In vivo reactive astrocyte imaging using [18F]SMBT-1 in tauopathy and familial Alzheimer's disease mouse models: A multi-tracer study

Background: Reactive astrocytes play an important role in the development of Alzheimer’s disease and primary tauopathies. Here, we aimed to investigate the relationships between reactive astrocytes. microgliosis and glucose metabolism with Tau and amyloid beta pathology by using multi-tracer imaging in widely used tauopathy and familial Alzheimer’s disease mouse models. Results: Positron


Introduction
Neuroinflammation characterized by microglial activation and reactive astrocytes is an important pathological process in the progression of Alzheimer's disease (AD) and primary tauopathies [1].
Astrocytes play important roles in maintaining hemostasis, regulating blood flow, and supporting neuronal metabolism [1].Astrocyte reactivity, as indicated by the plasma level of glial fibrillary acidic protein (GFAP), has been shown to be an early biomarker mediating the progression of AD and is correlated with the pathological hallmarks of AD, amyloid-β (Aβ) and tau pathology [2][3][4][5].Tau pathology epigenetically remodels neuron-glial cross-talk in AD [6].Tufted astrocytes and astrocytic plaques are among the morphological hallmarks of primary tauopathies, including corticobasal degeneration and progressive supranuclear palsy [7].Astrocytic 4R tau expression has been shown to drive astrocyte reactivity and dysfunction [7].In addition, microglia and astrocytes are involved in the elimination of tau oligomer-containing synapses in AD [8].In animal models of amyloidosis and tauopathy, chronic neuroinflammation has been shown to play an important role in neuroplasticity and cognitive function [9,10].Tau accumulation in astrocytes of the dentate gyrus has been shown to induce neuronal dysfunction, memory deficits and cerebrovascular changes in animal models [11].
Transcriptomic associations of hippocampal GFAP-positive astrocytes have been reported in PS2APP mice with amyloidosis and in a P301S mouse model of tauopathy [4].
Wild-type C57BL6 mice were obtained from Charles River, Germany, and Cavins Laboratory Animal
[ 18 F]FDG (1.48 GBq/ml) was prepared in the radiochemistry facility under Good Manufacturing Practices requirements.The identities of the aforementioned final products were confirmed by comparison with the high-performance liquid chromatography retention times of the nonradioactive reference compounds obtained by coinjection using a Luna 5 μm C18(2) 100 Å (250 mm×4.6 mm) column (Phenomenex).Acetonitrile and water (60:40) were used as solvents, and the flow rate was 1.0 mL/min.A radiochemical purity > 95% was achieved for all the tracers (SFig. 1 for [ 18 F]PM-PBB3).

Small animal PET
PET experiments were performed using a Siemens Inveon PET/computed tomography (CT) system (Siemens Medical Solutions, United States) [37].The mice underwent [ 18 F]SMBT-1, [ 18 F]florbetapir, [ 18 F]PM-PBB3, [ 18 F]DPA-714 and [ 18 F]FDG scans sequentially, with 2 days of rest between the two scans.Mice were anaesthetized using 5% isoflurane and maintained using 1.5% isoflurane in medical oxygen (0.3-0.5 L/min) at room temperature with an isoflurane vaporizer (Molecular Imaging J o u r n a l P r e -p r o o f Journal Pre-proof Products Company, USA) during the PET/CT procedure.The mice were positioned in a prone position on the heated imaging bed.
[ 18 F]SMBT-1 small animal PET was performed as described earlier, and a single dose of tracers (∼0.37 MBq/g body weight, 0.1-0.2mL) was injected into the animals through the tail vein [27].
Static PET/CT images were obtained 10 minutes after intravenous administration of [ 18 F]SMBT-1 at 60-70 minutes.PET/CT images were reconstructed using the ordered subsets expectation maximization 3D algorithm (OSEM3D), with a matrix size of 128×128×159 and a voxel size of 0.815 mm×0.815mm×0.796mm.Attenuation corrections derived from hybrid CT data were applied.To assess the specificity and selectivity of [ 18 F]SMBT-1 in detecting MAO-B in the mouse brain, a blocking experiment was performed in 7-month-old wild-type male mice.Three mice were treated daily with the MAO-B inhibitor selegiline (intraperitoneal injection (i.p.) of selegiline, 10 mg/kg mouse weight) for 5 days before in vivo PET [ 18 F]SMBT-1.Four age-matched WT mice (without selegiline treatment) were included in the control group for in vivo PET [ 18 F]SMBT-1.On the 6 th day, static PET/CT images were obtained of the mice in the selegiline-treated group and control group 10 minutes after intravenous administration of [ 18 F]SMBT-1 at 60-70 minutes.
Radioactivity is presented as the standardized uptake value (SUV) (decay-corrected radioactivity per

