Multifocal Cerebral Microinfarcts Modulate Early Alzheimer’s Disease Pathology in a Sex-Dependent Manner

APP/PS1 transgenic mice expressing the chimeric mouse/human Swedish amyloid precursor protein (APP695swe) and a mutant human presenilin1 (PS1-dE9) under the control of independent mouse prion promoter elements [B6.Cg-Tg(APPswe,PSEN1dE9)85Dbo/J] were used. C57BL/6J wildtype (WT) littermate mice were used as controls whenever necessary. CX3CR1^GFP/+ mice in which an enhanced green fluorescent protein (EGFP) sequence replacing the first 390 bp of the coding exon (exon 2) of CX3CR1 gene allowing GFP expression in peripheral monocytes [B6.129P2(Cg)-Cx3cr1^tm1Litt/J] were used to generate chimeric mice. Moreover, 5 months old male APP/PS1/CX3CR1^GFP/+ triple transgenic mice, which were obtained by crossbreeding APP/PS1 x CX3CR1^GFP/+ mice, were used to investigate the dynamics of CX3CR1^GFP/+ cells in the brain of APP/PS1 mice. All animals were originally purchased from The Jackson Laboratory (Bar Harbor, ME, USA) backcrossed, maintained on a pure C57BL/6J background, and genotyped using the protocols recommended by The Jackson Laboratory. Five months old male and non-ovariectomized female mice were used. Young animals were chosen to better reflect the early phases of AD and the consequences of microinfarcts on disease progression. Mice were housed for a maximum of 5 per cage under standard laboratory conditions (23°C, 50-60% humidity, 12:12 hour light and dark cycle), and were provided with standard laboratory chow and water ad libitum. Mice were randomized for the experimental groups and neurobehavioral tests, and the experimenter was blind to the genotype as well as the experimental group. All animal procedures and handling were performed according to the Canadian Council on Animal Care guidelines, as implemented by the Laval University Animal Welfare Committee.

Generation of chimeric mice was performed as previously described (18, 53, 54). Briefly, femurs were flushed from CX3CR1^GFP/+ donor mice in a sterile environment to obtain bone marrow cells. Cells were harvested using Dulbecco’s Phosphate Buffered Saline (DPBS; Wisent) and 5% fetal bovine serum (FBS). The extract was filtered using a 40 µm nylon filter and pelleted to remove any clumps. The cell pellet was collected with DPBS, then centrifuged and re-suspended in fresh DPBS. The recipient WT mice underwent a myelosuppressive chemotherapy regimen using busulfan (Otsuka America Pharmaceutical) and procytox (Baxter). The chemotherapy regimen was based on 2 injections per day (every 8 hours) of 10 mg/kg of busulfan (150 μl, intraperitoneally) for 4 days (a total of 80 mg/kg) followed by 1 injection per day of 100 mg/kg of procytox (150 μl, intraperitoneally) for 2 days (a total of 200 mg/kg). Mice were subcutaneously injected with 1 ml of 0.9% NaCl solution up to 1 week after procytox last injection to avoid dehydration. Mice were kept during the procedure in sterile cages and given previously irradiated food, and antibiotic treatment continued for 1 week following treatment. Chimerism was confirmed 8 weeks later by validating the presence of GFP⁺ myeloid cells in CD45⁺ cells in the blood circulation of CX3CR1^GFP/+ → WT chimeric mice using flow cytometry to validate chimerism success, indicating that the bone marrow cells were reconstituted with monocytes expressing CX3CR1^GFP/+ at 90% ± 1.78%.

Five months old male and female APP/PS1 mice, C57BL6/J mice, CX3CR1^GFP/+ → WT mice and APP/PS1/CX3CR1^GFP/+ mice were subjected to micro-occlusions (MO) to generate sporadic multifocal microinfarcts in the brain, as previously described (55). Briefly, mice were anesthetized under 1.5% isoflurane in 1.5 l/minute (95% O2) and body temperature was maintained between 36 and 37°C using a feedback-controlled heating system (Harvard Apparatus, QC, Canada). A midline neck incision was performed to expose the left common carotid artery (CCA), as well as the external carotid artery (ECA) and the pterygopalatine artery (PPA), which were temporarily blocked using a microvascular clip under a surgical microscope (Leica Microsystems, ON, Canada). Afterwards, 2500 sterilized FITC-tagged microspheres of 20 µm (Polysciences Inc., PA, USA) suspended in a 100 µl of PBS were slowly injected into the CCA using a 33G hypodermic needle (TSK Laboratory International, BC, Canada). The microspheres of 20 µm have been shown to allow occlusion of penetrating cerebral arterioles and capillaries (Figure 1A) (55). The ECA and PPA were temporarily blocked prior to injection to ensure that all microspheres are directed towards the brain. After injection, the ECA and PPA were unblocked, the needle was gently removed from the CCA and the bleeding was immediately stopped by applying pressure using bioabsorbable Gelfoam (Pfizer, NY, USA). APP/PS1 mice were euthanized 1 month following induction of surgery (mice were 6 months old), while CX3CR1^GFP/+ → WT mice were divided into 3 subgroups that were sacrificed 3 days, 1 week and 1 month post-surgery. APP/PS1/CX3CR1^GFP/+ mice were sacrificed 1 week post-surgery. Sham-operated mice underwent the same surgical procedure without microsphere injection. Distribution analysis of the microspheres showed that 9% were present in the hippocampus, 4% in the white matter, 5% in the striatum, 31% in the thalamus, 34% in the neocortex and 15% in other structures, producing microinfarcts in 7% of cases in the hippocampus, 5% in the white matter and striatum, 3% in the thalamus, 1% in the cortex, as previously described (Figure 1B) (55). The fluorescent microspheres are sterile and non-immunogenic, outlined by the absence of any direct impact on microglial reactivity per se (Figure 1C). Finally, the injected microspheres were distributed throughout the brain of APP/PS1 mice in which Aβ pathology is developing (Figure 1D).

