Retana et al. Alzheimer's Research & Therapy (2019) 11:42 https://doi.org/10.1186/s13195-019-0498-8 RESEARCH Open Access Peripheral administration of human recombinant ApoJ/clusterin modulates brain beta-amyloid levels in APP23 mice Sofía Fernández de Retana1, Paula Marazuela1, Montse Solé1, Guillem Colell1, Anna Bonaterra1, Jose Luis Sánchez-Quesada2,3, Joan Montaner1, Daniel Maspoch4,5, Mary Cano-Sarabia4,5 and Mar Hernández-Guillamon1* Abstract Background: ApoJ/clusterin is a multifunctional protein highly expressed in the brain. The implication of ApoJ in β- amyloid (Aβ) fibrillization and clearance in the context of Alzheimer’s disease has been widely studied, although the source and concentration of ApoJ that promotes or inhibits Aβ cerebral accumulation is not clear yet. ApoJ is abundant in plasma and approximately 20% can appear bound to HDL-particles. In this regard, the impact of plasmatic ApoJ and its lipidation status on cerebral β-amyloidosis is still not known. Hence, our main objective was to study the effect of a peripheral increase of free ApoJ or reconstituted HDL particles containing ApoJ in an experimental model of cerebral β- amyloidosis. Methods: Fourteen-month-old APP23 transgenic mice were subjected to subchronic intravenous treatment with rHDL- rApoJ nanodiscs or free rApoJ for 1 month. Aβ concentration and distribution in the brain, as well as Aβ levels in plasma and CSF, were determined after treatments. Other features associated to AD pathology, such as neuronal loss and neuroinflammation, were also evaluated. Results: Both ApoJ-based treatments prevented the Aβ accumulation in cerebral arteries and induced a decrease in total brain insoluble Aβ42 levels. The peripheral treatment with rApoJ also induced an increase in the Aβ40 levels in CSF, whereas the concentration remained unaltered in plasma. At all the endpoints studied, the lipidation of rApoJ did not enhance the protective properties of free rApoJ. The effects obtained after subchronic treatment with free rApoJ were accompanied by a reduction in hippocampal neuronal loss and an enhancement of the expression of a phagocytic marker in microglial cells surrounding Aβ deposits. Finally, despite the activation of this phagocytic phenotype, treatments did not induce a global neuroinflammatory status. In fact, free rApoJ treatment was able to reduce the levels of interleukin-17 (IL17) and keratinocyte chemoattractant (KC) chemokine in the brain. Conclusions: Our results demonstrate that an increase in circulating human rApoJ induces a reduction of insoluble Aβ and CAA load in the brain of APP23 mice. Thus, our study suggests that peripheral interventions, based on treatments with multifunctional physiological chaperones, offer therapeutic opportunities to regulate the cerebral Aβ load. Keywords: Clusterin, Apolipoprotein J, ApoJ, APP23, Reconstituted HDL, Alzheimer’s disease * Correspondence: Mar.hernandez.guillamon@vhir.org 1 Neurovascular Research Laboratory, Vall d’Hebron Research Institute, Universitat Autònoma de Barcelona, 08035 Barcelona, Spain Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 2 of 17 Background Aβ accumulation in the brain. APOE genotype is the Cerebral β-amyloidosis is characterized by the accumula- strongest genetic risk factor for the development of tion of β-amyloid (Aβ) protein in the brain parenchyma sporadic cerebral β-amyloidosis [16, 17], and although and in cerebral vessels, which are major features of Alzhei- its interaction with Aβ is not well understood, it has mer’s disease (AD) and cerebral amyloid angiopathy been proposed that apolipoprotein E (ApoE) influences (CAA), respectively [1, 2]. Both diseases are characterized Aβ fibrillization [18], together with its clearance across by devastating clinical presentations together with a high the BBB [19]. On the other hand, apolipoprotein J (ApoJ, prevalence in the elderly population. In this sense, AD is also known as clusterin) is one of the most highly the first cause of dementia affecting 24 million people expressed apolipoproteins in the brain after ApoE [20]. worldwide [3, 4]. AD patients show severe β-amyloidosis A crucial role of ApoJ in cerebral β-amyloidosis is sup- characterized by the presence of parenchymal Aβ deposits, ported by several lines of evidence, although its precise together with intracellular accumulation of hyperpho- role is not well determined. ApoJ is a heterodimeric sphorylated TAU protein (neurofibrillary tangles) triggering multifunctional protein. It has been described as an neuroinflammation, neurodegeneration, and dementia [2]. anti-inflammatory and antiapoptotic protein [21, 22] and CAA, on the other hand, is characterized by Aβ deposition a regulator of complement activation [23]. In addition, it in brain vessels causing dementia and being the leading is a very effective chaperone that is able to bind to par- cause of lobar intracerebral hemorrhage (ICH) in the eld- tially misfolded proteins. In this context, ApoJ is able to erly population [5]. The Aβ peptide is the product of the bind to Aβ and prevent its fibrillization [24] and toxicity amyloid precursor protein (APP) processed by β- and in vitro [25]. In AD brains, ApoJ is codeposited with γ-secretases, giving rise to peptides of different lengths, in- neuritic plaques and CAA-affected vessels, and ApoJ-Aβ cluding Aβ40 and Aβ42, which are the most interesting complexes have been detected in cerebrospinal fluid forms from a clinical point of view. Although both Aβ pep- (CSF) [26]. The role of ApoJ in AD was further sup- tides are prone to aggregate-forming amyloid fibrils that de- ported by a GWAS (genome-wide association study) that posit in the brain, Aβ42 preferentially deposits in the identified the CLU locus as a risk factor for developing parenchyma (forming neuritic plaques), whereas Aβ40 is this neurological disease [27]. More revealingly, the risk more commonly associated with CAA-affected vessels [6, factor allele rs11136000C was associated with low ex- 7]. pression of ApoJ, whereas the protective allele The underlying mechanisms explaining pathological rs11136000T caused higher ApoJ expression [28]. On the Aβ accumulation in the brain are not completely clear. other hand, elevated circulating levels of ApoJ in plasma Defective routes of Aβ elimination have been proposed have been related to the severity and prevalence of AD as an important factor leading to sporadic cerebral [29]. This last piece of evidence suggests that an increase β-amyloidosis [8]. In this context, the most relevant Aβ in ApoJ levels would occur as a protective response clearance pathways described to date include proteolytic against the aberrant accumulation of Aβ in the brain. degradation by extracellular proteases, such as However, there are conflicting results in the literature insulin-degrading enzyme (IDE) and neprilysin (NEP) regarding the protective role of ApoJ. This is the case in [9]; clearance across the blood-brain barrier (BBB) by a study showing that PDAPP mice lacking Clu (PDPP/ binding to specific receptors, such as LRP-1 and LRP-2, Clu) presented reduced fibrillary Aβ without altering among others [10]; and through perivascular drainage total levels of Aβ [30]. On the other hand, Wojtas and [11], promoting the mobilization of Aβ material along colleagues observed that knocking out Clu in APP/PS1 the basement membrane of cerebral arteries and capil- mice resulted in an exacerbation of vascular Aβ depos- laries to the cervical lymph nodes for its elimination. Fi- ition, whereas total Aβ and parenchymal Aβ deposition nally, Aβ can be eliminated through glial uptake by were remarkably reduced [31]. In this sense, more re- astrocytes, microglia, and myeloid cells derived from cently, it has been described that the Clu deficiency in peripheral monocytes followed by intracellular digestion the 5xFAD mouse AD model reduces the different Aβ [12–14]. Microglial cells are resident innate immune pools only in young mice, although this effect disappears cells that play an important role in mediating the neu- with age [32]. Nevertheless, the impact of increasing roinflammatory process in the brain and the phagocyt- ApoJ in the circulation and/or in the brain still needs to osis of toxic species such as Aβ material. In this context, be clarified. the importance of this dual role of microglia in cerebral From this background, our main objective was to β-amyloidosis is increasingly recognized [15]. At this understand the role of ApoJ in the development and juncture, therapeutic strategies stimulating these elimin- progression of cerebral β-amyloidosis by determining ation pathways have been proposed for AD. the potential effect of ApoJ-based therapies in an experi- Certain apolipoproteins, beyond their role in lipid and mental model of AD. Because ApoJ is present in plasma cholesterol metabolism, have been shown to modulate as free/nonlipidated form or associated with HDL Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 3 of 17 (high-density lipoproteins) particles (20% of circulating micelles. For the preparation of rHDL-rApoJ nanodiscs, ApoJ) [33], we also analyzed whether lipidation of ApoJ mixed micelles were incubated with free rApoJ at could induce functional modifications associated with 550:110:1 DMPC/CHOL/rApoJ molar ratio and three in- Aβ distribution in the brain. For this purpose, we ana- cubation cycles were performed at 4 °C and 37 °C to pro- lyzed the effect of peripheral administration of human mote the lipid-protein interaction. After incubation, the recombinant ApoJ (h-rApoJ) in nonlipidated (free rApoJ) rHDL-rApoJ nanodiscs self-assembly started with the and lipidated (reconstituted HDL nanodiscs formulated cholate removal through extensive dialysis against with rApoJ, rHDL-rApoJ nanodiscs) [34] structural 1000-fold excess TBS at 4 °C during 48 h in 10 KDa states, in terms of cerebral Aβ accumulation, neuronal SnakeSkin Dialysis Tubing with two buffer changes (1 × loss, and neuroinflammation in aged APP23 transgenic 109 overall dilution factor). Finally, dialyzed samples mice. were centrifuged at 16000×g for 30 min at 4 °C to elim- inate the unbound lipids. A purification step with potas- Methods sium bromide (KBr; Sigma Aldrich) density gradient Human recombinant ApoJ ultracentrifugation was needed to allow the elimination The production of functional human recombinant ApoJ of unbound free rApoJ. This step (100,000×g; 24 h, 1250 (rApoJ) has been previously reported by our group [34, mg/ml density of KBr), allowed the recollection of puri- 35]. Briefly, human embryonic kidney 293 T cells fied rHDL-rApoJ in the upper fractions (< 1250 mg/ml (HEK293T) were transfected with the pcDNA4.0™ vector density). These samples were dialyzed against PBS-2% containing the human APOJ cDNA (Abgent, Clairemont, sucrose to eliminate KBr. The protein concentration was San Diego, USA). Stable transfected cells were grown in determined by BCA (Thermo Scientific). rHDL-rApoJ HYPERFlask systems (Corning Inc., New York, USA) containing solutions were diluted in PBS to a final rApoJ followed by the recollection of cell supernatants for pro- concentration of 200 μg/ml, and aliquots were stored at tein purification with Ni-affinity chromatography with fast − 80 °C until its usage. protein liquid chromatography (FLPC; AKTA Purifier 100 system, GE Healthcare Bio-Sciences Corp., Piscataway, NJ, Subchronic in vivo administration of rHDL-rApoJ USA) in HiScreen Ni FF columns (GE Healthcare). The nanodiscs and free rApoJ purified protein was dialyzed overnight (ON) at 4 °C Mice were housed in a climate-controlled environment on against tris buffer saline (TBS) in 10 KDa SnakeSkin Dialy- a 12/12-h light/dark cycle with food and water available ad sis Tubing membranes (Thermo Fisher, Waltham, MA, libitum. The effect of the subchronic administration of both USA). We confirmed that the produced human rApoJ was rHDL-rApoJ nanodiscs and free rApoJ was studied in male a mature-secreted protein formed as a heterodimeric APP23 transgenic mice (Hemizygote B6, complex of two 40–45 kDa subunits [34]. For those D2-TgN[Thy-APPSWE]-23-Tg mice, Novartis, Basel, batches intended to the preparation of reconstituted HDL Switzerland) [36]. Male APP23 mice were used in order to particles, the resulting purified rApoJ was dialyzed against diminish variation in age-related Aβ deposit quantification, TBS (pH = 6.4). Finally, the protein concentration was de- since male and female APP23 mice differ in the age at termined by BCA assay (Thermo Fisher) and diluted to a which Aβ deposition begins. These transgenic mice overex- final rApoJ concentration of 200 μg/ml in PBS. Aliquots press the APP protein with the Swedish mutation (K670M/ were stored at − 80 °C until its usage. N671L) under the murine neuronal Thy1 promoter (thymo- cyte antigen-1). Hemizygous APP23 mice were backcrossed Preparation and purification of rHDL-rApoJ nanodiscs with C57BL/6 mice (Janvier Labs, Le Genest-Saint-Isle, A detailed protocol for the preparation and purification France), and the APP genotype was tested in the offspring of reconstituted HDL nanodiscs formulated with rApoJ by Transnetyx (Cordova, TN, USA). Wild-type (wt) and (rHDL-rApoJ nanodiscs) has been previously published APP23 mice were aged in the animal facility of our institu- by our group [34]. Briefly, the process started with lipid tion to obtain the final study cohort. APP23 mice received 8 mixture preparation with 1,2-dimyristoyl-sn-glycero-3-- intravenous (IV) doses of rHDL-rApoJ or free rApoJ (1 mg/ phosphocholine (DMPC, Avanti Polar Lipids, Alabaster, kg of rApoJ) or saline for 4 weeks (2 doses/week). The age AL, USA) and free cholesterol (CHOL, Sigma-Aldrich, (mean ± SD) at the beginning of the treatment for each Saint Louis, MO, USA) in chloroform solution (5:1 APP23 group was 14.00 ± 0.73 months old for the DMPC:CHOL molar ratio). The organic solvent was re- saline-treated mice (N = 7), 14.07 ± 0.75 months old for the moved under vacuum and nitrogen to afford a dry lipid rHDL-rApoJ-treated mice (N = 7) and 13.99 ± 0.47 months film, followed by rehydration with TBS 40 mM sodium old for the free rApoJ-treated mice (N = 7). For the adminis- deoxycholate (cholate, Sigma-Aldrich). This suspension tration of treatments, mice were anesthetized with isoflur- was incubated 30 min at 37 °C until achieving a clear so- ane (5% for induction and 2% for maintenance in medicinal lution containing the DMPC/CHOL/cholate mixed air/oxygen, Abbot Laboratories, Spain) followed by Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 4 of 17 retro-orbital sinus injection, selected as a reliable adminis- for both Aβ40 (KHB3481) and Aβ42 (KHB2442, Thermo tration route for IV delivery [37]. All animals completed the Fisher). Levels of human ApoJ in plasma of treated mice treatment, which was not associated with mortality or body were quantified with the commercial ELISA kit weight loss. All assays conducted in mouse brains after (3713-1HP, Mabtech, Nacka Strand, Sweden). Levels of treatment were evaluated in a blinded manner. the endogenous mouse ApoJ in brain homogenates and plasma in APP23 mice were determined using the Cerebrospinal fluid collection mouse Clusterin ELISA Kit (ab199079, Abcam, Cam- Thirty minutes after the last administration, mice were bridge, UK). Commercially available multiplexed ELISA anesthetized under isoflurane and CSF was sampled as was used to investigate the general inflammatory state in previously described [38]. Briefly, anesthetized mice were brain soluble fraction after treatments (Mouse cytokine placed in the stereotaxic instrument and a sagittal section magnetic 20-plex panel, LMC0006M, Invitrogen). Re- inferior of the occiput was done in the skin. Once the dura sults were analyzed in duplicate with Luminex technol- mater of the cistern magna was visualized under the ogy in a MAGPIX™ instrument, and samples with a microscope, it was penetrated with a glass capillary (inner coefficient of variation (CV) > 35% were excluded. In all diameter of 0.5 mm) and clear CSF ascended by capillarity. cases, data were corrected by the amount of total pro- Only completely clear CSF was considered for further tein determined by BCA. experiments. Immunohistochemistry Blood and brain collection A series of three paraffin-embedded sagittal sections Blood was collected through cardiac puncture, and were selected 800 μm apart, starting from the division blood samples were centrifuged in ethylenediaminetetra- between the hemispheres. For the detection of Aβ with cetic acid (EDTA) tubes to recollect the EDTA-plasma. conventional immunohistochemistry (IHC), all samples Then, after transcardial perfusion with 25 ml of cold PBS were deparaffinized for 1 h at 65 °C, rehydrated, and (pH = 7.4), brains were rapidly removed and divided into treated with 2% H2O2 and 10% methanol in PBS for 15 two hemispheres. One of the hemispheres was frozen at min. The samples were incubated for 1 h in blocking so- − 80 °C for homogenization, while the other was fixed in lution (PBS, 0.2% Triton-X, 10% fetal bovine serum 10% formalin (Diapath, Martinengo, Italy) at room (FBS, Millipore, Darmstadt, Germany) followed by ON temperature (RT) for 48 h before paraffin embedding. incubation with the anti-Aβ monoclonal antibody 4G8 (BioLegend, San Diego, CA, USA) diluted 1:10000 in Brain homogenates blocking solution. Next, slices were incubated with bio- One of the hemispheres of each mouse was homogenized to tinylated anti-mouse IgG (Vector Laboratories, Burlin- obtain soluble, membrane-associated, and insoluble fractions game, CA, USA) diluted 1:200 in blocking solution for 1 through sequential centrifugations. First, the brain tissue h at RT and with streptavidin-HRP (Vector Laboratories) was homogenized in 4 ml of TBS (pH = 7.4) supplemented diluted 1:200 in blocking solution for 1 h at RT. Finally, with cOmplete EDTA-free protease inhibitor cocktail diaminobenzidine (DAB; Dako, Denmark) was applied to (Roche Diagnostic, Mannheim, Germany) and samples were the samples until a brown end-product was observed. centrifuged at 8000×g for 30 min a 4 °C. The supernatant Contrast staining was performed by immersion of the was selected as the soluble fraction whereas the pellet was samples in Harris hematoxylin solution (Sigma-Aldrich). homogenized in TBS supplemented with TritonX-100 and To quantify the number of Aβ-positive deposits, we used protease inhibitors followed by centrifugation at 8000×g for the overview tool of a Leica LMD 6500 Laser Microdis- 30 min at 4 °C. The supernatant was selected as the section Microscope System to scan each brain slice in membrane-bound