J o u r n a l P r e -p r o o f
Journal Pre-proof cm 3 divided by the injected dose per gram body weight).The volume of interest was defined based on a mouse MRI T 2 -weighted image template as described earlier [27].For [ 18 F]PM-PBB3, [ 18 F]florbetapir, [ 18 F]DPA-714, and [ 18 F]FDG, the cerebellum was chosen since it is the commonly used reference region for calculating the brain regional SUVR.However, few studies on [ 18 F]SMBT-1 in animal models have been reported.In our recent study, we performed dynamic [ 18 F]SMBT-1 imaging over 90 minutes in WT mice and amyloidosis mouse models and performed ex vivo staining for the MAO-B distribution in mouse brain slice.We found that the cerebellum is suitable as a reference brain region for computing the SUVR as the signal is relatively low in the cerebellum.
Therefore, in this study, we used the cerebellum as a reference brain region for [ 18 F]SMBT-1 SUVR analysis [27].A mask was applied for signals outside the mouse brain for illustration.

Immunofluorescence staining
Mice were anaesthetized with tribromoethanol, perfused with ice-cold 0.1 M phosphate-buffered saline (PBS, pH 7.4) and 4% paraformaldehyde in 0.1 M PBS (pH 7.4), fixed for 36 h in 4% paraformaldehyde (pH 7.4) and subsequently stored in 0.1 M PBS (pH 7.4) at 4°C.For cryosectioning, the brain was placed in 30% sucrose in PBS (pH 7.4) until it sank.The brain was embedded in OCT gel (Tissue-Tek O.C.T., USA).Coronal brain sections (20 m) were cut around the bregma 0 to -2 mm using a Leica CM1950 cryostat (Leica Biosystem, Germany).Detailed information on the reagents and antibodies used is provided in STable 1.For MAO-B immunofluorescence labelling, the sections were blocked in blocking buffer containing 3% bovine serum albumin, 0.4% Triton X-100, and 5% normal goat serum in PBS (pH 7.4) for 2 h at room temperature.After washing with PBS 3×10 minutes, the sections were incubated with primary antibodies against MAO-B (1:20; Sc-515354; Santa Cruz Biotechnology) in blocking buffer overnight at 4°C.For A and phospho-Tau staining, the brains were fixed in paraffin according to a routine protocol.Coronal brain sections (3 m) were cut using a Leica RM2016 Microtome (Leica Microsystems, Germany).The sections were first washed in PBS 3×10 minutes, followed by antigen retrieval for 20 minutes in citrate buffer (pH 6.0) at room temperature.After antigen retrieval in citrate buffer at room temperature, the sections were

J o u r n a l P r e -p r o o f
Journal Pre-proof permeabilized and blocked in 3% bovine serum albumin for 30 minutes at room temperature with mild shaking.Paraffine-embedded sections were incubated overnight at 4°C with primary antibodies against A and phospho-Tau (Ser202, Thr205), as described earlier [38].The next day, the slices were washed with PBS 3×5 minutes, incubated with secondary antibody for 2 hours at room temperature and washed 3×5 minutes with PBS.The sections were incubated for 10 minutes in 4',6-diamidino-2phenylindole (DAPI) at room temperature and mounted with antifade mounting media [27].The brain sections were imaged at ×20 magnification using a Pannoramic MIDI slide scanner (3DHISTECH) using the same acquisition settings for all brain slices.The immunofluorescence images were analysed by using ImageJ (NIH, U.S.A.).The signal intensity in the cortex and hippocampus of the brain slices was quantified (normalized to the values in wild-type mice).