Animals underwent neurobehavioral tests prior to surgery (baseline) and at different time points post-surgery, as follow:

Novel object recognition (NOR) test: The test was used to assess recognition memory, as previously described (18, 52, 56). Briefly, in an open arena (45 cm x 45 cm) the mouse was free to explore two identical objects for 10 minutes in an acquisition phase. The animal was placed back to its home cage for 1 hour, before the 10 minutes test phase. In this recall phase, one of the objects was replaced by a new one. We evaluated the time spent on both objects to assess mice performances. Indeed, mice are very curious and attracted to novelty which will lead to a higher exploratory time on the novel object during the test phase. Trials were recorded by a camera placed above the arena and a blind experimenter scored by hand the collected videos. The arena and objects were cleaned with 70% alcohol between each mouse and trial. The task was performed at baseline, 1 week and 1-month post-surgery. A discrimination index ratio was calculated as follow: time novel object/(time novel object - time familiar object). A ratio = 0.5 means an equal exploration of the 2 objects, a ratio > 0.5 means a higher exploratory time for the novel object, and < 0.5 means a preference for the familiar object.

Open field (OF) test: The test was used to assess anxiety and general ambulatory ability as previously described (57). Briefly, in an open arena (45 cm x 45 cm) placed under a camera controlled by a computer, the mouse was free to move for 10 min. Assessment of the animals’ movements were recorded, tracked, and analyzed with the Anymaze software on which the arena was divided into different zones namely, the center, borders, and corners. Parameters evaluated were, the time spent in the center as it is at risk for mice that are prey and the distance moved. An increase in the time spent in the center would indicate a disinhibitory behavior whereas a decrease shows an anxious state. A decrease in the distance could either mean a freezing state, a typical anxiety behavior, or motor impairment.

Laser Speckle contrast imaging (LCSI) was used to evaluate the longitudinal variation of CBF following the induction of microinfarcts. LSCI was performed 24 hours after induction of microinfarcts to evaluate the impact on CBF in the acute phase, and 1 month afterwards in the same animals to assess the long-term consequences. Mice were anesthetized with ketamine/xylazine (1 mg/10 ml) and received 1 ml of Ringer’s lactate solution. Prior to surgery the head was shaved, lidocaine/bupivacaine solution applied on the incision sites (100 µl) and the ears (50 µl/ear), and the skin disinfected. MO- and sham-operated mice were placed on a stereotaxic frame (RWD Life Science Inc., CA, USA), the skull was then exposed by removing the skin using fine-tip forceps, and relative CBF was measured using 2D high-resolution laser blood flow imager (OMEGAZONE OZ-3, OMEGAWAVE, INC.). The OZ-3 system is equipped with a visible and near infra-red (NIR) CCD camera, which allows showing real images continuously, and to compare the color difference between real and blood flow images. Furthermore, the system is equipped with a measurement and analysis software that allows simultaneous quantification of the relative CBF in different regions of interest (ROI). The mean of relative CBF of each hemisphere in similar ROI was quantified, and a ratio of ipsilateral/contralateral values were computed. A ratio of 1 indicates that the relative CBF values are equal in both hemispheres, and a ratio below 1 would indicate a reduced CBF in the ipsilateral, and a ratio above 1 would outline a CBF overall deregulation.

Male and female WT and APP/PS1 mice were sacrificed 1-month post-surgery, while male and female CX3CR1^GFP/+ → WT mice were divided into 3 subgroups that were sacrificed 3 days, 1 week and 1 month post-surgery. Male APP/PS1/CX3CR1^GFP/+ mice were sacrificed 1 week post-surgery. Mice were deeply anesthetized with a solution containing ketamine (90 mg/ml) and xylazine (10 mg/ml) and sacrificed via intracardiac perfusion with 0.9% NaCl. For immunohistochemical analysis, the brains were retrieved, post-fixed in 4% paraformaldehyde (PFA) for 24 hours and transferred into 20% sucrose containing 4% PFA for a minimum of 24 hours. Brains were then cut into 25 μm coronal sections using a freezing microtome (Leica Biosystems, ON, Canada), and serial sections were collected in 12 well-plates in an antifreeze solution (30% glycerol, 30% ethylene glycol in 0,9% NaCl, phosphate buffer (PB)) and kept at -20°C for further use. For Western blot analysis, the brains were removed and dissected to separate the hippocampus and cortex from each hemisphere that were snap-frozen on dry ice and kept at -80°C until further use. Each region was then homogenized separately in a NP40 lysis buffer containing 1% of phosphatase inhibitors and 1% of protease inhibitors, sonicated, and then stored in -80°C until further use.

