Quantitative Brain Positron Emission Tomography in Female 5XFAD Alzheimer Mice: Pathological Features and Sex-Specific Alterations

Successful back-translating clinical biomarkers and molecular imaging methods of Alzheimer's disease (AD), including positron emission tomography (PET), are very valuable for the evaluation of new therapeutic strategies and increase the quality of preclinical studies. 18F-Fluorodeoxyglucose (FDG)–PET and 18F-Florbetaben–PET are clinically established biomarkers capturing two key pathological features of AD. However, the suitability of 18F-FDG– and amyloid–PET in the widely used 5XFAD mouse model of AD is still unclear. Furthermore, only data on male 5XFAD mice have been published so far, whereas studies in female mice and possible sex differences in 18F-FDG and 18F-Florbetaben uptake are missing. The aim of this study was to evaluate the suitability of 18F-FDG– and 18F-Florbetaben–PET in 7-month-old female 5XFAD and to assess possible sex differences between male and female 5XFAD mice. We could demonstrate that female 5XFAD mice showed a significant reduction in brain glucose metabolism and increased cerebral amyloid deposition compared with wild type animals, in accordance with the pathology seen in AD patients. Furthermore, we showed for the first time that the hypometabolism in 5XFAD mice is gender-dependent and more pronounced in female mice. Therefore, these results support the feasibility of small animal PET imaging with 18F-FDG- and 18F-Florbetaben in 5XFAD mice in both, male and female animals. Moreover, our findings highlight the need to account for sex differences in studies working with 5XFAD mice.


INTRODUCTION
Alzheimer's disease (AD), the most common form of dementia, affects more than 40 million people worldwide and is characterized by the decline of cognitive abilities and progressive memory loss (1). To date, there is still no cure available, and the current pharmacotherapeutic strategies do not target the underlying molecular processes of the disease (2,3).
To establish new treatments for AD, both preclinical models of the disease as well as reliable in vivo biomarkers are essential. Clinical biomarkers include measurement of Abeta42 and (phosphorylated) tau in cerebrospinal fluid, and also MRI and molecular imaging with positron emission tomography (PET). PET is a functional, non-invasive imaging technique that allows in vivo monitoring of pathological and physiological processes (4). 18 F-Fluorodeoxyglucose (FDG)-PET and 18 F-Florbetaben-PET are clinically established biomarkers for the diagnosis of AD capturing two key pathological features of the disease (5,6). FDG-PET can visualize cerebral glucose metabolism that reflects regional neuronal and synaptic activity and can therefore be used to assess synaptic dysfunctions and neuron loss in the brain (7). 18 F-Florbetaben-PET detects cerebral amyloid load (8). Successful back-translating clinical biomarkers and molecular imaging methods of AD, including PET, is very valuable for the evaluation of new therapeutic strategies and to increase the quality of preclinical studies. However, while PET is an established tool in the assessment of AD patients, its role in preclinical studies using AD mouse models remains unclear.
Several mouse models of AD are available so far showing different pathological hallmarks. Thereby AD models that express mutations in the amyloid precursor protein (APP) and/or presenilin (PS) genes have been widely used to study AD pathology, especially with regard to the Abeta pathology. A wellestablished and commonly used mouse model is the 5XFAD model, as these mice carry five APP and PS1 mutations that lead to an age-dependent aggressive amyloid plaque pathology, neuronal dysfunction, and neuron loss as well as behavioral deficits (9,10).
A recent study of our group showed that both 18 F-FDG-and 18 F-Florbetaben-PET are suitable tools for the assessment of AD pathologies in male 5XFAD mice in vivo, postulating its broader use in longitudinal preclinical therapy studies (11). However, only data on male 5XFAD mice have been published so far whereas possible sex differences in 18 F-FDG and 18 F-Florbetaben uptake have not been studied (12)(13)(14).
The aim of this study was to evaluate the suitability of 18 F-FDG-and 18 F-Florbetaben-PET in female 5XFAD mice and to assess possible sex differences.

