Altered glial expression of the cannabinoid 1 receptor in the subiculum of a mouse model of Alzheimer's disease

RESEARCH ARTICLE Altered glial expression of the cannabinoid 1 receptor in the subiculum of a mouse model of Alzheimer's disease Itziar Terradillos 1,2 | Itziar Bonilla-Del Río 1,2 | Nagore Puente 1,2 | Maitane Serrano 1,2 | Amaia Mimenza 1,2 | Leire Lekunberri 1,2 | Ilazki Anaut-Lusar 1,2 | Leire Reguero 1,2 | Inmaculada Gerrikagoitia 1,2 | Samuel Ruiz de Martín Esteban 3 | Cecilia J. Hillard 4 | María T. Grande 3 | Julián Romero 3 | Izaskun Elezgarai 1,2 | Pedro Grandes 1,2 1 Department of Neurosciences, Faculty of Medicine and Nursing, University of the Basque Country UPV/EHU, Leioa, Spain 2 Achucarro Basque Center for Neuroscience, Leioa, Spain 3 Faculty of Experimental Sciences, Universidad Francisco de Vitoria, Pozuelo de Alarc on, Spain 4 Department of Pharmacology and Toxicology, Neuroscience Research Center, Medical College of Wisconsin, Milwaukee, Wisconsin, USA Correspondence Pedro Grandes, Department of Neurosciences, Faculty of Medicine and Nursing, University of the Basque Country UPV/EHU, Leioa, Spain. Email: [email protected]us Funding information Eusko Jaurlaritza, Grant/Award Numbers: IT1230-19, IT1620-22; Ministerio de Ciencia e Innovaci on, Grant/Award Numbers: PID2019-107548RB-I00, PID2019-108992RB-I00; Research and Education Component of the Advancing a Healthier Wisconsin Endowment at the Medical College of Wisconsin Abstract The alteration of the endocannabinoid tone usually associates with changes in the expression and/or function of the cannabinoid CB 1 receptor. In Alzheimer's disease (AD), amyloid beta (Aβ)-containing aggregates induce a chronic inflammatory response leading to reactivity of both microglia and astrocytes. However, how this glial response impacts on the glial CB 1 receptor expression in the subiculum of a mouse model of AD, a brain region particularly affected by large accumulation of plaques and concomitant subcellular changes in microglia and astrocytes, is unknown. The CB 1 receptor localization in both glial cells was investigated in the subiculum of male 5xFAD/CB 2EGFP/f/f (AD model) and CB 2EGFP/f/f mice by immuno-electron microscopy. The findings revealed that glial CB 1 receptors suffer remarkable changes in the AD mouse. Thus, CB 1 receptor expression increases in reactive microglia in 5xFAD/CB 2EGFP/f/f , but remains constant in astrocytes with CB 1 receptor labeling rising proportionally to the perimeter of the reactive astrocytes. Not least, the CB 1 receptor localization in microglial processes in the subiculum of controls and closely surrounding amyloid plaques and dystrophic neurites of the AD model, supports previous suggestions of the presence of the CB 1 receptor in microglia. These findings on the correlation between glial reactivity and the CB 1 receptor expression in microglial cells and astrocytes, contribute to the understanding of the role of the endocannabinoid system in the pathophysiology of Alzheimer's disease. KEYWORDS astroglia, endocannabinoid system, immuno-electron microscopy, microglia, neurodegeneration Itziar Terradillos and Itziar Bonilla-Del Río share first authorship. Received: 2 August 2022 Revised: 23 October 2022 Accepted: 18 November 2022 DOI: 10.1002/glia.24312 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2022 The Authors. GLIA published by Wiley Periodicals LLC. 866 Glia. 2023;71:866–879. wileyonlinelibrary.com/journal/glia 1|INTRODUCTION Astrocytes and microglia activity associated with chronic inflammation induced by Aβ-containing aggregates results in abnormal morphology and proliferation of both glial cells in AD (Benito et al., 2003;McAlpine et al., 2021; Smit et al., 2021). The potential of cannabinoids to target several processes involved in AD pathogenesis is regarded as a therapeutic strategy (Casarejos et al., 2013; Chen et al., 2012;Eljaschewitsch et al., 2006; Talarico et al., 2019), in part because the main cannabinoid CB 1 receptor is localized in brain regions and cells, including glia, affected by the