Free floating brain sections were rinsed 3x with potassium phosphate buffer saline (KPBS) and incubated at room temperature in a blocking/permeabilization solution containing 4% normal goat serum (NGS), 1% BSA, 1% Triton X-100 in KBPS for 45 minutes. Brain sections were then incubated with primary antibodies diluted in KPBS solution containing 1% NGS, 1% BSA, 0.2% Triton X-100 overnight at 4°C. The following primary antibodies were used; anti-Aβ (6E10) (1:1000; mouse anti-human, Biolegend, 8030001), anti-ionized calcium binding adaptor molecule (IBA)-1 (1:1000, rabbit anti-mouse, WAKO, 019-19741), anti-cluster of differentiation (CD)-68 (1:1000, rat anti-mouse, Bio-Rad, Hercules, CA, USA, MCA1957) and anti-CD45 (1:500, rat anti-mouse, BD bioscience, 55376) and anti-nuclear receptor subfamily 4 group A member 1 (Nr4A1 or Nurr77) (1:250, rabbit, Abcam, ab13851). The next day, brain sections were rinsed 3x with KPBS, and incubated for 2 hours at room temperature in the dark with one of the following Cy3 or Cy5-conjugated secondary antibody; AffiniPure goat anti-rabbit IgG (H+L) (1:1000, Invitrogen, A10523) and AffiniPure goat anti-mouse IgG (H+L) (1:1000, Jackson Immunoresearch, 115-165-003); AffiniPure goat anti-rat IgG (H+L) (1:1000, Jackson Immunoresearch, 112-175-143). The brain sections were then rinsed 2x with KPBS and incubated with 4′,6-diamidino-2-phenylindole (DAPI) (1:10 000) for 5 minutes. The brain sections were mounted onto Superfrost^® Plus slides and cover-slipped with Fluoromount-G^® anti-fade medium (Sigma-Aldrich). Epifluorescence images were taken using Axio Observer microscope equipped with a module for optical sectioning (Apotome.2) and Axiocam 503 monochrome camera, and processed using ZEN Imaging Software (Carl Zeiss Canada, Toronto, ON, Canada). The density of IBA1 and CD68 immunolabeled cells as well as the number and volume of 6E10 immunolabeled Aβ plaques were assessed using unbiased computer-assisted stereological software (Stereologer; SRC Biosciences, FL, USA) (58).

Free floating brain sections were processed as specified before, and then incubated with a biotinylated goat anti-mouse Immunoglobulin G (IgG; H+L) (1:1000, Vectorlab, BA-9200). The next day, the brain sections were rinsed 3x with KPBS, and then incubated for 30 minutes at room temperature with Avidin-Biotin Peroxidase Complex (ABC) (Vectastain Elite Kit Standard). Brain sections were rinsed 3x with KPBS before incubation with 3,3’ 3,3′-diaminobenzidine tetrahydrochloride (DAB; Sigma-Aldrich), washed 3x with KPBS, and dried overnight at room temperature. The next day, brain sections were mounted onto Superfrost^® Plus slides and dehydrated via immersion in an increased concentration of ethanol (EthO) solution (50%, 70%, 75%, 95% and 100%), and finally immersed in Xylene solution 2x times for 3 minutes. The slides were then scanned, and the intensity of DAB signal in the ipsilateral as well as contralateral (i.e. background) cortex and hippocampus was analyzed using ImageJ software, as previously described (52). Ratio of intensity ipsilateral/contralateral was computed and values above 1 indicates an increased IgG intensity (i.e. infiltration) in the ipsilateral brain structures.

Free floating brain sections were rinsed 3x with KPBS, mounted onto Superfrost^® Plus slides, and dried overnight at room temperature prior to staining. The next day, the mounted slides were immersed in a cold 4% PFA solution for 45 minutes to fix the brain sections. The slides were then rinsed in KPBS for 10 minutes and incubated next for 1 hour in a 1% Thioflavin S solution diluted in MilliQ water (Millipore Sigma, ON, Canada). The slides were washed 2x in 100% EthO for 1 minutes and next in 80% EthO for 1 minute, and finally 3x in MilliQ water for 1 minutes. The slides were dried overnight and coverslipped with Fluoromount-G^® anti-fade medium (Sigma-Aldrich). The overall number of Thioflavin S-labeled vessels in the brain of APP/PS1 mice was counted using the fluorescence Axio Observer microscope (Carl Zeiss Canada).

Free floating brain samples were mounted onto Superfrost^® Plus slides and kept at -20°C to dry for 1 hour. The mounted slides were incubated for 15 minutes in 4% PFA at 4°C to fix the brain sections and washed with 0.1M PBS. The brain sections were dehydrated using EthO, and a hydrophobic barrier was drawn around the sections. RNAscope^® Fluorescent Multiplex Reagent Kit was used to perform the experiments following the manufacturer’s recommendations (Advanced Cell Diagnostics, Newark, CA, USA). Five drops of Protease III (provided in the kit) were added to entirely cover the brain sections, which were incubated at room temperature for 40 minutes. Following a series of washes with 0.1M PBS, 4 drops of mouse DKK1 probe (Advanced Cell Diagnostics) were added to entirely cover each brain section, and the slides were placed into a slide rack, and inserted in the humidity control tray of the HybEZ™ Oven for 2 hours at 40°C. The humidity control tray was removed from the HybEZ™ Oven, one slide at a time, and the excess of liquid was removed and rinsed with a 1X washing buffer. Four drops of Amp-1-FL were added to entirely cover each brain section, which were incubated again in the HybEZ™ Oven for 30 minutes at 40°C. This process was repeated with Amp-2-FL, Amp-3-FL, and Amp-4-FL-Alt B. Finally, 4 drops of DAPI (provided in the kit) were added to each brain section and kept for 30 seconds, followed by a series of washes with 1X washing buffer. The slides were next mounted with 80 μl of Fluoromount-G^® anti-fade medium, carefully cover-slipped, and placed at 4°C in the dark until analysis under the fluorescence Axio Observer microscope (Carl Zeiss Canada).

