Imbalance of Endocannabinoid/Lysophosphatidylinositol Receptors Marks the Severity of Alzheimer’s Disease in a Preclinical Model: A Therapeutic Opportunity

Simple Summary Alzheimer’s disease (AD) remains a major challenge for the healthcare system worldwide and, to date, no curative treatment is available. This disease is an irreversible progressive dementia that harms memory and cognitive functions, weakening the ability to carry out tasks by themselves. Among the potential targets for developing innovative therapies for AD, the endocannabinoid system has aroused much interest in the scientific community, since it is involved in multiple processes related to AD pathology. A major challenge to understand the role of the cannabinoid system in AD is to characterize how it contributes to the expression of a specific phenotype, from neuropathology to behavior. In the present study, we addressed this challenge by evaluating the expression of the endocannabinoid system in a transgenic mouse model of AD, bearing five familial AD mutations. Our data suggest that there is an association between the cannabinoid receptors and both the cognitive function and inflammatory response characterizing the disease. Moreover, this association is aggravated by genetic factors. From these data, the expression of endocannabinoid and G protein-coupled 55 receptors (GPR55), and endocannabinoid-related enzymes might be candidate markers for the detection of the severity of this neurodegenerative disease, eventually arising as potential therapeutic targets capable of modifying the course of this incapacitating dementia. Abstract Alzheimer’s disease (AD) is the most common form of neurodegeneration and dementia. The endocannabinoid (ECB) system has been proposed as a novel therapeutic target to treat AD. The present study explores the expression of the ECB system, the ECB-related receptor GPR55, and cognitive functions (novel object recognition; NOR) in the 5xFAD (FAD: family Alzheimer’s disease) transgenic mouse model of AD. Experiments were performed on heterozygous (HTZ) and homozygous (HZ) 11 month old mice. Protein expression of ECB system components, neuroinflammation markers, and β-amyloid (Aβ) plaques were analyzed in the hippocampus. According to the NOR test, anxiety-like behavior and memory were altered in both HTZ and HZ 5xFAD mice. Furthermore, both animal groups displayed a reduction of cannabinoid (CB1) receptor expression in the hippocampus, which is related to memory dysfunction. This finding was associated with indirect markers of enhanced ECB production, resulting from the combination of impaired monoacylglycerol lipase (MAGL) degradation and increased diacylglycerol lipase (DAGL) levels, an effect observed in the HZ group. Regarding neuroinflammation, we observed increased levels of CB2 receptors in the HZ group that positively correlate with Aβ’s accumulation. Moreover, HZ 5xFAD mice also exhibited increased expression of the GPR55 receptor. These results highlight the importance of the ECB signaling for the AD pathogenesis development beyond Aβ deposition.

the ECB system components and neuroinflammatory markers, as well as the β-amyloid accumulated in the hippocampus. The outcome of this study aims to highlight the association of the ECB with the 5xFAD phenotype as a basis for establishing the potential utility of the ECB system as a novel therapeutic target to treat AD.

Animals and Ethics Statement
We used 5xFAD (FAD: familial Alzheimer's disease) APP/PS1 double transgenic mice that co-express and co-inherit FAD mutant forms of human APP (the Swedish mutation: K670N, M671L; the Florida mutation: I716V; the London mutation: V717I) and PS1 (M146L; L286V) transgenes under transcriptional control of the neuron-specific mouse Thy-1 promoter (Tg6799 line) [39][40][41]. These mice co-express mutations of the human genes encoding the amyloid protein precursor (APP) and presinilin1 (PS1), increasing the production of 42-amino-acid β-amyloid (Aβ42). All experiments were done with females and males at 11 months of age and were realized in compliance with the ARRIVE guidelines [48] and in concordance with the European Communities Council Directives 2010/63/EU, Regulation (EC) No. 86/609/ECC (24 November 1986) and Spanish National and Regional Guidelines for Animal Experimentation (Real Decreto 53/2013). 5xFAD mice used were heterozygous (HTZ) (n = 7; four males and three females) and homozygous (HZ) (n = 12; seven males and five females) concerning the transgene, and nontransgenic wild-type (No-Tg) (n = 13; five males and eight females) littermate mice served as controls. We did not perform a sex analysis because of the reduced number of samples per genotype and because our main goal was to evaluate the differences according to heterozygous and homozygous conditions. Experimental protocols were approved by the Local Ethical Committee for Animal Research of the University of Malaga (CTS-8221, July 2016). Accordingly, all efforts were made to minimize animal suffering and to reduce the number of animals used.

