Quantitative Microglia Morphological Features Correlate with Diffusion MRI in 2-month-old 3xTg-AD Mice

Abstract

Microglia (MØ) morphologies are closely related to their functional state and have a central role in the maintenance of brain homeostasis. It is well known that inflammation contributes to neurodegeneration at later stages of Alzheimer’s Disease, but it is not clear which role MØ-mediated inflammation may play earlier in the disease pathogenesis. We have previously reported that diffusion MRI (dMRI) is able to detect early myelin abnormalities present in 2-month-old 3xTg-AD (TG) mice; since MØ actively participate in regulating myelination, the goal of this study was to assess quantitatively MØ morphological characteristics and its association with dMRI metrics patterns in 2-month-old 3xTg-AD mice. Our results show that, even at this young age (2-month-old), TG mice have statistically significantly more MØ cells, which are overall smaller and more complex, compared with age-matched normal control mice (NC). Our results also confirm that myelin basic protein is reduced in TG mice, particularly in fimbria (Fi) and cortex. Additionally, MØ morphological characteristics, in both groups, correlate with several dMRI metrics, depending on the brain region examined. For example, the increase in MØ number correlated with higher radial diffusivity (D_⊥) (r= 0.59, p= 0.008), lower fractional anisotropy (FA) (r= −0.47, p= 0.03), and lower kurtosis fractional anisotropy (KFA) (r= −0.55, p= 0.01) in the CC. Furthermore, smaller MØ cells correlate with higher axial diffusivity (D_∥) in the HV (r= 0.49, p= 0.03) and Sub (r= 0.57, p= 0.01). Our findings demonstrate, for the first time, that MØ proliferation/activation are a common and widespread feature in 2-month-old 3xTg-AD mice and suggest that dMRI measures are sensitive to these MØ alterations, which are associated in this model with myelin dysfunction and microstructural integrity abnormalities.

1. Introduction

Microglia are the immune cells of the brain and key players in brain injury and disease^1,2. Their distribution, morphology, and numbers are influenced by local cytoarchitecture and vary considerably^3,4. Microglial morphology consists of a distinct cell body (soma), from which elongated ramified processes are projected; however, the number, length, and complexity of branching of the processes can differ widely depending on the brain region, particularly in reactive situations^2,3. While white matter microglia show elongated soma and processes preferentially oriented along fiber tracts, in gray matter microglia exhibit various elaborate radially oriented arbors³. Microglia actively adapt their cell morphology and function in response to pathological situations, in a process called “deramification”^2,5. In this process, ramified microglia can transform progressively into an “activated state”, characterized by swollen ramified cells with a larger cell body and shorter, thicker processes. Alternatively, microglia can adopt a “reactive state” which are typically small, spherical cells, but may also exhibit rod-shape or amoeboid-like morphologies with high phagocytic activity.

Microglia morphologies are closely related to their functional state⁵ and have a central role in the maintenance of brain homeostasis under healthy conditions. In their dynamic surveilling state⁶, in which microglia are characterized by a ramified morphology, they scan their environment while having direct interactions with neurons, blood vessels and other glial cells, as they participate in pruning synapses, neurotransmitter signaling and synaptic transmission, regulating neuronal activity^7–11. They also actively participate in myelination, both during development and homeostatic myelin maintenance^12,13, through the phagocytosis of excess oligodendrocytes and myelin membrane that forms abnormally. In addition, microglia support oligodendrocyte survival and differentiation and regulate apoptosis of oligodendrocytes by secreting several trophic factors and other intercellular signaling molecules, including chemokines and cytokines^12,13.

Alterations in microglial morphology and function with age and their association with neurodegeneration are well demonstrated^14–16, and microglia-mediated neuroinflammation has been implicated in the pathogenesis of Alzheimer’s disease (AD)^17–21. Furthermore, there are indications that microglial pathology appears earlier in mouse models of AD²², with microglial numbers significantly increasing in the subiculum of 4-month-old 5xFAD mice²³. Additionally, microglia with the white matter (WM) associated gene signature (WAM), form more nodules containing myelin debris in 2-month-old 5xFAD mice²⁴ than in control mice. Nonetheless, the role of microglia dysfunction early in the AD disease process is not well understood.