5XFAD Transgenic Mice
5XFAD mice overexpress the 695 amino acid isoform of the human APP carrying the Florida (I716V), London (V717I), and Swedish (K670N/M671L) mutations under the control of the neuron-specific murine Thy1-promoter. In addition, human presenilin-1 (PS1), which carries the L286V and M146L mutations, is expressed under the control of the Thy-1 promoter (9). 5XFAD mice used in this study were kept on a C57Bl/6J genetic background (15) (Jackson Laboratories, Bar Harbor, ME, USA) and wild type (WT) littermates served as age-matched control animals.

F-FDG-PET/MRI
Seven-month-old female (n = 6) and male (n = 4) 5XFAD, and female (n = 6) C57Bl/6J WT mice were scanned with 18 F-FDG-PET/MRI (n = 4-6 per group) using a small animal 1 Tesla nanoScan PET/MRI (mouse whole body volume coil, diameter 35 mm, Mediso, Hungary) as previously described (11,16). Mice fasted overnight, and blood glucose levels were measured before tracer injection. Mean activity of 16.2 MBq 18 F-FDG (11.5-19.5 MBq) was injected into a tail vein with a maximum volume of 200 µl. The uptake period was 45 min in which mice were awake and kept warm. Afterward, mice were anesthetized with 1-2% isoflurane supplemented with oxygen (0.5 l/min, cp-pharma, Burgdorf, Germany). PET acquisition time was 20 min. Scans were performed in a single mouse model. Mice were kept on a heated bed (37 • C) during the scans and the respiratory rate was measured. MRI-based attenuation correction was conducted with the material map (matrix 144 × 144 × 163 with a voxel size of 0.5 × 0.5 × 0.6 mm 3 , repetition time: 15 ms, echo time: 2.032 ms, and a flip angle of 25 • ), and the PET images were reconstructed using the following parameters: matrix 136 × 131 × 315, voxel size 0.23 × 0.3 × 0.3 mm 3 .

F-Florbetaben-PET/MRI
Amyloid imaging was performed with the commercially available tracer 18 F-Florbetaben (NeuraCeq, Life Molecular Imaging, Berlin, Germany) as described previously (11). 18 F-Florbetaben-PET/MRI was performed on 7-month-old 5XFAD and C57Bl/6J wild type mice (n = 4-6 per group). 18 F-Florbetaben (9.1-21.8 MBq; mean 15.6 MBq) was injected into a tail vein with a maximum volume of 200 µl. PET acquisition time was 30 min starting after an uptake period of 40 min. Mice were anesthetized with isoflurane supplemented with oxygen during the whole procedure. PET images were reconstructed as described for 18 F-FDG-PET/MRI.

Image Analysis
Positron emission tomography images were analyzed using PMOD v3.9 (PMOD Technologies, Switzerland) as previously described (16). In brief, different volumes of interest (VOI), including whole brain volume as well as the amygdala, brain stem, cerebellum, cortex, hippocampus, hypothalamus, midbrain, olfactory bulb, septum/basal forebrain, striatum, and thalamus were defined using an MRI-based mouse brain atlas template. VOI were between 10 mm 3 (amygdala) and 0.3 cm 3 (cortex) in size. VOI statistics (kBq/cc) were generated for all brain areas mentioned above and standardized uptake values (SUV) were calculated [SUV = tissue activity concentration average (kBq/cc) × body weight (g)/ injected dose (kBq)] for semiquantitative analysis. SUV of 18 F-FDG-PET scans were corrected for measured blood glucose levels [Glc = SUV × blood glucose level (mg/dl)]. Blood glucose levels were between 116 and 209 mg/dl (mean 156 mg/dl) with no significant difference between the groups. 18 F-Florbetaben-PET scans were normalized to the cerebellum, and the obtained ratios (SUVr) were used for further analysis.

Behavior Experiments: Learning and Memory Performance
Animals (n = 10-12 per group) were kept on an inverted 12h/12-h dark/light cycle and tested during the dark period.

Morris Water Maze
The Morris water maze (MWM) test was used to assess spatial reference memory as previously described (17). During the final "probe trial, " the platform was removed from the pool, and the mice were allowed to swim freely for 1 min while their swimming path and swimming speed were recorded using an automated video tracking system (ANY-Maze, Stoelting Co, USA).