Novel Object Recognition Test (NOR)
The NOR test was used to analyze long-term memory. The test ( Figure 1) consisted of four open-field apparatus (40 × 40 cm, made of gray Plexiglas), and mice were placed in the center of the arena to explore freely for 5 min (habituation trial). Then, animals were returned to their cages for 5 min. In the acquisition trial, two identical objects were presented on the up and down corners (the left-right position was counterbalanced) of the apparatus, and mice explored both objects for 10 min. Twenty-four hours later, mice were returned to the arena and were presented with one familiar object and a novel one that allowed them to explore for 5 min [49]. We analyzed locomotion (distance moved; cm) and anxiety-like behavior (time spent in the center of the apparatus). During acquisition and retention trials, the total time spent exploring objects (i.e., touching them with the nose or forepaws, analyzed observationally) was recorded. To assess cognitive performance, the percentage of novelty preference was calculated as follows: (the time spent exploring novel object / the time spent exploring both objects) × 100 [50]. the β-amyloid accumulated in the hippocampus. The outcome of this study aims to highlight the association of the ECB with the 5xFAD phenotype as a basis for establishing the potential utility of the ECB system as a novel therapeutic target to treat AD.

Animals and Ethics Statement
We used 5xFAD (FAD: familial Alzheimer's disease) APP/PS1 double transgenic mice that coexpress and co-inherit FAD mutant forms of human APP (the Swedish mutation: K670N, M671L; the Florida mutation: I716V; the London mutation: V717I) and PS1 (M146L; L286V) transgenes under transcriptional control of the neuron-specific mouse Thy-1 promoter (Tg6799 line) [39][40][41]. These mice co-express mutations of the human genes encoding the amyloid protein precursor (APP) and presinilin1 (PS1), increasing the production of 42-amino-acid β-amyloid (Aβ42). All experiments were done with females and males at 11 months of age and were realized in compliance with the ARRIVE guidelines [48] and in concordance with the European Communities Council Directives 2010/63/EU, Regulation (EC) No. 86/609/ECC (24 November 1986) and Spanish National and Regional Guidelines for Animal Experimentation (Real Decreto 53/2013). 5xFAD mice used were heterozygous (HTZ) (n = 7; four males and three females) and homozygous (HZ) (n = 12; seven males and five females) concerning the transgene, and nontransgenic wild-type (No-Tg) (n = 13; five males and eight females) littermate mice served as controls. We did not perform a sex analysis because of the reduced number of samples per genotype and because our main goal was to evaluate the differences according to heterozygous and homozygous conditions. Experimental protocols were approved by the Local Ethical Committee for Animal Research of the University of Malaga (CTS-8221, July 2016). Accordingly, all efforts were made to minimize animal suffering and to reduce the number of animals used.

Novel Object Recognition Test (NOR)
The NOR test was used to analyze long-term memory. The test ( Figure 1) consisted of four openfield apparatus (40 × 40 cm, made of gray Plexiglas), and mice were placed in the center of the arena to explore freely for 5 min (habituation trial). Then, animals were returned to their cages for 5 min. In the acquisition trial, two identical objects were presented on the up and down corners (the left-right position was counterbalanced) of the apparatus, and mice explored both objects for 10 min. Twenty-four hours later, mice were returned to the arena and were presented with one familiar object and a novel one that allowed them to explore for 5 min [49]. We analyzed locomotion (distance moved; cm) and anxiety-like behavior (time spent in the center of the apparatus). During acquisition and retention trials, the total time spent exploring objects (i.e., touching them with the nose or forepaws, analyzed observationally) was recorded. To assess cognitive performance, the percentage of novelty preference was calculated as follows: (the time spent exploring novel object / the time spent exploring both objects) × 100 [50]. 24h Figure 1. Schematic representation of novel object recognition assay. Following habituation trial in the empty boxes for 5 min, mice were allowed to explore an identical pair of objects placed in the boxes for 10 min as the acquisition trial. After a 24 h stay in the home cage, the mice were returned to the arena where two objects, one familiar and one novel, were placed. The time that mice spent exploring the two Figure 1. Schematic representation of novel object recognition assay. Following habituation trial in the empty boxes for 5 min, mice were allowed to explore an identical pair of objects placed in the boxes for 10 min as the acquisition trial. After a 24 h stay in the home cage, the mice were returned to the arena where two objects, one familiar and one novel, were placed. The time that mice spent exploring the two objects was recorded. The arrow indicates the novel object in the box. Squares symbolize the object number one (familiar object); the circle symbolizes the object number two (novel object).