In the 3xTg-AD mouse model, which develops both amyloid-β (Aβ) and neurofibrillary tangles in a temporal and spatial pattern similar to that of human AD pathology^25–27, microglia activation has been reported to be first detected at 6 months of age^28,29. However, abnormal myelination patterns with changes in oligodendrocyte and myelin marker expression are the earliest pathological feature described in the 3xTg-AD mice, leading to WM disruption in the hippocampus and entorhinal cortex by 2 months of age^30,31. These WM changes are followed by the first signs of cognitive impairment, detected even before Aβ plaques appear³².

Diffusion MRI (dMRI) is widely used as a noninvasive means of detecting changes in brain tissue microstructure associated with pathology³³. Its sensitivity to microstructure stems from its ability to quantify the motion of water molecules on length scales commensurate with the size of neurons and glia^33,34. In prior studies, we have found dMRI metrics to change longitudinally with aging³⁵ and AD pathology progression³⁶ in several brain regions, along with evidence that they are sensitive to the larger number of cholinergic neurons reported in the basal forebrain (BF) of these young 3xTg-AD mice³⁵. Moreover, our work supports the ability of dMRI to detect the early myelin abnormalities described in 2-month-old 3xTg-AD mice³⁷, particularly in the corpus callosum, fimbria, and fornix. Since microglia actively participates in regulating myelination¹³, we hypothesize that dMRI will be able to detect and correlate with abnormal microglia morphology in these young mice. Therefore, our aim in this study is to quantitatively investigate morphological characteristics of microglia and their association with dMRI metrics for selected brain regions in young 2-month-old 3xTg-AD mice, which may help inform the application of dMRI to the in vivo assessment of inflammation in AD and other neuropathologies.

2. Methods

2.1. AD mouse model

All experimental procedures were approved by the Institutional Animal Care and Use Committee at the Medical University of South Carolina and conducted in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals. The 3xTg-AD model possesses the three mutations of human presenilin-1 M146V, human amyloid precursor protein Swedish mutation, and the P301L mutation of human tau, based on human mutations identified in familial AD^25,32. The 3xTg-AD (TG) [B6;129-Psen1tm1Mpm Tg(APPSwe, tauP301L)1Lfa/Mmjax; MMRRC Stock No: 34830-JAX|3xTg-AD] and age-matched controls (NC) [101045 B6129SF2/J] were purchased from The Jackson Laboratory. The control mice for the 3xTg-AD are the suggested controls for genetically engineered strains generated with 129-derived embryonic stem cells and maintained on a mixed B6;129 background. Only 2-month-old female mice were used in this study since the development of AD pathology is far less pronounced and more variable in male 3xTg-AD mice^38,39. All mice were housed in temperature- and humidity-controlled rooms on a 12-h light/dark cycle (lights on at 6:00 AM) in an accredited animal care facility. In this study, a total of 11 TG and 8 NC mice were used for analysis of microglia morphological characteristics and for performing the correlations with dMRI metrics.

2.2. MRI acquisition

Mice were anesthetized using an isoflurane set at the following percentages: 3% for induction, 2% during pilot scanning and data acquisition. An animal monitoring unit (SA instruments, Inc., model 1025, Stony Brook, NY) was used to record respiration and rectal temperature. Respiration was measured with a pressure transducer placed under the abdomen just below the ribcage. Body temperature was maintained using ventilated warm air, controlled by a feedback circuit between the heater and thermistor. After induction, mice were placed on a mouse holder and restrained using a mouse tooth bar (Bruker, T10146) and ear bars (Bruker, T10147) placed in the auditory canal. Compressed air was used as the carrier gas and delivered at a flow rate of 1L/min to a nose cone positioned around the tooth bar, where gases mixed with air and passed over the rodent’s nose. All animals were maintained at 37.0 ± 0.2 °C and respiration ranged between 60 and 80 breaths per minute during scanning.