Cross Maze
The cross maze was used to assess the working memory by analyzing the spontaneous alternation behavior of mice (18). The cross maze consists of four arms (30 cm length × 8 cm width × 15 cm height) arranged in 90 • angles extending from a central region (8 cm length × 8 cm width × 15 cm height). During a single trial, mice were allowed to move freely through the maze for 10 min. ANY-Maze video tracking software (Stoelting Co, USA) was used to record the sequence of arm entries and the distance traveled. An alternation was defined as successive entries into the four arms in overlapping quadruple sets (e.g., 1, 3, 2, 4 or 2, 3, 4, 1 but not 1, 2, 3, 1) (15,19). The alternation percentage was calculated as the percentage of actual alternations to the possible number of arm entries.

Novel Object Recognition
The novel object recognition task (NORT) is widely used to assess recognition memory and novelty preference in rodents. The test is based on the spontaneous tendency of rodents to explore a novel object more often than a familiar one (20). The NORT was performed in an open-field arena made of gray plastic (50 × 50 cm), and the animals were tracked using an automated video tracking system (ANY-maze, Stoelting Co, USA). A habituation phase 24 h before the actual test preceded the testing to avoid neophobic interference. During the habituation phase, each mouse was given 5 min to explore the testing arena and become habituated to the testing environment. Twenty-four hours after the habituation, animals were placed in the same box, containing two identical objects, and allowed to freely explore the box for 5 min. The following day, mice were placed back in the same arena, now with a familiar and novel object. The time spent with each object was recorded.

Statistical Analysis
Statistical analysis was performed using GraphPad Prism version 8 for Mac (GraphPad Software, San Diego, CA, USA). Differences between groups were tested with a paired t-test, unpaired ttest, or one-way analysis of variance (ANOVA) followed by Bonferroni multiple comparisons as indicated. Data are given as mean +/-SD. Significance levels are given as follows: * p < 0.05; * * p < 0.01; * * * p < 0.0001.

Female 5XFAD Mice Show Decreased Cerebral Glucose Metabolism Compared to WT Mice
Seven-month-old female 5XFAD and age-and sex-matched WT mice were scanned with 18 F-FDG-PET to determine cerebral glucose metabolism in vivo. 18 F-FDG uptake was measured in the whole brain VOI, and several brain region VOIs and glucose correction were performed (Figures 1A-L). 18 F-FDG uptake was detected in the brain of all mice and also regular extracranial uptake within Harderian glands, myocardium, brown adipose tissue, intestines, kidneys, and the urinary bladder. Seven-month-old female 5XFAD mice showed significantly lower 18 F-FDG uptake in the whole brain compared with WT mice (Figure 1M, unpaired t-test: p = 0.0015). Decreased 18 F-FDG uptake was detected within all brain regions ( Figure 1N, unpaired t-test: cortex: p = 0.0035; hippocampus: p < 0.0001; thalamus: p = 0.0029; cerebellum: p = 0.0032; forebrain: p = 0.002; hypothalamus: p < 0.0001; amygdala: p < 0.0001; olfactory bulb: p = 0.004; midbrain: p < 0.0001).

Female 5XFAD Mice Show Increased Amyloid Deposition Compared to WT Mice
To determine amyloid plaque deposition in vivo, 7-month-old 5XFAD mice and age-and sex-matched WT mice were examined with amyloid-PET using 18 F-Florbetaben (Figures 2A-D). Tracer uptake was measured in the whole brain and eight different brain areas. 18 F-Florbetaben uptake was normalized by cerebellar uptake as the reference region. 18 F-Florbetaben was detected in the brain in all tested mice. Extracranial distribution was physiological, showing 18 F-Florbetaben uptake within Harderian glands, intestines, and urinary bladder.