Brain Protein Extract
All mice were sacrificed at 11 months of age. The brain was removed and bisected down the midline; one hemibrain was used for histological procedures and stored in 4% paraformaldehyde (PFA) and the other hemibrain was kept in dry ice for storage at −80 • C for biochemical analysis. Hippocampus samples (17 mg per sample) were dissected and homogenized in 1 mL of cold RIPA lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.5% NaDOC, 1 mM EDTA, 1% Triton, 0.1% SDS, 1 mM Na 3 VO 4 , 1 mM NaF) supplemented with a protease cocktail (Hoffmann-La Roche, Basel, Switzerland). The suspension was incubated for 2 h at 4 • C, followed by centrifugation at 12,000 rpm for 15 min at 4 • C. The supernatant was transferred to a new clean centrifuge tube, and the Bradford colorimetric method was used to determine the concentration of the total protein. The protein extracts were diluted 1:1 in loading buffer (DTT 2X) and heated for 5 min at 99 • C before being subjected to electrophoresis.

Plaque and Microglial Cell Activation and Neuron Quantification in the Hippocampus
The number of labeled Aβ plaques, as well as GFAP-and Iba1-positive cells, was quantified in the hippocampus. Images were acquired with digital Camera DP70 (Olympus Iberia, S.A., Barcelona, Spain) connected to a microscope Olympus BX41. For immunostaining quantification, ImageJ software was used [52]. The images were binarized to 16 bit black, and a fixed intensity threshold was applied for each immunostaining. Eight mice per group and three sections per mouse at three different hippocampal levels were used.

Statistical Analyses
Statistical analysis was conducted in GraphPad Prism, version 8 (GraphPad Software, Inc., La Jolla, CA, USA). The Shapiro-Wilk test was used to assess normal distribution of data. Levene's test was used to analyze the assumption of homogeneity of variance. One-way and two-way ("genetic × object") analysis of variance (ANOVA) was assessed for behavioral tests, Western blot, and immunostaining quantification followed by a Tukey's multiple-comparisons test. The post hoc tests were conducted only if F in ANOVA achieved a p-value less than 0.05 and there was no statistically significant variance inhomogeneity. The analysis of two single groups was performed using Student's unpaired t-test. Correlation analysis was performed computing the value of the Pearson correlation coefficient, r (its value ranges from −1 to +1). All data were expressed as the mean ± standard error of the mean (SEM), and p-values less than 0.05 were considered significant.

Both Heterozygous and Homozygous 5xFAD Transgenic Mice Have Impairment in Novel Object Recognition
First, we evaluated long-term memory using the novel object recognition task. In the habituation trial, homozygous mice showed less time spent in the center of the arena compared to their heterozygous and nontransgenic littermates (one-way ANOVA: F (2,22) = 4.110, p = 0.0298; Tukey's test: p ≤ 0.05; Figure 2a), suggesting an anxiety-like behavior. In the acquisition trial, all animals explored both objects without presenting a marker preference for any ones (two-way ANOVA: "genotype" effect: F (2,44) = 0.8665, p = 0.4275; "object" effect: F (1,44) = 0.7382, p = 0.3949, Figure 2b), and there was no significant difference between genotypes in the total exploration time (one-way ANOVA: F (2,22) = 0.5010, p = 0.6127; Figure 2c). However, in the retention trial, where one of the objects was novel, both heterozygotes and homozygotes 5xFAD mice exhibited a lower percentage of novelty preference compared to control mice (one-way ANOVA: F (2,23) = 5.940, p = 0.0083; Tukey's test: p ≤ 0.05; Figure 2d). No significant differences were observed in locomotion (distance moved) during the habituation or retention trial between genotypes (one-way ANOVA: habituation trial: F (2,22)