disease. We have previously estimated that about 56% of the CB 1 receptor labeling localizes to GABAergic terminals, 12% to glutamatergic terminals, 6% to astrocytes, 15% to mitochondria and the rest to other cells and compartments to be determined (Bonilla-Del Rίoetal.,2019;Bonilla-DelRíoetal.,2021). The expression of CB 1 receptors in astrocytes seems to be regulated by multiple factors and can vary under different brain conditions, for example, transient receptor potential vanilloid 1 knock out mice show a significant decrease in CB 1 density in astrocytes (Egaña-Huguet et al., 2021), acute Δ-9-tetrahydrocannabinol exposure causes CB 1 increase in these glial cells (Bonilla-Del Río et al., 2021), or adolescent binge drinking significantly decreases CB 1 in astrocytes in the adult brain (Bonilla-Del Rίoetal.,2019). Although the impact that astroglial dysfunction may have in AD is still poorly understood (Smit et al., 2021; Verkhratsky & Nedergaard, 2018), pieces of evidence suggest that changes in the endocannabinoid system occur in astrocytes near AD lesions. Thus, astrocytes closely associated with Aβaggregates show more intermediate filament proteins and hypertrophy of cell bodies (Escartin et al., 2019;Smitetal.,2021). These astrocytes clear as well as degrade Aβaggregates and release pro-inflammatory molecules (Farina et al., 2007) which can be diminished by endocannabinoids acting on CB 1 receptors localized in these glial cells (MetnaLaurent & Marsicano, 2015). Also, high fatty acid amide hydrolase (FAAH) levels, the main degrading enzyme for the endocannabinoid anandamide, have been found in astrocytes around neuritic plaques (Abate et al., 2021;Benitoetal.,2003). However, the real impact of the progression of Alzheimer's disease on the expression of CB 1 receptors in astrocytes surrounding the lesions is unknown. Microglia is another important player in the pathogenesis of Alzheimer's disease. Cannabinoids prevent Aβ-induced neurodegeneration by reducing microglial activity. Both CB 1 and CB 2 receptorsexpressedinmicrogliaareinvolvedinthisactionasthey inhibit neuroinflammation by preventing reactive oxygen species (ROS) formation and cytokines release by these cells (Casarejos et al., 2013; Martín-Moreno et al., 2011; Ramírez et al., 2005; Talarico et al., 2019). Microglia also elicits a significant increase in endocannabinoid production that, in turn, activates more CB 1 and CB 2 receptors and their signaling cascades, amplifying the antiinflammatory and protective microglial phenotype (Duffy et al., 2021; Mecha et al., 2016). In fact, the number and Aβphagocytic capacity of microglial cells decrease in mice lacking CB 2 receptors (de Martín et al., 2022). There are pieces of evidence indicating that microglia constitutively expresses CB 1 receptors that mediate some of the cannabinoid effects in resting microglial cells (Kaplan, 2013; Navarro et al., 2018; Ribeiro et al., 2013;Stella,2009;Thionetal.,2018), and regulate neuroinflammation in a sex-dependent manner (De Meij et al., 2021). CB 1 receptors have been noticed in cultured microglia (Carlisle et al., 2002; Facchinetti et al., 2003;Klegeris et al., 2003; Molina-Holgado et al., 2002; Sinha et al., 1998; Stefano et al., 1996; Waksman et al., 1999;Walteretal.,2003), and specific anti-CB 1 antibodies detected some scattered CB 1 signal in microglia in the hypothalamic arcuate nucleus of females (De Meij et al., 2021). CB 1 receptor expression increases in many inflammatory and neurodegenerative diseases like AD (Bisogno & Di Marzo, 2010; Ribeiro et al., 2013); however, very little is known about the localization and expression of cannabinoid receptors in glial cells in AD. We hypothesize in this study that CB 1 receptor expression in glia is altered in the subiculum of a mouse model of AD, a brain region particularly affected by large accumulation of plaques, as a result of concomitant subcellular changes in microglia and astrocytes. Our findings show that CB 1 receptors in microglial cells suffer remarkable modifications in 5xFAD/CB 2EGFP/f/f mice, a murine model of AD recently reported to have an increase in CB 2 receptors in microglia related to dystrophic neurites (Ruiz de Martín Esteban et al., 2022), similarly to the CB 2 rise observed in plaque-associated microglia (Benito et al., 2003,2007). 