Memory Deficits in Female 7-Month-Old 5XFAD Mice
Spatial reference memory was assessed in female 5XFAD and WT mice in the probe trial of the Morris water maze. WT mice displayed a significant preference for the target quadrant as indicated by the percentage time spent in the different quadrants of the pool (Figure 4A, One-way ANOVA, WT: p < 0.001 targets vs. all other quadrants). In contrast, 5XFAD mice showed no preference for any of the quadrants (Figure 4A). The absence of a preference for the target quadrant compared to the remaining quadrants demonstrates that 5XFAD mice display a deficit in spatial reference memory. Swimming speed was not altered in 5XFAD mice (Figure 4B, unpaired t-test: p = 0.3651).
Recognition memory was tested using the NORT. Female 7-month-old 5XFAD mice showed an impaired recognition memory as they displayed no preference for the novel object ( Figure 4D, paired t-test, 5XFAD: p = 0.6999). In contrast, WT animals showed a clear preference for the novel object and intact recognition memory (Figure 4D, paired t-test, WT: p < 0.0001). During the training phase, WT and 5XFAD spend ∼50% of the time with both identical objects ( Figure 4C, paired t-test, WT: p = 0.8355; 5XFAD: p = 0.7825).
In addition, female 5XFAD mice showed an impaired spatial working memory in the cross maze. The alternation rate was used as an indicator of spatial working memory impairment in this task, and 5XFAD mice displayed a significantly reduced alternation rate compared with the same-aged wild type animals ( Figure 4E, unpaired t-test, p = 0.0112). The reduced alternation rate was not due to a decrease in overall exploration behavior as mice covered the same distance ( Figure 4F, unpaired t-test, p = 0.3207).