The in vivo MRI experiments were all performed on a 7T Bruker Biospin 30 cm bore scanner (BioSpec 70/30 USR) running Paravision version 5.1. This system is equipped with a 12 cm high performance B-GA 12S2 gradient and shim coil set, capable of generating a maximum gradient amplitude of 440 mT/m and a slew rate of 4570 T/m/s. A 86 mm 1H quadrature volume coil (T128038) was used for signal transmission, and an actively decoupled phase array coil (four channel receiver T11765) was used for signal reception. For dMRI data acquisition, we employed a diffusional kurtosis imaging (DKI) protocol^40–42. DKI is an extension of diffusion tensor imaging (DTI) that includes dMRI data acquired with stronger diffusion weightings than is typical for DTI and requires at least 2 different nonzero b-values⁴¹. DKI provides all the diffusion measures available with DTI plus additional measures related to the kurtosis of the water diffusion displacement probability density function⁴². The main imaging parameters were: a 2-shot SE-EPI sequence with TR/TE = 3750/32.6 ms, δ/Δ = 5/18ms, slice thickness = 0.7 mm, 15 slices with no gap, data matrix = 128×128, image resolution = 156×156 μm², 2 signal acquisitions, 10 b-value = 0 (b0) images, followed by 30 diffusion encoding gradient directions with 4 b-values for each gradient direction (0.5, 1.0, 1.5, and 2.0 ms/μm²) and fat suppression flip angle = 105°. Total acquisition time = 33 minutes.

2.3. DKI post-processing and Image analysis

DKI post-processing was performed using Diffusional Kurtosis Estimator (DKE)⁴³ (http://nitrc.org/projects/dke). Post-processing included motion correction by aligning all DWIs to the first b0 image using SPM12 (Wellcome Trust Centre for Neuroimaging, UCL, UK). Parametric maps were obtained by fitting dMRI signal measurements to the DKI signal model for each voxel using a linearly constrained weighted linear least squares fitting algorithm, which generated the diffusion and kurtosis tensors. These two tensors were then used to calculate parametric maps for several diffusion metrics^42,44. Figure 1 illustrates, for a 3xTg-AD mouse, the maps for all diffusion metrics utilized in this study. In our analysis, we examined the four diffusion tensor (DT) parameters of mean diffusivity (MD), axial diffusivity (D_∥), radial diffusivity (D_⊥), and fractional anisotropy (FA). MD corresponds to the diffusivity averaged over all diffusion directions, D_∥ corresponds to the diffusivity in the direction of the principal diffusion tensor eigenvector, D_⊥ corresponds to the diffusivity averaged over all diffusion directions perpendicular to the principal diffusion tensor eigenvector, and FA quantifies the anisotropy of the diffusion tensor. We similarly considered the four diffusional kurtosis (DK) metrics of mean kurtosis (MK), axial kurtosis (K_∥), radial kurtosis(K_⊥), and kurtosis fractional anisotropy (KFA). These are kurtosis analogs of the DT metrics that quantify diffusional non-Gaussianity and provide complementary information about the diffusion dynamics^42,44. All four b-values (0.5, 1.0, 1.5, 2.0 ms/μm²), for each gradient direction, were used in calculating the diffusion metrics. It is worth noting that, due to the inclusion of non-Gaussian effects, DKI yields more accurate estimates of DT metrics than does conventional DTI⁴⁵, as well as enabling a more comprehensive assessment of the diffusion microenvironment in brain tissue.

Figure 1.

Representative diffusion MRI parametric maps of all diffusion measures, for a single anatomical slice, from a 3xTg-AD mouse. The top row shows the DT measures of mean diffusivity (MD), axial diffusivity (D_∥), radial diffusivity (D_⊥) and fractional anisotropy (FA). The bottom row shows the DK measures of mean kurtosis (MK), axial kurtosis (K_∥), radial kurtosis (K_⊥) and kurtosis fractional anisotropy (KFA). Scale bars: 0–1 for FA and KFA; 0–2 μm²/ms for MD, D_∥ and D_⊥; 0–3 for MK, K_∥, and K_⊥.

Regions of interest (ROIs) were defined for corpus callosum (CC), fimbria (Fi), hippocampus [including dorsal (HD), ventral (HV) and subiculum (Sub) areas], and cortex [ cingulate (Ctx_Cg) and retrosplenial cortex (Ctx_Rsp)], which are all brain regions relevant to AD pathology. All ROIs were manually drawn on the averaged b0 image by a neuropathologist (MFF), using ImageJ (http://imagej.nih.gov/ij//)⁴⁶. Anatomical guidelines for outlining these regions were determined by comparing anatomical structures in the MRI slices with a standard mouse atlas⁴⁷ and verified with the FA maps to ensure correct anatomical location and to avoid contamination of unintended tissue or cerebrospinal fluid (CSF). Figure 2 shows FA maps for a 3xTg-AD mouse with the selected ROIs drawn. The regional values for the dMRI measures were obtained by averaging all voxels within an ROI except those with MD > 1.5 μm²/ms, which were excluded to minimize the effect of CSF contamination.

