of microglia in the frontal cortex. a Representative images of the frontal cortex (primary somatosensory area) from coronal sections prepared from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx) stained for CX3CR1. Scale bars correspond to 100 μm. b Quantitative analysis of CX3CR1-positive cells. Data are means ± SE of n ≥ 5 replicates per group Mol Neurobiol (2019) 56:8628–86428632 memory impairments present in 5xFAD mice. However, after the treatment, 5xFAD/CCL2KO and CCL2KO mice per- formed more arm entries than mice from other groups, although these differences were not statistically significant; this suggests that the lack of CCL2 may stimulate a more explorative behavior. Mol Neurobiol (2019) 56:8628–8642 8633 Effects of Reboxetine Treatment and CCL2 Removal on the Expression of Pro-inflammatory Genes The mRNA levels of different factors known to mediate the inflammatory response were quantified in cortex samples ob- tained from WT, 5xFAD, CCL2KO, and 5xFAD/CCL2KO mice after the conclusion of their reboxetine or vehicle treat- ments. The largest differences were found for IL1β and MIP1α. The analysis of these results allowed us to determine the effect of the treatment on the neuroinflammatory status of the mice, as well as the alterations caused in this regard by the suppression of CCL2. The larger mRNA concentrations found for both cytokines in 5xFAD mice confirm the existence of a neuroinflammatory status in these mice that was largely re- duced by the reboxetine treatment (Fig. 2a, b). The deletion of CCL2 from 5xFAD mice also reduced the expression of IL1β and MIP1α. This indicates that CCL2 contributes to the maintenance of the cortical inflammatory response characteristic of 5xFAD mice. Interestingly, reboxetine reduction of MIP1α mRNA was also observed in 5xFAD/CCL2KO mice (Fig. 2b). Therefore, we can conclude that this effect of reboxetine is independent, at least in part, of CCL2 activity. Additionally, CCL2KO mice have a substantial elevation in the expression of IL1β (almost 30%) that was prevented by the reboxetine treatment (Fig. 2a). The analysis of CCL2 mRNA differences in WT and 5xFAD samples demonstrates that reboxetine treatment in- duces the expression of CCL2 in WT mice (Fig. 2c). This is in agreement with our previous study in which another inhib- itor of NA reuptake such as reboxetine was administered to WT mice [8]. Distribution of Microglia Since microglia is one of the main cell types responsible for the propagation of the neuroinflammatory response, we analyzed if the genetic alterations and the reboxetine treatment here used modify the morphology and distribution of microg- lia in the frontal cortex. For this purpose, we used antibodies directed against the CX3CL1 receptor, CX3CR1, which is exclusively expressed by microglia within the CNS and used as a marker for these cells [18]. As shown in Fig. 3, 5xFAD and 5xFAD/CCL2KO mice present areas with clusters of CX3CR1-positive cells. Reboxetine treatment seems to reduce the density of microglia in 5xFAD mice even in the absence of CCL2, but no signifi- cant differences could be detected through the quantification of CX3CR1-positive cells. Alterations of GFAP Expression and Synthesis Having found the effects of reboxetine and CCL2 suppression on the expression of neuroinflammation mediators, we decid- ed to confirm if these alterations also affect to the activation of astrocytes since this fact constitutes a characteristic feature of AD and contributes to the pathogenesis of this disease [19]. This way, we observed that the removal of CCL2 results in a large reduction of GFAP expression (Fig. 4a) and synthesis (Fig. 4b, c) in 5xFAD mice. Interestingly, and in agreement with the differences ob- served for IL1β, CCL2KO mice show a significant elevation in the production of GFAP compared with WTones (Fig. 4a– c). The presence of GFAP in the primary somatosensory area was also analyzed by immunohistochemistry. This way, we confirmed that CCL2 removal reduces the synthesis of GFAP in 5xFAD mice (Fig. 4d, e). Also, while no significant differ- ences could be detected by PCR or Western