Statistics
Two-way ANOVA with Dunnett's post hoc test was used for comparisons between multiple groups (GraphPad Prism 9.0, CA, USA).Two-way ANOVA with Sidak's post hoc test was used for comparisons between the blocking and baseline wild-type groups.A p value less than 0.05 was considered to indicate statistical significance.The data are shown as the mean ± standard deviation.Nonparametric Spearman's rank analysis was performed to assess the correlation between the SUVRs of different tracers.

Increased brain regional [ 18 F]SMBT-1 SUVRs in 5×FAD and rTg4510 mice
[ 18 F]SMBT-1 is a relatively novel tracer for imaging MAO-B.Its specificity for MAO-B in the human brain has been demonstrated by an in vivo blocking study using selegiline [24].To assess the in vivo specificity of [ 18 F]SMBT-1 in the mouse brain, we performed an in vivo blocking experiment using the MAO-B inhibitor selegiline, which has been used in earlier blocking studies [24,42].We chose the 5-day selegiline 10 mg/kg (i.p. injection) regimen based on earlier studies using a high dose of selegiline [43,44].[ 18 F]SMBT-1 PET was performed on 7-month-old wild-type mice treated with selegiline and compared with those in the baseline group (7-month-old wild-type mice without treatment).Partial (25-30%) blocking was observed in the cortex, cerebellum and hippocampus of wild-type mice treated with selegiline (n=3 blocking group) compared to wild-type mice without treatment (n=4 baseline group, SFig.2).

J o u r n a l P r e -p r o o f
Journal Pre-proof month-old 5×FAD mice (n=6) or 7-month-old rTg4510 mice (n=6) compared with 3-month-old rTg4510 mice (n=6).

Immunofluorescence staining revealed pathologies in the mouse brain
Immunofluorescence staining was performed on the mice after they underwent in vivo imaging (Figs. 7, 8).Phospho-Tau (p-Tau)-positive inclusions were detected in the cortex and hippocampus of 7-month-old rTg4510 mice, while no specific positive p-Tau signal was detected in the brains of agematched WT mice.Immunofluorescence staining (Figs. 7e-i) revealed A deposits in the cortex, hippocampus and thalamus of 7-month-old 5×FAD mice, while no specific positive A signals were detected in the brains of age-matched WT mice.The distribution of MAO-B in the brain slices of 7month-old 5×FAD mice appeared comparable to that in the brain slices of WT mice (SFig.3).No significant increase in the level of MAO-B fluorescence intensity was detected in the cortex and hippocampus of the brain slices from 7-month-old rTg4510 mice compared with those from 7-monthold WT mice (Fig. 8).

Discussion
In the present study, we showed increased regional [ 18  Journal Pre-proof brains of 5×FAD mice [9,30,41,[51][52][53][54]. The distribution pattern of [ 18 F]PM-PBB3 in rTg4510 mice corroborates the ex vivo staining results and the findings of previous studies in which [ 11 C]PBB3, [ 18 F]PM-PBB3 and [ 18 F]PI-2620 were used to detect increased uptake in tau-loaded brain regions [39,40,49,50,55,56]: Increased tau was found in cortical regions as well as in the hippocampus at 7 months and in association with the known pathological pattern of tau.
[ 18 F]SMBT-1 has shown high selectivity for MAO-B and low nonspecific binding in the brain, with more than 85% blocked by selegiline across the brain in humans in an earlier study [24].Here, we found that the blocking effect in the mouse brain was only partial (less than 30%) when using selegiline, which is much less than that in humans.In addition, MAO-B might not be a specific marker for astrocyte reactivity due to its expression in astrocytes as well as in histaminergic and serotonergic neurons.In comparison, I2 binding sites (I 2 BSs) exhibit specific expression in reactive astrocytes of the brain and may be a better biomarker for PET imaging of astrocyte reactivity.Sofar, an I 2 BS tracer, [ 11 C]BU99008, has been evaluated in vivo [57].Here, we found increased regional levels of [ 18 F]SMBT-1 in 7-month-old rTg4510 and in 3-, and 7-month-old 5×FAD mice.No prior astrocytosis or MAO-B PET imaging studies have been reported in tauopathy mouse models or in the 5×FAD mouse line.These findings are consistent with the findings of previous ex vivo staining and in vivo microscopy studies [58] revealing increased GFAP staining in the brains of 6-month-old rTg4510 mice [59].Our finding of increased [ 18 F]SMBT-1 expression in 5×FAD mice is consistent with the findings of a recent imaging study in which [ 11 C]acetate PET was used to image reactive astrocytes from 5×FAD mice of similar age [23] and ex vivo characterization of elevated neuroinflammatory marker levels [9,60].We found differences in the distribution pattern between the ex vivo MAO-B immunofluorescence staining pattern and in vivo [ 18 F]SMBT-1 signal in the brain: greater immunofluorescence was found in the hippocampus than in the cortical region of the mouse brain slices, opposite to the in vivo pattern.In addition, no significant difference was found between ex vivo MAO-B immunofluorescence between different mouse groups.One probable reason is that MAO-B is an enzyme, and the activity measured in vivo might be different from the level measured ex vivo.In earlier studies, both association of in vivo MAO-B signal (such as [ 18 F]F-DED) with ex