2|MATERIAL AND METHODS 2.1 |Ethics statement The protocols for animal care and use were approved by the Committee of Ethics for Animal Welfare of the University of the Basque Country (M20/2020/109) and were in accordance to the European Communities Council Directive of September 22, 2010 (2010/63/ EU) and Spanish regulations (Real Decreto 53/2013, BOE 08-02-2013). Efforts were made to minimize the number and suffering of animals. 2.2 |Experimental animals Experiments were done in 6.5–7.5-month-old male CB 2EGFP/f/f mice, controls and co-expressing five AD mutations (5xFAD; Oakley et al., 2006) previously used in our laboratory for the localization of CB 2 receptors (Ruiz de Martín Esteban et al., 2022). The CB 2EGFP/f/f mice were generated at the Genoway facilities (Lyon, France) by designing a targeting strategy consisting of the insertion of an enhanced green fluorescent protein (EGFP) reporter gene, preceded by an internal ribosomal entry site sequence (IRES) into the 30 untranslated region (UTR) of the mouse cnr2 gene. This resulted in expression of the reporter gene (EGFP) under the control of the TERRADILLOS ET AL.867 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License mouse endogenous cnr2 promoter, and transcription of the same bicistronic mRNA as the CB 2 receptor protein. In addition, these mice co-expressed 5xFAD mutations at the same time. The 5xFAD mice with a C57BL/6J background were purchased from Jackson Laboratory (Bar Harbor, Main, USA). To obtain the co-expression, 5xFAD mice were mated with CB 2EGFP/f/f mice for at least five generations to generate 5xFAD/CB 2EGFP/f/f mice (L opez et al., 2018). This 5xFAD model does not seem to express the mutation that causes the neurofibrillary degeneration (Oblak et al., 2021), but it may occur indirectly through neuronal degeneration and Aβ1–42 deposits. 2.3 |Brain tissue processing Mice were anesthetized with ketamine/xylazine (100 mg/10 mg/kg body weight, intraperitoneal injection) and subsequently perfused transcardially at room temperature (RT) through the left ventricle. First with phosphate buffered saline (PBS) 0.1 M (pH 7.4) for 20 s, and then with the fixative solution composed of 4% formaldehyde, 0.2% picric acid and 0.1% glutaraldehyde in PBS 0.1 M (pH 7.4) for 10–15 min, with a fixative solution volume of 80 ml per mouse. The brains were then removed from the skull and post-fixed in the fixative solution for approximately 1 week at 4C. Subsequently, they were stored in 1:10 diluted fixative solution at 4C with 0.025% sodium azide. Brain vibrosections were cut coronally at 50 μm and stored with 1 ml of phosphate buffer (PB) 0.1 M (pH 7.4) with 0.025% sodium azide at 4C. 2.4 |Immunohistochemistry for light microscopy Brain sections containing the subiculum were collected in PB 0.1 M (pH 7.4) at RT, pre-incubated with a blocking solution of 10% bovine serum albumin (BSA), 0.1% sodium azide and 0.5% triton X-100 in 1x tris-buffered saline (TBS 1x) (pH 7.4) for 30 min at RT, and incubated with the following primary polyclonal antibodies: guinea pig anti-CB 1 receptor (1:100; CB 1 -GP-Af530; AB_2571593, Frontier Institute Co., ltd), rabbit anti-ionized calcium-binding adapter molecule 1 (Iba1, 1:500; 019-19741; AB_839504, FUJIFILM Wako Pure Chemical Corporation), rabbit anti-glutamate aspartate transporter 1 (Anti-A522 [GLAST] EAAT1, 0.3 μg/ml; Ab#314; AB_231456, kindly gifted by Prof. Niels Christian Danbolt, University of Oslo). They were prepared in blocking solution and gently shaken for 2 days at 4C or 1 day at RT. Then, sections were washed with 1% BSA and 0.5% triton X-100 in TBS 1x for 30 min, and incubated with a biotinylated antiguinea pig (1:200, Biotin-SP-AffiniPure Goat Anti-Guinea Pig IgG; AB_2337394, Jackson Immuno Research), or biotinylated anti-rabbit secondary antibody (1:200, Biotin-SP-AffiniPure