blot, immunohis- tochemistry allowed us to detect a significant reduction in the GFAP staining in the samples from reboxetine-treated 5xFAD mice. This technique did not allow us to detect significant differences for CCL2KO mice, although a different pattern of staining from that found for WT mice was observed. The greater increases found for CCL2KO mice by PCR and Western blot could be due to the inclusion of a larger cortical area in the sample evaluated, while the histochemical analyses were only focused on the primary somatosensory area. COX2 Expression and Synthesis Given the relevant role of COX2 in the progression of AD [20], we have also analyzed the effect of the reboxetine treat- ment on the expression of COX2 and PGE2 and if this is affected by the suppression of CCL2. This way, we observed that, in WT animals, reboxetine elevates the expression of COX2 (Fig. 5a, b). The comparison of the mRNA levels between 5xFAD/ CCL2KO and 5xFAD samples indicates that the expression of COX2 and microsomal prostaglandin E synthase-1 �Fig 4 Reboxetine treatment and CCL2 removal reduce GFAP expression and synthesis. a GFAP mRNA concentrations were analyzed in brain cortex samples from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KO mice treated with vehicle or reboxetine. Data are means ± SE of n ≥ 5 replicates per group. ***p < 0.001, *p < 0.05 vs. Wv. φφφp < 0.001 vs. 5xFAD. b Protein levels of GFAP and β-actin were analyzed by Western blot in brain cortex samples from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KOmice treated with vehicle or reboxetine (Rbx). The gels shown are representative of three separate experiments. c Densitometric analysis of the bands. AU: arbitrary units relative to Wv. ***p < 0.001, **p < 0.01 vs. WT. d Representative images of the frontal cortex (primary somatosensory area) from coronal sections prepared from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KOmice treated with vehicle or reboxetine (Rbx) and stained for GFAP (green). Scale bar corresponds to 100 μm. e Quantitative analysis of GFAP-positive cells per field. Data are means ± SE of n ≥ 5 replicates per group. ***p < 0.001 vs. WT. φφφp < 0.001 vs. 5xFAD Mol Neurobiol (2019) 56:8628–86428634 (mPGES1), inducible enzyme that catalyzes the synthesis of PGE2, is significantly reduced by the removal of CCL2 (Fig. 5a, b) in 5xFAD mice. The analysis of COX2 protein accumulation by Western blot does not allow to notice differences as large as those observed by PCR. However, the effects of reboxetine and of CCL2 suppression on COX2 seem to be confirmed also at the protein synthesis level (Fig. 5c, d). Interestingly, while no elevation was detected for COX2 mRNA in CCL2KO mice, the concentration of COX2 protein seems to be slightly in- creased in them. Accumulation of Aβ Oligomers and Plaques NA has been shown to play a relevant role in the accumu- lation of Aβ characteristic of AD [21], and CCL2 also seems to modulate the formation of Aβ plaques and the phagocytosis of these aggregates by glial cells [22]. Based Fig. 5 Regulation of COX2 expression and synthesis. a GFAP and b mPGES1 mRNA concentrations were analyzed in brain cortex samples from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx). Data are means ± SE of n ≥ 5 replicates per group. ***p < 0.001, *p < 0.05 vs. WT. φφp < 0.01, φp < 0.05 vs. 5xFAD. c Protein levels of COX2 and β-actin were analyzed by Western blot in brain cortex samples from WT, 5xFAD, 5xFAD+ CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx). The gels shown are representative of three separate experiments. d Densitometric analysis of the bands. AU: arbitrary units relative to WT Mol Neurobiol (2019) 56:8628–8642 8635 on this, we have also analyzed if the treatment with reboxetine, in addition to the anti-inflammatory effects ob- served, can modify the accumulation of Aβ in 5xFAD mice. This way, immunohistochemistry of Aβ (Fig. 6a) demon- strates that the mice treated with reboxetine had a lower density of plaques (Fig. 6b). Moreover, 5xFAD/CCL2KO Fig. 6 Regulation of amyloid β (Aβ) accumulation. Representative im- ages of the frontal cortex (primary