Comparison of microglial cell population expansion in the Tg344-AD rat and APP SWE /PS1 ΔE9 mouse model of Alzheimer's disease

Abstract

Background

Chronic microglial activation is a key characteristic of Alzheimer's disease (AD). In mouse models of AD, microglial activation is considered associated with microglial cell population expansion.

Objective

To elucidate species- and gene-dosage-dependent differences and similarities in microglial cell population expansion in response to amyloid-β (Aβ) plaque pathology.

Methods

The total number of Iba1⁺ microglia in the neocortex of hemizygous APP_swe/PS1_ΔE9 (APP/PS1) mice and hemi- and homozygous TgF344-AD rats, carrying the same human mutations, was estimated by use of stereological techniques. Furthermore, microglial cluster formation was assessed. Proliferation was assessed in the mouse model.

Results

A significant two-fold increase in microglia was observed in the neocortex of hemizygous APP/PS1 mice at 18 months and of homozygous TgF344-AD rats at 17 months. In comparison, the number of microglia in hemizygous TgF344-AD rats, remained constant from 4 to 17 months. Microglial clusters formed prior to the increase in microglial numbers in both species. The clusters were typically small, surrounding the smaller-sized Aβ plaques, occasionally also containing recently proliferated microglia. The Aβ plaque loads were comparable in hemizygous TgF344-AD rats and APP/PS1 mice, and two-three-fold higher in homozygous TgF344-AD rats. The microglial population remained constant across ages in wild types in both species.

Conclusions

Transgenic mouse and rat AD models show significant differences in microglial population expansion, with a more restrained expansion in the rat. However, in both species and regardless of gene-dosage, population expansion is preceded by microglial clustering around Aβ plaques, indicating that cluster-formation is a key event in AD neuropathology.

Keywords

Alzheimer's disease amyloid-beta microglia microglial clusters

Introduction

Alzheimer's disease (AD) is a progressive neurodegenerative disorder and the leading cause of dementia.¹ Increasing evidence points to a critical role of microglial cells, which are the parenchymal macrophages of the central nervous system (CNS),² in the pathogenesis of AD. Most recently, single nucleotide mutations in innate immunity genes, several of them expressed in microglia, have been shown to increase the risk of developing AD.³ These findings have been strengthened by pathway analysis, suggesting a key contribution of innate immunity to AD pathogenesis.^4,5

In the adult CNS, the microglial population slowly proliferates, regulated by colony-stimulating factor 1 receptor (CSF1R) signaling. An increase in proliferating microglia and upregulation of the CSF1R has been observed in postmortem samples from patients with AD, and in several AD animal models.^6,7 Specifically, studies in the APP_swe/PS1_ΔE9 (APP/PS1) mouse AD model suggest that the clusters of microglia that form around amyloid-β (Aβ) plaques are at least partly formed by microglia generated by CSF1R-dependent proliferation.⁸ In line with these findings, there has been detected an approximately two-fold increase in microglial density in the neocortex and hippocampus of 14- and 18-month-old APP/PS1 mice compared to wild type mice.^8,9 A similar expansion of the microglial population has also been demonstrated in the neocortex and hippocampus in the 5xFAD mouse model.¹⁰ Despite these findings, the total number of microglia in the temporal cortex appears to remain largely constant throughout the clinical disease course in patients with AD.¹¹

Mouse models of AD recapitulate only part of the complex brain environment seen in patients with AD.^12–16 The microglial population in the human brain is more heterogeneous than in mice. The higher degree of microglial heterogeneity in human brains as compared to mouse brains is considered to be a result of the longer life span of humans.¹⁷ However, microglial turnover in the human brain is consistent with findings in mice, as the microglial population in human and mouse brains is replenished several times during a lifetime.^18,19

Comparing the microglial responses to AD pathology in rat and mouse models can provide insights into translationally relevant species differences but also reinforce the disease-related significance of specific findings. The TgF344-AD rat model²⁰ carries the same human mutations as the more widely used APP/PS1 mouse model of AD.²¹ TgF344-AD rats manifest age-dependent cerebral amyloidosis that precedes tauopathy, gliosis, apoptotic loss of neurons in the cerebral cortex, and cognitive disturbance.²⁰ This contrasts with mouse models of cerebral amyloidosis, which have been criticized for not demonstrating robust tauopathy, neuronal loss, or cognitive disturbance.^20–23

The aim of the present study was to compare the microglial population response to A $β$ plaque pathology, a key neuropathological characteristic of AD,²⁴ in the TgF344-AD rat and the APP/PS1 mouse model during disease development. We used stereological estimates of cell numbers in a defined region of the brain unaffected by shrinkage or expansion of the tissue.^25,26 Accordingly, we estimated and compared the number of Iba1⁺ microglial cells in the neocortex of hemi- and homozygous TgF344-AD rats and hemizygous APP/PS1 mice of ages ranging from 3 to 18 months. The results of the analysis provide novel insight into gene-dosage- and species-dependent differences and similarities in the microglial population response to Aβ plaque pathology. The results also contribute novel insight in the sequence of the formation of microglial clusters around $A β$ plaques compared to the microglial cell population expansion, and the spatial distribution of proliferating microglia in APP/PS1 mice.