Donkey Anti-Rabbit IgG; AB_2340593, Jackson Immuno Research) diluted in the washing solution for 1 h on a shaker at RT. They were washed with 1% BSA and 0.5% triton X-100 in TBS 1x (30 min). Tissue was incubated with the avidin-biotin peroxidase complex (ABC; 1:50, Elite, Vector Laboratories, Burlingame, CA, USA) prepared in the washing solution, for 1 h at RT. Samples were washed with 1% BSA and 0.5% triton X-100 in TBS 1x (3 x 1 min) and lastly with PB 0.1 M (pH 7.4) and 0.5% triton X-100 (2 x 10 min). Labeling was revealed with 0.05% diaminobenzidine (DAB) in PB 0.1 M (pH 7.4) containing 0.5% triton X-100 and 0.01% hydrogen peroxide for 3.5 min at RT. This was followed by washes in PB 0.1 M (pH 7.4) with 0.5% triton X-100 (3 x 1 min, 2 x 10 min). Tissue sections were mounted on gelatinized slides, dried and dehydrated in graded ethanol for 5 min each. After rinsing with xylene (3 x 5 min), the slides were cover slipped with DPX. The subiculum was examined and photographed with a Zeiss AxioCam light microscope coupled to a Zeiss AxioCam HRc camera. 2.5 |Immunohistochemistry for electron microscopy The protocol is already published (Puente et al., 2019). Four to five sections per brain containing the subiculum were selected. They were preincubated in a blocking solution (1 ml/well) of 10% BSA, 0.02% saponin and 0.1% sodium azide in TBS 1x (pH 7.4), for 30 min on the shaker at RT. Tissue was then incubated with a primary guinea pig anti-CB 1 receptor antibody (1:100) in combination with a rabbit anti-GLAST antibody (0.3 μg/ml) or a rabbit anti-Iba1 antibody (1:500). The solution contained 10% BSA in TBS 1x, 0.1% sodium azide and 0.004% saponin. Incubation was performed on a shaker for 2 days at 4Cfollowedby washes in 1% BSA/TBS 1x. Then, sections were incubated with 1.4 nm gold-conjugated goat anti-guinea pig IgG antibody (Fab fragment, 1:100, #2055, Nanoprobes, Inc., Yaphank, NY, USA). They were also incubated with biotinylated anti-rabbit IgG antibody (1:200) diluted in 1% BSA/TBS 1x with 0.004% saponin on a shaker for 4 h at RT. Tissue was washed in 1% BSA/TBS 1x on a shaker at RT and incubated with ABC (1:50) prepared in washing solution for 1.5 h at RT. Sections were washed in 1% BSA/TBS 1x, kept overnight at 4C and post-fixed with 1% glutaraldehyde in TBS 1x (1 ml/well) for 12 min at RT. Then, they were washed in double distilled water and gold particles were silverintensified with the HQ Silver kit (#2012, Nanoprobes, Inc., Yaphank, NY, USA) in the dark for 12 min at RT. After intensification, the sections were washed first in double distilled water and then with PB 0.1 M (pH 7.4) for 30 min. The biotinylated antibody was revealed with 0.05% DAB prepared in PB 0.1 M (pH 7.4) containing 0.5% triton X-100 and 0.01% hydrogen peroxide for 3.5 min at RT, followed by several washes in PB 0.1 M (pH 7.4). They were osmicated (1% osmium tetroxide in PB 0.1 M, pH 7.4) in the dark for 20 min, washed in PB 0.1 M (pH 7.4), dehydrated in graded ethanol, cleared in propylene oxide, pre-embedded in 1:1 propylene oxide/Epon 812 resin on a shaker overnight at RT and embedded in pure Epon 812 resin. Immunogold labeling was visualized with a light microscope in sections containing the subiculum,andtissueportionswithgoodandconsistentCB 1 receptor labeling were identified and trimmed down for ultrathin sectioning. The procedure has already been described in detail (Bonilla-Del Rίoetal.,2019; Bonilla-Del Río et al., 2021; Gutiérrez-Rodríguez et al., 2018; Puente et al., 2019). Three to four semi-thin sections (0.7 μm-thick) were obtained with a histo-diamond knife (Diatome USA) and stained with 1% toluidine blue. To further standardize the conditions, only the 868 TERRADILLOS ET AL. 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License first five ultrathin sections were cut (50 nm-thick) with an ultradiamond knife (Diatome USA), collected onto nickel mesh grids and counterstained with 2.5% lead citrate for 20 min at RT. Electron micrographs were randomly taken with a Hamamatsu FLASH digital camera inserted in a transmission electron microscope (JEOL JEM 1400 Plus). Sampling was always carefully and accurately done using the following anatomical coordinates to delimit the subiculum: interaural, 0.40/0.00 mm; bregma, 3.40/3.80 mm (Franklin & Paxinos, 2008). 