somatosensory area) from coronal sections prepared from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx) stained for aAβ (red) and c Fluoro-Jade C (green). Scale bars correspond to 100 μm. b, d Number of Aβ plaques per field and percentage of the total area occupied by Aβ plaques. Data are means ± SE of n ≥ 5 replicates per group. ***p < 0.001 vs. WT. φφφp < 0.001 vs. 5xFAD. e Protein levels of Aβ and β-actin were analyzed byWestern blot in brain cortex samples fromWT, 5xFAD, 5xFAD+CCL2KO, and CCL2KOmice treated with vehicle or reboxetine (Rbx). The gels shown are representative of three separate experiments. f Densitometric analysis of the bands. AU: arbitrary units relative to WT. ***p < 0.001, **p < 0.01 vs. WT. φφφp < 0.001 vs. 5xFAD. g Representative image of the whole Western blot membrane Mol Neurobiol (2019) 56:8628–86428636 mice presented an even lower number of plaques, but reboxetine treatment did not seem to affect the formation of plaques in these mice. An alternative technique, based on the use of Fluoro-Jade C [16], was used to also label Aβ plaques. This allowed us to detect more precisely the plaques and to use a lower Mol Neurobiol (2019) 56:8628–8642 8637 microscope magnification. The images obtained and the counting of plaques confirmed the results obtained by regular immunohistochemistry (Fig. 6c, d). Additionally, in order to compare the concentration of Aβ in tissue homogenates, Western blots were performed. The labeling with a specific antibody directed against residues 1– 14 of Aβ generated two bands of approximately 50 and 60 kDa for all samples, being the lighter one more intense but of a similar intensity for all samples (Fig. 6g). Instead, clear differences could be observed for the 60-kDa bands. The relevance of Aβ aggregates of this size in AD has been previously described [23]. The comparison of these bands reveals that reboxetine treatment also reduces the accumula- tion of this form of Aβ aggregates (Fig. 6e, f). And contrarily to the effect observed on Aβ plaques, the deletion of CCL2 did not seem to lower significantly the concentration of these oligomers. Furthermore, reboxetine treatment caused a large reduction of this form of Aβ in 5xFAD/CCL2KO mice while it did not have a detectable effect in the number of plaques. Axonal Damage Different factors could be responsible for the neuronal degra- dation that leads to the cognitive deficit associated to AD, but the accumulation of Aβ and the sustained activation of glial cells are considered to be among the main ones [24]. Therefore, having confirmed that the occurrence of these two factors can be prevented by reboxetine treatment and CCL2 ablation, we analyzed next if this translates into a mod- ification of the neuronal damage present in 5xFAD mice. First, we performed immunohistochemical analyses to compare the concentration of non-phosphorylated neurofila- ment H (SMI32) in the brain cortex because this provides an indirect index of axonal damage [25–27]. The intense staining of neuronal axons in the brain cortex from 5xFAD mice con- firms that these mice bear a substantial neuronal degeneration. Also, in agreement with the aforementioned observations, the administration of reboxetine prevented this neuronal alter- ation. A similar protection was provided by the genetic removal of CCL2. However, the images obtained from 5xFAD/CCL2KO mice with or without reboxetine treatment do not present detectable differences (Fig. 7a, b). The detection of anti-neurofilament H non-phosphorylated (SMI32) labeling in brain cortex homogenates by Western blot allowed us to compare more accurately the degree of axonal damage and confirmed the existence of the differences observed by immunohistochemistry (Fig. 7c, d). Both techniques also demonstrate a considerable increase in the signal provided by the CCL2KO samples. This indi- cates that the mere suppression of the CCL2 gene leads to a sizeable neuronal damage. Analysis of Neuronal Cell Death In order to elucidate if the alterations observed translate into a loss of cells, we analyzed the levels of the cell death marker annexin V. The comparison of the image