Exploration time Acquisition trial
Total distance moved Figure 2. Impaired memory in both heterozygous (HTZ) and homozygous (HZ) 5xFAD (FAD: family Alzheimer's disease) transgenic mice at 11 months of age on the novel recognition memory. (a) HZ 5xFAD mice spent less time in the center throughout the habituation trial. No differences were found (b) in exploring both objects or (c) in the total exploration time. (d) Both HTZ and HZ 5xFAD mice exhibited impaired memory in the retention trial. (e) There were no differences in locomotion between genotypes in the habituation and retention trial, but HZ 5xFAD mice exhibited hyperactivity in the acquisition trial compared to No-Tg and HTZ. One-way ANOVA and Tukey's test were performed: (*) p < 0.05 vs. nontransgenic wild-type (No-Tg) group; ($) p < 0.05 between HTZ and 5xFAD group. All data are displayed as the mean ± standard error of the mean (SEM).

Hippocampal Alteration of the Expression of CB1, CB2, and GPR55 Receptors in Homozygous 5xFAD Transgenic Mice
We next analyzed the expression of the cannabinoid receptors (CB1, CB2, and GPR55 receptors) by Western blotting in the hippocampus of 5xFAD transgenic mice, as it is a key structure for learning and memory function. Figure 3a shows the quantification of Western blot membranes (Figure 3b) after incubating with the selected antibodies labeling those lipid receptors. Intriguingly, HTZ and HZ 5xFAD Figure 2. Impaired memory in both heterozygous (HTZ) and homozygous (HZ) 5xFAD (FAD: family Alzheimer's disease) transgenic mice at 11 months of age on the novel recognition memory. (a) HZ 5xFAD mice spent less time in the center throughout the habituation trial. No differences were found (b) in exploring both objects or (c) in the total exploration time. (d) Both HTZ and HZ 5xFAD mice exhibited impaired memory in the retention trial. (e) There were no differences in locomotion between genotypes in the habituation and retention trial, but HZ 5xFAD mice exhibited hyperactivity in the acquisition trial compared to No-Tg and HTZ. One-way ANOVA and Tukey's test were performed: (*) p < 0.05 vs. nontransgenic wild-type (No-Tg) group; ($) p < 0.05 between HTZ and 5xFAD group. All data are displayed as the mean ± standard error of the mean (SEM).

Hippocampal Alteration of the Expression of CB1, CB2, and GPR55 Receptors in Homozygous 5xFAD Transgenic Mice
We next analyzed the expression of the cannabinoid receptors (CB1, CB2, and GPR55 receptors) by Western blotting in the hippocampus of 5xFAD transgenic mice, as it is a key structure for learning and memory function. Figure 3a shows the quantification of Western blot membranes ( Figure 3b) after incubating with the selected antibodies labeling those lipid receptors. Intriguingly, HTZ and HZ 5xFAD animals exhibited a clear reduction in the expression of CB1 receptors (one-way ANOVA: F (2,14) = 7.948, p = 0.0049; Tukey's test: p ≤ 0.05; Figure 3a,b) compared to the No-Tg mice, suggesting that this decrease may be related to memory dysfunction. Regarding the expression of CB2 receptors, there was a remarkable increase in the expression in homozygous 5xFAD mice compared to the No-Tg group (one-way ANOVA: F (2,14) = 3.868, p = 0.0460; Tukey's test: p ≤ 0.05; Figure 3a,b).    In addition to CB1 and CB2 receptors, the GPR55 receptor is proposed as a cannabinoid receptor involved in the modulation of neuroinflammation [31]. As shown in the Western blot analysis, GPR55 is highly expressed in homozygous 5xFAD mice (one-way ANOVA: F (2,13) = 15.15, p = 0.0004; Tukey's test: p ≤ 0.01; Figure 3a,b) compared to HTZ and No-Tg mice. We also performed an immunohistochemistry assay to analyze the expression of GPR55 in the hippocampus, especially in the HZ 5xFAD mice (Figure 3c). Qualitatively, the higher presence of GPR55 was observed in HZ 5xFAD mice in both dentate gyri, CA1 and CA3, compared to the control group. These results validated the data found by Western blot (see Figure S1, for GPR55 immunofluorescence staining).
To establish a relationship between the main changes found in endocannabinoid receptors and the memory-related behavioral test, we performed correlation studies (Table 1) Table 1. Pearson's correlation between endocannabinoid receptors (CB1, CB2, and GPR55) and behavioral parameters related to memory measured by novel object recognition test in 5xFAD transgenic mice.