Twenty µg of proteins harvested either from the cortex or the hippocampus were mixed with 1X sodium dodecyl sulfate (SDS)-loading buffer and heated for 10 minutes at 95° C. Samples were run on 8–16% gradient precast acrylamide gels (Bio-Rad) and subjected to electrophoresis using Criterion^® cell (Bio-Rad). Following migration, resolved protein bands were transferred onto a 0.45 µm polyvinylidene fluoride (PVDF) membrane (EMD Millipore) for 30 minutes at 200V on ice using Criterion^® blotter (Bio-Rad). The PVDF membranes were washed 3x with a 0.1M tris buffer saline (TBS) solution containing 0.5% Tween-20 (TBS-T; Sigma-Aldrich) for 10 minutes and blocked in TBS-T with 5% (w/v) skim milk for 30 minutes at room temperature. The PVDF membrane was then incubated overnight at 4°C with different primary antibodies diluted at 1:1000 in TBS-T solution. The following primary antibodies were used; mouse anti-rabbit ATP Binding Cassette Subfamily B Member 1 (ABCB1) (1:1000, Santa Cruz Biotechnology, sc-8313), mouse anti-rabbit LDL Receptor Related Protein 1 (LRP1) (1:1000, Abcam, AB-92544), mouse anti-rabbit beta-site APP cleaving enzyme 1 (BACE1) (1:1000, Abcam, ab108394), and anti-mouse actin (1:1000, EMD Millipore, MAB 1501). Primary antibodies were detected with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson Immunoresearch, West Grove, PA, USA) that were diluted 1:5000 in TBS-T and revealed by enhanced chemiluminescence plus (ECL) solution (Bio-Rad). β-actin were used to ensure equal protein loading. Blots were revealed using Clarity western (ECL) substrate (Bio-Rad) and digitized using Biorad chemidoc XRS+ (Biorad, Montreal, Qc, Canada). Digitized blots were densitometrically analyzed with ImageJ software corrected for protein loading by means of β-actin, and expressed as relative values comparing different groups, as described (58).

The levels of soluble Aβ₄₀ and Aβ₄₂ in the contralateral and ipsilateral hemisphere of mice were quantified using Human Amyloid β 1-42 and 1-40 ELISA kits (EMD Millipore, EZHS SET). The experimental procedure for the detection of Aβ₄₂ and Aβ₄₀ was performed as recommended by the manufacturer’s instructions. Absorbance was obtained using a microtiter plate reader (SpectraMax 340PC, Molecular Devices), and analyzed using SOFTmax Pro3.1.1 software (Molecular Devices), as previously described (18).

Blood samples were collected from the submandibular vein into ethylene-diamine-tetra-acetic acid (EDTA) coated vials (Sarstedt, Montréal, QC, Canada) for the analysis of circulating monocytes and neutrophils. The experimental procedure was performed as previously described (59). Briefly, 50 μl of total blood was incubated at room temperature for 20 minutes with 1 ml of ACK (ammonium, chloride, potassium) lysing buffer to get rid of red blood cells. Samples were then washed with 3 ml of DPBS and centrifuge at 400 g for 8 minutes à 4°C. The supernatant was removed, and cells were resuspended with 100 µl of blockage solution, 1 µl CD16/CD32 antibody (BD bioscience) diluted in 100 µl of DPBS per tube and incubated for 10 min on ice. Then, 100 µl of antibody mix added in each tube which comprises cluster of differentiation (CD)-45 FITC (1:100, BD Pharmigen, 553079), cluster of differentiation (CD)-11b Pe-Cy7 (1:100, EBioscience, 25-0112), Ly6C V450 (1:100, BD Bioscience, 560594), Ly6G Pe (1:100, BD bioscience, 551461) and live/dead (1:50, Life technologies, L23105↗) for 30 minutes on ice in the dark. The remaining cells were next washed with 3 ml DPBS, centrifuged for 8 minutes at 400 g, resuspended in 300 µl of DPBS, and complemented with 50 µl of 123count eBeads™ Counting Beads (Invitrogen, 01-1234-42). Samples were processed using a LSR II flow cytometer, and data was acquired using BD FACS Diva software (Version 6.1.2, BD Bioscience), and analyzed using FlowJo software v10 (Tree Star; Ashland, OR, USA).

Data are presented as the mean ± standard error of the mean (SEM). For comparison between the 2 main different experimental groups, data was analyzed using standard two-tailed unpaired t-test. For ratio analysis, data was analyzed using a one sample t-test to compare the mean of experimental groups with 0.5 that represents a probability ratio of 50%. P-value < 0.05 was considered statistically significant (95% confidence interval). Statistical analyses were carried out using the GraphPad Prism Version 8.0 for Mac OS X (GraphPad Software).