stacks acquired through confocal microscopy as well as Western blots per- formed for annexin V indicates that reboxetine treatment and CCL2 removal provide a substantial protection against the development of cell death in the brain cortex of 5xFAD mice (Fig. 8b). The immunohistochemistry data indicates that reboxetine administration to 5xFAD/CCL2KOmice increases the presence of annexin V in the tissue analyzed. However, this is not confirmed by Western blot. Also, according to the results obtained with SMI32, a sig- nificant increase in the concentration of annexin V was detect- ed for CCL2KOmice. This parallelism suggested that some of the apoptotic cells detected could be neurons. Therefore, dou- ble labeling was performed in 5xFAD and CCL2KO samples for annexin V and a specific neuronal marker such as NeuN. The colocalization detected confirmed that neuronal cells were included among those undergoing apoptosis (Fig. 9). Discussion The results obtained in this study demonstrate the potential of reboxetine as an alternative strategy to prevent the develop- ment of some alterations characteristic of AD. Previous works demonstrate that the administration of different inhibitors of NA transporters such as atomoxetine or NA precursors such as L-DOPS reduces the inflammatory changes, the accumula- tion of Aβ, and the neuronal damage displayed by different models of AD [4, 28]. In addition, beneficial biochemical and behavioral results have been obtained in similar models using different strategies to elevate brain noradrenaline levels, in- cluding the administration of an α2-adrenergic antagonist [27] or increasing the concentration of tyrosine hydroxylase (the rate-limiting enzyme in NA synthesis) in the locus coeruleus [29]. Therefore, we consider that the results obtain- ed, while far from being transferable to humans, contribute to �Fig 7 Reboxetine and CCL2 removal reduce axonal damage. a Representative images of the frontal cortex (primary somatosensory area) from coronal sections prepared from WT, 5xFAD, 5xFAD+ CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx) and stained with anti-neurofilament H non-phosphorylated antibody (SMI32 green). Scale bar corresponds to 100 μm (top panel) and 25 μm (lower panel). b Quantitative analysis of SMI32-positive cells. Data are means ± SE of n ≥ 5 replicates per group. ***p < 0.001 vs. WT. φφφp < 0.001 vs. 5xFAD. c Levels of SMI32-labeled and β-actin proteins were analyzed byWestern blot in brain cortex samples fromWT, 5xFAD, 5xFAD+CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx). The gels shown are representative of three separate experiments. d Densitometric analysis of the bands. AU: arbitrary units relative to WT. *p < 0.05 vs. Wv. φp < 0.05 vs. 5xFAD Mol Neurobiol (2019) 56:8628–86428638 support the interest of performing studies on the potential use of drugs that increase NA transmission in AD and other neu- rodegenerative diseases [30]. Most interestingly, given the age of the 5xFAD mice at the onset of the pharmacological treat- ment, we conclude that reboxetine may be effective even in advanced stages of AD. However, a recent study demonstrates that reboxetine treatment elevates tau phosphorylation in an animal model of AD based on the expression of mutant (P301L) tau protein [31]. Therefore, while reboxetine safety for humans has been verified in clinical trials, additional Fig. 8 Reboxetine and CCL2 removal reduce cell death in 5xFAD brain cortex. a Representative confocal image stacks of the frontal cortex (primary somatosensory area) from coronal sections prepared from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx) and stained for annexin V (red). Scale bar corresponds to 20 μm. b Quantitative analysis of annexin V–positive cells. Data are means ± SE of n ≥ 5 replicates per group. ***p < 0.001, **p < 0.01 vs. WT. φφφp < 0.001 vs. 5xFAD. θθθp<0.001 vs. 5xFAD/CCL2KO. c Protein levels of annexin V and β-actin were analyzed by Western blot in brain cortex samples from WT, 5xFAD, 5xFAD+CCL2KO, and CCL2KO mice treated with vehicle or reboxetine (Rbx). The gels shown are representative of three separate