Methods

Animal models and experiments

APP/PS1 mouse model

APP_swe/PS1_ΔE9 (APP/PS1) transgenic (Tg) mice expressing early onset AD (EOAD) mutations driven by the mouse prion protein promoter²⁷ and littermate wild type (WT) mice were bred and maintained on a hybrid (C57BL/6 C3H/HeN; B6C3) background by breeding hemizygous APP/PS1 male with B6C3 female mice, thereby generating hemizygous APP/PS1 mice and WT mice, which were distinguished by genotyping.²⁸ Mice were bred and housed at the University of Southern Denmark. A group of APP/PS1 and WT mice was housed at Translational Neuropsychiatric Unit, Risskov, from the age of 8 months.

TgF344-AD rat model

The TgF344-AD rat model, developed by Cohen et al.²⁰ expresses the same EOAD mutations as the APP_swe/PS1_ΔE9 mouse model.²¹ Hemi- and homozygous TgF344-AD rats and WT littermate rats were generated by crossing hemizygous TgF344-AD male and female rats. Offspring was genotyped by digital PCR to distinguish between hemi- and homozygous TgF344-AD and WT rats. The TgF344-AD rat colony was maintained in the animal facility at Aalborg University.

BrdU injection

Eighteen-month-old WT and APP/PS1 mice received 90 mg/kg BrdU (Sigma Aldrich) i.p. at 2, 12, and 22 h prior to tissue collection.

Tissue collection and section sampling

Tissue collection

Animals were anesthetized with a lethal dose of pentobarbital before being transcardially perfused with 25 ml (rats) or 10 ml (mice) of phosphate-buffered saline (PBS), followed by 50 ml (rats) or 20 ml (mice) of 4% paraformaldehyde (PFA). Brains used for stereological analysis of Iba1⁺ microglia were removed and immersed in 4% PFA and stored at 4°C for 24 h, followed by storage in 1% PFA for another 24 h. All mouse brains were processed immediately after immersion fixation, whereas rat brains were stored in de Olmos cryoprotective solution at −12°C for up to 36 months until they were processed. The mouse brains used for co-expression analysis of Iba1 with Aβ or BrdU were PFA-perfused, sucrose-immersed, frozen and processed into 20-μm-thick cryostat sections. The mouse brains used for stereological estimation of neuronal numbers were processed as outlined in Babcock et al.²⁸

Section sampling

Brains to be used for stereological analysis of Iba1⁺ microglia were sectioned into 50-µm-thick sections using a vibratome, arranged in parallel series, and stored in de Olmos solution at −12°C. All mouse brains were sectioned frontally. Rat brains were sectioned either frontally or horizontally.

Immunohistochemistry

Antibodies

The antibodies used were biotinylated mouse anti-human Aβ1-16 (clone 6E10, 2 μg/mL, BioLegend^®), rabbit anti-Iba1 (2 μg/mL, Wako), and rabbit-anti Ki67 (2 μg/mL, Novus Biologicals). For control, biotinylated mouse IgG1 (Life Technologies) and rabbit Ig (Dako) were used.

Pretreatment

All samples were washed in PBS at 4°C for two days to remove the sucrose in the De Olmos solution in which they were stored. All rat material was subjected to antigen retrieval either by heating sections in an oven at 60°C in TEG buffer for 3 h (Iba1, Ki67) or in Citrate buffer (Ki67) (10 mM citrate, pH 6.0) at RT for 3 h. For Aβ staining, sections were pretreated with 70% formic acid for 30 min (rats) or 15 min (mice) at RT.

General procedures

Free-floating vibratome sections were rinsed in PBS for 10 min. Endogenous peroxidases were blocked using a solution of PBS, methanol, and hydrogen peroxide (ratio 8:1:1). Sections were rinsed in PBS containing 1% Triton (PBS-T) before blocking non-specific binding in PBS containing 10% FBS for 30 min. Next, the sections were incubated for 1 h at RT and then overnight (48 h for Iba1 in the rat sections to ensure penetration), with the primary antibody in PBS containing 10% FBS at 4°C. The following day, after adjusting to RT, the sections were rinsed in PBS-T for 3 × 15 min. In the case of Aβ sections were then incubated with streptavidin-conjugated horseradish peroxidase (SA-HRP, GE Healthcare, UK), and in the case of Iba1 with anti-rabbit Envision + HRP-labelled polymer (Dako). In both cases for 1 h at RT. After a final rinse of 3 × 15 min in PBS-T, sections were developed using 0.05% 3,3′diaminobenzidine (DAB) with 0.01% H₂O₂ and rinsed in PBS for 5 min. After mounting on gelatine-coated glass slides, the sections were dehydrated in graded ethanol, cleared in xylene, and cover-slipped with Pertex^®. No staining was observed when primary antibodies were replaced with isotype controls. Sections stained for Iba1 were in general counterstained with toluidine blue (TB).