J o u r n a l P r e -p r o o f
Journal Pre-proof vivo MAO-B expression in GFAP-positive astrocytes [18], and lack of in vivoex vivo association [17,19] has been reported in animal models.Possible reasons for the different observation include the choice of antibody for staining, different quantification, as well as both cellular sources of MAO-B (astroglial and neuronal).
The time course of microglial changes has been well characterized by in vivo TSPO imaging in different mouse models.including PS2APP mice, APP/PS1 mice, and the TgF344 rat model [18,61,62] and ex vivo by using staining [63,64] and transcriptomics.Our observation of increased [ 18 F]DPA-714 in 7-month-old 5×FAD and rTg4510 mice is in line with previous reports.Increased cerebral uptake has been shown using [ 18 F]DPA-714 [65] at 8 months and [ 11 C]PBR28 [66] at 6 months in 5×FAD mice.Moreover, TSPO colocalizes with iba1-positive reactive microglia but not GFAP-positive astrocytes in the brains of 6-month-old 5×FAD mice [66].Increased levels of [ 18 F]DPA-714 [49], [ 11 C]AC-5216 [50], and [ 18 F]FEBMP [67] have been reported in rTg4510 mice at 7 months of age.However, the exact mechanism by which TSPO signalling occurs is still unclear: TSPO is known to be expressed on both microglia and astrocytes.A previous study showed that astrocytic TSPO upregulation occurs before microglial TSPO upregulation in TgF344 rats [68].
Another study suggested that TSPO upregulation is selective for proinflammatory polarized astrocytes and microglia through the use of profiling and immunostaining [69].Notably, an earlier study indicated sex differences in microglial activation by TSPO imaging in Aβ but not in tau animal models [70].
We showed that, compared with those in age-matched controls, there was reduced [ 18 F]FDG uptake in the brains of rTg4510 mice but not in those of 7-month-old 5×FAD mice.These findings are in line with the observed hypometabolism in the brains of 7-month-old rTg4510 mice [49,71].Perfusion, functional resting-state brain network and metabolic connectivity changes have been reported in rTg4510 mice [58,72,73].Inconsistent FDG results have been reported in 5×FAD mice.A recent characterization study revealed no overt metabolic changes caused by [ 18 F]FDG uptake in 5×FAD mice at 4, 6 or 12 months of age [74].Another study showed some edgewise differences, but no J o u r n a l P r e -p r o o f Journal Pre-proof regions showed differences in 5×FAD mice at 6 months of age for either sex [75].In addition, an increased cerebral level of [ 18 F]FDG was reported in 5×FAD mice at 11 months of age [30].In contrast, two other studies reported decreased [ 18 F]FDG in 5×FAD mice at 7 and 12 months of age [76] and at 13 months [77] (no difference was observed in 5×FAD mice).One possible reason for this complexity is that the increased activity and metabolic changes in microglia in amyloidosis or tau mouse brains are associated with overall [ 18 F]FDG glucose uptake [78,79].The inconsistency in the [ 18 F]FDG PET results in the rodent brain might stem from the variability in temperature, anaesthesia depth, physiological condition of the mice, sex of the mice, etc.It has also been suggested that the use of the cerebellum as a reference region may result in enhanced hypermetabolism in AD mouse models, as the cerebellum might also be subject to metabolic changes [51].
Previous study revealed a positive correlation between in vivo uptake of [ 18 F]SMBT-1 with amyloid and tau tracer uptakes in AD patients and control cases, indicating an association between MAO-B alteration with AD pathologies [25].Here we found a positive correlation between regional SUVR of [ 18 F]SMBT-1 and [ 18 F]PM-PBB3 in rTg4510 mice (p<0.0001,r=0.4945).It is noted that there is no correlation between [ 18 F]SMBT-1 and [ 18 F]PM-PBB3 within a single region.Our data did not indicate a correlation between regional SUVR of [ 18 F]SMBT-1 and [ 18 F]florbetapir in 5×FAD mice.Similarly, we also did not observe correlation between regional SUVR of [ 18 F]SMBT-1 and [ 18 F]florbetapir in our recent study using another amyloidosis mouse model (APP/PS1 mice) [27].