2.6 |Antibodies specificity Experiments were always performed under the same conditions. In addition, negative controls omitting the primary antibodies were done. Furthermore, the CB 1 receptor antibody was tested in CB 1 /brain tissue (Marsicano et al., 2002) by double pre-embedding immunogold (CB 1 ) and immunoperoxidase (Iba1 or GLAST) method for electron microscopy (Figure 1). The anti-A522 (EAAT1 [GLAST]) antibody (Ab#314) targeting the C-terminal residues 522–541 of rat EAAT1 (Hu et al., 2020) was used to identify astrocytic compartments. GLAST was restricted to astrocytes and localized intracellularly with no detectable labeling in nerve terminals, as previously described (Lehre et al., 1995; Schmitt et al., 1997). Furthermore, GLAST distribution is very similar in rodents and humans (Li et al., 2012), so the use of anti-GLAST antibody is a good approach to label astrocytes in the AD model. The specificity of the Iba1 antibody has been confirmed in previous studies (Delcambre et al., 2016; Szabo & Gulya, 2013). GLAST can be expressed in both microglia and astrocytes under certain conditions (Beschorner et al., 2007). Experiments were conducted to figure this out. Strikingly, GLAST and Iba1 were not seen to colocalize in a subicular area of 3506 μm 2 analyzed in CB 2EGFP/f/f and of 4743 μm 2 in 5xFAD/CB 2EGFP/f/f (data not shown). Taken together, we could reasonably conclude that GLAST and Iba1 are selective markers for astrocyte and microglia, respectively, in the 6.5–7.5 month-oldCB 2EGFP/f/f and 5xFAD/CB 2EGFP/f/f mice studied. 2.7 |Quantitative and statistical assessment To ensure homogeneous labeling between all samples, only the first 1.5 μm from the section surface of each specimen were considered for the analysis. Area, perimeter, number of processes and CB 1 receptor expression in astrocytes, were studied in 5596 μm 2 of five CB 2EGFP/f/f , and in 7681 μm 2 of seven 5xFAD/CB 2EGFP/f/f mice. In addition, the area, perimeter and number of microglial processes were measured in 7900 μm 2 of seven CB 2EGFP/f/f , and in 10,793 μm 2 of 10 5xFAD/CB 2EGFP/f/f mice. For the study of CB 1 receptors in microglia, 6345 μm 2 in five CB 2EGFP/f/f and 7268 μm 2 in seven 5xFAD/ CB 2EGFP/f/f mice, were analyzed. CB 1 receptor labeling in astrocytes and microglia was assessed in GLASTand Iba1-immunopositive processes, respectively. The proportion of cell compartments with CB 1 receptor signal was then tabulated. Positive labeling was considered when at least one immunoparticle was within 30 nm of the membrane studied. CB 1 receptor density (particles/μm membrane) was also determined by counting gold particles in the positive compartments. Membrane length (perimeter) was measured with the Image-J software (NIH; SCR_003070). All values were given as mean ± S.E.M. using a statistical software package (GraphPad Prism 8, SCR_002798, GraphPad Software, Inc., San Diego, USA). The normality test (Kolmogorov–Smirnov normality test) was always applied before statistics was done. Data were analyzed by non-parametric or parametric tests: Mann–Whitney U test or Student's Unpaired t-test (*p< .05). Minor contrast and brightness adjustments were made to the figures using Adobe Photoshop (Adobe Photoshop, SCR_014199, CS3, Adobe Systems, San Jose, CA, USA) and Gimp (GNU Image Manipulation Program, SCR_003182). 