Immunofluorescence

Free-floating vibratome sections were stained with a combination of primary antibodies applied simultaneously to detect Aβ plaques (clone 6E10, Biolegend), microglia (Iba1, Wako), and BrdU (rat anti-BrdU, clone BU1/75, ICR1. Abcam). The sections were kept in the dark after adding the secondary reagents. In the case of the double-immunofluorescense staining for Aβ and Iba1 the procedure followed the general protocol, with the primary antibodies being detected using a mix of secondary reagents. Alexa Fluor 488-conjugated donkey anti-rabbit IgG (Invitrogen) and Streptavidin Alexa Fluor 594 (Invitrogen) were applied to detect Iba1 and Aβ. For the triple-immunofluorescense staining for Iba1, Aβ and BrdU, sections were washed in 2x Saline-Sodium Citrate buffer (SSC) and then incubated in 2xSSC containing 49% formamide at 60°C for 2 h, in 2xSSC alone at 60°C for 2 h, and then 2N HCL in TBS at 37°C for 30 min before rinsing in 0.1 M Sodium Borate buffer (pH 8.5) and TBS. AlexaFluor 350-conjugated donkey anti-rabbit (Invitrogen), Streptavidin TRITC (Serotec), and AlexaFluor 488-conjugated goat anti-rat (Invitrogen) were used as the secondary fluorescence reagents. Lastly, sections were immersed in DAPI in PBS for 10 min and mounted on gelatine-coated glass slides and covered with fluorescence mounting media (Dako). No staining was observed when primary antibodies were replaced with the relevant isotype or IgG controls.

Stereology

Delineation of the neocortex

The neocortex was delineated following the anatomical boundaries laid out in Lyck et al.²⁹ for frontal sections in both species, while the horizontally sectioned rat brains were delineated using the Waxholm space atlas of the Sprague Dawley rat brain.

Equipment

For delineation and counting an Olympus BX 50-microscope fitted with a U-PMTVC Japan color camera (Olympus, Germany). A Proscan Prior motorized specimen stage, and a Heidenhain MT12 microcator connected to a PC equipped with CAST-2 Software (TVisiopharm, Denmark). All analyses were performed blinded.

Microglial counting

Single Iba1⁺ microglia were identified and counted when the nucleus of the Iba1⁺ cell first came into focus while focusing through the disector, following the rules for optical disector counting.²⁵ Clusters of microglia were defined as 3 or more Iba1⁺ cells, with their Iba1⁺ cell bodies being less than one cell-body-diameter apart. The Iba1⁺ cells were counted using a x100 oil-immersion objective (numerical aperture (NA) = 1.30, mice) or a x60 oil-immersion objective (NA = 1.35, rat) (Olympus, Germany) in a single hemisphere with an optical disector height of 10 μm in sections with a final mean thickness of 16.1 μm ± 0.2 μm (rats) or 25.6 μm ± 0.3 μm (mice) (mean ± SEM). Iba1⁺ cells were counted in 7–15 parallel sections separated by 600 μm (or, in 2 brains, 300 μm apart (mice)). The counting frame area was 987.6 μm², and the step area was 493,800 µm² or 987,500 µm² (rats) or 160,000 μm² (mice). The total number of Iba1⁺ cells was estimated as done in West et al.³⁰ Mathematical methods were employed to account for missing sections, missing areas, or missing frames to ensure any microglia that may have been in those missing areas were accounted for.

Neuronal counting

Neurons were identified in TB-stained 50-μm-thick horizontal cryostat sections from 18-month-old APP/PS1 and WT mice by their large nucleus surrounded by intensely stained Nissl substance. Cells were counted using a x100 oil-immersion objective (Olympus, Germany) in a single hemisphere with an optical disector height of 8 μm in sections with a final mean thickness of 16.4 μm ± 0.1 μm (mean ± SEM). Neurons were counted in 11–13 parallel sections separated by 500 μm, using a counting frame area of 494 μm², and a step area of 490,000 μm². The counting frame area was 987.6 μm², and the step area was 160,000 μm². The total number of cells was estimated as done by West et al.³⁰

Coefficient of error (CE)

CEs were calculated as previously described by West et al.³⁰ For all Iba1 estimates the mean CE was 0.08 (range 0.05–0.12) (rats) and 0.11 (0.07–0.14) (mice).

Volume estimation

The neocortical volume (V_neo) was estimated using the following equation: Vneo = A_neo x d x 2, where A_neo is the area of the neocortex determined by multiplying the number of frames by the step size from the stereological analysis³⁰, and (d) is the distance between sections (600 µm). The constant 2 is included to account for the neocortical volume of both hemispheres.

Aβ plaque load analysis

Digital images were obtained using an Olympus VS200 microscope scanner (Evident) with an x40 objective. Regions of Interest (ROIs) were defined using annotation tools to delineate the neocortex. Stain vectors recognizing DAB were set up to ensure accurate identification and differentiation of stained areas by QPath. Thresholding was applied to identify and segment Aβ plaques, setting a threshold level to distinguish stained plaques from the background. The measurement tools were then used to calculate the area covered by Aβ plaques within the defined ROIs, providing quantitative data on the percentage of the area occupied by Aβ plaques, i.e., % Aβ plaque load.