experiments. dDensitometric analysis of the bands. AU: arbitrary units relative to Wv. **p < 0.01 vs. WT Mol Neurobiol (2019) 56:8628–8642 8639 studies would be necessary to elucidate its beneficial or detri- mental actions in AD patients. Another goal of this study was to elucidate the role of CCL2 in the anti-inflammatory actions of NA, because we had previously observed the modulation of this chemokine by NA [7, 15]. Our data indicate that, in the conditions ana- lyzed, the loss of CCL2 reduced the memory deficits, the expression of pro-inflammatory cytokines, the accumulation of Aβ plaques, and the neuronal damage caused by the muta- tions carried by 5xFAD mice. This is in agreement with some of the studies performed by one research group which dem- onstrated that overexpression of CCL2 in APP transgenic mice increases the accumulation of Aβ and memory deficits [32, 33]. Surprisingly, this group has described that CCL2 deficiency in APP/PS1 or PS1 mice also results in alterations similar to the ones they observed when CCL2 was overexpressed [34, 35]. The protection provided by CCL2 removal in other conditions in which the CNS was exposed to inflammatory injuries [36–40] is in agreement with our observations. Nevertheless, we consider that the divergences between our data and those mentioned could be due, among other factors, to the different types of APP mice used which may result in altered patterns of Aβ production. The differences in the concentration of MIP1α mRNA be- tween 5xFAD/CCL2KO mice treated with vehicle and those treated with reboxetine suggest that this effect of NA is inde- pendent of CCL2. This is particularly remarkable in the case of Aβ detection byWestern blot, for which reboxetine showed a large reduction in the presence and in the absence of CCL2, while CCL2 removal had no apparent effect. Therefore, while the concentration of NA may influence the production of CCL2 in the brain, and this may mediate NA neuroprotective and anti-inflammatory actions [7], some of these actions do not require the presence of CCL2. Nevertheless, reboxetine’s apparent lack of effect in 5xFAD/CCL2KOmice in the case of Aβ plaques, anti-neurofilament H non-phosphorylated (SMI32), or GFAP could be due to the large reduction provid- ed in all those cases by the removal of CCL2 which hinders the detection of a potential additive or synergistic effect of reboxetine. The increase in the expression of COX2 and mPGES1 caused by reboxetine in WTanimals supports previous results Fig. 9 Identification of damaged cells with a neuronal marker. Representative images of the frontal cortex (primary somatosensory area) from coronal sections prepared from 5xFAD and CCL2KO mice treated with vehicle and stained for annexin V (red), NeuN (green), and DAPI (blue). The images shown are representative of three separate experiments Mol Neurobiol (2019) 56:8628–86428640 obtained by other authors in primary microglia cultures [41] and by us in cultured astrocytes [15]. In those studies, we observed that directly adding NA to the cells induces the ex- pression and synthesis of COX2. However, we also observed a dual effect of NA according to which it induces the produc- tion of certain cytokines in basal conditions but prevents it when astrocytes are activated by LPS. This did not apply for COX2 which was induced by NA in the absence and in the presence of LPS. Instead, here we observed a similar duality of effects, since reboxetine induces COX2 and mPGES1 in WT animals but has the opposite effect in 5xFAD. Therefore, some of the effects of increased brain NA levels may depend on the presence of an inflammatory stimulus such as the for- mation of Aβ aggregates forced in 5xFAD mice. A different behavior of isolated astrocytes in comparison to in vivo con- ditions could explain the lack of a dual effect in cultured cells. This could also be due to the various sources of the COX2 detected in brain homogenates which include astrocytes but also other cell types with a different response to NA. The results presented here indicate that the removal of CCL2 in 5xFAD mice reduces memory deficits, neuronal damage, and