Figure 2:

Representative fractional anisotropy (FA) maps of a 2-month-old 3xTg-AD mouse illustrating examples of the region of interest (ROI): corpus callosum (CC - red), fimbria (Fi - purple), dorsal (HD - blue) and ventral (HV - yellow) hippocampus, subiculum (Sub – light green), cingulate cortex (Ctx_Cg – orange) and retrosplenial cortex (Ctx_Rsp - dark green).

2.4. Histological protocol

Following the MRI scan, mice from each group (TG, n = 11; NC, n = 8) were perfused with ice-cold phosphate buffered saline (PBS) followed by paraformaldehyde (4%) in PBS. Mice brains were then removed and post-fixed overnight in the same fixative, and subsequently transferred to a storage buffer PBS with 0.02% sodium azide until final processing by NeuroScience Associates (Knoxville, TN, USA). The brains were embedded into a gelatin matrix, frozen and sectioned from one solid block (MultiBrain^® Technology). Sections of 25 μm thickness were taken in the coronal plane, wet-mounted on 2% gelatin-subbed slides, and stained using ionized calcium binding adaptor molecule (Iba1) (Abcam ab178846, Cambridge, MA, USA) aimed at microglia (MØ) identification. To assess myelin and axonal abnormalities, subsequent sections were stained for myelin basic protein (MBP; Abcam ab40390, Cambridge, MA, USA), and neurofilament (NF; Neurofilament-200, Sigma Aldrich N4142, Burlington, MA, USA). Histological slices were digitized in one session using an Olympus BX53 microscope (Olympus Corporation of the Americas, Center Valley, PA) with QImaging digital camera and QCapture Suite. Images (40x magnification) were acquired for MØ and images (4x magnification) were acquired for the MBP and NF, under the same image light and camera settings to avoid possible image intensity variation, and then saved as 8-bit RGB images.

The quantification of the immunoreactivity for both MBP and NF was done by averaging the mean grayscale intensity measurements from all pixels in each ROI, which were manually drawn based on the mouse brain atlas⁴⁷ by a neuropathologist (MFF) on the same regions as for the dMRI data, in both hemispheres. MBP and NF values were expressed as optical density (OD) = log (max intensity/mean intensity)³⁷, where max intensity = 255 for the 8-bit images.

For the 2D-morphological MØ evaluation, IBA-1 labeled 40x images were used and depending on the ROI, either four (CC, Fi, HD, Sub) or six (HV, Ctx_Cg and Ctx_Rsp) images for each mouse were analyzed, generating a large number of images for each ROI (e.g.,18 mice and 6 HV images/per mouse = 108 images). Figure 3 illustrates the series of steps performed in order to analyze the MØ morphological aspects, which included: 1) conversion of RGB image to grayscale using maximum intensity projection (MIP) to retain intensity features from each channel; 2) noise removal with 5 × 5 low-pass Weiner filter; 3) removal of uneven background illumination with a morphological top-hat filter; 4) removal of small noise-like features with fast-Fourier transform (FFT) bandpass filter, where the lower and upper bounds were 50 and 100 Hz respectively; 5) image sharpening with unsharpen mask using a Gaussian lowpass filter with σ = 3; and 6) rescaling the image from 0 to 1. Additionally, a marker-controlled watershed segmentation algorithm⁴⁸ was employed to balance over- and under-segmentation of overlapping cellular regions. Since the slice thickness of each histological section is 25 μm, the 2D images contain partial branches, fragmented and/or overlapping MØ cells; thus, to overcome this issue, we also excluded cells with only two endpoints, cells with branch length and/or sum of all branch lengths less than 5μm, and when the pixels area occupied by the cell was less than 100 μm².

Figure 3:

Image processing and segmentation strategy for microglia quantification. The original image (A, enlargement Ai) was first converted to grayscale with maximum intensity projection and inverted, followed by denoising, removal of uneven background illumination, fast Fourier transform bandpass filtering, image sharpening, and rescaling to prepare a processed image (B, enlargement Bi) for segmentation. Soma locations were marked as regional minima to produce a composite image for watershed segmentation (C, enlargement Ci), which yielded a segmented and labelled image (D, enlargement Di) where the cell perimeter defined by the blue boundary, and the skeletonized microglia by the red line.