Statistics

Data are presented as mean ± SEM. Groups of WT and APP/PS1 mice of different ages were analyzed by two-way ANOVA followed by Bonferroni's multiple comparison test. Groups of WT and TgF344-AD rats of different ages or gene-dosage were analyzed using one-way ANOVA followed by Holm-Sidak multiple comparison test. An unpaired two-tailed Student's t-test was used for comparisons between two groups. Statistical significance was considered at p < 0.05, with significance levels indicated as follows: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Statistical analyses were performed using Prism v10.2.3 (GraphPad Software).

Results

Microglial response to Aβ pathology in TgF344-AD rats and APP/PS1 mice

Stereological analysis of Iba1⁺ microglia in the neocortex of hemizygous TgF344-AD rats indicated that the number of microglial cells is constant from age 4 to 17 months. Additionally, there was no significant difference in the number of Iba1⁺ microglia when comparing hemizygous TgF344-AD rats to WT controls at age 4 and 17 months, respectively (Figure 1(a)). In comparison, in 17-month-old homozygous TgF344-AD rats, a significant two-fold increase in Iba1⁺ microglia was observed compared to both 17-month-old WT and hemizygous TgF344-AD rats (p < 0.0001, both comparisons). Despite there was no increase in microglial numbers in hemizygous TgF344-AD rats at 17 months, a significant increase in the number of microglial clusters, defined as three or more closely apposed Iba1⁺ cells, was observed (p = 0.0035) (Figure 1(b)). The microglia in clusters comprised 8.6% of microglia in hemizygous TgF344-AD rats at 17 months of age, 8.3% in homozygous TgF344-AD rats and clusters were virtually absent in WT rats of both ages (Figure 1(b) and (c)).

Figure 1.

Microglial cluster formation and population expansion in TgF344-AD rats and APP/PS1 Tg mice. (a) Stereological analysis of age- and gene-dosage-dependent response on the Iba1⁺ microglial population in the neocortex of TgF344-AD rats reveals no significant difference in Tg hemizygous rats at any age, nor over time, compared to age-matched WT controls. In comparison, a significant increase is observed in 17-month-old homozygous Tg rats. Data are presented as mean ± SEM, n = 4–7 per group. Asterisks indicate the result of Holm-Sidak's multiple comparison test (****p < 0.0001). (b) The percentage of Iba1⁺ microglial cells occurring in clusters, defined as ≥3 apposed cells, is significantly increased in 17-month-old hemi- and homozygous TgF344-AD rats. Data are presented as mean ± SEM, n = 4–6 per group. Asterisks indicate the result of Holm-Sidak's multiple comparison test (*p < 0.05, **p < 0.01). (c) Immunohistochemical staining for Iba1⁺ microglia, counterstained with TB, shows microglial clusters in 17-month-old hemi- and homozygous TgF344-AD rats. Microglial cells in clusters had partially retracted processes. Scalebar: 50 µm. (d) Stereological analysis of the age-dependent expansion of the Iba1⁺ microglial population in the neocortex of APP/PS1 Tg mice shows a significantly increased number of Iba1⁺ microglia in 18-month-old Tg mice, but not in WT mice or in younger Tg mice. In APP/PS1 Tg mice, bars represent the number of clustered (black) and non-clustered (hatched) cells. Data are presented as mean ± SEM. Asterisks indicate the result of Bonferroni multiple comparison test (***p < 0.001). (e) The percentage of microglial cells occurring in clusters, defined as ≥3 apposed Iba1⁺ cells, is significantly increased at 8- and 18-months of age in APP/PS1 Tg. Data are presented as mean ± SEM. Asterisks indicate the result of Bonferroni multiple comparison test (***p < 0.001). (f) Iba1⁺ microglial clustering increases with age in APP/PS1 Tg mice. The TB counterstaining slightly reduced the Iba1 staining intensity but was superior for the stereological cell counting. Scale bar: 50 μm.

Stereological analysis of Iba1⁺ microglia in the neocortex of hemizygous APP/PS1 mice showed a significant two-fold increase in Iba1⁺ microglia at 18 months compared to age-matched WT mice and young APP/PS1 mice (p < 0.001, both comparisons) (Figure 1(d)). The total number of Iba1⁺ microglia remained constant with age in the neocortex of WT mice. Further, no significant increase in Iba1⁺ microglia was observed in 4-month-old and 8-month-old APP/PS1 mice (Figure 1(d)). However, in 8-month-old APP/PS1 mice, microglial clusters were observed. Clusters comprised 7.5% of microglia at 8 months and 21% of microglia at 18 months, whereas clusters were virtually absent in WT mice of all ages (Figure 1(e) and (f)).

In both TgF344-AD rats and APP/PS1 mice, Iba1⁺ microglia within clusters exhibited the morphology of activated cells, characterized by partially retracted processes and hypertrophic cell bodies (Figure 1(c) and (f)).