DISCUSSION
Alzheimer's disease is one of the most challenging diseases of the century. To investigate new therapies, longitudinal evaluation of their effects in vivo is essential and relies heavily on diseasespecific biomarkers. Biomarkers, including molecular imaging with 18 F-FDG-and amyloid-PET, have shown their value in the clinical diagnosis of AD (21). However, only a limited number of preclinical studies with PET are available, even though modern small animal PET scanners are able to detect the same molecular pathologies that are seen in humans. Both, 18 F-FDGand amyloid-PET can detect early changes of the disease, making them valuable targets for in vivo imaging of disease progression FIGURE 2 | 18F-Florbetaben uptake in 7-month-old female 5XFAD mice and age-and sex-matched WT mice. (A) Mouse brain atlas with predefined brain regions projected on MRI images in coronal view. (B) 18F-Florbetaben-PET images of a 7-month-old WT mouse. (C) 18F-Florbetaben-PET images of a 7-month-old female 5XFAD mouse. (D) 18F-Florbetaben-PET images of a 7-month-old male 5XFAD mouse. 18F-Florbetaben-PET detected increased amyloid deposition in 7-month-old female and male 5XFAD mice compared to WT. (E) 7-month-old female 5XFAD mice showed significantly higher SUVr in the whole brain compared with 7-month-old female WT mice. (F) Differences could be detected in the cortex, hippocampus, thalamus, amygdala, and midbrain. Unpaired t-test; *p < 0.05; **p < 0.01, ***p < 0.001; data presented as mean +/-SD. and efficacy of new therapeutic strategies. Advances in small animal PET have enabled imaging in AD mouse models, allowing FIGURE 3 | Sex differences in 7-month-old 5XFAD mice. (A,C) 18F-FDG uptake in 7-month-old female 5XFAD mice showed significantly lower SUVglc in the whole brain compared to 7-month-old male 5XFAD mice. Differences could be detected in the hippocampus, thalamus, cerebellum, and midbrain. (B) 7-month-old female 5XFAD mice did not show significant differences in 18F-Florbetaben uptake compared to 7-month-old male 5XFAD mice in the whole brain or in (D) any of the analyzed brain regions. Unpaired t-test; *p < 0.05; **p < 0.01; data presented as mean +/-SD.
longitudinal follow-up studies (22,23) and a better translation of findings from bench to bedside.
Neuronal dysfunctions can be detected using 18 F-FDG-PET as 18 F-FDG uptake in neurons is mainly determined by synaptic activity (24). AD patients typically show a pattern of decreased 18 F-FDG uptake in the posterior cingulate cortex proceeding to posterior temporal and parietal cortex areas and eventually to the frontal lobe (25). However, the use of 18 F-FDG as a preclinical biomarker to study AD processes in animal models has been arguable, as inconsistent results with hyper-, hypo-, and normal metabolism have been described in transgenic AD mouse models (12,(26)(27)(28).
The 5XFAD mouse model of AD is a widely used model of amyloidosis with five AD-linked mutations. 5XFAD mice show a massive plaque pathology, intraneuronal Abeta, robust microgliosis, and inflammatory processes as well as synaptic and neuronal loss and memory deficits (9,10,15,29,30). Unbiased stereology revealed significant neuron loss in the subiculum by 9 months and neuron layer 5 of the cortex by 12 months in female 5XFAD mice (15,29). The robust appearance of histological, (F) There was no difference in the total distance traveled between 5XFAD and WT mice in the cross maze. (A) One-way ANOVA followed by Bonferroni multiple comparisons (B,E,F) unpaired t-test, (C,D) paired t-test. *p < 0.05; ***p < 0.001; Data presented as mean ± SD. molecular, and behavioral AD hallmarks makes the 5XFAD model popular in biomedical research, and the "Alzheimer's Disease Preclinical Efficacy Database" (https://alzped.nia.nih. gov/) shows that ∼10% of all AD studies that work with animal models use this strain. Therefore, it is essential that reliable biomarkers are available to monitor the success of a possible therapy in 5XFAD mice. As 5XFAD mice recapitulate major features of AD, 18 F-FDG-and amyloid-PET should theoretically be useful in vivo biomarkers for 5XFAD animals. However, only a few PET studies with inconsistent findings, which largely focus on male mice, have been performed in 5XFAD mice (11,13,31,32). The current study, which is the first preclinical study using PET/MRI with different tracers in female 5XFAD mice, showed distinct reduced 18 F-FDG uptake in the whole brain as well as in all the nine analyzed brain regions including the hippocampus, cortex, thalamus, and hypothalamus. To our knowledge, only one other study analyzed cerebral glucose metabolism in female 5XFAD mice (33). Son et al. focused on the hippocampus and frontal lobe and described a regional hypometabolism in these two brain regions in 9.5-month-old female 5XFAD mice. In line with these findings, Xiao et al. (34) demonstrated decreased 18 F-FDG uptake in the hippocampus, cerebral cortex, and olfactory bulb of 6-month-old male 5XFAD mice. Similarly, DeBay et al. (32) confirmed a hypometabolism in the cerebral cortex, hippocampus, and basal ganglia in 5-month-old male 5XFAD mice. In addition, Macdonald et al. (13) detected lower 18 F-FDG uptake in the whole brain of 13-month-old male 5XFAD mice. In contrast, a study by Rojas et al. (31) showed contradictory findings with increased 18 F-FDG uptake in the whole brain of 11-month-old 5XFAD mice, without stating the sex, using the cerebellum as the reference region. As discussed before (11), this normalization technique seems to heavily influence study results. In contrast to human glucose metabolism in the cerebellum (which is relatively preserved in disease progression), metabolism in the cerebellum of transgenic mouse models can also be influenced, possibly resulting in higher uptake values (27,(35)(36)(37). Studies in other AD mouse models using normalized cerebral FDG uptake to the cerebellum also showed higher 18 F-FDG uptake in the brain (27,36,37). This factor could explain the results by Rojas et al. (31) in 5XFAD mice.