The analysis was performed with a MATLAB script (R2022a; 9.12.0.1884302) based on the widely used fractal analysis (FracLac plugin) for ImageJ^49–53 and skeletonize analysis (Analyze Skeleton plugin)^54,55. The MATLAB script facilitated the analysis of the large number of images containing many MØ cells by automatically running the image processing and segmentation steps, labeling MØ cells and generating the parameters values. Nevertheless, after the image processing and segmentation steps were performed, all images and labeled MØ cells were reviewed, blinded to the mouse group, by a neuropathologist (MFF), so that only individual and complete MØ cells, not in contact with the edges of the image, fragmented, or in cell clusters were counted (total number of MØ) and included for analysis. For this study, we assessed the number of MØ cells and employed fractal analysis in order to report on cell complexity (fractal dimension), elongation of microglia in WM ROIs (span ratio) and MØ cell size (density) in 2-month-old 3xTg-AD and control mice. Figure 4 shows the pictorial representation and definition of the fractal analysis parameters examined for the microglia morphological analysis.

Figure 4:

Pictorial representation and definition of the fractal analysis parameters examined for the microglia morphological analysis.

2.5. Statistical analyses

Summary statistics are expressed as group-averaged means ± standard error of the mean (SEM) for dMRI metrics and morphological measurements (MBP/NF/MØ). Two-tailed, two-sample t-tests, assuming unequal variances, were performed to assess group differences in each ROI for all measurements. To evaluate the relationship between measures of microglia morphology and dMRI changes, Spearman correlation coefficients were calculated. Correlations were done for all mice pooled together due to the small number of mice in each group. A p-value of ≤ 0.05 was considered to be statistically significant. Given the exploratory nature of this study, no adjustments for multiple comparison corrections were performed. Analyses were conducted using GraphPad Prism (v.9.1.2).

3. Results

In the present study, we performed a quantitative morphological evaluation of the MØ cells and how MØ number and morphological changes, as determined by fractal analysis parameters, correlates with MBP and NF degree and dMRI metrics in different brain regions of 2-month-old 3xTg-AD mice. MØ results for each ROI are presented in Figure 5. Bar graphs show the distribution of the data points for each mouse group and means ± standard error of the mean (SEM) for the following parameters: A) the number of MØ; B) span ratio; C) fractal dimension; D) density. Since span ratio is a measure of MØ elongation, it is relevant and presented only for WM ROIS (CC and Fi). We also note that the density parameter, which is an indicator of MØ cell size, is calculated by taking into consideration the cell perimeter and the convex hull area (Fig. 4); therefore, a higher density value reveals smaller MØ cells.

Figure 5:

Bar graphs illustrating the microglia morphological quantitative analysis (Iba1 stain) for each ROI; graphs show the distribution of the data points for each mouse group and means ± standard error of the mean (SEM) for the following microglia morphological parameters: A) Number of microglia cells; B) Span ratio; C) Fractal dimension; D) Density. Region of interest (ROI) abbreviations: corpus callosum (CC), fimbria (Fi), dorsal (HD) and ventral (HV) hippocampus, subiculum (Sub), cingulate cortex (Ctx_Cg) and retrosplenial cortex (Ctx_Rsp); p-values at the level of * p ≤0.05; ** p ≤0.01; ***p ≤0.001

Our results demonstrate that, even at this young age (2-month-old), the number of MØ cells is increased in TG mice, statistically significant for almost all ROIs examined, except HD (Fig.5A). TG mice MØ cells also presented with significantly different morphological characteristics compared with NC mice, depending on the brain region examined. In the CC, despite a trend for less elongated (span ratio; Fig.5B), more complex (fractal dimension; Fig.5C) and smaller (density; Fig.5D) cells, no statistically significant group differences were observed for these parameters. In the Fi, MØ cells were significantly less elongated (span ratio; Fig.5B) and more complex (fractal dimension; Fig.5C) in the TG mice. In the HD, although TG mice appears to have more MØ cells than NC mice, this difference did not reach statistical significance (Fig.5A); however, MØ cells were significantly smaller (density; Fig.5D) in TG mice compared to NC mice. In the HV, Sub, Ctx_Cg and Ctx_Rsp, TG mice had significantly more MØ cells, which were more complex (fractal dimension; Fig.5C) and smaller (density; Fig.5D) than did NC mice.