Alteration in the Endocannabinoid Production and Degradation Pathways in Homozygous 5xFAD Transgenic Mice
We performed a Western blot analysis to study the expression of the enzymes DAGL and NAPE-PLD involved in the biosynthesis of the endocannabinoids 2-AG and AEA. Heterozygous 5xFAD transgenic mice increased the expression of DAGLα (one-way ANOVA: F (2,14)  Regarding the degradation enzymes, we did not observe differences between genotype in the expression of fatty acid amide hydrolase (FAAH) enzyme (one-way ANOVA: F (2,14) = 0.7668, p = 0.4830; Figure 4b,d). However, the expression of monoacylglycerol lipase (MAGL) enzyme was decreased in the homozygous 5xFAD group (one-way ANOVA: F (2,13) = 4.059, p = 0.0427; Tukey's test: p ≤ 0.05; Figure 4b,d) compared to the control group.
To study whether there were more production or degradation enzymes, we performed a ratio between both enzymes. First, we compared the 2AG production and degradation enzymes (DAGLα versus MAGL) and it showed an increase in DAGLα in homozygous 5xFAD transgenic mice (one-way ANOVA: F (2,13) = 3.818, p = 0.0496; Tukey's test: p ≤ 0.05; Figure 4c). On the other hand, the ratio of AEA production and degradation enzymes (NAPE-PLD versus FAAH) did not show a significant difference among groups (one-way ANOVA: F (2,14) = 0.7735, p = 0.4802; Figure 4c).
The potential relationship of the ECB receptors (CB1, CB2, and GPR55) and neuroinflammation response exhibited by animals was explored ( Table 2

Neuroinflammatory Response in Both Heterozygous and Homozygous 5xFAD Transgenic Mice-Stronger in the Homozygous Group
To evaluate the β-amyloid burden, the number of labeled Aβ plaques was quantified in the hippocampus of 5xFAD transgenic mice. Both HTZ and HZ 5xFAD transgenic mice exhibited a similar accumulation of β-amyloid in the hippocampus (Figure 6), and it was significantly higher than No-Tg mice which had an absence of plaques (one-way ANOVA: F (2,10) = 70.03, p < 0.0001; Tukey's test: p ≤ 0.001; Figure 6). In the same sense, we did not observe differences between heterozygous and homozygous 5xFAD transgenic mice in Aβ1-40 although it was significantly different from the control group (one-way ANOVA: F (2,11) = 66.57, p < 0.0001; Tukey's test: p ≤ 0.001; Figure 6). Interestingly, there was a higher accumulation of Aβ1-42 peptide in homozygous 5xFAD mice than heterozygous mice (one-way ANOVA: F (2,12) = 80.68, p < 0.0001; Tukey's test: p ≤ 0.001; Figure 6) suggesting a more severe neuropathological phenotype.