3|RESULTS 3.1 |Glial morphology in the subiculum of CB 2EGFP/f/f and 5xFAD/CB 2EGFP/f/f mice Changes in staining density were detected in microglial and astrocytic cells (Figure 2a–d). Microglia identified by Iba1 in CB 2EGFP/f/f FIGURE 1 Subiculum of CB 1 knock out mice (CB 1 /). Double immunogold and immunoperoxidase method for electron microscopy. Antibodies were tested in CB 1 /mice (a–d). Simultaneous labeling for CB 1 (gold) and Iba1 (a and b; DAB immunodeposits in brown) or GLAST (c and d; DAB in pink). No CB 1 receptor signal is observed. Scale bars: 50 nm TERRADILLOS ET AL.869 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License (Figure 2a,a1) occupied a much larger area in 5xFAD/CB 2EGFP/f/f (Figure 2b,b1). This increase seemed to relate to both the number of microglial cells and the thickness of their processes (Figure 2a1,b1). In the electron microscope, only scattered microglial processes were observed in controls (Figure 3a,b), while numerous Iba1-positive processes surrounding plaques (Figure 3c,f) and dystrophic neurites (Figure 3c1,c3-f) were seen in 5xFAD/CB 2EGFP/f/f mice. The area of the microglial processes was significantly increased in 5xFAD/ CB 2EGFP/f/f (0.3229 ± 0.05282 μm 2 ; ***p< .0001) relative to CB 2EGFP/f/f (0.1000 ± 0.01754 μm 2 ;Figure5a), as it was the total microglial area per sample normalized to 100 μm 2 (5xFAD/CB 2EGFP/f/f : 2.074 ± 0.5156 μm 2 ; ***p< .0001; CB 2EGFP/f/f : 0.3485 ± 0.06955 μm 2 ; Figure 5b). In addition, a significant increase in the perimeter of the microglial prolongations was detected in 5xFAD/CB 2EGFP/f/f (2.200 ± 0.1248 μm; ***p< .0001) versus CB 2EGFP/f/f (1.260 ± 0.1036 μm; Figure 5c) and reflected in the perimeter of the total microglial processes per sample normalized to 100 μm 2 (15.73 ± 1.929 μmin5xFAD/ CB 2EGFP/f/f ; 4.334 ± 0.5345 μminCB 2EGFP/f/f ; ***p< .0001; Figure 5d). Finally, significant changes were also noticed in the number of microglial processes (7.260 ± 0.6304/100 μm 2 in 5xFAD/CB 2EGFP/f/f ; 3.413 ± 0.4092/100 μm 2 in CB 2EGFP/f/f ;***p< .0001; Figure 5e). As to astrocytes, GLAST staining seen in CB 2EGFP/f/f (Figure 2c,c1) was more intense in 5xFAD/CB 2EGFP/f/f (Figure 2d,d1). By using GLASTDAB, the area, perimeter and number of astrocytic elements were analyzed in the electron microscope (Figure 4). Astrocytic processes surrounded dystrophic neurites and plaques in 5xFAD/CB 2EGFP/f/f (Figure 4b–d) and showed a significant increase in their area (5xFAD/ CB 2EGFP/f/f : 0.2598 ± 0.01853 μm 2 ;CB 2EGFP/f/f : 0.1565 ± 0.006515 μm 2 ; ***p< .0001; Figure 5a). However, no differences were observed in the total area per sample occupied by astrocytic processes normalized to 100 μm 2 (5xFAD/CB2 EGFP/f/f : 8.993 ± 0.8664 μm 2 ;CB2 EGFP/f/f : 7.415 ± 0.6552 μm 2 ;p: .1711; Figure 5b). There was also a great increase in the perimeter of the astrocytic processes (5xFAD/CB 2EGFP/f/f : 2.833 ± 0.08486 μm; CB 2EGFP/f/f : 2.116 ± 0.04741 μm; ***p< .0001; Figure 5c). Nevertheless, no differences in the total perimeter of astrocytic processes persampleweredetected(5xFAD/CB 2EGFP/f/f : 99.89 ± 8.087 μm; CB 2EGFP/f/f : 100.1 ± 6.811 μm; p: .9820; Figure 5d). Consistent with these results, a significantdecrease in the numberof astrocytic processes in 5xFAD/CB 2EGFP/f/f (35.52 ± 2.661 per 100 μm 2 )relativetoCB 2EGFP/f/f (47.33 ± 2.709 per 100 μm 2 ) was noticed (**p: .0036; Figure 5e). Overall, our analyses in the subiculum of the Alzheimer's model reveal that the larger area of the microglia correlates with an increase in size and number of their processes, while astrocytic projections are fewer but bigger (Figure 5a–e). 3.2 |CB 1 receptors in CB 2EGFP/f/f and 5XFAD/ CB 2EGFP/f/f subiculum The CB 1 receptor labeling observed in CB 2EGFP/f/f (Figure 2e,e1) was more patchy in 5xFAD/CB 2EGFP/f/f showing many delimited round areas with much lower or insignificant staining, probably corresponding to neuritic plaques surrounded by a neuropil with more CB 1 receptor immunoreactivity (Figure 2f,f1). FIGURE 2 Subiculum of CB 2EGFP/f/f and 5xFAD/CB 2EGFP/f/f mice showing Iba1, GLAST and CB 1 immunostaining. Avidin-biotin peroxidase method for light microscopy. Iba1 in CB 2EGFP/f/f (a, a1) is drastically increased in 5xFAD/CB 2EGFP/f/f microglia (b, b1). GLAST in CB 2EGFP/f/f (c, c1) is weaker than in 5xFAD/CB 2EGFP/f/f (d, d1). Dense CB 1 receptor staining in CB 2EGFP/f/f (e, e1) changes to a weaker and more patchy appearance in 5xFAD/CB 2EGFP/f/f (f, f1). Scale bars: 200 μm(a–f) and 50 μm (a1–f1) 870 TERRADILLOS ET AL. 