Ann Neurol: Correlation between pathological spread of Alzheimer's disease and activation of microglia

Alzheimer disease (AD) is often characterized by β-like starch polypeptide (Aβ) plaques and neurofibrillary tau tangles, and there is growing evidence that neuroinflammation also plays a key role
driven by microglial activation.
Aβ and tau pathology appears to travel along the path of highly connected brain regions, but it remains difficult to figure
out whether microglial activation follows a similar distribution pattern.
A recent ANNALS of Neurology research article titled "Microglial Activation and Connectivity in Alzheimer Disease and Aging" evaluated the correlation
between this connectivity and microglial activation patterns 。 The authors included 32 Aβ-positive early AD subjects (18 females and 14 males) and 18 Aβ-negative age-matched healthy controls (10 females and 8 males) from the prospective ActiGliA (Aging and Alzheimer's Disease Midbrain Network Activity, Amyloid, and Microglia Activity) and Magnetic Resonance Imaging (MRI) activated by TSPO ligands [18F]GE-180 microglia, to measure its functional and structural connections
at rest.
It was found that the interregional covariance between TSPO-PET and normalized uptake ratios was preferentially distributed along functionally highly connected brain regions, while MRI structural connections showed a weak
link with microglial activation.
Compared with the control group, bilateral anteromedial TSPO-PET tracer uptake was increased in patients with AD, while higher TSPO-PET uptake was associated with cognitive impairment and dementia severity, and its distribution was related
to disease stage.

background

Extracellular β amyloid polypeptide (Aβ) plaques and intracellular nerve fiber tau tangles are pathological features of Alzheimer disease (AD), causing a series of brain changes and triggering cognitive decline and dementia
.
Recent genetic, molecular, and clinical evidence suggests that neuroimmune mechanisms are associated with AD risk and promote disease progression
.
Symptomatic AD patients exhibit changes in cerebrospinal fluid (CSF) and pro- and anti-inflammatory proteins in the blood, including soluble trigger receptor 2 (sTREM2)—markers of activated microglia expressed on myeloid cells), often surrounding Aβ plaques as the brain's primary innate immune response
.
This suggests that neuroinflammation and microglial activation play a key role
in the pathogenesis of AD.

In vivo, the activation of microglia (which is located on the outer mitochondrial membrane and overexpressed on activated immune cells) can be measured using a tracer targeting mitochondrial 18 kDa translocation protein (TSPO)
using positron emission tomography (PET).
TSPO-PET scans show increased
microglial activation in different neurodegenerative diseases, including AD, corticobasal ganglia syndrome, and progressive supranuclear palsy.
Neuroinflammation, extracellular Aβ plaques, and intracellular tau neurofibrils appear to be closely related
.
In the microglia population, two different expression profiles have recently been found to be associated
with Aβ loading or tissue phosphorylated tau, respectively.
Colocalization of Tau and activated microglia has been demonstrated, and Aβ plaque-dependent microglia have been identified
.
Furthermore, in AD mouse models, TSPO-PET binding is associated with increased Aβ accumulation, and microglial activation may also act as a link
between Aβ and tau pathology.

Several studies in AD have demonstrated that changes in functional magnetic resonance imaging (fMRI) ligation are associated
with signature protein aggregation.
Highly connected pivotal regions show increased tau accumulation on PET imaging (and Aβ pathologic shows exacerbation of this phenotypic feature), supporting a model
of the diffusion of tau molecules across neurons.
There is increasing evidence that the spread of tau tangles in the cortex is accompanied by the activation
of microglia.
Both pathological types appear to follow a Braak-stage-like distribution pattern along a network of microglial activation, suggesting that Aβ and microglia work together to initiate the spread
of tau in the brain.

Fig.
1 Voxelwise analysis
between Alzheimer's disease (AD) and cognitively normal control group (CN).
The scatterplot shows individual and mean differences
in the standardized uptake ratio (SUVR) of mitochondrial 18 kDa translocation protein (TSPO) between AD patients and CN.
The B plot shows the distribution of the average SUVR TSPO-PET binding pattern within the AD group
.