Aβ pathology in TgF344-AD rats and APP/PS1 Tg mice

Quantification of the Aβ plaque load in the sections stained immunohistochemically with the 6E10 antibody showed a higher % Aβ plaque load in homozygous (30.4%) compared to hemizygous TgF344-AD rats (p < 0.01) (Figure 2(a)), and a comparable % Aβ plaque load in hemizygous TgF344-AD rats and APP/PS1 mice. Specifically, Aβ plaques occupied 11.7% of the neocortex in hemizygous TgF344-AD rats compared to 16.6% in APP/PS1 mice (Figure 2(a)). Aβ plaques were dispersed throughout all cortical layers in both hemizygous 17-month-old TgF344-AD rats and the 18-month-old APP/PS1 mice (Figure 2(b)). As expected, no Aβ plaques were detected in the neocortex of age-matched WT mice or rats (data not shown).

Figure 2.

Amyloid-β plaque load and microglial clustering and proliferation in TgF344-AD rats and APP/PS1 Tg mice. (a) Immunohistochemical staining for 6E10⁺ Aβ reveals no significant difference in % Aβ plaque load between 17-month-old hemizygous TgF344-AD rats and hemizygous APP/PS1 Tg mice, but a significantly higher % Aβ plaque load in 17-month-old homozygous TgF344-AD rats. Bars represent the means, n = 4 per group. Asterisks indicate the result of Holm-Sidak's multiple comparison test (*p < 0.05. ***p < 0.01). (b) Immunohistochemical labeling of Aβ shows Aβ plaques in all cortical layers in 17-month-old hemizygous TgF344-AD rats, 17-month-old homozygous TgF344-AD rats, and 18-month-old hemizygous APP/PS1 Tg mice. Scale bar: 200 µm. (c) Double immunofluorescence staining for Iba1 (green) and Aβ (red) demonstrates that Iba1⁺ microglial cells cluster around 6E10⁺ Aβ plaques in hemizygous 17-month-old TgF344-AD rats and hemizygous 18-month-old APP/PS1 mice. Scale bar: 100 μm. (d) Triple immunofluorescence staining for BrdU (green), Iba1 (blue), and Aβ (red) in a 20-μm-thick cryostate section from BrdU-injected 18-month-old APP/PS1 mouse. Proliferating BrdU ⁺ Iba1⁺ cells are occasionally associated with small deposits of Aβ (arrowheads, inserts) but are also often distant from Aβ plaques (arrows). Scale bar: 50 μm (overviews), 75 μm (inserts).

Spatial relationship between microglial cells and Aβ plaques

To elucidate the spatial relationship between microglial cells and Aβ plaques, double immunofluorescence staining for Iba1 and Aβ (6E10) was conducted in hemizygous 17-month-old TgF344-AD rats and 18-month-old APP/PS1 mice. Analysis revealed clustering of Iba1⁺ microglia around Aβ plaques in both animal models (Figure 2(c)). Moreover, the clustering microglia exhibited a morphology consistent with those observed in the immunohistochemically stained sections. These were characterized by hypertrophic cell bodies and partially retracted Iba1⁺ processes, which in the double immunofluorescence staining could be seen to radiate deeply into the 6E10⁺ Aβ plaques (Figure 2(c)). The clusters were typically small, surrounding or associated with the smaller-sized Aβ plaques.

Location of proliferating microglia relative to Aβ plaque-associated microglial clusters

Given the observed increase in microglial numbers, we investigated the occurrence and location of proliferating microglia in relation to Aβ plaques in the neocortex of 17-month-old TgF344-AD rats and 18-month-old APP/PS1 mice as compared to their WT controls. Unfortunately, the immunohistochemical staining for the proliferation marker Ki67 in the rat samples was unsuccessful, despite trying different demasking protocols. However, in the neocortex of BrdU-injected 18-month-old WT and APP/PS1 mice, BrdU⁺ cells co-expressing Iba1 were observed (Figure 2(d)). In the APP/PS1 mice, these BrdU⁺Iba1⁺ cells were both observed between (Figure 2(d)) and in the immediate vicinity of small Aβ deposits (Figure 2(d), insert). This indicates that the majority of actively proliferating microglia in the 18-month-old APP/PS1 mice are not integrated into the Aβ plaque-associated microglial clusters.

Neocortical volume and neuronal numbers

Despite pronounced Aβ pathology, no significant reduction in neocortical volume was observed in 17-month-old hemi- and homozygous TgF344-AD rats compared to age-matched WT controls (Figure 3(a)). Similarly, no global neuronal loss was detected in 18-month-old APP/PS1 mice relative to age-matched WT controls (Figure 3(b)), consistent with findings by Babcock et al.,²⁸ showing a constant volume up to 24 months in APP/PS1 mice.

Figure 3.

No evidence of cortical atrophy or global neuronal loss in TgF344-AD rats and APP/PS1 Tg mice. (a) Neocortical volume is not significantly altered in 17-month-old hemi- and homozygous TgF344-AD rats compared to age-matched WT controls. Bars represent the means. Statistical analysis is performed using One-Way ANOVA. (b) No significant difference in the numbers of neocortical neurons is observed in 18-month-old APP/PS1 mice compared to age-matched WT controls. Bars represent the means. Statistical analysis is performed using Student's t-test.