Discussion
The present study allowed us to holistically analyze the expression of the ECB system in the hippocampus of a genetic model of human AD, under conditions of hetero-and homozygosis. AD pathophysiology is well known for the accumulation of senile plaques and reactive gliosis, among other events that develop especially in the hippocampus [1,2]. The present data firmly support the recent publications on the association of the ECB system and AD [53], which adds a set of new biomarkers related to the severity of the phenotype to the classical neuropathological analysis. In this sense, the major findings of our study are the differences observed in the expression pattern of CB1, CB2, and GPR55 receptors, as well as the MAGL degrading enzyme, according to the transgenic load in the 5xFAD transgenic mouse model at 11 months of age. Here, our data show ongoing ECB system alterations in the hippocampus reflected by an elevated neuroinflammatory response induced by β-amyloid burden consistent with memory impairment aggravated by homozygosity. To our knowledge, we are the first to report an association between cannabinoid molecules and the neuropathological hallmarks of AD studied in the 5xFAD line, taking special consideration of the transgenic load: heterozygous versus homozygous condition. This is a relevant finding since the heterozygosity condition might allow space for analyzing disease-modifying factors (i.e., the impact of accelerating factors that promote more severe disease progression).
In the 5xFAD mice model (homozygous condition), it has been demonstrated that memory deficit occurs from 5-6 months of age [39] and increases substantially with age [46,47], caused by the accumulation of β-amyloid plaques. However, the biological mechanisms involved in the detriment of pathology are still unclear. Our data show that both HTZ and HZ 5xFAD mice exhibited a significant decrease in the percentage of novelty in NOR indicating hippocampus-dependent memory impairment at 11 months of age. Moreover, there are only a few studies that focused on heterozygosity as a useful tool to understand AD. In this context, Richard et al. (2015) informed that 5xFAD mice, when bred to homozygosity, presented a significant age-dependent motor phenotype and spatial reference memory assessed by Morris water maze compared to the heterozygous condition at 5 months of age [54]. Otherwise, our data revealed that HTZ 5xFAD mice exhibited an impaired memory in NOR at 11 months of age, and it could be associated with the enhanced number of Aβ deposits throughout the hippocampus, as shown by our results analyzed using immunohistochemistry. In this sense, it is important to note that the transgenic load affects the ECB system in a different way, exacerbating emotional and memory function in this transgenic mouse model of AD. Therefore, the 5xFAD mouse model (heterozygous and homozygous condition) could elucidate the role(s) of cannabinoid and noncannabinoid receptors in the development of AD. Potential experimental analysis of new cannabinoid receptor ligands, FAAH inhibitors, or even DAGL inhibitors might benefit from this approach with heterozygous animals.
Physiological changes in the brain, for instance, a dysregulation of the endocannabinoid receptors, may explain AD-related behavioral deficits. In the present study, we described that the increase in CB2 and GPR55 receptors in the hippocampus of HZ 5xFAD mice is involved in the anxiogenic response in the NOR test. Moreover, the reduction in CB1 together with the increase in GPR55 receptor, both aggravated by the transgenic load, was associated with hyperactivity when HZ 5xFAD mice were exposed to two identical objects in the acquisition trial of the NOR test. Furthermore, memory impairment was associated with this unbalanced expression of CB1, CB2, and GPR55 receptors found in HTZ and HZ 5xFAD mice. In the literature, recent evidence suggests that the stimulation of the CB2 receptor modulates neuronal function, as well as emotional behavior and memory formation [23]. Additionally, the upregulation of CB2 observed only in homozygosity is in line with previous publications that use the 5xFAD mice model [7], and it was corroborated in AD patients [21]. Regarding CB1, previous reports described that these receptors are highly expressed throughout the brain such as the hippocampus, with a notable presence on multiple neuronal populations (for a review, see [12]). In this context, data available suggest that CB1 receptors play an integral role in learning and memory showing a special relevance on anxiety-like behavior [14,55] and long-term (acquisition, consolidation, and retrieval) memory-relative responses [23,[56][57][58][59][60]. Therefore, pharmacological modulation of endocannabinoid receptors may be a potential target for the treatment of AD, and attention has been paid to CB2 receptor-specific agonists because of their lack of psychotropic properties compared to among the CB1 receptor agonists [61]. In this sense, several studies support that the pharmacological treatment with the specific CB2 agonist JWH-133 ameliorates cognitive function and long-term recognition memory decline in AD-model animals, such as AβPP/PS1 mice [27,30,62], and APP 2576 mice [29] at 6 and 11 months of age, respectively. Although these results are promising, detailed studies are still needed.