Deposition of Aβ in the brain constitutes one of the key pathological hallmarks of AD (8). In APP/PS1 mice, AD-like pathology is associated with Aβ depositions and cognitive decline that begin around the age of 4 months, progressing over time to become apparent in histological and neurobehavioral tests around 6 months of age (60–62). Although microinfarcts have been shown to occur at the early stages of AD, the consequences on Aβ pathology remain unclear. As such, we investigated the impact of multifocal cerebral microinfarcts induced using MO on early Aβ pathology in young male and female APP/PS1 mice to better appreciate the progression of AD-like pathology before it gets overwhelmed with advanced age (Figure 2A). For this purpose, brain sections were immunolabeled with a 6E10 antibody that recognize human Aβ and were processed for stereological analysis. Interestingly, we found out that the overall baseline density of Aβ plaques was lower in the cortex and hippocampus of sham-operated male APP/PS1 mice in comparison to females (Figures 2B–D), whereas the mean volume of individual Aβ plaques was more important in the cortex and hippocampus of sham-operated male APP/PS1 mice in comparison to females (Figures 2B, E, F). This observation suggests that Aβ pathology is fundamentally different in male and female APP/PS1 mice, with males having a lower number of large Aβ plaques, while females exhibiting a higher number of relatively small Aβ plaques. Furthermore, our analysis indicated that the density of Aβ plaques remained unchanged (Figures 2B, G, H), whereas the mean volume of individual Aβ plaques was decreased (Figures 2B, I, J), in the cortex as well as the hippocampus of MO-operated male APP/PS1 mice. The early induction of multifocal microinfarcts did not influence the levels of soluble Aβ₄₀ and Aβ₄₂ in the brain of male APP/PS1 mice (Figures 2K, L). On the other hand, the density of Aβ plaques was reduced (Figures 2B, M, N), whereas the mean volume of individual Aβ plaques did not change (Figures 2B, O, P), in the cortex as well as the hippocampus of MO-operated female APP/PS1 mice. As observed in males, multifocal microinfarct induction did not affect the levels of soluble Aβ₄₀ and Aβ₄₂ in the brain of MO-operated female APP/PS1 mice (Figures 2Q, R). The analysis suggests that microinfarcts attenuate parenchymal Aβ pathology by reducing the size of individual Aβ plaques in the brain of male APP/PS1 mice, while decreasing the overall number of Aβ deposits in the brain of females. We next evaluated the impact of multifocal microinfarcts on the frequency of CAA associated with vascular Aβ deposition in the brain of male and female APP/PS1 mice by examining the number of Thioflavin S-labeled arteries (Figure 2S). No changes were detected in the frequency of CAA in the brain of sham-operated (Figure 2T) and MO-operated (Figure 2U) male APP/PS1 mice. Similarly, the frequency of CAA in the brain of sham-operated female APP/PS1 mice remained unchanged (Figure 2V), whereas it was decreased in MO-operated mice (Figure 2W). Our results suggest that multifocal microinfarcts attenuated early Aβ pathology in APP/PS1 mice in a sex-dependent manner by differentially reducing the density and volume of parenchymal Aβ plaques as well as deposition of vascular Aβ aggregates.

Some observations indicated that the presence of microinfarcts in young AD patients is associated with worsening of the clinical outcomes. The impact of microinfarct occurrence at the early stages of AD on cognitive decline remains elusive. As in APP/PS1 mice the cognitive decline is progressive, the consequences of microinfarct induction of cognitive functions were evaluated using different neurobehavioral paradigms to assess recognition memory using the NOR test (Figure 3A) and to examine willingness to explore and anxiety-like behavior using the OF test (Figure 3B). Analysis of NOR discrimination index showed that 1 week after microinfarct induction, MO-operated male WT mice and APP/PS1 mice exhibited an impaired recognition memory in comparison to sham-operated mice (Figure 3C), an effect that was maintained for 1 month in sham-operated APP/PS1 mice in which AD-like pathology normally progressed, and MO-operated WT mice and APP/PS1 mice (Figure 3D). Furthermore, 1 week after microinfarct induction, NOR discrimination analysis indicated that, similarly to males, the recognition memory was impaired in MO-operated female WT mice and APP/PS1 mice in comparison to sham-operated mice (Figure 3E). Interestingly, 1 month after microinfarct induction, the recognition memory was recovered in MO-operated female WT mice and APP/PS1 mice, as well as sham-operated WT mice, except in sham-operated APP/PS1 mice in which AD-like pathology normally progressed (Figure 3F). The OF test analysis indicated that the time spent in the center was high in MO-operated male APP/PS1 mice in comparison to sham-operated mice, indicative of a reduced state of vigilance translated by an increased disinhibitory behavior, whereas no differences were observed between sham- and MO-operated WT mice (Figure 3G). In parallel, the distance travelled by MO-operated WT mice and APP/PS1 mice significantly increased in comparison to sham-operated mice, with an exaggerated effect in young APP/PS1 mice (Figure 3H), outlining a reduced state of vigilance as well. Interestingly, the patterns of the time spent in the center (Figure 3I) and the travelled distance (Figure 3J) were maintained in MO-operated WT mice and APP/PS1 mice 1 month later, indicating a longitudinal preserved effect of microinfarcts. On the other hand, the time spent in the center remained unchanged among sham- and MO-operated female WT mice and APP/PS1 mice (Figure 3K). The distance travelled was increased specifically in MO-operated female WT mice, whereas it remained unchanged in MO-operated female APP/PS1 mice (Figure 3L). In contrast to young males, the time spent in the center remained unchanged 1 month later in MO-operated female WT mice and APP/PS1 mice in comparison to sham-operated mice (Figure 3M). Furthermore, the distance travelled by MO-operated female WT mice and APP/PS1 mice increased in comparison to sham-operated littermates, with female APP/PS1 mice showing a moderate increase (Figure 3N). Our findings suggest that microinfarcts caused a rapid and prolonged memory impairment in young male APP/PS1 mice, accompanied by a reduced state of vigilance and an increased disinhibitory behavior, independently upon AD-like pathology, whereas in female APP/PS1 mice, these effects were mild and transient.