Discussion

In this study we used state-of-the-art stereological methods to quantitatively assess the total number of Iba1⁺ cells in the neocortex of hemi- and homozygous TgF344-AD rats and hemizygous APP/PS1 mice at various ages. This enabled an unbiased estimation of the total number of microglial cells and of microglial clusters throughout the progression of Aβ plaque pathology.

The total number of microglial cells in the neocortex of TgF344-AD hemizygous rats remained constant from 4 to 17 months of age. Furthermore, a comparison of the 17-month-old hemizygous TgF344-AD and WT rats showed no statistically significant effect of genotype. These findings are comparable to observations in the temporal cortex of AD patients, where stereology-based quantitative analysis of microglia double-labeled for Iba1 and major histocompatibility complex (MHC) class II antigen showed no significant difference between non-demented and patients with AD. In addition, no increase in the microglial population was observed with the clinical progression of the disease.¹¹ In contrast, we did observe a significant increase in microglial clusters in both the 17-month-old hemizygous and homozygous TgF344-AD rats compared to 17-month-old WT rats. Double immunofluorescence staining for Iba1 and Aβ showed that these microglial clusters typically surrounded or were associated with the smaller-sized Aβ plaques. This finding is consistent with an increase in the density of activated microglia in the vicinity of the dense-core plaques observed by others in both the currently used and other mouse AD models,^8,31 as well as in patients with AD.^32,33

Using stereology, we found that expansion of the microglial population occurred subsequent to the clustering of Iba1⁺ microglia around Aβ plaques. Our finding of proliferating microglia both between and integrated into the Iba1⁺ microglial clusters in the APP/PS1 mouse complements previous reports of proliferating microglia in the immediate vicinity of Aβ plaques,⁸ as well as of microglial clustering resulting from microglia migrating to the plaques.³⁴ Proliferating microglia have also been reported near Aβ plaques in other mouse studies.^35–37 However, the number of proliferating microglia in the AD mouse models remains low compared to the transiently high numbers of proliferating microglia that can be seen after acute injury.²⁶

Notably, a significant increase in the number of microglial cells was observed in 17-month-old TgF344-AD homozygous versus hemizygous rats, whereas a significant increase in the number of microglial cells was observed in the hemizygous APP/PS1 mice at 18 months. Considering both the two-three-fold higher % Aβ plaque load in the homozygous compared to the hemizygous TgF344-AD rats, and the comparable % Aβ plaque load in the hemizygous TgF344-AD rats and hemizygous APP/PS1 mice, these findings suggest that microglial population expansion is influenced both by gene-dosage and species. However, although the TgF344-AD rat AD model may in this regard better mirror the microglial response taking place in the brain of patients with AD,¹¹ an increased density of microglia expressing the proliferation marker Ki67 has been observed in temporal cortex samples from AD patients in other studies.⁸ Additionally, there is evidence that microglia undergo replicative senescence in the brain of AD patients, as demonstrated for the APP/PS1 mouse model of AD,³⁸ where prevention of microglial senescence appears to reduce Aβ plaque pathology.

Like Aβ plaque pathology, cerebral atrophy is a key neuropathological feature of AD. In patients with AD, cortical atrophy is generally considered a marker of neurodegeneration, however, there are significant regional differences.³⁹ Focusing on the neocortex, imaging studies show that it is mainly the inferior parietal cortex that is affected in patients with AD.³⁹ This could explain why stereological studies have shown an absence of a global loss of neocortical neurons in AD.⁴⁰ Here, we detected no global loss of neocortical neurons in the APP/PS1 mouse model, which is consistent with our former finding of a constant neocortical volume up to 24 months of age in APP/PS1 mice (28). Similarly, the neocortical volume in homozygous TgF344-AD rats remained unchanged compared to hemizygous TgF344-AD and WT rats, indicating that there is no major global loss of neocortical neurons in these 17-month-old TgF344-AD rats. Interestingly, Cohen et al. reported a 29% decrease in the number of neurons in the neighboring cingulate cortex of 16-month-old TgF344-AD rats but did not report the volume of the analyzed tissue.²⁰ For simplicity, we in our studies included the cingulate cortex in our estimations.²⁹ However, the presently used stereological designs do not allow detection of a minor volume reduction or neuronal loss in a subregion with the size of the cingulate cortex. Further investigations are required to obtain a better understanding of the relationship between neuronal and glial changes and the preservation of cortical volume in both AD models as well as in patients with AD.

From the combined analyses of the clusters and the microglial population response in the neocortex of the TgF344-AD rat and the APP/PS1 mouse model, it becomes evident that the formation of microglial clusters represents the earliest detectable microglial response using a stereological approach. The expansion of the total microglial population occurs later when Aβ plaque pathology is more pronounced. In the TgF344-AD rat model, the microglial clusters appeared in both the hemizygous and homozygous TgF344-AD rats, whereas only the homozygous rats showed an increase in the total microglial population at 17 months of age. The conservation of microglial cluster formation across species and across the differences in gene-dosages, points to an important role of the cellular events that are associated with microglial cluster formation.