fraction whereas the pellet was homoge- full. Overview images were labeled with codes to ensure nized in 5 M guanidine-HCl (pH = 8, Sigma-Aldrich) sup- blind quantification. The number of Aβ deposits larger plemented with protease inhibitors. Samples were shaken than 50 px2 in the cortex and the area (px2) occupied for 3 h at RT followed by centrifugation at 8000×g for 30 per deposit were quantified with ImageJ software. Three min at 4 °C, and the supernatant was selected as the insol- different deep sections per brain were analyzed, and a uble fraction. Soluble, membrane-bound, and insoluble frac- mean of the values determined in those sections was ob- tions were aliquoted and frozen at − 80 °C. The protein tained. Data are expressed as the number of Aβ-positive concentration of each fraction was quantified by BCA assay deposits per total area (px2) and the mean size of those (Thermo Fisher). Aβ-positive deposits. Representative images were ob- tained with a conventional microscope (BX61, Olympus, Enzyme-linked immunosorbent assay (ELISA) Tokyo, Japan). Aβ levels in plasma, CSF, and soluble and insoluble brain An immunohistochemistry protocol was also per- fractions were quantified with commercial ELISA kits formed for the quantification of NeuN-positive cells in Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 5 of 17 the cortex and hippocampus (CA1, CA2/CA3, and poly- solution was incubated ON at 4 °C. Next, secondary morphic layer of dentate gyrus (poDG)) of APP23 and antibody anti-rabbit-AlexaFluor488 (Thermo Fisher) di- age-matched wt mice in the first deep-section slice. Be- luted 1:500 in blocking solution was incubated 1 h at RT fore the blocking step, an antigen retrieval step was per- followed by incubation with Tomato lectin-DyLight564 formed by immersing the slices in citrate buffer (10 mM (Vector Laboratories) diluted 1:100 in blocking solution sodium citrate, 0.05% TWEEN-20, pH = 6) at 95 °C for 1 h at RT. On the other hand, the localization of CD68 30 min. The NeuN antibody (Millipore) was diluted and Aβ was analyzed in paraffin-embedded brain sec- 1:200 in blocking solution. The quantification of tions. After the deparaffination and rehydration steps, NeuN-positive cells in the different areas was performed brain slices were treated with 70% formic acid for 10 as previously described [39]. Briefly, for each sample and min followed by 3 washes of 5 min with PBS-1% brain region, three different images were taken at × 40 Tween20. Then, the samples were incubated in blocking magnification using an LMD6500 microscope (Leica, solution for 1 h at RT followed by ON incubation at 4 °C Wetzlar, Germany). NeuN-positive cells were counted in with mouse anti-human Aβ monoclonal antibody (4G8) di- three different squares in every image measuring 150 × luted 1:1000 and rabbit anti-mouse CD68 (Abcam) diluted 150 μm in cortical regions, whereas squares measuring 1:50 or rabbit anti-Iba1 (ionized calcium binding adapter 100 × 100 μm were used in hippocampal regions. The re- molecule-1, Abcam) diluted 1:100 in blocking solution. sults are expressed as the number of NeuN-positive cells Next, secondary antibodies anti-rabbit-AlexaFluor-488 and per square millimeter. anti-mouse-AlexaFluor694 (Thermo Fisher) diluted 1:500 in blocking solution for 1 h at RT. Finally, contrast staining Thioflavin S staining was done with DAPI before mounting the preparations. Im- Thioflavin S (ThS) staining was performed to detect fi- ages were obtained with a conventional fluorescence micro- brillar Aβ deposits in brain slices. After deparaffination scope (BX61, Olympus). Finally, human ApoJ was analyzed and rehydration, a series of 3 selected sagittal sections in brain sections of treated animals. To this end, after the 800 μm apart were immersed in ThS (Sigma-Aldrich) 1% rehydration, antigen retrieval, and blocking steps, brain solution in 75% ethanol for 30 s. The excess of ThS was slices were incubated ON with mouse-anti-human ApoJ drained, and the sections were immersed in ThS 0.1% in (BD Bioscience, San Jose, CA, USA) diluted 1:100 in block- 75% ethanol for 1 min. Finally, the slices were dehy- ing solution. AlexaFluor-594 was used as a secondary anti- drated, and DAPI (4′,6-diamidino-2-phenylindole) was body (1:500), and DAPI was used as contrast staining. used for counterstaining before the preparations were Images were obtained with a confocal spectral microscopy mounted. The number of ThS-positive deposits, the size (FV1000, Olympus). of those deposits, and the number of ThS-positive brain vessels were quantified as previously described for the Western blot quantification of Aβ deposits by immunohistochemistry. For the immunodetection of different protein profiles, The count of ThS-positive deposits larger than 50 px2 40 μg of soluble or membrane-bound brain homogenate and the analysis of the area (px2) occupied per deposit fractions or 1 μl of mouse plasma was loaded in 10% were quantified with ImageJ software. Data are acrylamide SDS-PAGE under reducing conditions (5% expressed as the number of CAA-affected vessels β-mercaptoethanol and sample heating at 95 °C). Then, (ThS-positive brain vessels), the number of ThS-positive gels were transferred onto nitrocellulose membranes deposits per total area (px2) and the mean size of using Trans-Blot Turbo transfer system (Bio-Rad, ThS-positive deposits. Representative images were ob- Hercules, CA, USA) at 1.3 A, 25 V for 10 min and block- tained with a conventional fluorescence microscope ing was done in 10% non-fatted milk for 1 h at RT. (BX61, Olympus, Tokyo, Japan). Membranes were incubated in tested specific antibodies diluted in blocking solution at 4 °C ON. Suitable second- Immunofluorescence ary HRP-labeled antibodies diluted in blocking solution A cluster of differentiation 68 (CD68) as a marker of were incubated 1 h at RT. The processing of membranes phagocytosis and Tomato lectin (TL) as vascular and continued with the incubation with PierceECL Western microglia marker [40] were analyzed through immuno- blotting Luminol/Enhancer and Stable Peroxidase solu- fluorescence in paraffin-embedded brain sections. To tions (Thermo Fisher). The relative amount of the tar- this end, after deparaffination, rehydration, and antigen geted protein was calculated from the quantification of retrieval with citrate buffer (10 mM sodium citrate, the mean intensity with the ImageJ software. Immuno- 0.05% Tween20, pH = 6, 30 min at 95 °C), slices were blotting of β-actin, β-tubulin, or glyceraldehydes blocked in PBS pH = 7.4 supplemented with 10% FBS 3-phosphate dehydrogenase (GADPH) was used as load- and 1% Tween 20 for 1 h at RT. Rabbit anti-mouse ing controls. The primary antibodies used were rabbit CD68 antibody (Abcam) diluted 1:50 in blocking anti-CD68 (1:1000, Abcam), rabbit anti-GFAP (Glial Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 6 of 17 Fibrillary Acidic Protein, 1:1000, Dako), rabbit anti-Iba1 one-to-one comparisons. The Benjamini-Hochberg false (1:1000, Abcam), mouse anti-LRP1 (1:1000, Abcam), discovery rate (FDR) correction was used to control for mouse anti-human APP (1:1000, Millipore), mouse multiple testing in the multiplexed analysis of brain in- anti-sAPPα (Soluble APP fraction-α, 1:1000, Clontech, flammatory status. A p value < 0.05 was considered sta- Mountain View, CA, USA), mouse anti-sAPPβ (Soluble tistically significant. APP fraction-β, 1:1000, Clontech), rabbit-anti RAGE (1:1000, Abcam), goat-anti-mouse ApoJ (1:1000, RD Sys- Results tems, Minneapolis, MN, USA), mouse anti-β-actin APP23 mice are characterized by increasing accumula- (1:10000, Sigma-Aldrich) and mouse anti-GADPH tion of both human Aβ40 and Aβ42 in the brain. This ris- (1:10000, Invitrogen). The following HRP-labeled sec- ing Aβ deposition became especially evident in very old ondary antibodies were used: Anti-rabbit-HRP (1:2000, APP23 mice (24 months old), where a 20-fold increase GE Healthcare Biosciences, Little Chalfont, UK), in insoluble Aβ40 and a 10-fold increase in insoluble anti-mouse-HRP (1:2000, GE Healthcare Bioscience), Aβ42 were observed in comparison with the correspond- and anti-goat-HRP (1:5000, Sigma-Aldrich). ing Aβ load in 15-month-old APP23 mice (Fig. 1a, b). Interestingly, this growing Aβ accumulation was accom- Microglial cell cultures panied by an increase in mouse ApoJ (m-ApoJ) levels in The mouse microglial cell line BV-2 was obtained from the the brain (Fig. 1c). Twenty-four-month-old APP23 mice European Collection of Authenticated Cell Culture presented significantly higher m-ApoJ brain levels than (ECACC). BV-2 cells were seeded (30.000 cells/well) in age-matched wt mice, whereas the difference between poly-L-lysine (Sigma-Aldrich)-coated sterile coverslips in genotypes was nonsignificant in 15-month-old mice. In 24-well plates and incubated at 37 °C in RPMI 1640 medium addition, as expected, this increase in brain m-ApoJ was (Thermo Fisher) supplemented with 10% fetal bovine serum exacerbated with age, as significant differences in (FBS, Thermo Fisher) and 1% penicillin-streptomycin m-ApoJ levels were detected between 15- and (Thermo Fisher) for 24 h. For treatments, the medium was 24-month-old APP23 mice. Plasma levels of endogenous changed to RPMI containing 0.1% FBS and cells were incu- m-ApoJ were also determined, although no