Regarding the noncannabinoid GPR55 receptor, we have to highlight that this receptor is primarily an LPI receptor, and we did not analyze the biochemical pathway for the synthesis or degradation of this bioactive lipid tightly related to endogenous cannabinoids. However, since GPR55 activity can be modulated by both endogenous and synthetic cannabinoids, we included it in the analysis. The GPR55 receptor has been previously described as being involved in spatial learning [33] and memory dysfunction [63], and GPR55 knockout mice showed impaired movement coordination [64]. Few studies investigated its pharmacological modulation to improve cognition [33], synapsis plasticity [65], and neural stem-cell proliferation [66], which may present a new target for the treatment of neurodegenerative diseases such as AD. An interesting hypothesis for the importance of GPR55 is not only its location on immune cells but specifically its intracellular location in the lysosomal compartment. Intracellular delivery of LPI can activate Ca 2+ release from these organelles through a GPR55-dependent mechanism. This calcium release might disrupt the pH homeostasis of the lysosomal compartment because of the well-known mutual interaction of Ca 2+ and H + in these organelles [67,68]. Since lysosomal de-acidification is a relevant process associated with AD [69], the GPR55 modulation of AD progression hypothesis has to be considered in futures studies. It is currently unknown whether GPR55 modulates Aβ production and trafficking because there are no studies with selective GPR55 ligands (most of the studies were done using nonselective ligands such as the GPR18/GPR55 ligand abnormal cannabidiol). In any case, the contribution of GPR55 to the neuropathology of AD is supported by additional reports suggesting that LPI deposits are enriched in the outer layer of the amyloid plaque, facilitating the interaction with microglia/macrophages and probably modulating neuroinflammation [70]. In this sense, it is important to note that the association of LPI/GPR55 with degenerative diseases displays growing relevance. As an example, recent studies identified a crucial role for LPI signaling in metabolic disorders leading to nonalcoholic steatohepatitis, a degenerative disorder of the liver that leads to steatosis and fibrosis in which both GPR55 and LPI are upregulated [71]. Our findings set in place its expression in the hippocampus as a marker of severity, helping to clearly differentiate HTZ from HZ animals. Its association with the emotional and memory performance of HZ animals is very interesting, as well as its inverse correlation with neuroinflammatory markers from microglia, suggesting a potential anti-inflammatory role. Taken together, these data expose the importance of stimulating or inhibiting ECB receptors to ameliorate behavioral changes related to AD such as anxiety, memory deficit, or coordination of locomotion.
Endocannabinoids, e.g., AEA or 2-AG, are synthesized and released by neurons acting as neurotransmitters. Nonetheless, ECB differentiates from other neurotransmitters because (i) retrograde signaling is their principal mode of acting as messengers [72], and (ii) they do not accumulate within synaptic vesicles [73]. Interestingly, neuronal damage boosts the production of ECB, providing a mechanism of protection against toxicity [74]. Once released upon demand, AEA and 2-AG activate the presynaptic cannabinoid receptors, and then they are rapidly inactivated by the action of specific degradation enzymes (FAAH for AEA, and MAGL for 2-AG) [75]. When analyzing production/degradation enzymes, we have seen that, overall, the activity of the enzymes involved in biosynthesis and degradation of ECBs is altered; there is an increase in production rather than in degradation in the HZ group. These findings may be related to the neuroinflammation observed in aged 5xFAD animals, which is clearly associated with the genetic load. The balance between production and degradation, measured as the DAGL/MAGL (for 2-AG) and NAPE-PLD/FAAH (for anandamide and other acylethanolamines) ratios favors ECB production. The activation of DAGLα and the reduction in MAGL activity may be increasing the 2-AG levels, thus leading to the desensitization of CB1 receptors [76], which could explain the reduced CB1 found in our samples. Several related works aligned with this hypothesis have been published supporting this idea, such as the work of Mulder et al. (2011) who determined the alteration of 2-AG signaling during late stages in the transgenic APdE9 mice [77]. Furthermore, Altamura et al. (2015) found that AD patients present high plasma 2-AG levels compared to controls [78]. Here, we propose an alteration of 2-AG signaling in the transgenic 5xFAD mice due to the combination of impaired MAGL degradation and increased DAGL levels, which is also a marker of severity, since it is only observed in HZ animals. Nonetheless, it will be necessary to measure the release of 2-AG in the hippocampus of these animals to confirm whether this marker is functionally relevant. The overall significance of this overproduction of endogenous cannabinoids could probably reflect an anti-inflammatory response to fight the damage associated with β-amyloid deposition.