In summary, the results of our study document the occurrence of gene-dosage- as well as species-dependent differences in microglial population expansion to Aβ plaque pathology in TgF344-AD rats and APP/PS1 mice. At the same time, the concensus finding that microglial population expansion is preceded by microglial clustering around Aβ plaques, across species and gene-dosages, suggests cluster-formation to be a key event in AD neuropathology. Collectively, the study thereby underscores the strength of comparative analyses of the microglial population response to increase insight into specific cellular events.

Footnotes

Acknowledgements

We want to thank technician Sussanne Petersen for the genotyping of the mouse AD model. Chemist Poul Henning Madsen, Department of Molecular Diagnostics, Aalborg University Hospital, is acknowledged for assistance with the genotyping of the rat AD model. Animal technicians Helle Christiansen at Aalborg University Animal Facility, and Majken Lyhne Jensen at the Biomedical Laboratory, University of Southern Denmark, are acknowledged for their support with the animal handling.

ORCID iDs

Thomas Krøigård

Malene P Stoico

Bente Finsen

Ethical considerations

Animal experiments were conducted under permission from the National Danish Animal Research Committee (Mice: 2005/561-1068, 2011/561-1950, 2013-15-2934-00814 and 2012-15-2935-00023) and (Rats: 2019-15-0201-00215).

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Author contribution(s)

Julie Hansen: Formal analysis; Investigation; Methodology; Writing – original draft; Writing – review & editing.

Rowda Mohamed: Formal analysis; Investigation.

Laura Ilkjær: Formal analysis; Investigation.

Thomas Krøigård: Formal analysis; Investigation.

Mithula Sivasaravanaparan: Investigation.

Malene P Stoico: Formal analysis.

Alicia A. Babcock: Data curation; Supervision; Writing – review & editing.

Mark J. West: Conceptualization; Resources; Writing – review & editing.

Ove Wiborg: Conceptualization; Resources; Writing – review & editing.

Bente Finsen: Conceptualization; Data curation; Formal analysis; Project administration; Resources; Supervision; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Sygeforsikring Danmark (OW), the Danish Alzheimer Research Foundation (BF), and the Danish Medical Research Council (AB, BF).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data presented in this study are available from the corresponding author upon reasonable request.

References

Batsch

Mittelman

and Alzheimer's Disease International . World Alzheimer Report 2012. Overcoming the stigma of dementia. London, UK: Alzheimer's Disease International, 2012.

Simon

Obst

Gomez-Nicola

. The evolving dialogue of microglia and neurons in Alzheimer's disease: microglia as necessary transducers of pathology. Neuroscience 2019; 405: 24–34.

Kunkle

Grenier-Boley

Sims

, et al. Genetic meta-analysis of diagnosed Alzheimer's disease identifies new risk loci and implicates Aβ, tau, immunity and lipid processing. Nat Genet 2019; 51: 414–430.

Lambert

Grenier-Boley

Chouraki

, et al. Implication of the immune system in Alzheimer's disease: evidence from genome-wide pathway analysis. J Alzheimers Dis 2010; 20: 1107–1118.

Bellenguez

Kucukali

Jansen

, et al. New insights into the genetic etiology of Alzheimer's disease and related dementias. Nat Genet 2022; 54: 412–436.

Akiyama

Nishimura

Kondo

, et al. Expression of the receptor for macrophage colony stimulating factor by brain microglia and its upregulation in brains of patients with Alzheimer's disease and amyotrophic lateral sclerosis. Brain Res 1994; 639: 171–174.

Gomez-Nicola

Fransen

Suzzi

, et al. Regulation of microglial proliferation during chronic neurodegeneration. J Neurosci 2013; 33: 2481–2493.

Olmos-Alonso

Schetters

Sri

, et al. Pharmacological targeting of CSF1R inhibits microglial proliferation and prevents the progression of Alzheimer's-like pathology. Brain 2016; 139: 891–907.

Sivasaravanaparan

Olesen

Severino

, et al. Efficacy of chronic paroxetine treatment in mitigating amyloid pathology and microgliosis in APPSWE/PS1DeltaE9 transgenic mice. J Alzheimers Dis 2022; 87: 685–699.

10.

Spangenberg

Lee

Najafi

, et al. Eliminating microglia in Alzheimer's mice prevents neuronal loss without modulating amyloid-beta pathology. Brain 2016; 139: 1265–1281.

11.

Serrano-Pozo

Gomez-Isla

Growdon

, et al. A phenotypic change but not proliferation underlies glial responses in Alzheimer disease. Am J Pathol 2013; 182: 2332–2344.

12.

Galatro

Holtman

Lerario

, et al. Transcriptomic analysis of purified human cortical microglia reveals age-associated changes. Nat Neurosci 2017; 20: 1162–1171.

13.

Gosselin

Skola

Coufal

, et al. An environment-dependent transcriptional network specifies human microglia identity. Science 2017; 356: eaal3222.

14.

Olah

Menon

Habib

, et al. Single cell RNA sequencing of human microglia uncovers a subset associated with Alzheimer's disease. Nat Commun 2020; 11: 6129.

15.