significant bated with or without rApoJ (0.1 μM) for 24 h as a differences between genotypes were found in 15- or pre-treatment. Then, rApoJ was removed by washing the 24-month-old mice (Fig. 1d). wells with RPMI 0.1% FBS and 0.5 μM human Aβ(1–40) The endogenous increase in m-ApoJ in APP23 mice at HiLyteTM Fluor 488-labeled (AnaSpec, Fremont, CA, USA) late pathological stages could be interpreted as a physio- was added for 3 h to the corresponding wells. Next, cells logical protective mechanism to promote reduction and were washed twice with PBS containing 5% bovine serum clearance of Aβ in the brain. In this regard, we won- albumin (BSA, Sigma-Aldrich) and fixed with 4% parafor- dered whether an earlier increase in ApoJ levels could be maldehyde (PFA, Sigma-Aldrich) for 15 min. After fixation, a beneficial strategy to delay the Aβ deposition detected cells were washed twice with PBS and levels of 488-Aβ(1– over time. Therefore, our main objective was to study 40) in the cells were analyzed by measuring the fluorescence the effect of peripheral administration of free rApoJ or (Ex/Em = 503/528 nm) using a Synergy Mx microplate lipidated rApoJ (rHDL-rApoJ nanodiscs) in APP23 mice. reader (Biotek). To visualize the cellular structure, a To this end, APP23 mice received 8 intravenous admin- phalloidin-tetramethylrhodamine B isothiocyanate (phalloi- istrations (over the course of 1 month) of free rApoJ (1 din-TRITC, Sigma-Aldrich) actin staining was used (50 μg/ mg/kg), rHDL-rApoJ (1 mg/kg), or saline, with 2 infu- ml in PBS for 20 min). After washing twice with PBS, the sions/week (Fig. 2a). The administered dose of 1 mg/kg coverslips were mounted on slides using FluoroshieldTM of human rApoJ would theoretically lead to a mean in- with DAPI (Sigma-Aldrich). Images were obtained with con- crease of 20% in total circulating ApoJ levels considering focal spectral microscopy (FV1000, Olympus). the m-ApoJ levels determined in plasma. The equality of the initial amounts of infused rHDL-rApoJ and free Statistical analysis rApoJ was confirmed by Western blot, as shown in GraphPad Prism 6 and SPSS Statistics 17.0 software Fig. 2b, where the α- and β-chains of rApoJ were ob- were used for the statistical analysis. The normality of served at ≈ 37 KDa. Next, the circulating concentration distributions was assessed with the Shapiro-Wilk nor- of human rApoJ (h-rApoJ) was quantified by ELISA. A mality test. The differences were assessed using a t test significant increase in plasma levels of h-rApoJ was ob- or one-way ANOVA with an LSD (least significant dif- served in both rHDL-rApoJ- and free rApoJ-treated ference) post hoc test when appropriate. When data did mice, whereas human ApoJ, as expected, was not de- not fit a normal distribution, the Kruskal–Wallis test tected in the saline group (Fig. 2c). In addition, no sig- with a Dunn’s post hoc was used for multiple compari- nificant differences in h-ApoJ between the rHDL-rApoJ- sons, and the Mann–Whitney test was used for and free rApoJ-treated groups were found in plasma 30 Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 7 of 17 Fig. 1 Levels of cerebral insoluble Aβ and mouse ApoJ expression in APP23 transgenic mice over time. a Cerebral accumulation of insoluble Aβ40 in 15- and 24-month-old APP23 and wt mice. b Cerebral accumulation of insoluble Aβ42 in 15- and 24-month-old APP23 and wt mice. c Mouse ApoJ (m-ApoJ) cerebral expression in 15- and 24-month-old APP23 and wt mice. d Mouse ApoJ (m-ApoJ) levels in the plasma of 15 and 24-month-old APP23 mice. Data are expressed as the mean ± SEM. N = 3–4/group. *p < 0.05; **p < 0.01; *p < 0.001 min after administration. On the other hand, the levels Aβ42 levels were under the lower limit of detection (< 6 of h-rApoJ in brain homogenates were undetectable by pg/ml). Interestingly, we observed lower levels of both ELISA (< 70 pg/ml) 30 min after the last administration insoluble Aβ40 and Aβ42 in mice treated with free rApoJ of the subchronic treatment (Fig. 2d). However, im- than in the saline-treated group (insoluble Aβ40 1.01 ± munofluorescence analysis revealed specific staining for 0.12 vs. 2.70 ± 0.44 μg/mg and insoluble Aβ42 0.27 ± 0.02 h-ApoJ in occasional vessels from the brains of both vs. 0.60 ± 0.11 μg/mg in rApoJ-treated group compared rHDL-rApoJ- and free rApoJ-treated mice. Brains from to the saline-treated group, p < 0.01) (Fig. 3b, c). How- saline-treated mice did not present human ApoJ staining ever, after treatment with rHDL-rApoJ nanodiscs, the in- (Fig. 2e). On the other hand, h-rApoJ-based peripheral soluble Aβ40 levels remained unaltered, whereas lower treatments did not influence the expression of m-ApoJ insoluble Aβ42 brain levels were found in comparison in brain homogenates or m-ApoJ levels in plasma, as de- with the saline-treated group (insoluble Aβ42 0.38 ± 0.07 tected by ELISA (Additional file 1: Figure S1). vs. 0.60 ± 0.11 μg/mg in rHDL-treated group compared The efficacy of subchronic rApoJ-based treatments in to the saline-treated group, p < 0.05) (Fig. 3b, c). Regard- APP23 mice was first evaluated in terms of cerebral Aβ ing the cerebral Aβ distribution, the number of load. Levels of Aβ40 and Aβ42 in soluble and insoluble Aβ-affected vessels and the number of parenchymal de- fractions of brain homogenates were quantified by posits were quantified by ThS staining and immunohis- ELISA after the treatments (Fig. 3a–c). The levels of sol- tochemistry with a specific anti-Aβ antibody (Fig. 3d–f ). uble Aβ40 were not significantly different among treat- Brains from mice treated with rHDL-rApoJ and free ments, although a trend towards a reduction in free rApoJ presented a significantly lower number of rApoJ-treated mice was observed (p = 0.09). Soluble CAA-affected arteries as detected by ThS staining (the Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 8 of 17 Fig. 2 Subchronic intravenous treatment of APP23 mice with human recombinant ApoJ (h-rApoJ). a Schematic timeline representing the experimental design of the study based on the subchronic administration of APP23 mice with rHDL-rApoJ nanodiscs, free rApoJ or saline. b Immunodetection (anti-h- ApoJ) of treatment samples before being intravenously infused. c Plasmatic levels of h-ApoJ in treated mice 30 min after the last administration detected by ELISA (N = 7/group). d Levels of h-ApoJ in brain homogenates from treated mice detected by ELISA (N = 7/group). e Immunofluorescent detection of h-ApoJ in paraffin-embedded brain sections. **p < 0.01; ***p < 0.001 reduction percentages were 64.9% and 54.8% vs. poDG (Fig. 4). Importantly, the treatment with free saline-treated group, respectively) (Fig. 3d). Although no rApoJ was able to significantly ameliorate the neuronal differences were detected in the parenchymal Aβ depos- loss in both CA1 and poDG in comparison to the ition load after rApoJ-based treatments (Fig. 3e, f ), the saline-treated group (Fig. 4). mean size of fibrillar Aβ accumulations (ThS-positive Because an increase in fibrillary Aβ in the brain is as- deposits) was reduced in those mice treated with free sociated with low levels of particular Aβ species in the rApoJ (the reduction percentage in free rApoJ vs. CSF of AD and CAA patients (Aβ42 and Aβ40, respect- saline-treated group was 17.5%) (Fig. 3e). Finally, the de- ively [42, 43]), we analyzed the Aβ levels in CSF 30 min position load and size of parenchymal Aβ deposits did after the last treatment in APP23 mice. We observed a not differ significantly among groups as shown by a spe- significant increase in CSF Aβ40 levels in mice treated cific immunostaining against Aβ, although a clear ten- with free rApoJ compared to the rest of the groups dency towards smaller deposits after both rApoJ-based (17.10 ± 1.83 vs. 10.46 ± 1.47 ng/ml in the saline-treated treatments was again observed (Fig. 3f ). group, p < 0.01) (Fig. 5a). However, Aβ40 levels in plasma Neuronal loss is one of the neuropathological hall- were not different among groups 30 min after the last in- marks in AD [41], and it has been described that the fusion (Fig. 5b). Circulating levels of Aβ42 in APP23 APP23 mice model shows a neuronal loss in the CA1 re- mice were undetectable in all cases. In addition, we ex- gion of the hippocampus and in the poDG from plored the expression of different brain receptors previ- 12 months of age [39]. Thus, we measured the effect of ously described to mediate the transport of Aβ across rHDL-rApoJ and free rApoJ treatments over the neuro- the BBB [10]. In this context, the expression of LRP-1 or degeneration in the cortex and hippocampus through RAGE did not exhibit expression differences among NeuN immunostaining. In comparison with groups (Fig. 5c). Moreover, we also eliminated the possi- age-matched wt mice, we found that 15-month-old bility that the decrease in insoluble Aβ in the brain and APP23 mice exhibited a neuronal reduction in CA1 and the increase in CSF Aβ levels obtained after peripheral Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 9 of 17 Fig. 3 Cerebral Aβ levels in 15-month-old APP23 mice treated subchronically with saline, rHDL-rApoJ nanodiscs, or free rApoJ. a Cerebral soluble Aβ40 levels. b Cerebral insoluble Aβ40. c Cerebral insoluble Aβ42, detected by ELISA. d Representative images showing ThS-positive vessels and the corresponding