Results for the quantification of MBP and NF immunoreactivity in WM ROIs are presented in Table 1A, showing significantly reduced MBP in the Fi of the TG mice, with trends for NF decrease in the Fi, and trends of both NF and MBP decrease in the CC. Table 1B, shows Spearman’s correlation coefficient (r) with statistically significant p-values for the association between several MØ morphological parameters and MBP and NF immunoreactivity degree in these two WM ROIs.

Table 1A:

Quantitative analysis of the myelin basic protein (MBP) and neurofilament (NF) immunoreactivity; group-averaged means ± standard error (SEM) in each white matter brain region (ROI). 3xTg-AD (TG) and normal control (NC) mice; corpus callosum (CC); fimbria (Fi).

ROI	Variable	Mean TG	Mean NC	Mean Difference	SEM TG	SEM NC	p-value
CC	MBP	34.58	38.00	−3.42	0.97	1.51	0.062
CC	NF	30.50	33.41	−2.91	1.02	2.19	0.206
Fi	MBP	32.84	37.37	−4.53	0.82	1.73	0.019
Fi	NF	29.53	32.68	−3.15	0.96	1.24	0.057

Table 1B:

Spearman’s correlation values (r) and p-values measuring the association between MØ morphological parameters and myelin basic protein (MBP) and neurofilament (NF) immunoreactivity in each white matter brain region (ROI). corpus callosum (CC); fimbria (Fi); both groups pulled together (ALL).

ROI	Group	MØ Metric	MBP/NF	r	p-value
CC	ALL	Fractal	NF	−0.592	0.008
Fi	ALL	Fractal	MBP	−0.522	0.022
Fi	ALL	Span Ratio	MBP	0.513	0.025
CC	ALL	Cells n	NF	−0.497	0.030
CC	ALL	Span Ratio	MBP	0.496	0.031
CC	ALL	Density	NF	−0.459	0.050

Table 2 displays the Spearman’s correlation coefficient (r) with statistically significant p-values for the association between MØ morphological parameters and dMRI metrics ranked by higher p-value for each ROI. Our results indicate that MØ morphological characteristics are significantly correlated with several dMRI metrics, depending on the brain region examined. We highlight that more MØ cells in the CC correlates with higher K_⊥ and lower FA and KFA.

Table 2:

Spearman’s correlation values (r) and p-values measuring the association between MØ morphological parameters and diffusion MRI (dMRI) metrics for each region of interest (ROI), ranked by higher p-value; region of interest (ROI) abbreviations: corpus callosum (CC), fimbria (Fi), ventral (HV) hippocampus, subiculum (Sub), cingulate cortex (Ctx_Cg) and retrosplenial cortex (Ctx_Rsp). Mean diffusivity (MD), axial diffusivity (D_||), radial diffusivity (D_⊥), fractional anisotropy (FA), radial kurtosis (K_⊥), and kurtosis fractional anisotropy. NC: control mice; TG: 3xTg-AD mouse; ALL: both groups pulled together.

ROI	Group	MØ Metric	dMRI Metric	r	p-value
CC	ALL	Cells n	D⊥	0.590	0.008
Sub	ALL	Density	D∥	0.573	0.010
CC	ALL	Cells n	KFA	−0.549	0.015
Ctx Cg	ALL	Density	KFA	0.539	0.017
Fi	ALL	Span Ratio	FA	0.514	0.024
CC	ALL	Span Ratio	K⊥	0.513	0.025
Fi	ALL	Cells n	D∥	−0.469	0.043
CC	ALL	Cells n	FA	−0.496	0.031
Fi	ALL	Span Ratio	KFA	0.490	0.033
HV	ALL	Density	D∥	0.487	0.034
Ctx Rsp	ALL	Cells n	D∥	0.476	0.039
Sub	ALL	Density	MD	0.456	0.050

A representative of the Iba1 stain (40× objective) is shown on Figure 6 exemplifying the MØ morphological aspects for each group for Fi and Sub. The images display the Fi-NC MØ cells with a more elongated and smaller soma compared with Fi-TG MØ cells, which shows an increase in number of less elongated MØ, with a larger soma and shorter branches. In the Sub-TG, MØ number is also increased and characterized by cells with a larger cell body and shorter, thicker branches. These aspects reflect an activated MØ state in these regions.

