Sex- and age-specific sensitivities of the endocannabinoid system in Alzheimer's disease revealed by PET imaging with [18F]FMPEP- d 2 and [18F]MAGL-2102

Abstract

The endocannabinoid system is a critical brain signaling pathway that is dysregulated in various brain disorders, including Alzheimer's disease (AD). Cannabinoid-targeted therapies and imaging approaches have gained increasing interest; however, the biological impact of the endocannabinoid system in disease needs further validation. We aimed to study changes in cannabinoid receptor 1 (CB1) and monoacylglycerol lipase (MAGL), components of endocannabinoid signaling and degradation, in a mouse model of AD by PET imaging. Methods: [¹⁸F]FMPEP-d ₂ and [¹⁸F]MAGL-2102 were produced on a commercial radiosynthesis module. PET-CT images with both tracers were acquired in a knock-in mouse model of AD bearing mutated human amyloid precursor protein (App^NL-G-F ) at 3 ages, and compared to wild-type mice. Excised brains were used for in vitro autoradiography with [¹⁸F]FMPEP-d ₂ and [¹⁸F]MAGL-2102, immunofluorescence, and western blotting. Male wild-type and 5xFAD mice were chronically treated with MAGL inhibitor JZL184 and imaged with [¹⁸F]MAGL-2102 two days after ending treatment. Results: PET imaging showed sex-, age- and genotype-dependent changes in CB1 and MAGL availability. At 4-months (early-stage β-amyloid pathology), female App^NL-G-F mice had lower CB1 availability, and MAGL availability was increased in male App^NL-G-F _, compared to wild-types_. At 8-months, no genotype differences in CB1 were observed, yet MAGL availability was reduced in App^NL-G-F frontal cortex, and male App^NL-G-F mice exhibited higher MAGL than transgenic females brain-wide. At 12-months (late-stage β-amyloid pathology), significantly lower uptake of [¹⁸F]FMPEP-d ₂ was observed in App^NL-G-F compared to wild-type, with no changes in [¹⁸F]MAGL-2102 binding. App^NL-G-F plaque staging was confirmed by Thioflavin-S staining. Imaging findings were supplemented by autoradiography, immunofluorescence, and western blots. [¹⁸F]MAGL-2102 availability was responsive to target engagement of the MAGL inhibitor JZL184 in wild-type and 5xFAD mice. Conclusions: The present study showed dynamic age-, sex- and pathology-related changes in CB1 and MAGL availability from early-stage β-amyloid pathology, suggesting that the endocannabinoid system is a useful target for diagnostics and treatment of AD. Finally, these results highlight that endocannabinoid sex differences should be considered in diagnostics and drug development.

Keywords: Alzheimer's disease; PET; cannabinoid receptor 1; endocannabinoid system; monoacylglycerol lipase.

Figure 1

A) AUCs of [¹⁸F]FMPEP-d₂ TACs for App^NL-G-F and wt mice (m/f) at 4-, 8- and 12-months, for whole brain, caudate putamen, frontal cortex and hippocampus. Mean±SEM, *P < 0.05 and #P < 0.10 for genotype comparisons. B) Average summed (0-120 min) images of [¹⁸F]FMPEP-d₂ in brain of wt and App^NL-G-F mice of both sexes (m/f) at 4-, 8- and 12-months of age. Shown are sagittal images at the right side of brain (2 mm from middle line). Abbreviations: AUC: area under the curve; m/f: male, female; NLGF: App^NL-G-F; SUV: standardized uptake value; TAC: time activity curve; wt: wild-type.

Figure 2

A) AUCs of [¹⁸F]MAGL-2102 TACs for App^NL-G-F and wt mice (m/f) at 4-, 8- and 12-months, for whole brain, caudate putamen, frontal cortex and hippocampus. Mean±SEM, #P < 0.10, *P < 0.05 for App^NL-G-F sex comparisons. B) Average summed (0-120 min) images of [¹⁸F]MAGL-2102 in brain of wt and App^NL-G-F mice of both sexes (m/f) at 4-, 8- and 12-months of age. Shown are sagittal images at the right side of brain (2 mm from middle line). Abbreviations: AUC: area under the curve; m/f: male, female; NLGF: App^NL-G-F; SUV: standardized uptake value; TAC: time activity curve; wt: wild-type.