Currently, neuroinflammation is one of the hallmarks of AD pathology acting as a vicious cycle [79]. Cytotoxicity stimulatory factors such as Aβ deposition lead to microglial activation, releasing inflammatory and neurotoxic factors. This, in turn, causes progressive neuronal loss and degeneration, secreting neurotoxic factors and entering again into the AD vicious cycle. After evaluating inflammatory markers, we have seen that there is a substantial neuroinflammatory response in the hippocampus of these animals. Both Western blot results and immunohistochemistry staining show the accumulation of GFAP-and Iba1-positive cells within the hippocampus, being higher in the HZ than in the HTZ group. As expected, a higher transgenetic load leads to a more robust inflammatory response. GFAP is the main astrocytic intermediate filament considered to be a highly specific marker for glia [80], and Iba1 is a microglia/macrophage-specific calcium-binding protein [81]. Both have been previously studied in AD mouse models, suggesting that AD pathogenesis is not restrained to neurons but actively interacts with micro-and astroglia, triggering an innate immune response [82]. Another inflammatory marker analyzed was iNOS, which was highly expressed in the hippocampal cells of HZ mice. Interestingly, this oxidative stress marker has also been found in AD human samples, being involved in the pathogenesis of neuronal degeneration of the disease [83]. Here, we report the correlation between the expression of cannabinoid receptors and the substantial increase in neuroinflammation. Results show that a lower number of CB1 and an increasing number of CB2 receptors, both aggravated by the transgenic load, are related to the neuroinflammatory response. Overall, CB1 and GPR55 negatively correlate with the inflammatory markers, and CB2 positively correlates with them. This means that an increased genetic load in the animal model also increases the levels of CB2 receptor expression, which in turn is related to high levels of neuroinflammation since these receptors are mainly located in activated microglia. From these analyses, we can propose ECB receptors as biomarkers for inflammatory responses associated with AD. The potential utility of cannabinoid receptor ligands to modulate these inflammatory responses places them as new targets for developing therapies capable of acting as disease modifiers that retard the progression of the disease. A proof of concept of this neuroinflammation modulation is granted in this 5xFAD mouse model.
On the other hand, a negative correlation between CB1 expression and amyloid plaques (Aβtotal, Aβ40, and Aβ42) in the hippocampus of 5xFAD mice was reported. These results are consistent with other studies affirming that CB1 activity is higher at earlier AD stages and decrease at advanced stages [84]. Nevertheless, the increase in CB2 receptor in our HZ 5xFAD mice positively correlated with Aβ accumulation and senile plaque score. This is a response to excessive neuroinflammation induced by Aβ deposition, indicating an enhanced glial reactivity as confirmed by expanding previously published literature [7,19,21,85]. It is important to note that pharmacological evidence suggests that the CB2 receptor can modulate Aβ and hyperactivity tau levels [61], although there are conflicting results. The chronic treatment of a specific agonist of CB2, JWH-133, failed to reduce Aβ in the hippocampus and cortex of 5xFAD mice [27], while prolonged oral administration of this agonist reduced significantly cortical β-amyloid in Tg App 2576 mice [29]. These differences could be due to the animal model used and the severity of the disease. Moreover, Vázquez et al. (2015) found that, by blocking CB1, inflammation worsened in 5xFAD mice but no quantification of Aβ was specified [36]. Therefore, more experiments are needed to clarify the functionality of these receptors on the Aβ-clearance. Remarkably, the increased expression of the GPR55 receptor, depending on the complete transgene load, cannot be explained by Aβ accumulation, and it is independent of neuroinflammation.
Finally, an important limitation of the present study is that the 5xFAD model is an amyloid deposition model that lacks a very relevant pathophysiological mechanism in AD: the deposition of hyperphosphorylated tau protein, responsible for neurofibrillary tangles. Further studies must be addressed in preclinical models of tau deposition to clarify the role of the endogenous cannabinoid system in AD, where both Aβ accumulation and tau deposition contribute to the loss of neurons and, subsequently, to the devastating dementia characterizing the disease.

Conclusions
Overall, the results of the present study suggest that the endocannabinoid system is involved in modulating memory impairment and anxiety-like behavior, as well as pathological changes in the 5xFAD mouse model such as inflammation. This modulatory response is related to the imbalance of the expression pattern of CB1, CB2, and the GPR55 receptors in the hippocampus. This imbalance is clearly aggravated by the transgenic load, making these receptors candidate markers for the severity of the disease. Here, we described for the first time the alteration of the ECB system in 5xFAD mice, taking the heterozygosity and homozygosity into special consideration, shedding light on the unknown neurobiological changes of the AD. However, more studies are needed to elucidate the role(s) of the ECB system and provide a new therapeutic target for the treatment of neurodegenerative diseases such as AD. In future work, we plan to investigate the release and dynamics of ECB in the brain of these animals through the development of the disease and to include a more complete protein-protein interaction pathway according to the literature. Understanding the evolution of the ECB system in these animals at an early stage will help to understand when and how to use new cannabinoid-based drugs to modify the course of the disease. Finally, using models where tau deposition is also present will help to better clarify the efficacy of this newly proposed ECB-based therapeutic approach to Alzheimer's disease.