Figure 6:

Representative images of the Iba1 stain (40× objective) for fimbria (Fi) and subiculum (Sub), for both normal control (NC) and 3xTg-AD (TG) mice exemplifying the microglia morphological aspects of each group. The images display the Fi-NC MØ cells with a more elongated and smaller soma compared with Fi-TG, which shows an increase in number of less elongated MØ, with a larger soma and shorter branches. In the Sub-TG, MØ number is also increased and characterized by cells with a larger cell body and shorter, thick branches.

4. Discussion

The morphology and distribution of MØ cells varies among different regions of the brain^2,3, reflecting its sensitivity to the surrounding microenvironment. In this study, we focused our analysis on the comparison of MØ morphological features between NC and TG mice and explored how alterations in MØ morphology are associated with dMRI measures.

When we examined the two WM ROIs (CC and Fi), where in normal conditions MØ cells would have an elongated soma and branches preferentially oriented along fiber tracts^2,3, TG mice had significantly more MØ cells (Fig.5A) in both regions, but despite a trend to have smaller and complex MØ cells, no group differences were observed in the CC. However, in the Fi, MØ cells were less elongated (Fig.5B) and more complex (Fig.5C), reflecting an activated MØ state (Fig. 6) in this WM region, which correlated with less MBP in these mice (Table 1A–B). Despite no significant MØ morphological group differences in the CC, the increase of MØ cells, mainly complex and small cells, correlated with less NF, whereas more elongated MØ cells, reflecting normal conditions, correlated with more MBP. Thus, it is likely that both MØ morphological changes and myelin defects present in the TG mice at this young age are combining factors driving these correlations. When considering the dMRI metrics (Table 2), the increase of MØ cells in the CC correlated with higher in D_⊥ and lower FA and KFA, while more elongated MØ, which is the normal cells aspect in WM, correlated with higher K_⊥. In the Fi, more typical elongated MØ cells (Fig.5B) correlated with higher FA and KFA (Table 2), showing that MØ elongation in this region is associated with microstructural anisotropy as quantified by dMRI. However, the increase in the number of MØ cells was associated with lower D_∥, probably reflecting a more restricted diffusion environment. As we previously reported in these two WM regions³⁷, TG mice have lower FA, KFA and K_⊥, and higher diffusivity compared with NC mice supporting the notion that dMRI can detect early WM abnormalities representing a combination of myelin abnormalities and the presence of activated MØ cells in young 2-month-old 3xTg-AD mice.

In the HD, TG mice had smaller MØ cells (Fig.5D) than NC mice, but no correlations with dMRI metrics were observed. As seen in Fig. 5 (A, C, D) and Table 2, the TG mice HV and Sub had significantly more MØ cells than NC mice, which were smaller and more complex, all features of activated MØ cells, and the presence of small MØ cells correlated with higher D_∥. These results suggest that the presence of small MØ in the HV and Sub are associated with a less restricted diffusion environment, which is an indication of early microstructural abnormalities in the hippocampus of young TG mice. Indeed, our previous results from a large dMRI data set³⁷ shows that TG mice has statistically significant higher MD, D_∥ and D_⊥ in all regions of the hippocampus, which are likely related to changes in the structural integrity of the hippocampal complex, associated with early MØ-mediated neuroinflammation reported in young 3xTg-AD mice^56,57.

In the cortex (Ctx_Cg and Ctx_Rsp), TG mice had more MØ cells than NC mice, which were smaller and more complex (Fig.5 A, C, D). In the Ctx_Cg, small cells correlated with higher KFA, and in the Ctx_Rsp more MØ cells correlated with higher D_∥ (Table 2). Quantification of MBP immunoreactivity in both cortices showed significantly lower levels (Ctx_Cg; p= 0.008 and Ctx_Rsp; p= 0.002) in the 3xTg-AD mice compared to the NC, with no changes in the NF levels. Together these changes demonstrate early MØ activation associated with myelin dysfunction in the cortex of young 3xTg-AD mice.