To our knowledge, no studies on sex differences in 18 F-FDG-PET in 5XFAD mice or other transgenic mouse models of AD are available so far. We could previously demonstrate a significant reduction in glucose metabolism in 7-and 12-monthold male 5XFAD mice in several cortical areas even before the onset of memory defecits (11). In this study, we show for the first time that the hypometabolism in 5XFAD mice is genderdependent and more pronounced in female mice. Our results demonstrated a lower glucose metabolism in the whole brain as well as in the hippocampus, thalamus, cerebellum, and midbrain of 7-month-old female 5XFAD mice compared with age-matched male 5XFAD mice.
Although the reason is still unknown, AD has a genderspecific epidemiological profile, disproportionately affecting women both in prevalence and in severity (38,39). A number of female transgenic AD mouse models also display an earlier onset of AD pathologies as well as a more severe pathology than their male counterparts (40)(41)(42). Moreover, behavior deficits are often more severe in female mice (43,44). Sex-based differences in 5XFAD mice have also been demonstrated in several studies ranging from differences in odor detection and motor impairment to working memory deficits (45)(46)(47). A gene ontology analysis of the hippocampi of 4-month-old 5XFAD mice detected sex-specific patterns with female 5XFAD mice, which showed more differentially expressed genes compared with males including higher levels of human APP and PS1 mRNA expression. Interestingly, the majority of molecular changes that were more severely affected in female 5XFAD mice were associated with the immune system (48).
Another pathological hallmark of AD that can be detected by PET is cerebral amyloid depositions. Several PET-tracers that are able to visualize amyloid burden have been developed and are used in clinical routine including 11 C-labeled Pittsburgh Compound-B ( 11 C-PIB), 18 F-Flutemetamol, 18 F-Florbetapir, and 18 F-Florbetaben (21,(49)(50)(51)(52). Amyloid-PET has also been demonstrated as a valuable tool in preclinical studies and can be used to measure therapeutic outcomes via amyloid load in vivo (53)(54)(55)(56). The presence of a distinctive plaque pathology in 5XFAD mice has been well-described in vitro with extracellular amyloid plaques beginning at 2 months in the fifth layer of the cortex and progressively increasing with age and spreading to the cortex, hippocampus, and subiculum (9,15,57). However, only a few studies on the use of amyloid-PET in 5XFAD mice are available so far, lacking data on female 5XFAD mice. Available studies show higher cerebral uptake of several different amyloid tracers ( 11 C-PIB, 18 F-Florbetapir, 18 F-FC119S, and 18 F-Florbetaben) compared with WT mice. Rojas and colleagues reported a higher uptake of the amyloid radiotracers 11 C-PIB and 18 F-Florbetapir in 11-month-old 5XFAD mice, whereas Frost et al. showed elevated uptake of 18 F-Florbetapir in 14-monthold mice, both without stating the sex (31,58). In addition, Oh et al. demonstrated elevated levels of 18 F-FC119S in the cortex, hippocampus, and thalamus of 5.5-month-old male 5XFAD mice (14). We previously showed an increased 18 F-Florbetaben tracer uptake in male 5XFAD mice (11). Here we could confirm that 18 F-Florbetaben is also a valuable tracer for female 5XFAD as our findings demonstrate increased cerebral uptake of 18 F-Florbetaben in 7-month-old female 5XFAD mice.
In contrast to 18 F-FDG uptake, 18 F-Florbetaben-PET did not show significant sex-differences between male and female 5XFAD mice. This is in line with the original manuscript characterizing the 5XFAD model, as Oakley et al. only noted a trend toward greater plaque deposition in young female AD animals relative to males of the same age, but this trend declined with age (9).
In contrast, several studies demonstrated higher levels of APP and Aβ42 as well as an increased plaque pathology in female 5XFAD mice (48,59,60). In line with our findings, Biechele and colleagues also detected no sex differences in 18 F-Florbetaben uptake in App NL−G−F mice (61).
Limitations of our study include possible partial volume effect due to analysis of relatively small VOIs. However, this effect might have been avoided as the analyzed volumes were above the suggested threshold in small animal scanners of 9 mm 3 [the smallest VOI (amygdala) was 10 mm 3 ] (62-64). Furthermore, there is no standard for small animal brain imaging which can result in high inter-study variations, especially for FDG studies. Therefore, to avoid inconsistencies between different labs, standardization of imaging protocols including fasting protocols, equipment, and animal handling would be beneficial (12,65). In addition, it might be interesting to study additional age groups for female 5XFAD mice in the future. However, we decided to focus on 7-month-old mice as female 5XFAD mice show a robust Abeta pathology as well as severe memory deficits at this age. Furthermore, this age is often used as an endpoint in treatment studies in 5XFAD mice (66)(67)(68)(69)(70)(71). Thus, our present findings underline that PET imaging with 18 F-FDG-and 18 F-Florbetaben could provide a valuable and robust preclinical evaluation tool of therapeutic strategies modulating the AD pathology in 5XFAD mice.
Our results support the feasibility of small animal PET imaging with 18 F-FDG-and 18 F-Florbetaben in the 5XFAD mouse model of AD in both male and female animals. Moreover, our findings highlight the need to account for sex differences in studies working with 5XFAD mice.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
TF, CI, and NB performed experiments. CB and YB designed the project, performed experiments, analyzed data, and wrote the manuscript. All authors contributed to revising the manuscript and approved the final version.

FUNDING
This work was supported by the Alzheimer Stiftung Göttingen to CB and YB, and in part by the Deutsche Forschungsgemeinschaft (DFG) for PET/MRI use (INST 335/454-1FUGG). We acknowledge support by the Open Access Publication Funds of the Göttingen University.