Figure 3

Representative images of ARG (baseline and blocking with cold self) with [¹⁸F]FMPEP-d₂ (A) and [¹⁸F]MAGL-2102 (B) in sagittal brain sections of 12-month App^NL-G-F mice. IF in sequential sections for CB1 (A; red, left) and MAGL (B; red, right) at 10x (whole brain), 20x (hippocampus, cortex, caudate putamen) and striatal 40x images demonstrating CB1, but not MAGL, specificity around plaques (Thioflavin-S, green). C) Hippocampal CB1 IF quantification reveals a loss in App^NL-G-F mice at 4-months, no change at 8-months, and a loss again at 12-months (mean +/- SEM; n = 5-6/genotype/age). All images in panel A are from the same female 12-month App^NL-G-F mouse. All images in panel B are from the same male 12-month App^NL-G-F mouse. Abbreviations: ARG: autoradiography; CB1: cannabinoid receptor 1; GCL: granular cell layer; HP: hippocampus; IF: immunofluorescence; m/f: male, female; MAGL: monoacylglycerol lipase; NLGF: App^NL-G-F; PCL: pyramidal cell layer; ThioS: Thioflavin-S; wt: wild-type.

Figure 4

A) Representative blot of CB1 and β-actin. B) In 4-month App^NL-G-F mice, CB1 protein levels are significantly reduced in the frontal cortex and hippocampus, and trending down in the striatum. C) At 8-months, App^NL-G-F mice exhibit reduced frontal cortex CB1, compensation in the hippocampus, and no change in the striatum. D) At 12-months, CB1 is significantly reduced in the App^NL-G-F frontal cortex and striatum, but not the hippocampus. Sex differences were detected in the 8-month frontal cortex and 12-month hippocampus, with less CB1 protein in males in both cases. Normalized protein values are relative to exposure time and should not be compared between gels and graphs. Within each region, 4- and 12-month mice were on the same gels; see Figure S12 for age comparisons. Mean+/-SEM (n = 5-6/genotype/age). Abbreviations: CB1: cannabinoid receptor 1; kDa: kilodalton; m/f: male, female; NLGF: App^NL-G-F; wt: wild-type.

Figure 5

A) Representative blot of MAGL and β-actin. B) 4-month App^NL-G-F mice exhibit no differences in frontal cortical or striatal MAGL protein levels, with a trend to higher hippocampal MAGL levels. C) No genotype differences were detected at 8-months in any region. D) At 12- months, female App^NL-G-F mice exhibit increased MAGL protein levels. In the hippocampus, no changes were observed at 12-months. Striatal MAGL is reduced in 12-month App^NL-G-F mice. Normalized protein values are relative to exposure time and should not be compared between gels and graphs. Within each region, 4- and 12-month mice were on the same gels; see Figure S12 for age comparisons. Mean+/-SEM (n = 5-6/genotype/age). Abbreviations: gen: genotype; kDa: kilodalton; m/f: male, female; MAGL: monoacylglycerol lipase; NLGF: App^NL-G-F; wt: wild-type.

Figure 6

A) Representative summed [¹⁸F]MAGL-2102 PET images (0-60 min) and B) SUV curves demonstrating reduced MAGL availability following JZL184 treatment (MAGL inhibition) in 6-month wt and 5xFAD male mice (n = 3/genotype/treatment). 5xFAD mice exhibited higher MAGL availability than wt with or without JZL184. Mean+/-SEM. Abbreviations: ID/cc: injected dose per cubic centimetre; MAGL: monoacylglycerol lipase; PET: positron emission tomography; SUV: standardized uptake value; wt: wild-type.