quantification in brain sections. e Representative images showing ThS-positive deposits, quantification of the number of ThS- positive deposits per area (px2), and quantification of the mean size of those deposits. f Representative images of anti-Aβ (4G8)-positive deposits in the brain parenchyma, quantification of the number of Aβ-positive deposits per area (px2), and quantification of the mean size of those deposits. Data are expressed as the mean ± SEM. N = 7/group. *p < 0.05; **p < 0.01 rApoJ treatment could be explained by an alteration of ApoJ-treated mice (p < 0.05), whereas the expression of APP transgene expression or the Aβ peptide production the resident microglial marker Iba-1 was not signifi- process. Specifically, we confirmed that the expression cantly different among groups (Fig. 6b). Iba1 and CD68 of APP was unaltered in the brains of mice receiving dif- localization was determined by immunofluorescence, ferent treatments, and the processing of APP mediated and the signal was observed in the periphery of paren- by α-secretase, β-secretase, and γ-secretase was also un- chymal Aβ-positive deposits, showing elevated CD68 ex- changed among groups (Additional file 1: Figure S2). pression levels in mice treated with free rApoJ (Fig. 6c, The neuroinflammatory state of mice that received d). Furthermore, double staining of CD68 with tomato ApoJ-based treatments was first analyzed by exploring lectin (TL) confirmed the presence of this lysosomal the expression levels of specific cellular markers in brain marker in microglial cells (Fig. 6e). Additionally, a multi- homogenates. On this front, we determined that the plexed ELISA was performed to detect the global inflam- presence of the activated astrocytic protein GFAP did matory state in the brain parenchyma (the complete list not differ after treatments (Fig. 6a). On the other hand, of molecules analyzed can be found in Additional file 1: we detected a 2.3-fold increase in the levels of the lyso- Table S1). Despite the phagocytic phenotype shown by somal marker CD68 in brain homogenates from free microglial cells after the free rApoJ treatment, we did Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 10 of 17 Fig. 4 Neuronal loss in 15-month-old wt mice and APP23 mice treated subchronically with saline, rHDL-rApoJ nanodiscs or free rApoJ. a Representative images of NeuN-positive cells in CA1 region of the hippocampus and in poDG of APP23-treated mice and wt mice. b Quantification of NeuN-positive cells per square millimeter in the different studied regions (cortex, CA1, CA2/CA3, and poDG). Abbreviations: Ctx: Cortex; poDG: polymorphic layers of the dentate gyrus. Data are expressed as the mean ± SEM. N = 6–7/group. *p < 0.05 (comparison between WT and saline-APP23 groups) and p < 0.05 (comparison among APP23-treated groups) not observe an increase in the levels of prevalent brain Discussion cytokines and inflammatory molecules, such as the clas- Despite the high socioeconomic impact of AD and sical mediator interleukin 1-β (IL-1β) or tumor necrosis CAA, these diseases lack an effective disease-modifying factor-α (TNF-α). In fact, we detected reduced levels of therapy. The involvement of ApoJ in AD is widely ac- interleukin 17 (IL17) and KC (keratinocyte chemo- knowledged, although its underlying mechanism in this attractant chemokine) in the brains from mice treated pathology is not yet elucidated. Therefore, the principal with both rHDL-rApoJ and free rApoJ. However, after aim of our study was to test the potential protective role FDR correction, only levels of IL17 and KC in free of peripheral ApoJ in an experimental model of AD/ r-ApoJ-treated brains remained statistically different CAA. To this end, we first considered the need to evalu- than the corresponding levels in the saline-treated group ate the effect of the progressive Aβ deposition on cere- (Table 1). bral endogenous mouse ApoJ levels. In fact, we Finally, the potential direct effect of free rApoJ on micro- confirmed that, in the APP23 mouse model, the accu- glial phagocytic activity was tested in vitro. BV2 microglial mulation of both insoluble Aβ40 and Aβ42 was accom- mouse cells were pretreated with or without rApoJ (0.1 μM) panied by an increase in mouse ApoJ expression in the for 24 h and then incubated with HyLite Fluor 488-labeled brain. This increase in ApoJ has also been described in Aβ40 (0.5 μM) for 3 h. Confocal images revealed that AD and CAA patients [44, 45], although it is not clear rApoJ-pretreated cells retained higher fluorescent-labeled whether the increase in brain ApoJ levels promotes or Aβ40 levels than non-pretreated cells (Fig. 7a), indicating prevents the Aβ accumulation and toxicity. In this higher Aβ uptake in cells challenged with rApoJ. The pres- regard, the elevated circulating levels of ApoJ observed ence of fluorescent-labeled Aβ40 inside the cells was corrob- in AD patients have been proposed to be a protective re- orated in confocal microscope z-stack images (Fig. 7b). In sponse against the pathologic Aβ deposition [27]. To addition, quantification of HyLite Fluor 488-labeled Aβ40 in- confirm this hypothesis, we treated APP23 mice with ternalized in cultured cells was determined fluorometrically. rApoJ at the time that mouse ApoJ levels elevation was Results confirmed that rApoJ-pretreated cells significantly not yet significant. This approach allowed us to study enhanced the Aβ phagocytic activity compared with the efficacy of an early increase in peripheral ApoJ in non-pretreated cells (Fig. 7c). terms of Aβ deposition, neuronal loss, and Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 11 of 17 Fig. 5 Effect of rHDL-rApoJ nanodiscs or free rApoJ treatment on circulating Aβ levels. a CSF levels of Aβ40 detected by ELISA. b Plasmatic Aβ40 levels detected by ELISA. c Expression of Aβ-transporting receptors (LRP-1 and RAGE) determined by Western blot. Data are expressed as the mean ± SEM. N = 5–7 group. *p < 0.05; **p < 0.01 neuroinflammation. Indeed, the strategy was to increase of insoluble Aβ42 accumulation. When we analyzed the the circulating levels of ApoJ in mice following two dif- distribution of Aβ, we observed a reduced number of ferent treatments: intravenous administration of free Aβ-affected vessels in both rApoJ-based treatments. This rApoJ or rHDL-rApoJ nanodiscs. Several lines of evi- reduction was in accordance with a previous published dence support the relevance of cholesterol and HDL me- study [31], where authors described a marked exacerba- tabolism to AD [46], and we previously observed the tion of CAA in APP/PS1/Clu−/− mutant mice. The au- increased ability of rHDL-rApoJ nanodiscs to promote thors proposed that the lack of Clu impaired an efficient the efflux of cholesterol esters in vitro [34]. Therefore, Aβ clearance across the BBB, promoting the accumula- we believed that the peripheral administration of tion of Aβ in perivascular spaces. However, this same rHDL-rApoJ could shed light on the effect of lipidation study showed a clear reduction of parenchymal Aβ accu- on the functionality of ApoJ in cerebral β-amyloidosis. mulation and total Aβ in APP/PS1 mice lacking Clu After 1 month of subchronic administration of treat- [31]. After this, another study showed that 5xFAD/ ments based on human rApoJ, we confirmed the pres- Clu−/− mice had less Aβ cerebral accumulation only in ence of h-rApoJ in plasma. However, we were not able young mice, although this effect disappeared with age to detect h-rApoJ in the brains of mice treated with [32]. Beyond the valuable information obtained from rHDL-rApoJ or free rApoJ. It has previously been de- these knockout models, we consider that our approach, scribed that ApoJ is able to cross the BBB [47]. Based on based on the peripheral administration of rApoJ, allowed this, we expected that the infused rApoJ could enter the us to study the role of human ApoJ in contrast to mur- brain parenchyma, but the human protein was detected ine ApoJ and to focus specifically on the effect of a cir- by immunofluorescence only on the luminal side of oc- culating increase in ApoJ levels. As mentioned, our casional leptomeningeal vessels in both rApoJ-based results obtained in aged APP23 mice treated with rApoJ treatment groups. In fact, we previously observed spe- partly agreed with the effect observed in the APP/PS1/ cific vascular accumulation of rHDL-rApoJ nanodiscs in Clu−/− mutant mice described by Wojtas et al. [31], since very old (28 months) APP23 mice [34]. After the corre- we detected a significant reduction in the CAA load sponding subchronic peripheral treatment in APP23 after treatments. Thus, it can be speculated that the mice, the cerebral Aβ load was evaluated. Free localization of human rApoJ in brain vessels after the rApoJ-treated mice showed significant prevention of peripheral subchronic treatment directly stimulated the both insoluble Aβ40 and Aβ42 accumulation in compari- clearance of Aβ from perivascular drainage spaces redu- son with the saline-treated group. In turn, the treatment cing the Aβ vascular load. However, we also observed with rHDL-rApoJ nanodiscs resulted in the prevention that the peripheral increase of human rApoJ reduced the Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 12 of 17 Fig. 6 (See legend on next page.) Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 13 of 17 (See figure on previous page.) Fig. 6 Levels of astrocytic and microglial markers in APP23 mice treated subchronically with saline, rHDL-rApoJ nanodiscs, or free rApoJ. a Expression of the reactive astrocyte marker GFAP in mouse APP23-brain homogenates after treatments. b Expression of microglial markers (CD68 and Iba-1) in mouse APP23-brain homogenates after treatments. Data are expressed as the mean ± SEM. N = 6–7/group. *p < 0.05. c Representative immunofluorescence images showing the localization of Iba1 and d representative immunofluorescence images showing the localization of CD68 around anti-Aβ (4G8)-positive deposits in brain slices from APP23 mice. e Representative immunofluorescence images showing the localization of CD68 with tomato lectin (TL, as a microglial marker) in brain slices from APP23 mice total insoluble Aβ in the brain, which was accompanied neuronal loss in CA1 and poDG of the hippocampus. by a decrease in the size of the fibrillar Aβ deposits. Due to the association between Aβ deposits and neuro- Interestingly, these effects were observed without detect- degeneration [50], the reduced Aβ levels observed in free ing the human rApoJ protein inside the brain. In fact, rApoJ-treated mice could explain the amelioration of intravenous infusion of free ApoJ induced an increase in neuronal death. Once again, as shown by the effect on Aβ levels in CSF, which might reflect an Aβ mobilization Aβ load after treatments, rHDL-rApoJ was less effective process from the brain parenchyma. On the other hand, than the free protein, suggesting that the structural it has recently been proposed that intraventricular ad- modification of ApoJ due to its lipidation may alter its ministration of an ApoJ mimetic peptide promotes the ability to activate some of the protective pathways within amelioration of both parenchymal and vascular Aβ accu- the brain. The inflammatory state of mouse brains after mulation and improved the cognitive function if the ApoJ-based treatments was also determined. Treatment Tg6799 mouse model [48]. These results can be under- with free rApoJ or rHDL-ApoJ did not induce changes stood as complementary to ours, as they focused only on in the reactivity of astrocytes. Likewise, microglial prolif- the intracerebral pool of ApoJ. Therefore, to our know- eration was not altered among the treatment groups, but ledge, our study is the first to administrate human a significant increase in the expression of brain CD68 in recombinant ApoJ, produced in mammalian cells, in a free rApoJ-treated mice was observed. CD68 was local- preclinical model of AD. We believe that this is espe- ized surrounding Aβ parenchymal deposits in cells posi- cially relevant due to the high structural complexity of tive for the TL dye, which stains microglia and other ApoJ and the importance of its full-length form for its cells, such as macrophages and endothelial cells [40, 51]. functionality [49]. In this context, CD68 is a lysosomal protein that has Another beneficial effect of subchronic IV treatment been widely used as a marker of phagocytic activity [52]. with free rApoJ in APP23 mice was the prevention of Accordingly, we demonstrated that rApoJ had a direct Table 1 Effect of rHDL-rApoJ nanodiscs or free rApoJ on brain inflammatory status, analyzed by multiplexed ELISA. Data are expressed as the mean ± SEM. N = 5–7/group. *p < 0.05 and **p < 0.01 refer to statistical differences vs. the saline-treated group determined by one-way ANOVA test (plus multiple comparison test). The false discovery rate (FDR) correction was used to control for multiple testing in the multiplexed analysis of brain inflammatory status Inflammatory APP23 saline APP23 rHDL- APP23 free One-way ANOVA (3 rHDL-rApoJ vs saline FDR- Free rApoJ vs saline FDR- markers (pg/mg) rApoJ (pg/mg) rApoJ (pg/mg) groups) p value corrected p value corrected p value FGF-basic 984.4 ± 154.6 895.6 ± 113.3 560.2 ± 151.4 p = 0.108 p = 0.787 p = 0.174 IL-1β 85.28 ± 12.17 50.86 ± 3.75 72.91 ± 11.43 p = 0.080 p = 0.122 p = 0.448 IL2 151.0 ± 13.09 116.7 ± 14.22 125.2 ± 14.49 p = 0.222 p = 0.198 p = 0.324 IL5 141.7 ± 26.35 130.61 ± 27.54 148.0 ± 21.31 p = 0.893 p = 0.819 p = 0.874 IL6 203.1 ± 22.15 177.7 ± 28.82 176.3 ± 19.5 p = 0.656 p = 0.583 p = 0.448 IL13 21.31 ± 3.40 17.68 ± 2.34 12.63 ± 3.22 p = 0.171 p = 0.552 p = 0.183 IL17 20.01 ± 2.33 13.9 ± 1.37* 10.23 ± 1.08** p = 0.004 p = 0.122 p = 0.017 IFN-γ 127.8 ± 15.53 89.81 ± 13.28 94.08 ± 10.68 p = 0.123 p = 0.137 p = 0.194 IP10 54.4 ± 4.34 44.48 ± 5.11 46.71 ± 5.05 p = 0.330 p = 0.289 p = 0.388 KC 1558 ± 115.5 1206 ± 112.0* 1043 ± 104.6** p = 0.012 p = 0.122 p = 0.029 MIG 81.65 ± 8.21 57.81 ± 6.91 61.48 ± 8.72 p = 0.090 p = 0.122 p = 0.194 TNF-α 127.2 ± 13.46 125.2 ± 9.43 144.7 ± 17.79 p = 0.620 p = 0.937 p = 0.448 VEGF 17.67 ± 2.36 14.27 ± 2.35 10.72 ± 1.22 p = 0.080 p = 0.399 p = 0.129 MCP-1 36.1 ± 3.93 24.31 ± 2.84 28.19 ± 3.12 p = 0.060 p = 0.122 p = 0.194 Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 14 of 17 Fig. 7 Effect of rApoJ treatment on Aβ phagocytic activity of microglial cultured cells. a Representative micrographs corresponding to BV2 cells pre-incubated during 24 h with/without rApoJ (0.1 μM) and then treated for 3 h with HiLyte Fluo 488-Aβ40 (0.5 μM). Blue arrows indicate cells containing HiLyte Fluo 488-Aβ40 (green). Cell morphology, visualized by TRITC-phalloidin staining, and cell nuclei with Dapi are shown in gray. b Confocal image showing the z-stack of a single cell containing HiLyte Fluo 488-Aβ40 (green), stained with TRITC-phalloidin (red), and Dapi (blue). c Fluorometric quantification of HiLyte Fluo 488-Aβ40 in fixed BV2 cells (Ex/Em = 503/528 nm). Data is expressed as mean ± SEM of 4 independent experiments performed in duplicates. *p < 0.05; ***p < 0.001 role enhancing the Aβ uptake by microglial cells in vitro, myeloid cells in pathological conditions is under discus- as previously proposed [53]. However, infiltrated mono- sion, since it has been observed that, for example, cul- cytes have previously demonstrated a higher ability than tured microglia produce high levels of neurotoxic microglial cells to phagocytose Aβ peptide [14, 54]. Be- cytokines in response to Aβ stimulation [58]. To test cause resident microglia and peripheral macrophages whether the enhancement of phagocytic pathways was consist of heterogeneous populations that in many ways accompanied by an increased release of inflammatory are indistinguishable entities [55], the detailed mecha- molecules, we analyzed the overall inflammatory state of nisms implicated in the stimulation of phagocytosis in brains after treatments. We did not observe an increase in brain myeloid cells by peripheral ApoJ merit exploration the cerebral levels of classical proinflammatory molecules in future studies. after treatments. In fact, reduced neuroinflammation was As mentioned, it has been extensively demonstrated observed in the brains of rApoJ-treated mice in comparison that brain myeloid cells participate in Aβ elimination with saline-treated APP23 mice, as shown by the significant through phagocytosis and that stimulation of the im- reductions in IL17 and KC (the mouse homolog of CXCL1) mune response in the CNS ameliorates Aβ deposition levels, especially in free rApoJ-treated mouse brains. In this [56, 57]. Therefore, promoting the phagocytic ability of context, higher expression of IL17 has been linked to BBB these cell types could be a potential strategy to prevent disruption and neurotoxicity [59, 60], and KC chemokine cerebral Aβ deposition. Nevertheless, the exact role of levels in the brain have been associated with memory Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 15 of 17 deficits [61]. This last result agrees with previous studies Funding that demonstrated anti-inflammatory and anti-apoptotic This work has been funded by Instituto de Salud Carlos III (ISCIII) [PI17/00275, PI14/01134, and PI13/00364], co-financed by the European Regional Develop- properties of ApoJ/Clu [62]. ment Fund (FEDER), and Fundació La Marató de TV3 [40/U/2014]. The Neuro- In summary, beyond the accepted implication of ApoJ vascular Research Laboratory is part of the INVICTUS+ network, ISCIII, Spain in cerebral β-amyloidosis, our study is the first to dem- [RD16/0019/0021]. J.L.S-Q is a member of the CIBER of Diabetes and Metab- olism (CIBERDEM), ISCIII, Spain. M.H-G is supported by the Miguel Servet onstrate that the intravenous administration of human programme, ISCIII, Spain [CPII17/00010]. ICN2 acknowledges the support of recombinant ApoJ is safe and effective reducing the the Spanish MINECO through the Severo Ochoa Centers of Excellence levels of cerebral insoluble Aβ and the CAA load. In- Programme, under Grant SEV-2013-0295. deed, our results confirm that peripheral interventions Availability of data and materials offer low invasive opportunities to regulate the cerebral The datasets used and/or analyzed during the current study are available β-amyloid load. An important limitation of our study is from the corresponding author on reasonable request. the lack of functional analysis after treatments, even Authors’ contributions though we have demonstrated that free rApoJ treatment MH-G designed the study. MC-S, JM, and DM and SF-R participated in the prevents