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 3.3 |Microglial CB 1 receptor localization in CB 2EGFP/f/f and 5XFAD/CB 2EGFP/f/f subiculum CB 1 receptor particles were localized to membranes of Iba1-positive microglial processes in both CB 2EGFP/f/f (Figure 3b2) and 5xFAD/ CB 2EGFP/f/f mice (Figure 3c–f). The analysis revealed a significant increase in CB 1 -positive microglial processes in 5xFAD/CB 2EGFP/f/f (0.9942 ± 0.1259 CB 1+ processes/100 μm 2 ) compared to CB 2EGFP/f/f (0.3254 ± 0.07758 CB 1+ processes/100 μm 2 ; ***p< .0001; Figure 6a, left). In addition, a strike increase in the proportion of CB 1 -positive FIGURE 3 Double pre-embedding immunogold (CB 1 ) and immunoperoxidase (Iba1) method for electron microscopy in the subiculum of CB 2EGFP/f/f and 5xFAD/CB 2EGFP/f/f mice. In CB 2EGFP/f/f , a few slender Iba1 positive microglial elements are observed (a, b) (DAB immunodeposits in brown). However, thick processes of reactive microglia appear in 5xFAD/CB 2EGFP/f/f (c–f) surrounding plaques (in purple; c, f) and dystrophic neurites (in turquoise; c–f). Notice membrane CB 1 particles (orange arrows) in microglial processes of CB 2EGFP/f/f (b2) and 5xFAD/CB 2EGFP/f/f (c–f), with particular abundance in 5xFAD/CB 2EGFP/f/f .CB 1 receptor labeling is also in membranes of excitatory terminals (green arrows and profiles in a1, b2 and d1), inhibitory terminals (yellow arrows and profiles in b1, d2) and mitochondria (blue arrows and profiles in a1 and d2), in both mutants. Scale bars: 2 μm TERRADILLOS ET AL.871 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License elements was detected in 5xFAD/CB 2EGFP/f/f mice (6.27 ± 1.15%; CB 2EGFP/f/f : 3.79 ± 2.10%; **p: .0033; Figure 6b, left). However, CB 1 receptor density in the positive microglial processes was significantly reduced in 5xFAD/CB 2EGFP/f/f (69.44 ± 7577 particles/100 μmof membrane; CB 2EGFP/f/f : 135.5 ± 24.78/100 μm of membrane; **p: .0023; Figure 6c, left). Furthermore, significant differences in the total number of microglial CB 1 particles per 100 μm 2 were observed between 5xFAD/CB 2EGFP/f/f (1.31 ± 0.18 particles) and CB 2EGFP/f/f FIGURE 4 Double pre-embedding immunogold (CB 1 receptor) and immunoperoxidase (GLAST) method for electron microscopy in the subiculum of CB 2EGFP/f/f and 5xFAD/CB 2EGFP/f/f mice. GLAST-positive astrocytic processes (DAB immunodeposits in pink) seen in CB 2EGFP/f/f (a) are thicker in 5xFAD/CB 2EGFP/f/f (b–d). Observe astrocytic elements surrounding dystrophic neurites (turquoise in b–d) and close to a plaque (purple in d) in the 5xFAD/CB 2EGFP/f/f .CB 1 particles (red arrows) localize to plasma membranes of GLAST-positive astrocytic processes in CB 2EGFP/f/f (a) and 5xFAD/CB 2EGFP/f/f (b–d). Typical CB 1 receptor labeling is also observed in membranes of excitatory terminals (green arrows and profiles in a2, b1), inhibitory terminals (yellow arrows and profiles in a2, b2) and mitochondria (blue arrows and profiles in a2), in both CB 2EGFP/f/f and 5xFAD/CB 2EGFP/f/f . Scale bars: 2 μm 872 TERRADILLOS ET AL. 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License (0.48 ± 0.13 particles; ***p: .0009; Figure 6d, left). Finally, there were not differences in the number of CB 1 particles per positive microglial process between both mutants (5xFAD/CB 2EGFP/f/f : 1.333 ± 0.08347 particles/process; CB 2EGFP/f/f : 1.450 ± 0.2112 particles/process; p: .7736; Figure 6e, left). 3.4 |Astroglial CB 1 receptor localization in CB 2EGFP/f/f and 5XFAD/CB 2EGFP/f/f subiculum The CB 1 receptor was localized to membranes of GLAST-positive astrocytic processes in both mutants (Figure 4a–d), as previously reported (Bonilla-Del Rίo et al., 2019; Bonilla-Del Río et al., 2021; Bosier et al., 2013; Gutiérrez-Rodríguez et al., 2018; Han et al., 2012). No significant differences were detected in the number of CB 1 - positive astrocytic prolongations between both mice (CB 2EGFP/f/f : 8.661 ± 0.8977 CB 1+ processes/100 μm 2 ;5xFAD/CB 2EGFP/f/f :7.967± 1.224 CB 1+ processes/100 μm 2 ;p: .3094; Figure 6a, right). Likewise, the percentage of CB 1 -positive astrocytic branches was statistically