The primary objective of this study is to explore whether connected brain regions show associated TSPO-PET tracer uptake patterns (Figures 1, 2) to determine the relevance
of cranial neuroimmune responses to AD brain structural and functional network connections.
The data used in the authors' study, including TSPO-PET microglia activation, diffusion tensor imaging (DTI) structural connections, and MRI functional connections, address their hypotheses
from prospective brain networks, amyloid, and microglia activity in aging and Alzheimer's disease (Acti Gli A).
The authors speculate that microglial activation in AD spreads along connected brain regions, similar to the observed spatial transmission pattern
of tau pathology.

Figure 2: Overview of the analysis process

(1) After pretreatment, the mitochondrial 18 kDa translocation protein (TSPO)-positron emission tomography (PET) standardized uptake ratio (SUVR) map was segmented
according to the Brainnetome (BN) atlas.
(2) Next, the segmentation-based TSPO-PET SUVR of each subject is linked, vectorized, and z-transformed to eliminate intersubject differences in tracer uptake, and (3) interregion pairwise correlation is then performed to calculate the TSPO-PET covariance adjacency matrix
for each diagnostic group.
(4) In order to predict the TSPO-PET covariance from the structure/functional connectivity, the mean adjacency connectivity matrix obtained by TSPO-PET and stationary functional magnetic resonance imaging and diffusion tensor imaging magnetic resonance imaging was vectorized, and linear regression analysis was performed
.
(5) The average TSPO vector is calculated, and (6) is associated with
the average adjacency matrix of the functional and structural connection matrices.
(7) Test the association
between TSPO-PET and functional/structural connectivity and TSPO SUVR correlation vectors in linear regression analysis.
(8) Identify TSPO-PET hot and cold spot regions by ranking the average TSPO SUVR intake vector for 246 BN regions
.
(9) The relationship between TSPO-PET uptake and hot and cold spot areas in function/structural connectivity in all target regions was analyzed by regression, and the function/structural connectivity between TSPO-PET uptake and hot spot/cold spot areas of functional/structural connection target regions was
analyzed.

Table 1, sociodemographic, clinical, and biomarker characteristics of participants

outcome

Differences in sociodemographic and clinical data between diagnostic groups are shown in Table 1
.
There were no differences
in age, sex, weight and years of schooling.
Patients with AD had significantly worse cognitive performance (CERAD-NB and MMSE) (p < 0.
001) and a higher sum of CDR (SOB) score compared
with the cognitively normal control group.
As expected, AD patients showed higher concentrations of pathological CSF biomarkers (lower Aβ42 and Aβ42/40 ratios, higher p-tau181 and t-tau)
compared to controls.
Patients with AD had higher levels of sTREM2 compared with the control group, but this difference was not statistically significant in analysis of variance adjusted for age, sex, and education (see Table 1).

The authors analyzed the association
between TSPO-PET intake and CSF sTREM2 concentration.
There was a significant quadratic association between TSPO-PET tracer uptake in the above regions (these regions were identified as sites of microglial activation specific to AD) and CSF sTREM2 in AD patients (R2 = 0.
39, p < 0.
02); For the whole cohort (R2 = 0.
05, p < 0.
49) and the normal control alone (R2 = 0.
12, p < 0.
48; Figure 3) No link between
these two variables was found.

Fig.
3, TSPO-PET and cerebrospinal fluid analysis results of dementia patients (AD) and CN

Differences in imaging and clinical variables between patients and controls

Compared with Aβ-negative healthy controls, TSPO-PET tracer uptake in bilateral anteromedial temporal lobes was significantly increased in the Aβ-positive patient group (p < 0.
001; Figure 1A), which supports the concept of
increased microglial activation in AD.
Average TSPO-PET uptake was highest
in the occipital region except for elevated TSPO-PET uptake in the inferior temporal gyrus/middle temporal sphere, (anterior) wedge lobe, and posterior cingulate gyrus.
Consistent with the view of AD as a network disorder, patients with AD experienced decreased
structural and functional connectivity to healthy controls at both the whole brain level and within the default mode network.
To link microglial activation to clinical outcomes, the authors analyzed the relationship
between TSPO tracer uptake in the anteromedial temporal lobe and dementia severity and cognitive impairment 。 The results showed anteromedial temporal lobe TSPO-PET SUVR versus CDRSOB (R2 = 0.
39, p < 0.
001), CERAD-NB overall score (R2 = 0.
20, p < 0.
017), and MMSE score (R2 = 0.
15, p = 0.
049; Figure 3) showed a secondary association, which indicated that AD activated the brain's immune system
in stages.