neuronal loss in the hippocampus in old design of the study. SF-R performed most of the experiments, collected and analyzed data, and wrote the manuscript. MS, PM, and AB participated ac- APP23 mice. Further studies testing the efficacy of rApoJ tively in the experiments, performing Aβ immunohistochemistry in mice treatment using different models of AD will be relevant slices and western blots. MS and AB performed the microglial in vitro experi- to analyze its impact on behavior and cognition. We ments. GC was responsible for the mice colony care. MC-S and DM designed and produced the rHDL-rApoJ nanodiscs. JLS-Q did the purification of the suggest that peripheral rApoJ promotes the amelioration rHDL particles through KBr density ultracentrifugation. All authors read and of cerebral β-amyloidosis acting at the vascular level and approved the final manuscript. upregulating specific phagocytic pathways in myeloid Ethics approval cells. However, the exact molecular pathway activated by Animal use for this study was approved by the Ethics Committee for Animal peripheral ApoJ mediating the stimulation of the phago- Experimentation of Vall d’Hebron Research Institute (CEEA 78/13) and cytic phenotype needs to be elucidated in future studies. conducted in compliance with Spanish legislation and the Directives of the European Union. Additional file Consent for publication Not applicable. Additional file 1: Figure S1. Effect of rHDL-rApoJ or free rApoJ treat- Competing interests ment in the levels of mouse ApoJ in brain homogenates and plasma The authors declare that they have no competing interests. from treated mice, determined by ELISA. Data are expressed as the mean ± SEM. N = 3–4/group. Figure S2. Effect of rHDL-rApoJ or free rApoJ treatment in the expression of full-length APP protein and the processing Publisher’s Note of APP through α-, β-, and γ secretases. a) Representative Western blots Springer Nature remains neutral with regard to jurisdictional claims in showing the relative amount of full-length APP, sAPPα, sAPPβ, and CTF- published maps and institutional affiliations. APP. Relative quantification of b) full-length APP, c) sAPPα, d) sAPPβ, and e) CTF-APP levels after treatments. Data are expressed as the mean ± Author details 1 SEM. N = 6/group. Table S1. List of abbreviation of tested inflammatory Neurovascular Research Laboratory, Vall d’Hebron Research Institute, molecules by multiplexed ELISA. (DOCX 302 kb) Universitat Autònoma de Barcelona, 08035 Barcelona, Spain. 2Cardiovascular Biochemistry Group, Research Institute of the Hospital de Sant Pau (IIB Sant Pau), Barcelona, Spain. 3CIBER of Diabetes and Metabolism (CIBERDEM), ISCIII, Abbreviations Madrid, Spain. 4Catalan Institute of Nanoscience and Nanotechnology (ICN2), AD: Alzheimer’s disease; ApoE: Apolipoprotein E; ApoJ: Apolipoprotein J; CSIC and the Barcelona Institute of Science and Technology, , Campus UAB, APP: Amyloid precursor protein; Aβ: β-Amyloid; BBB: Blood-brain barrier; Bellaterra, Barcelona, Spain. 5Institució Catalana de Recerca i Estudis Avançats BSA: Bovine serum albumin; CAA: Cerebral amyloid angiopathy; CD68: Cluster (ICREA), 08100 Barcelona, Spain. of differentiation 68; CHOL: Free cholesterol; Cholate: Sodium deoxycholate; Clu: Clusterin; CNS: Central nervous system; CSF: Cerebrospinal fluid; Received: 19 December 2018 Accepted: 23 April 2019 DAB: Diaminobenzidine; DAPI: 4′,6-Diamidino-2-phenylindole; DMPC: 1,2- Dimyristoyl-sn-glycero-3-phosphocholine; EDTA: Ethylenediaminetetracetic acid; ELISA: Enzyme-linked immunosorbent assay; FBS: Fetal bovine serum; References FDR: False discovery rate; FLPC: Fast protein liquid chromatography; free 1. Revesz T, Holton JL, Lashley T, Plant G, Frangione B, Rostagno A, Ghiso J. rApoJ: Nonlipidated recombinant ApoJ; GAPDH: Glyceraldehydes 3- Genetics and molecular pathogenesis of sporadic and hereditary cerebral phosphate dehydrogenase; GFAP: Glial fibrillary acidic protein; amyloid angiopathies. Acta Neuropathol. 2009;118(1):115–30. GWAS: Genome-wide association study; HDL: High-density lipoproteins; 2. Thal DR, Rüb U, Orantes M, Braak H. Phases of a beta-deposition in the HEK293T: Human embryonic kidney 293 T cells; h-rApoJ: Human recombinant human brain and its relevance for the development of AD. Neurology. 2002; ApoJ; Iba1: Ionized calcium-binding adapter molecule-1; ICH: Intracerebral 58(12):1791–800. hemorrhage; IDE: Insulin-degrading enzyme; IHC: Immunohistochemistry; 3. Wilson D, Charidimou A, Werring DJ. Advances in understanding IL17: Interleukin 17; IL-1β: Interleukin 1-β; IV: Intravenous; KBr: Potassium spontaneous intracerebral hemorrhage: insights from neuroimaging. Expert bromide; KC: Keratinocyte chemoattractant chemokine; LSD: Least significant Rev Neurother. 2014;14(6):661–78. difference; m-ApoJ: Mouse ApoJ; NEP: Neprilysin; ON: Overnight; 4. Reitz C, Brayne C, Mayeux R. Epidemiology of Alzheimer disease. Nat Rev poDG: Polymorphic layer of the dentate gyrus; px: Pixel; rHDL- Neurol. 2011;7(3):137–52. rApoJ: Reconstituted HDL formulated with rApoJ; RT: Room temperature; 5. Attems J, Jellinger K, Thal DR, Van Nostrand W. Review: sporadic cerebral sAPPα: Soluble APP fraction-α; sAPPβ: Soluble APP fraction-β; TBS: Tris buffer amyloid angiopathy. Neuropathol Appl Neurobiol. 2011;37(1):75–93. saline; ThS: Thioflavin S; Thy: Thymocyte antigen-1 promoter; TL: Tomato 6. Sinha S, Lieberburg I. Cellular mechanisms of beta-amyloid production and lectin; TNF-α: Tumor necrosis factor-α; WT: Wild-type secretion. Proc Natl Acad Sci. 1999;96(20):11049–53. Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 16 of 17 7. Fryer JD, Simmons K, Parsadanian M, Bales KR, Paul SM, Sullivan PM, Bossù P, Piccardi P, Annoni G, Seripa D, Galimberti D, Hannequin D, Licastro Holtzman DM. Human apolipoprotein E4 alters the amyloid-beta 40:42 ratio F, Soininen H, Ritchie K, Blanché H, Dartigues JF, Tzourio C, Gut I, Van and promotes the formation of cerebral amyloid angiopathy in an amyloid Broeckhoven C, Alpérovitch A, Lathrop M, Amouyel P. Genome-wide precursor protein transgenic model. J Neurosci. 2005;25(11):2803–10. association study identifies variants at CLU and CR1 associated with 8. Mawuenyega KG, Sigurdson W, Ovod V, Munsell L, Kasten T, Morris JC, Alzheimer’s disease. Nat Genet. 2009;41(10):1094–9. Yarasheski KE, Bateman RJ. Decreased clearance of CNS beta-amyloid in 28. Nelson AR, Sagare AP, Zlokovic BV. Role of clusterin in the brain vascular Alzheimer’s disease. Science. 2010;330(6012):1774. clearance of amyloid-β. Proc Natl Acad Sci U S A. 2017;114(33):8681–2. 9. Jha NK, Jha SK, Kumar D, Kejriwal N, Sharma R, Ambasta RK, Kumar P. 29. Schrijvers EMC, Koudstaal PJ, Hofman A, Breteler MMB. Plasma clusterin and Impact of insulin degrading enzyme and neprilysin in Alzheimer’s disease the risk of Alzheimer disease. JAMA. 2011;305:1322–6. biology: characterization of putative cognates for therapeutic applications. J 30. DeMattos RB, O’dell MA, Parsadanian M, Taylor JW, Harmony JA, Bales KR, Alzheimers Dis. 2015;48(4):891–917. Paul SM, Aronow BJ, Holtzman DM. Clusterin promotes amyloid plaque 10. Deane R, Bell RD, Sagare A, Zlokovic BV. Clearance of amyloid-beta peptide formation and is critical for neuritic toxicity in a mouse model of across the blood-brain barrier: implication for therapies in Alzheimer’s Alzheimer’s disease. Proc Natl Acad Sci U S A. 2002;99(16):10843–8. disease. CNS Neurol Disord Drug Targets. 2009;8(1):16–30. 31. Wojtas AM, Kang SS, Olley BM, Gatherer M, Shinohara M, Lozano PA, Liu CC, 11. Hawkes CA, Jayakody N, Johnston DA, Bechmann I, Carare RO. Failure of Kurti A, Baker KE, Dickson DW, Yue M, Petrucelli L, Bu G, Carare RO, Fryer JD. perivascular drainage of β-amyloid in cerebral amyloid angiopathy. Brain Loss of clusterin shifts amyloid deposition to the cerebrovasculature via Pathol. 2014;24(4):396–403. disruption of perivascular drainage pathways. Proc Natl Acad Sci U S A. 12. Yang CN, Shiao YJ, Shie FS, Guo BS, Chen PH, Cho CY, Chen YJ, Huang FL, 2017;114(33):E6962–71. Tsay HJ. Mechanism mediating oligomeric Aβ clearance by naïve primary 32. Oh SB, Kim MS, Park S, Son H, Kim SY, Kim MS, Jo DG, Tak E, Lee JY. microglia. Neurobiol Dis. 2011;42(3):221–30. Clusterin contributes to early stage of Alzheimer’s disease pathogenesis. 13. Jones RS, Minogue AM, Connor TJ, Lynch MA. Amyloid-β-induced astrocytic Brain Pathol. 2018. https://doi.org/10.1111/bpa.12660. phagocytosis is mediated by CD36, CD47 and RAGE. J NeuroImmune 33. de Silva HV, Stuart WD, Duvic CR, Wetterau JR, Ray MJ, Ferguson DG, Albers Pharmacol. 2013;8(1):301–11. HW, Smith WR, Harmony JA. A 70-kDa apolipoprotein designated ApoJ is a 14. Simard AR, Soulet D, Gowing G, Julien JP, Rivest S. Bone marrow-derived marker for subclasses of human plasma high density lipoproteins. J Biol microglia play a critical role in restricting senile plaque formation in Chem. 1990;265(22):13240–7. Alzheimer’s disease. Neuron. 2006;49(4):489–502. 34. Fernández-De-Retana S, Cano-Sarabia M, Marazuela P, Sánchez-Quesada JL, 15. Salter MW, Stevens B. Microglia emerge as central players in brain disease. Garcia-Leon A, Montañola A, Montaner J, Maspoch D, Hernández-Guillamon Nat Med. 2017;23(9):1018–27. M. Characterization of ApoJ-reconstituted high-density lipoprotein (rHDL) 16. Nicoll JAR, McCarron MO. APOE gene polymorphism as a risk factor for cerebral nanodisc for the potential treatment of cerebral β-amyloidosis. Sci Rep. amyloid angiopathy-related hemorrhage. Amyloid. 