Although the impact of CBF deregulation and vascular permeability have been shown to constitute an important risk factor for AD (6, 63, 64), little is known about how these parameters are impacted following the occurrence of microinfarcts. Using LCSI to assess the regional changes in relative CBF (Figure 4A), our analysis showed that the overall CBF value in the ipsilateral hemisphere is reduced 24 hours after microinfarct induction in MO-operated male WT mice and APP/PS1 mice (Figure 4B), indicating an acute impairment of neurovascular coupling associated with cerebral hypoperfusion. Interestingly, 1 month later, the absolute relative CBF value in the ipsilateral hemisphere of MO-operated male WT mice and APP/PS1 mice increased beyond those of the contralateral hemisphere (Figure 4C), outlining a possible dysfunction of the autoregulation process over time associated with cerebral hyperperfusion. Like males, LCSI analysis indicated that the CBF in the ipsilateral hemisphere is reduced 24 hours after microinfarct induction in MO-operated female WT mice and APP/PS1 mice (Figure 4D). However, we found that the CBF was completely recovered in the ipsilateral hemisphere of MO-operated female WT mice and APP/PS1 mice after 1 month (Figure 4E). Next, the impact of microinfarcts on BBB permeability was assessed 1 month later using IgG immunohistology (Figure 4F). Our analysis showed that overall IgG infiltration into the ipsilateral hemisphere increased in MO-operated WT mice and APP/PS1 mice independently upon the biological sex (Figure 4G), outlining the presence of permeable BBB at this stage. Interestingly, IgG infiltration in the ipsilateral cortex was increased specifically in MO-operated females and to a lesser extent in MO-operated males, thus highlighting a more potent impact vascular permeability in the brain of female compared to male mice (Figure 4H). Finally, we found that IgG infiltration into the ipsilateral hippocampus was more potent in MO-operated female APP/PS1 mice (Figure 4I). Our results indicate that the microinfarcts caused a prolonged CBF deregulation in the ipsilateral hemisphere of male APP/PS1 mice, independently upon AD-like pathology, whereas the CBF alterations in the ipsilateral hemisphere of female APP/PS1 mice were acute and associated to permeable BBB.

It has been previously shown that several mechanisms associated with Aβ clearance and production are impaired in AD (35, 52). As such, the effects of microinfarcts on the expression of some key markers, including ABCB1 and LRP1, which are both involved in facilitating Aβ elimination from the brain to the periphery across the BBB (65), as well as BACE1, which plays a key role in generating Aβ peptides via the cleavage of APP (66, 67), was evaluated using Western blot analysis. Our data indicated that the protein levels of ABCB1, LRP1, and BACE1 remained unchanged in the ipsilateral hemisphere of MO-operated male (Figures 5A, B) as well as female (Figures 5C, D) WT mice and APP/PS1 mice 1 month after microinfarct induction. Our results suggest that the mechanisms implicated in Aβ vascular elimination or parenchymal production are not affected by the microinfarcts 1 month later.

At the early stages of AD, microglial cell activation slows disease progression via elimination of Aβ through phagocytosis (65, 68). Reports have indicated that microglial cells are activated in response to microinfarcts within the lesion sites (55). Moreover, microglial cell activation has been proposed to be influenced by biological sex, and that microglia isolated from females adopt a neuroprotective, phagocytic phenotype in response to ischemic challenge (69). Using immunohistochemical analysis (Figures 6A, B), we investigated the impact of multifocal microinfarct induction on microglial cell behavior in the brain of male and female APP/PS1 mice. Our immunofluorescence analysis shows that microglial cell overall density (IBA1 immunolabeled) remained unchanged in the cortex and hippocampus of MO-operated male APP/PS1 mice 1 month after induction of multifocal cerebral microinfarcts (Figure 6C). It has been previously shown that microglia are recruited to Aβ plaques (18, 65, 70). Our colocalization investigations indicated that the number of microglial cells (IBA1) surrounding Aβ plaques (6E10) was not affected 1 month after the induction of microinfarcts in MO-operated male APP/PS1 mice (Figure 6D). Similarly, the overall density of microglia remained unchanged in the cortex and hippocampus of MO-operated female APP/PS1 mice (Figure 6E). However, we observed a specific reduction in the number of microglial cells surrounding Aβ plaques in the cortex of MO-operated female APP/PS1 mice (Figure 6F). Next, microglial cell relative phagocytic capacity was evaluated by assessing expression of the lysosomal receptor CD68 (Figure 6G). Our analysis shows that CD68 expression increased in the ipsilateral cortex and hippocampus of MO-operated male (Figure 6H) and female (Figure 6I) APP/PS1 mice. Furthermore, the number of IBA1-immunolabeled cells expressing CD68 increased in the cortex (Figure 6J) and hippocampus (Figure 6K) of MO-operated APP/PS1 mice, exhibiting similar colocalization patterns between males and females. Our findings indicate that the microinfarcts triggered microglial activation and phagocytic capabilities similarly in the brain of male and female APP/PS1 mice, which may account for the attenuated Aβ pathology.