Figure 4, TSPO-PET covariance matrix and functional/structural connectivity matrix for dementia patients (AD) and CN

Correlation between functional and structural connections and TSPO-PET

To address their main hypothesis, namely the distribution of neuroinflammation based on connectivity, the authors assessed whether interconnected brain regions in AD patients showed varying degrees of microglial activation changes
.
The authors generated adjacency matrices of functional and structural connections obtained from 246 BN regions using the control and AD patient groups (Figures 4B, C), and calculated the TSPO-PET-derived covariance matrix for cognitively normal controls and AD patients using the same region (see Figure 4A).

For comparison, Figure 1B shows the mean TSPO-PET SUVR binding pattern within the AD group and evaluates the association
of functional and structural connections with the covariance of TSPO-PET tracer uptake in subsequent regression analysis.
The AD group (β = 0.
35, p < 0.
001) and the control group (β = 0.
47, p

Figure 5, Association of TSPO-PET function/structural connectivity between dementia patients (AD) and CN with microglial activation

Network connectivity prediction for TSPO-PET tracer uptake

The authors further explore whether the uptake of TSPO-PET tracer in a specific region can predict TSPO levels in the connected region (following a similar approach recently proposed for correlation between functional connectivity and tau PET).

。 Results In the AD group (β = 0.
83, p < 0.
001) and healthy control group (β = 0.
57, p < 0.
001; Figure 6A), functional connections were associated with higher levels of TSPO-PET uptake in the target region and vice versa
in seed regions with higher levels of TSPO-PET uptake 。 Structural connectivity between the two groups showed some weak association (AD:β = 0.
36, p < 0.
001; control: β = 0.
37, p < 0.
001; see Figure 6B).

Figure 6, Connectivity vs.
TSPO-PET uptake

The authors also analyze how functional connections between seed regions with high TSPO-PET tracer uptake and seed regions with low TSPO uptake correlate
with TSPO-PET uptake in all target regions.
The region of interest is determined
by all other regions in the BN map except hot and cold spots.
Analysis of hot spot seed regions showed a strong positive correlation
in cognitively normal controls (β = 0.
79, p < 0.
001) and AD patients (β = 0.
74, p < 0.
001).
In the cold spot seed zone analysis, two groups (cognitively normal control: β = -0.
22, p < 0.
03; AD patients: β = -0.
25, p = 0.
002; Figure 7A) shows a negative correlation
.
The structural connectivity-based analysis did not find any significant positive or negative associations (see Figure 7B).

Fig.
7, Relationship between TSPO-PET uptake and tempooccipital hot spot and basal ganglia cold spot function/structural connection

brief summary

In addition to Aβ and tau deposition, activation of microglia is increasingly recognized as a third pathological feature
of AD.
Using the third-generation TSPO-PET tracer [18F]GE-180, the authors explored whether connectivity in early AD patients and healthy controls was associated
with microglial activation characteristics.
One of the key findings is that functional connections are spatially related to microglial activation, which the authors speculate travels along highly connected brain region pathways, similar to the transmission pattern
of tau pathology.
Another important finding was that the increased activation of microglia in early AD patients was mainly located in the temporal lobe compared to healthy controls, and the correlation between TSPO-PET tracer uptake and clinical disease severity varied with disease stage, supporting the stage-specific role
of the brain immune system.

Original:

Rauchmann BS, Brendel M, Franzmeier N, Trappmann L, Zaganjori M, Ersoezlue E, Morenas-Rodriguez E, Guersel S, Burow L, Kurz C, Haeckert J, Tatò M, Utecht J, Papazov B, Pogarell O, Janowitz D, Buerger K, Ewers M, Palleis C, Weidinger E, Biechele G, Schuster S, Finze A, Eckenweber F, Rupprecht R, Rominger A, Goldhardt O, Grimmer T, Keeser D, Stoecklein S, Dietrich O, Bartenstein P, Levin J, Höglinger G, Perneczky R.
Microglial Activation and Connectivity in Alzheimer Disease and Aging.
Ann Neurol .
2022 Nov; 92(5):768-781.
doi: 10.
1002/ana.
26465

PDF file download