2001;8(Suppl 1):51–5. 2017;7(1):14637. 17. Strittmatter WJ, Saunders AM, Schmechel D, Pericak-Vance M, Enghild J, 35. Merino-Zamorano C, Fernández-de Retana S, Montañola A, Batlle A, Saint-Pol J, Salvesen GS, Roses AD. Apolipoprotein E: high-avidity binding to beta- Mysiorek C, Gosselet F, Montaner J, Hernández-Guillamon M. Modulation of amyloid and increased frequency of type 4 allele in late-onset familial amyloid-β1–40 transport by ApoA1 and ApoJ across an in vitro model of the Alzheimer disease. Proc Natl Acad Sci U S A. 1993;90(5):1977–81. blood-brain barrier. J Alzheimers Dis. 2016;53(2):677–91. 18. DeMattos RB, Cirrito JR, Parsadanian M, May PC, O'Dell MA, Taylor JW, 36. Sturchler-Pierrat C, Staufenbiel M. Pathogenic mechanisms of Alzheimer’s Harmony JA, Aronow BJ, Bales KR, Paul SM, Holtzman DM. ApoE and disease analyzed in the APP23 transgenic mouse model. Ann N Y Acad Sci. clusterin cooperatively suppress Abeta levels and deposition: evidence that 2000;920:134–9. ApoE regulates extracellular Abeta metabolism in vivo. Neuron. 2004;41(2): 37. Yardeni T, Eckhaus M, Morris HD, Huizing M, Hoogstraten-Miller S. Retro- 193–202. orbital injections in mice. Lab Anim (NY). 2011;40(5):155–60. 19. Kim J, Basak JM, Holtzman DM. The role of apolipoprotein E in Alzheimer’s 38. Liu L, Duff K. A technique for serial collection of cerebrospinal fluid from the disease. Neuron. 2009;63(3):287–303. cisterna magna in mouse. J Vis Exp. 2008;10(21). https://doi.org/10.3791/960. 20. DeMattos RB, Brendza RP, Heuser JE, Kierson M, Cirrito JR, Fryer J, Sullivan 39. Shang J, Yamashita T, Zhai Y, Nakano Y, Morihara R, Fukui Y, Hishikawa N, Ohta PM, Fagan AM, Han X, Holtzman DM. Purification and characterization of Y, Abe K. Strong impact of chronic cerebral hypoperfusion on neurovascular astrocyte-secreted apolipoprotein E and J-containing lipoproteins from unit, cerebrovascular remodeling, and neurovascular trophic coupling in wild-type and human apoE transgenic mice. Neurochem Int. 2001;39(5–6): Alzheimer’s disease model mouse. J Alzheimers Dis. 2016;52(1):113–26. 415–25. 40. Villacampa N, Almolda B, González B, Castellano B. Tomato lectin histochemistry 21. Schwarz M, Spath L, Lux CA, Paprotka K, Torzewski M, Dersch K, Koch-Brandt C, for microglial visualization. Methods Mol Biol. 2013;1041:261–79. Husmann M, Bhakdi S. Potential protective role of apoprotein J (clusterin) in 41. Grignon Y, Duyckaerts C, Bennecib M, Hauw J-J. Cytoarchitectonic atherogenesis: binding to enzymatically modified low-density lipoprotein alterations in the supramarginal gyrus of late onset Alzheimer’s disease. reduces fatty acid-mediated cytotoxicity. Thromb Haemost. 2008;100(1):110–8. Acta Neuropathol. 1998;95(4):395–406. 22. Song HB, Jun HO, Kim JH, Yu YS, Kim KW, Min BH, Kim JH. Anti-apoptotic 42. Mattsson N, Lönneborg A, Boccardi M, Blennow K, Hansson O. Geneva task effect of clusterin on cisplatin-induced cell death of retinoblastoma cells. force for the roadmap of Alzheimer’s biomarkers. Clinical validity of Oncol Rep. 2013;30(6):2713–8. cerebrospinal fluid Aβ42, tau, and phospho-tau as biomarkers for 23. McDonald JF, Nelsestuen GL. Potent inhibition of terminal complement Alzheimer’s disease in the context of a structured 5-phase development assembly by clusterin: characterization of its impact on C9 polymerization. framework. Neurobiol Aging. 2017;52:196–213. Biochemistry. 1997;36(24):7464–73. 43. Van Etten ES, Verbeek MM, Van Der Grond J, Zielman R, van Rooden S, van 24. Beeg M, Stravalaci M, Romeo M, Carrá AD, Cagnotto A, Rossi A, Diomede L, Zwet EW, van Opstal AM, Haan J, Greenberg SM, van Buchem MA, Wermer Salmona M, Gobbi M. Clusterin binds to Aβ1-42 oligomers with high affinity MJ, Terwindt GM. β-Amyloid in CSF: biomarker for preclinical cerebral and interferes with peptide aggregation by inhibiting primary and amyloid angiopathy. Neurology. 2017;88(2):169–76. secondary nucleation. J Biol Chem. 2016;291(13):6958–66. 44. Oda T, Pasinetti GM, Osterburg HH, Anderson C, Johnson SA, Finch CE. 25. Yerbury JJ, Poon S, Meehan S, Thompson B, Kumita JR, Dobson CM, Wilson Purification and characterization of brain clusterin. Biochem Biophys Res MR. The extracellular chaperone clusterin influences amyloid formation and Commun. 1994;204(3):1131–6. toxicity by interacting with prefibrillar structures. FASEB J. 2007;21(10):2312–22. 45. Miners JS, Clarke P, Love S. Clusterin levels are increased in Alzheimer’s 26. Howlett DR, Hortobágyi T, Francis PT. Clusterin associates specifically with disease and influence the regional distribution of Aβ. Brain Pathol. 2017; Aβ40 in Alzheimer’s disease brain tissue. Brain Pathol. 2013;23(6):623–32. 27(3):305–13. 27. Lambert J-C, Heath S, Even G, Campion D, Sleegers K, Hiltunen M, 46. Reitz C, Tang M-X, Schupf N, Manly JJ, Mayeux R, Luchsinger JA. Association Combarros O, Zelenika D, Bullido MJ, Tavernier B, Letenneur L, Bettens K, of higher levels of high-density lipoprotein cholesterol in elderly individuals Berr C, Pasquier F, Fiévet N, Barberger-gateau P, Engelborghs S, De Deyn P, and lower risk of late-onset Alzheimer disease. Arch Neurol. 2010;67(12): Mateo I, Franck A, Helisalmi S, Porcellini E, Hanon O, European Alzheimer’s 1491–7. Disease. Initiative Investigators, de Pancorbo MM, Lendon C, Dufouil C, 47. Zlokovic BV, Martel CL, Mackic JB, Matsubara E, Wisniewski T, McComb JG, Jaillard C, Leveillard T, Alvarez V, Bosco P, Mancuso M, Panza F, Nacmias B, Frangione B, Ghiso J. Brain uptake of circulating apolipoproteins J and E Retana et al. Alzheimer's Research & Therapy (2019) 11:42 Page 17 of 17 complexed to Alzheimer’s amyloid beta. Biochem Biophys Res Commun. 1994;205(2):1431–7. 48. Qi XM, Wang C, Chu XK, Li G, Ma JF. Intraventricular infusion of clusterin ameliorated cognition and pathology in Tg6799 model of Alzheimer’s disease. BMC Neurosci. 2018;19(1):2. https://doi.org/10.1186/s12868-018-0402-7. 49. Dabbs RA, Wilson MR. Expression and purification of chaperone-active recombinant clusterin. PLoS One. 2014;9(1):e86989. 50. Calhoun ME, Wiederhold KH, Abramowski D, Phinney AL, Probst A, Sturchler-Pierrat C, Staufenbiel M, Sommer B, Jucker M. Neuron loss in APP transgenic mice. Nature. 1998;395(6704):755–6. 51. Andjelkovic AV, Nikolic B, Pachter JS, Zecevic N. Macrophages/microglial cells in human central nervous system during development: an immunohistochemical study. Brain Res. 1998;814(1–2):13–25. 52. Hoogland ICM, Houbolt C, van Westerloo DJ, van Gool WA, van de Beek D. Systemic inflammation and microglial activation: systematic review of animal experiments. J Neuroinflammation. 2015;12:114. 53. Yeh FL, Wang Y, Tom I, Gonzalez LC, Sheng M. TREM2 binds to apolipoproteins, including APOE and CLU/APOJ, and thereby facilitates uptake of amyloid-beta by microglia. Neuron. 2016;91(2):328–40. 54. Malm T, Koistinaho M, Muona A, Magga J, Koistinaho J. The role and therapeutic potential of monocytic cells in Alzheimer’s disease. Glia. 2010; 58(8):889–900. 55. Lai AY, McLaurin J. Clearance of amyloid-β peptides by microglia and macrophages: the issue of what, when and where. Future Neurol. 2012;7(2): 165–76. 56. Michaud J-P, Hallé M, Lampron A, Thériault P, Préfontaine P, Filali M, Tribout-Jover P, Lanteigne AM, Jodoin R, Cluff C, Brichard V, Palmantier R, Pilorget A, Larocque D, Rivest S. Toll-like receptor 4 stimulation with the detoxified ligand monophosphoryl lipid A improves Alzheimer’s disease- related pathology. Proc Natl Acad Sci U S A. 2013;110(5):1941–6. 57. Yamanaka M, Ishikawa T, Griep A, Axt D, Kummer MP, Heneka MT. PPARγ/ RXRα-induced and CD36-mediated microglial amyloid-β phagocytosis results in cognitive improvement in amyloid precursor protein/presenilin 1 mice. J Neurosci. 2012;32(48):17321–31. 58. Kim SU, de Vellis J. Microglia in health and disease. J Neurosci Res. 2005; 81(3):302–13. 59. Zenaro E, Pietronigro E, Della Bianca V, Piacentino G, Marongiu L, Budui S, Turano E, Rossi B, Angiari S, Dusi S, Montresor A, Carlucci T, Nanì S, Tosadori G, Calciano L, Catalucci D, Berton G, Bonetti B, Constantin G, Piacentino G, Marongiu L, Budui S, Turano E, Rossi B, Angiari S, Dusi S, Montresor A, Carlucci T, Nanì S, Tosadori G, Calciano L, Catalucci D, Berton G, Bonetti B, Constantin G. Neutrophils promote Alzheimer’s disease-like pathology and cognitive decline via LFA-1 integrin. Nat Med. 2015;21(8):880–6. 60. Kebir H, Kreymborg K, Ifergan I, Dodelet-Devillers A, Cayrol R, Bernard M, Giuliani F, Arbour N, Becher B, Prat A. Human TH17 lymphocytes promote blood-brain barrier disruption and central nervous system inflammation. Nat Med. 2007;13(10):1173–5. 61. Silva B, Sousa L, Miranda A, Vasconcelos A, Reis H, Barcelos L, Arantes R, Teixeira A, Rachid MA. Memory deficit associated with increased brain proinflammatory cytokine levels and neurodegeneration in acute ischemic stroke. Arq Neuropsiquiatr. 2015;73(8):655–9. 62. Cunin P, Beauvillain C, Miot C, Augusto JF, Preisser L, Blanchard S, Pignon P, Scotet M, Garo E, Fremaux I, Chevailler A, Subra JF, Blanco P, Wilson MR, Jeannin P, Delneste Y. Clusterin facilitates apoptotic cell clearance and prevents apoptotic cell-induced autoimmune responses. Cell Death Dis. 2016;7:e2215.