neuroinflammation. Accordingly, analyses performed on samples obtained from human donors indi- cate that CCL2 plays a relevant role in the progression of AD [42, 43] and its inhibition protects neurons against Aβ [44]. Nevertheless, our data also indicate that the sup- pression of CCL2, in the absence of additional mutations, leads to neurodegeneration and to a neuroinflammatory status. This seems to contradict our own observations and those of others who have demonstrated the beneficial consequences of removing CCL2 in mice exposed to dif- ferent types of CNS injuries, as mentioned above [36–40]. However, contrarily to some studies in which it is con- cluded that certain mutations have no gross abnormalities after testing 5–6-week-old mice, our experiments were performed on relatively old mice (6 months), which allowed us to determine if the absence of CCL2 has some delayed effects impossible to determine at younger ages. This way, based on our observations, we propose that CCL2 constitutes a relevant target for the reduction of the damage associated to AD. But, given the pro- neurodegenerative consequences its removal has demon- strated to have, the inhibition of its activity in healthy conditions for long periods or in a non-local manner may be detrimental in the long run. In summary, this study demonstrates that reboxetine re- duces the neuroinflammatory and neurodegenerative alter- ations characteristic of Alzheimer’s disease in an animal mod- el based on the overproduction of amyloid beta. Additionally, it is also demonstrated here that, while the presence of CCL2 clearly contributes to the progression of the damage caused by the mutations present in 5xFAD mice, its absence is deleteri- ous in healthy conditions. Funding Information This work was supported by the UCM (PR26/16- 20278), the Spanish Ministry of Science (SAF2017-86620-R), and Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM). ILG was supported by a Fellowship from the Spanish Ministry of Science. MGP was supported by a Fellowship from the European Youth Employment Initiative (YEI). BGB and JRC are Ramón y Cajal fellows (Spanish Ministry of Science). Compliance with Ethical Standards All experimental protocols adhered to the guidelines of the Animal Welfare Committee of the Universidad Complutense of Madrid, Spain (PROEX 052/17), and according to European Union laws (2010/63/EU). Conflict of Interest The authors declare that they have no conflict of interest. References 1. Gannon M, Che P, Chen Y, Jiao K, Roberson ED, Wang Q (2015) Noradrenergic dysfunction in Alzheimer’s disease. Front Neurosci 9:220 2. Heneka MT, Ramanathan M, Jacobs AH, Dumitrescu-Ozimek L, Bilkei-Gorzo A, Debeir T, Sastre M, Galldiks N et al (2006) Locus ceruleus degeneration promotes Alzheimer pathogenesis in amy- loid precursor protein 23 transgenic mice. J Neurosci 26(5):1343– 1354 3. Hammerschmidt T, Kummer MP, Terwel D, Martinez A, Gorji A, Pape HC, Rommelfanger KS, Schroeder JP et al (2013) Selective loss of noradrenaline exacerbates early cognitive dysfunction and synaptic deficits in APP/PS1 mice. Biol Psychiatry 73(5):454–463 4. Kalinin S, Polak PE, Lin SX, Sakharkar AJ, Pandey SC, Feinstein DL (2012) The noradrenaline precursor L-DOPS reduces pathology in a mouse model of Alzheimer’s disease. Neurobiol Aging 33(8): 1651–1663 5. Dello RC, Boullerne AI, Gavrilyuk V, Feinstein DL (2004) Inhibition of microglial inflammatory responses by norepinephrine: effects on nitric oxide and interleukin-1beta production. J Neuroinflammation 1(1):9 6. Braun D, Madrigal JL, Feinstein DL (2014) Noradrenergic regula- tion of glial activation: molecular mechanisms and therapeutic im- plications. Curr Neuropharmacol 12(4):342–352 7. Madrigal JL, Leza JC, Polak P, Kalinin S, Feinstein DL (2009) Astrocyte-derived MCP-1 mediates neuroprotective effects of nor- adrenaline. J Neurosci 29(1):263–267 8. Madrigal JL, Garcia-Bueno B, Hinojosa AE, Polak P, Feinstein DL, Leza JC (2010) Regulation of MCP-1 production in brain by stress and noradrenaline-modulating drugs. J Neurochem 113(2):543– 551 9. Yadav A, Saini V, Arora S (2010) MCP-1: chemoattractant