Sankowski

Böttcher

Masuda

, et al. Mapping microglia states in the human brain through the integration of high-dimensional techniques. Nat Neurosci 2019; 22: 2098–2110.

16.

Srinivasan

Friedman

Etxeberria

, et al. Alzheimer's patient microglia exhibit enhanced aging and unique transcriptional activation. Cell Rep 2020; 31: 107843.

17.

Streit

Khoshbouei

Bechmann

. Dystrophic microglia in late-onset Alzheimer's disease. Glia 2020; 68: 845–854.

18.

Askew

Olmos-Alonso

, et al. Coupled proliferation and apoptosis maintain the rapid turnover of microglia in the adult brain. Cell Rep 2017; 18: 391–405.

19.

Reu

Khosravi

Bernard

, et al. The lifespan and turnover of microglia in the human brain. Cell Rep 2017; 20: 779–784.

20.

Cohen

Rezai-Zadeh

Weitz

, et al. A transgenic Alzheimer rat with plaques, tau pathology, behavioral impairment, oligomeric aβ, and frank neuronal loss. J Neurosci 2013; 33: 6245–6256.

21.

Jankowsky

Slunt

Ratovitski

, et al. Co-expression of multiple transgenes in mouse CNS: a comparison of strategies. Biomol Eng 2001; 17: 157–165.

22.

Hsiao

Chapman

Nilsen

, et al. Correlative memory deficits, Abeta elevation, and amyloid plaques in transgenic mice. Science 1996; 274: 99–103.

23.

Holcomb

Gordon

McGowan

, et al. Accelerated Alzheimer-type phenotype in transgenic mice carrying both mutant amyloid precursor protein and presenilin 1 transgenes. Nat Med 1998; 4: 97–100.

24.

Selkoe

. Alzheimer's disease: genes, proteins, and therapy. Physiol Rev 2001; 81: 741–766.

25.

Wirenfeldt

Dalmau

Finsen

. Estimation of absolute microglial cell numbers in mouse fascia dentata using unbiased and efficient stereological cell counting principles. Glia 2003; 44: 129–139.

26.

Wirenfeldt

Dissing-Olesen

Anne Babcock

, et al. Population control of resident and immigrant microglia by mitosis and apoptosis. Am J Pathol 2007; 171: 617–631.

27.

Jankowsky

Fadale

Anderson

, et al. Mutant presenilins specifically elevate the levels of the 42 residue beta-amyloid peptide in vivo: evidence for augmentation of a 42-specific gamma secretase. Hum Mol Genet 2004; 13: 159–170.

28.

Babcock

Ilkjær

Clausen

, et al. Cytokine-producing microglia have an altered beta-amyloid load in aged APP/PS1 Tg mice. Brain Behav Immun 2015; 48: 86–101.

29.

Lyck

Krøigård

Finsen

. Unbiased cell quantification reveals a continued increase in the number of neocortical neurones during early post-natal development in mice. Eur J Neurosci 2007; 26: 1749–1764.

30.

West

Ostergaard

Andreassen

, et al. Estimation of the number of somatostatin neurons in the striatum: an in situ hybridization study using the optical fractionator method. J Comp Neurol 1996; 370: 11–22.

31.

Stalder

Phinney

Probst

, et al. Association of microglia with amyloid plaques in brains of APP23 transgenic mice. Am J Pathol 1999; 154: 1673–1684.

32.

Serrano-Pozo

Mielke

Gomez-Isla

, et al. Reactive glia not only associates with plaques but also parallels tangles in Alzheimer's disease. Am J Pathol 2011; 179: 1373–1384.

33.

Itagaki

McGeer

Akiyama

, et al. Relationship of microglia and astrocytes to amyloid deposits of Alzheimer disease. J Neuroimmunol 1989; 24: 173–182.

34.

Bolmont

Haiss

Eicke

, et al. Dynamics of the microglial/amyloid interaction indicate a role in plaque maintenance. J Neurosci 2008; 28: 4283–4292.

35.

Liu

Condello

Schain

, et al. CX3CR1 In microglia regulates brain amyloid deposition through selective protofibrillar amyloid-beta phagocytosis. J Neurosci 2010; 30: 17091–17101.

36.

Bondolfi

Calhoun

Ermini

, et al. Amyloid-associated neuron loss and gliogenesis in the neocortex of amyloid precursor protein transgenic mice. J Neurosci 2002; 22: 515–522.

37.

Bornemann

Wiederhold

Pauli

, et al. Abeta-induced inflammatory processes in microglia cells of APP23 transgenic mice. Am J Pathol 2001; 158: 63–73.

38.

Sankar

, et al. ULK1 Phosphorylation of striatin activates protein phosphatase 2A and autophagy. Cell Rep 2021; 36: 109762.

39.

Baumeister

Gellersen

Polk

, et al. Disease stage-specific atrophy markers in Alzheimer's disease. Alzheimers Dement 2025; 21: e70482.

40.

Regeur

Jensen

Pakkenberg

, et al. No global neocortical nerve cell loss in brains from patients with senile dementia of Alzheimer's type. Neurobiol Aging 1994; 15: 347–352.