Monocytes play an important role in AD (19, 71). The inflammatory monocyte subset infiltrates the brain and differentiates into microglia-like cells that contribute to Aβ clearance via phagocytosis (14, 71). The patrolling monocyte subset contributes to the clearance of Aβ vascular micro-aggregates (18). A progressive defective production of Ly6C^low CX3CR1^high and Ly6C^high CX3CR1^low monocyte subsets, coinciding with cognitive decline, has been reported in APP/PS1 mice (71). Using flow cytometry analysis, we analyzed the response of the different monocyte subsets and neutrophils (Figure 7A) to multifocal microinfarcts in male and female WT mice and APP/PS1 mice. Our analysis showed that 48 hours after microinfarct induction, an increased frequency of total monocytes was observed in the blood circulation of MO-operated male WT mice, which was absent in MO-operated male APP/PS1 mice (Figure 7B). Nonetheless, monocyte frequency in MO-operated male APP/PS1 mice was significantly lower in comparison to MO-operated male WT mice (Figure 7B). Analysis of the different monocyte populations showed a reduced frequency of Ly6C^low monocytes in the blood circulation of MO-operated male WT mice in comparison to sham-operated mice, whereas no changes were observed in sham- and MO-operated male APP/PS1 mice (Figure 7C). The frequency of Ly6C^int monocytes significantly increased in the blood circulation of MO-operated male WT mice and APP/PS1 mice in comparison to sham-operated mice (Figure 7D), associated with a decreased frequency of Ly6C^high monocytes (Figure 7E). Additionally, MO-operated female WT exhibited higher total monocyte frequency in the blood circulation in comparison to female MO-operated APP/PS1 mice, which was absent in MO-operated female APP/PS1 mice (Figure 7F). Interestingly, monocyte frequency in MO-operated female APP/PS1 mice was reduced to a lesser extent in comparison to MO-operated female WT mice (Figure 7F). Moreover, the frequency of Ly6C^low monocytes was elevated in the blood circulation of MO-operated female WT mice in comparison to sham-operated mice, whereas in contrast to males, the frequency of Ly6C^low monocytes increased in the blood circulation of female MO-operated APP/PS1 mice (Figure 7G). Similar to males, the frequency of Ly6C^int monocytes increased in the blood circulation of MO-operated female WT mice and to a lesser extent of MO-operated female APP/PS1 mice in comparison to sham-operated mice (Figure 7H), accompanied with a decreased frequency of Ly6C^high monocytes (Figure 7I). Finally, the frequency of Ly6G^high neutrophils was increased in the blood circulation of MO-operated male WT, but to a lesser extent in MO-operated APP/PS1 mice (Figure 7J). Similarly, MO-operated female WT mice exhibited an increased frequency in circulating neutrophils, which was absent in MO-operated female APP/PS1 mice (Figure 7K). Moreover, MO-operated female APP/PS1 mice had a lower frequency of neutrophils in the blood circulation compared to the MO-operated WT females (Figure 7K). Our results suggest that microinfarcts stimulated the generation of Ly6C^int monocytes and induced a specific increase in Ly6C^low CX3CR1^high monocytes in the blood circulation of female APP/PS1 mice.