with a role beyond immunity: a review. Clin Chim Acta 411(21–22): 1570–1579 10. Semple BD, Kossmann T, Morganti-KossmannMC (2010) Role of chemokines in CNS health and pathology: a focus on the CCL2/ CCR2 and CXCL8/CXCR2 networks. J Cereb Blood Flow Metab 30(3):459–473 11. Deshmane SL, Kremlev S, Amini S, Sawaya BE (2009) Monocyte chemoattractant protein-1 (MCP-1): an overview. J Interf Cytokine Res 29(6):313–326 12. Eugenin EA, D’Aversa TG, Lopez L, Calderon TM, Berman JW (2003)MCP-1 (CCL2) protects human neurons and astrocytes from NMDA or HIV-tat-induced apoptosis. J Neurochem 85(5):1299– 1311 Mol Neurobiol (2019) 56:8628–8642 8641 13. Godefroy D, Gosselin RD, Yasutake A, Fujimura M, Combadiere C, Maury-Brachet R, Laclau M, Rakwal R et al (2012) The chemo- kine CCL2 protects against methylmercury neurotoxicity. Toxicol Sci 125(1):209–218 14. Zisman DA, Kunkel SL, Strieter RM, Tsai WC, Bucknell K, Wilkowski J, Standiford TJ (1997) MCP-1 protects mice in lethal endotoxemia. J Clin Invest 99(12):2832–2836 15. Hinojosa AE, Caso JR, Garcia-Bueno B, Leza JC, Madrigal JL (2013) Dual effects of noradrenaline on astroglial production of c h emok i n e s a n d p r o - i n f l amma t o r y med i a t o r s . J Neuroinflammation 10:81 16. Gutierrez IL, Gonzalez-Prieto M, Garcia-Bueno B, Caso JR, Leza JC, Madrigal JLM (2018) Alternative method to detect neuronal degeneration and amyloid beta accumulation in free-floating brain sections with Fluoro-Jade. ASN Neuro 10:1759091418784357 17. Oakley H, Cole SL, Logan S, Maus E, Shao P, Craft J, Guillozet- Bongaarts A, Ohno M et al (2006) Intraneuronal beta-amyloid ag- gregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J Neurosci 26(40):10129–10140 18. Cardona AE, Pioro EP, Sasse ME, Kostenko V, Cardona SM, Dijkstra IM, Huang D, Kidd G et al (2006) Control of microglial neurotoxicity by the fractalkine receptor. Nat Neurosci 9(7):917– 924 19. Chun H, Lee CJ (2018) Reactive astrocytes in Alzheimer’s disease: a double-edged sword. Neurosci Res 126:44–52 20. Woodling NS, Colas D, Wang Q, Minhas P, Panchal M, Liang X, Mhatre SD, Brown H et al (2016) Cyclooxygenase inhibition tar- gets neurons to prevent early behavioural decline in Alzheimer’s disease model mice. Brain 139(Pt 7):2063–2081 21. Kalinin S, Gavrilyuk V, Polak PE, Vasser R, Zhao J, Heneka MT, Feinstein DL (2007) Noradrenaline deficiency in brain increases beta-amyloid plaque burden in an animal model of Alzheimer’s disease. Neurobiol Aging 28(8):1206–1214 22. Naert G, Rivest S (2013) A deficiency in CCR2+ monocytes: the hidden side of Alzheimer’s disease. J Mol Cell Biol 5(5):284–293 23. Barghorn S, Nimmrich V, Striebinger A, Krantz C, Keller P, Janson B, Bahr M, Schmidt M et al (2005) Globular amyloid beta-peptide oligomer - a homogenous and stable neuropathological protein in Alzheimer’s disease. J Neurochem 95(3):834–847 24. Craft JM, Watterson DM, Van Eldik LJ (2006) Human amyloid beta-induced neuroinflammation is an early event in neurodegener- ation. Glia 53(5):484–490 25. Fowler JH, McCracken E, Dewar D, McCulloch J (2003) Intracerebral injection of AMPA causes axonal damage in vivo. Brain Res 991(1–2):104–112 26. Gilgun-Sherki Y, Panet H, Holdengreber V, Mosberg-Galili R, Offen D (2003) Axonal damage is reduced following glatiramer acetate treatment in C57/bl mice with chronic-induced experimen- tal autoimmune encephalomyelitis. Neurosci Res 47(2):201–207 27. Kalinin S, Polak PE, Madrigal JL, Gavrilyuk V, Sharp A, Chauhan N, Marien M, Colpaert F et al (2006) Beta-amyloid-dependent ex- pression of NOS2 in neurons: prevention by an alpha2-adrenergic antagonist. Antioxid Redox Signal 8(5–6):873–883 28. Heneka MT, Nadrigny F, Regen T, Martinez-Hernandez A, Dumitrescu-Ozimek L, Terwel D, Jardanhazi-Kurutz D, Walter J et al (2010) Locus ceruleus controls Alzheimer’s disease pathology by modulating microglial functions through norepinephrine. Proc Natl Acad Sci U S A 107(13):6058–6063 29. Braun D, Feinstein DL (2019) The locus coeruleus neuroprotective drug vindeburnol normalizes behavior in the 5xFAD transgenic mouse model ofAlzheimer's disease. Brain Res 1702:29–37 30. Chalermpalanupap