It is well known that inflammation contributes to neurodegeneration at later stages of AD^15–17,21. Likewise, alterations in MØ number and activation have been reported in AD mouse models, particularly in the hippocampus and the cerebral cortex, mostly alongside Aβ deposition^{22,23, 29, 58–60}. However, it is not clear which role MØ-mediated inflammation may play in early disease pathogenesis. One suggestion is that MØ may have a promoting role early in AD pathogenesis^58–60, but may be beneficial at a later stage by preventing the spread of neurotoxic plaques^62–64. Additionally, activated MØ have been classified with two distinctive phenotypes, namely, the M1 (pro-inflammatory or neurotoxic) associated with expression of pro-inflammatory cytokines, and M2 (non-inflammatory or neuroprotective) characterized by the production of anti-inflammatory cytokines^65,66. Thus far, in the 3xTg-AD mice, most studies describe an increase in MØ number starting at 6 months of age, preceding significant extracellular amyloid plaque deposition, but mirroring the regional and temporal distribution of pathology observed in AD brains^28,29,61,62. However, no detailed MØ morphology or inflammatory phenotype has been described in this model at this young age. Therefore, our study is the first to report the presence of activated MØ cells in young 2-month-old 3xTg-AD mice, which at this age have increased soluble Aβ42 levels⁶⁵ and shows the beginning of intraneuronal Aβ accumulation^25–27, both known to contribute to MØ activation^65,66. We also give, for the first time, evidence supporting the sensitivity of dMRI metrics to these MØ alterations associated with myelin dysfunction and microstructural integrity abnormalities in this AD mice.

In fact, only a few dMRI studies have investigated whether alterations in MØ number would significantly affect dMRI measures, reporting that the increase in MØ number is associated with changes in dMRI metrics of axonal integrity and structural connectivity^68–69. However, important methodological factors distinguish our study from these prior investigations (e.g., ex-vivo imaging, drug induced microglia depletion and rat model of spine damage), making it difficult to directly compare results. Moreover, only one of those three studies involved studying a different mouse model of AD⁶⁸, and although they reported clustered activated MØ associated with Aβ and changes in diffusion and functional MRI, no correlations between imaging and morphological data were presented.

In summary, our findings indicate that early MØ proliferation/activation, coupled with myelin abnormalities are a common and widespread feature in 2-month-old 3xTg-AD mice. Furthermore, our results suggest that dMRI is able to detect these early MØ and myelin morphological abnormalities. This is particularly apparent in WM regions, where we previously reported FA and KFA is significantly lower in the CC and Fi of the TG mice³⁷, and in this study both regions’ MØ morphological parameters correlated with several dMRI metrics. This is also evident in the hippocampus, where we reported increase in diffusivity metrics in the TG mice³⁷, and in this study the presence of more activated MØ cells in this region correlated with higher D_∥. However, our results should be considered in the context of some methodological limitations. First, for both the morphological analysis and for the Spearman correlation with dMRI metrics, there was a relatively small number of mice in each group. Also, we have not performed adjustments for multiple comparisons, but rather presented the uncorrected results to provide a broad overview and allowing readers to draw their own conclusions based on these. Thus, these results should be thoughtfully interpreted. Further studies are needed to assess markers for MØ phenotype and cytokine production, which could help associate MØ morphological changes with their inflammatory profiles¹⁶. Finally, the role of increased soluble Aβ in MØ activation in these young 3xTg-AD mice should be investigated, as well as which role MØ-mediated inflammation, at this early age, plays in myelin dysfunction and in the progression of AD pathology.

6. Conclusion

In this study, we have demonstrated in 2-month-old 3xTg-AD mice the presence of activated MØ cells, which are associated with myelin dysfunction and microstructural integrity abnormalities and found significant correlations between their morphological characteristics and several dMRI metrics. Since the imaging methods employed here are easily translatable to clinical MRI, the results of this study are also relevant to human AD. Consequently, dMRI holds promise for early detection of neuroinflammation in AD and for monitoring the effects of future therapeutic interventions aimed to reduce microglial neurotoxicity.

Abbreviations:

AD

Alzheimer’s disease

WM

white matter

MØ

Microglia

DK

diffusional kurtosis

DKI

diffusional kurtosis imaging

DT

diffusion tensor

DTI

diffusion tensor imaging

FA

fractional anisotropy

MD

mean diffusivity

D_‖

axial diffusivity

D_⊥

radial diffusivity

KFA

kurtosis fractional anisotropy

MK

mean diffusional kurtosis

K_‖

axial kurtosis

K_⊥

radial kurtosis

CC

corpus callosum

Fi

fimbria

HD

dorsal hippocampus

HV

ventral hippocampus

Sub

subiculum

Ctx_Cg

cingulate cortex

Ctx_Rsp

retrosplenial cortex