In addition to resident activated microglia, CD68 is highly expressed as well in peripheral mononuclear phagocytes, namely CX3CR1^high monocytes (25, 72). Moreover, Ly6C^low CX3CR1^high monocytes have been previously shown to transiently infiltrate the brain to promote neuroprotection in response to excitotoxicity (72). As our findings above suggest that not all CD68-reactive cells are resident microglia, we were interested in investigating the brain dynamics of Ly6C^low CX3CR1^high monocytes that differentially respond to cerebral microinfarction. For this purpose, chimeric mice enabling the specific tracking of Ly6C^low CX3CR1^high monocytes, were generated in male and female WT mice, which were chosen to eliminate any confounding effects associated with Aβ pathology. This was achieved by subjecting male and female WT mice to a chemotherapy regimen to ablate the bone marrow-derived myeloid cells followed by the transplantation of bone marrow-derived cells isolated from respectively male and female CX3CR1^GFP/+ mice (CX3CR1^GFP/+ → WT chimeric mice) (Figure 8A), as previously described (18). Flow cytometry analysis indicated that the reconstituted bone marrow cells comprised 90% ± 1.78% of CX3CR1^GFP/+ monocytes. As IBA1 could be expressed as well in infiltrating macrophages, its reactivity in the brain of male and female CX3CR1^GFP/+ → WT chimeric mice was investigated at different time points representing the acute (3 days), sub-acute (1 week) and chronic (1 month) phases after microinfarct induction (Figure 8B). Our immunofluorescence analysis shows that the number of IBA1⁺ cells increased beginning of day 3, peaking at 1 week, and returning to the basal levels 1 month later in the ipsilateral cortex of MO-operated male (Figure 8C) and female (Figure 8D) CX3CR1^GFP/+ → WT chimeric mice, thus exhibiting temporal transient increase independently upon biological sex. Similar IBA1 reactivity patterns were observed in the hippocampus of MO-operated male (Figure 8E) and female (Figure 8F) CX3CR1^GFP/+ → WT chimeric mice. We next assessed the infiltration of Ly6C^low CX3CR1^high monocytes into the brain of MO-operated CX3CR1^GFP/+ → WT chimeric mice, and we found out that CX3CR1^GFP/+ cells were massively recruited to the ipsilateral brain between day 3 and week 1, whereas they were not detectable in the ipsilateral brain 1 month later (Figure 8G). Stereological analysis indicated that GFP⁺ cells were detected at day 3, peaked at week 1, and were almost absent in the ipsilateral cortex of MO-operated male (Figure 8H) as well as female (Figure 8I) CX3CR1^GFP/+ → WT chimeric mice 1 month later. However, in the ipsilateral hippocampus, our analysis showed that the density of GFP⁺ cells was higher at day 3 and begun to decrease from week 1 until 1 month later in MO-operated male CX3CR1^GFP/+ → WT chimeric mice (Figure 8J), whereas in MO-operated females, GFP⁺ cell density was higher at 1 week (Figure 8K). Finally, some infiltrated GFP⁺ cells (i.e. CX3CR1 monocytes) were found to express IBA1 (Figure 8L), as well as and CD68 (Figure 8M), outlining their potential phagocytic capacity. Analysis of GFP⁺ and IBA1⁺ colocalization shows that the density of GFP⁺ cells expressing IBA1 peaked at 1 week in the ipsilateral cortex of MO-operated male (Figure 8N) and female (Figure 8O) CX3CR1^GFP/+ → WT chimeric mice. Interestingly, although similar colocalization pattern was observed in the ipsilateral hippocampus of MO-operated male CX3CR1^GFP/+ → WT chimeric mice (Figure 8P), higher density of GFP⁺ cells expressing IBA1 was observed in females (Figure 8Q). Our findings suggest that the multifocal cerebral microinfarcts promoted the infiltration of circulating Ly6C^low CX3CR1^high monocytes that exhibited a phagocytic potential more potently in the hippocampus of female mice.

As Ly6C^low CX3CR1^high monocytes were shown here to efficiently infiltrate the brain after multifocal microinfarcts in absence of Aβ pathology, we investigated next the behavior of infiltrated cells in the brain of APP/PS1 mice. For this purpose, we used APP/PS1/CX3CR1^GFP/+ triple transgenic mice in which Ly6C^low CX3CR1^high monocytes as well as resident microglia are GFP⁺. Brain infiltrated phagocytic Ly6C^low CX3CR1^high monocytes were discriminated from resident microglia through their expression of Nurr77, a transcription factor implicated in the survival of bone marrow derived Ly6C^low monocytes (18). Our immunofluorescence analysis indicates that GFP⁺ cells expressing Nurr77 (i.e. CX3CR1 monocytes) are specifically recruited to the lesion site (Figure 9A) in the brain MO-operated male APP/PS1/CX3CR1^GFP/+ mice 1 week after microinfarct induction. Interestingly, our analysis suggests that CX3CR1^GFP/+ cells expressing CD45 (i.e. infiltrating immune cells) co-localized with Aβ plaques immunolabeled with 6E10 (Figure 9B). As no major differences were observed in the infiltration rate of Ly6C^low CX3CR1^high monocytes in the brains of males and females, only males were used for this experiment. Our results suggest that the infiltrated CX3CR1^high monocytes are possibly contributing to the remodeling of injured tissue and to Aβ clearance, which may account for the attenuated Aβ pathology.

As the changes reported here in Aβ pathology do not correlate with cognitive decline following the induction of microinfarcts, we were interested in identifying the possible underlying mechanisms. DKK1 has been shown to be induced by ischemic/hypoxic insults (73), to be implicated in AD pathogenesis (52), and to mediate cognitive decline independently of Aβ pathology via inhibition of synaptic transmission (74). For this purpose, the expression of DKK1, an endogenous inhibitor of the canonical Wnt pathway was evaluated in the cortex of mice 1 month after microinfarct induction using FISH RNAscope assay (Figure 10A). Our analysis indicated that intensity DKK1 mRNA transcripts increased in the ipsilateral cortex of MO-operated male WT mice, which was exacerbated in MO-operated male APP/PS1 mice (Figure 10B). Interestingly, DKK1 mRNA levels remained low in the cortex of MO-operated female APP/PS1 mice (Figure 10C). We next investigate the expression pattern of DKK1 mRNA in the hippocampus 1 month after microinfarct induction (Figure 10D) and found out that DKK1 mRNA levels increased in the ipsilateral hippocampus of MO-operated male WT mice, which was exacerbated as well in MO-operated male APP/PS1 mice (Figure 10E). Finally, we found out that DKK1 mRNA levels remained unchanged in the ipsilateral hippocampus of MO-operated female WT mice, and significantly decreased in the hippocampus of MO-operated female APP/PS1 mice (Figure 10F). Our findings suggest that DKK1 expression is potently induced in the brain of male APP/PS1 mice after microinfarcts, which may account for the exaggerated cognitive decline in males.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.