Journal Pre-proof
The limitations of the current study include the lack of a longitudinal (instead of cross-sectional) study design.In addition, the scans performed in this study were static, and dynamic scans could provide more accurate quantification at the expense of a much longer scanning time.Moreover, as a multitracer study design, animals were subjected to anaesthesia 4 or 5 times for all imaging exams.
Although the mice were allowed to rest for 2 days between scans, anaesthesia might have affected tracer uptake [82].Moreover, a recent study showed that anaesthesia using sevoflurane reduced the in vivo binding of [ 11 C]L-deprenyl-D2 to MAO-B in nonhuman primates [83].It is possible that isoflurane could also affect the tracer retention of [ 18 F]SMBT-1 in mice during PET scans.Further study is needed to elucidate the influence of different anaesthetics on MAO-B PET tracer uptake in rodents.
In conclusion, the present study provided in vivo imaging evidence for regional reactive astrocytes in the brains of 5×FAD and rTg4510 mice, as did Aand tau deposition, hypoglucose metabolism and activated microglia.[ 18 F]SMBT-1 imaging in these two models might be useful for in vivo evaluation of treatments targeting astrocytes for tauopathy and AD.

Ethics approval and consent to participate
This study in animal models was carried out in compliance with the ARRIVE guidelines 2.0.All the experimental protocols were approved by the Institutional Animal Care and Ethics Committee of Fudan University.All methods were carried out in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals.
Consent to participate: Not applicable.J o u r n a l P r e -p r o o f

Consent for Publication: Not Applicable
Journal Pre-proof Co., Ltd., of Changzhou.Mice were housed in ventilated cages inside a temperature-controlled room under a 12-h dark/light cycle.Pelleted food and water were provided ad libitum.Paper tissue and shelters were placed in cages for environmental enrichment.Animal model studies were performed J o u r n a l P r e -p r o o f Journal Pre-proof following the American Research Institute (ARRIVE) guidelines 2.0.The PET imaging and experimental protocol were approved by the Institutional Animal Care and Ethics Committee of Fudan University and performed in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals.
F]SMBT-1 and [ 18 F]DPA-714 expression in the brains of 7-month-old rTg4510 mice and 5×FAD mice, respectively, along with tau and A pathology, compared to age-matched wild-type mice.The [ 18 F]SMBT-1 and [ 18 F]DPA-714 distributions appeared to be divergent both in the rTg4510 mice and in the 5×FAD mice.The use of PET has provided a quantitative in vivo imaging platform for tracking pathological amyloid and tau depositions and resultant pathophysiological changes in mouse models of AD and tauopathy.The [ 18 F]florbetapir pattern in 5×FAD mice observed in our study is in line with the ex vivo staining of A deposits and is consistent with the known Aβ plaque distribution pattern and earlier in vivo [ 18 F]florbetapir, [ 18 F]florbetaben, [ 18 F]FC119S, and [ 11 C] PIB imaging results in the J o u r n a l P r e -p r o o f

Fig. 7 Fig. 8
Fig. 7 Representative images of phospho-tau and amyloid-beta staining in the brains of 7-