similar between 5xFAD/CB 2EGFP/f/f (21.24 ± 2.37%) and CB 2EGFP/f/f (17.75 ± 1.21%; p: .2303; Figure 6b, right). There was neither differences in CB 1 receptor density in the astrocytic positive processes (5xFAD/CB 2EGFP/f/f : 29.15 ± 2.220 particles/100 μmofmembrane; CB 2EGFP/f/f : 37.57 ± 2.970 particles/100 μm of membrane; p: .2209; Figure 6c, right), nor in the number of astrocytic CB 1 particles per 100 μm 2 (5xFAD/CB 2EGFP/f/f : 12.78 ± 2.174 particles; CB 2EGFP/f/f : 11.63 ± 1.265; p: .6716; Figure 6d, right). However, the number of CB 1 particles per positive astrocytic processes was significantly higher in 5xFAD/CB 2EGFP/f/f (1.603 ± 0.05081 particles/process) than in CB 2EGFP/f/f (1.343 ± 0.03909 particles/process; ***p: .0005; Figure 6e,right). Altogether, the number of microglial processes expressing CB 1 receptors increases and the larger astrocytic profiles have more CB 1 receptors in the subiculum of the Alzheimer's mouse model (Figure 6a–e). FIGURE 5 Morphological parameters of microglia and astrocytes in the subiculum of CB 2EGFP/f/f and 5xFAD/ CB 2EGFP/f/f mice. (a) Microglial and astrocytic processes area. (b) Glial area (microglia and astrocytes) normalized to 100 μm 2 . (c) Microglial and astrocytic processes perimeter. (d) Glial perimeter (microglia and astrocytes) normalized to 100 μm 2 . (e) Number of glial processes (microglia and astrocytes) in 100 μm 2 . Data were analyzed by non-parametric or parametric tests (Mann–Whitney U-test or Student's t-test). Mann–Whitney Utest or Student's t-test. p< .05*; p< .01**; p< .001***; p< .0001****. All data are represented as mean ± SEM TERRADILLOS ET AL.873 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 4|DISCUSSION We detected in the electron microscope the presence of plaques and a multitude of dystrophic neurites in the 5xFAD/CB 2EGFP/f/f mice that confirms the usefulness of this animal model for studying the pathophysiology of AD (Ruiz de Martín Esteban et al., 2022). We also observed an overt microglial and astrocytic reactivity with an increase in the area and perimeter of their processes, demonstrating the existence of significant alterations in the subicular cytoarchitecture. Then we studied the expression of the major cannabinoid CB 1 receptor in glial cells in the subiculum of 5xFAD/ CB 2EGFP/f/f and CB 2EGFP/f/f mice. The main findings were that CB 1 receptor expression conspicuously changes in microglial cells but receptor density remains steady in astrocytes despite the reactivity of the astrocytic processes in the AD mouse. Not least, the localization of CB 1 receptors in microglial processes in the subiculum of controls and closely surrounding amyloid plaques and dystrophic neurites in the subiculum of the AD model, supports the presence of CB 1 in microglia. The discreet amount of CB 1 receptors in astrocytes has been revealed accurately by immuno-electron microscopy (Bonilla-Del Río et al., 2021; Gutiérrez-Rodríguez et al., 2018; Puente et al., 2019), a technique that has also been proven in this study to be optimal for the localization of CB 1 receptors in microglia. FIGURE 6 Statistical assessment of the CB 1 receptor localization in subicular astrocytes and microglia of CB 2EGFP/f/f and 5xFAD/CB 2EGFP/f/f mice. (a) Number of microglial (left) and astrocytic (right) CB 1 positive processes per 100 μm 2 . (b) Percentage of CB 1 positive microglial (left) and astrocytic (right) processes. (c) CB 1 density in positive microglial (left) and astrocytic (right) elements per 100 μm. (d) Microglial (left) and astrocytic (right) CB 1 particles per 100 μm 2 . (e) CB 1 receptor labeling per microglial (left) and astrocytic (right) process. Data were analyzed by nonparametric or parametric tests (Mann–Whitney U-test or Student's t-test). Mann–Whitney U-test or Student's t-test. p< .05*; p< .01**; p< .001***; p< .0001****. All data are represented as mean ± SEM 874 TERRADILLOS ET AL. 10981136, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/glia.24312 by Readcube (Labtiva Inc.), Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License