T, Kinkead B, Hu WT, Kummer MP, Hammerschmidt T, Heneka MT, Weinshenker D, Levey AI (2013) Targeting norepinephrine in mild cognitive impairment and Alzheimer’s disease. Alzheimers Res Ther 5(2):21 31. Koppel J, Jimenez H, Adrien L, Chang H, Malhotra AK, Davies P (2019) Increased tau phosphorylation follows impeded dopamine clearance in a P301L and novel P301L/COMT-deleted (DM) tau mouse model. J Neurochem 148(1):127–135 32. Kiyota T, Yamamoto M, Xiong H, Lambert MP, Klein WL, Gendelman HE, Ransohoff RM, Ikezu T (2009) CCL2 accelerates microglia-mediated Abeta oligomer formation and progression of neurocognitive dysfunction. PLoS One 4(7):e6197 33. Yamamoto M, Horiba M, Buescher JL, Huang D, Gendelman HE, Ransohoff RM, Ikezu T (2005) Overexpression of monocyte che- motactic protein-1/CCL2 in beta-amyloid precursor protein trans- genic mice show accelerated diffuse beta-amyloid deposition. Am J Pathol 166(5):1475–1485 34. Kiyota T, Gendelman HE, Weir RA, Higgins EE, Zhang G, Jain M (2013) CCL2 affects beta-amyloidosis and progressive neurocognitive dysfunction in a mouse model of Alzheimer’s dis- ease. Neurobiol Aging 34(4):1060–1068 35. Kiyota T, Morrison CM, Tu G, Dyavarshetty B,Weir RA, ZhangG, Xiong H, Gendelman HE (2015) Presenilin-1 familial Alzheimer’s disease mutation alters hippocampal neurogenesis and memory function in CCL2 null mice. Brain Behav Immun 49:311–321 36. Howe CL, LaFrance-Corey RG, Goddery EN, Johnson RK, Mirchia K (2017) Neuronal CCL2 expression drives inflammatory monocyte infiltration into the brain during acute virus infection. J Neuroinflammation 14(1):238 37. Hughes PM, Allegrini PR, Rudin M, Perry VH, Mir AK, Wiessner C (2002) Monocyte chemoattractant protein-1 deficiency is protec- tive in a murine stroke model. J Cereb Blood Flow Metab 22(3): 308–317 38. Janssen K, Rickert M, Clarner T, Beyer C, KippM (2016) Absence of CCL2 and CCL3 ameliorates central nervous system grey matter but not white matter demyelination in the presence of an intact blood-brain barrier. Mol Neurobiol 53(3):1551–1564 39. Rankine EL, Hughes PM, Botham MS, Perry VH, Felton LM (2006) Brain cytokine synthesis induced by an intraparenchymal injection of LPS is reduced in MCP-1-deficient mice prior to leu- cocyte recruitment. Eur J Neurosci 24(1):77–86 40. Semple BD, Bye N, Rancan M, Ziebell JM, Morganti-Kossmann MC (2010) Role of CCL2 (MCP-1) in traumatic brain injury (TBI): evidence from severe TBI patients and CCL2-/- mice. J Cereb Blood Flow Metab 30(4):769–782 41. Schlachetzki JC, Fiebich BL, Haake E, de Oliveira AC, Candelario- Jalil E, Heneka MT, Hull M (2010) Norepinephrine enhances the LPS-induced expression of COX-2 and secretion of PGE2 in pri- mary rat microglia. J Neuroinflammation 7:2 42. Sokolova A, Hill MD, Rahimi F, Warden LA, Halliday GM, Shepherd CE (2009) Monocyte chemoattractant protein-1 plays a dominant role in the chronic inflammation observed in Alzheimer’s disease. Brain Pathol 19(3):392–398 43. Westin K, Buchhave P, Nielsen H, Minthon L, Janciauskiene S, Hansson O (2012) CCL2 is associated with a faster rate of cognitive decline during early stages of Alzheimer’s disease. PLoS One 7(1): e30525 44. Severini C, Passeri PP, Ciotti M, Florenzano F, Possenti R, Zona C, Di MA, Guglielmotti A et al (2014) Bindarit, inhibitor of CCL2 synthesis, protects neurons against amyloid-beta-induced toxicity. J Alzheimers Dis 38(2):281–293 Publisher’s Note Springer Nature remains neutral with regard to juris- dictional claims in published maps and institutional affiliations. Mol Neurobiol (2019) 56:8628–86428642 Reboxetine... Abstract Introduction Material and Methods Mouse Models Reboxetine Treatment Y Maze mRNA Analysis Western Blot Immunohistochemistry Fluorescent Labeling of Aβ Plaques Statistical Analysis Results CCL2 Removal Reduces Memory Loss in 5xFAD Mice Effects of Reboxetine Treatment and CCL2 Removal on the Expression of Pro-inflammatory Genes Distribution of Microglia Alterations of GFAP Expression and Synthesis COX2 Expression and Synthesis Accumulation of Aβ Oligomers and Plaques Axonal Damage Analysis of Neuronal Cell Death Discussion References