Cerebrospinal Fluid sTREM-2,GFAP,and β-S100 in Symptomatic Sporadic Alzheimer’s Disease: Microglial,Astrocytic,and APOE Contributions Along the Alzheimer’s Disease Continuum

Abstract

Background:

Many transversal mechanisms act synergistically at different time-points in the cascade of Alzheimer’s disease (AD), since amyloid-β (Aβ) deposition, tau pathology, and neuroinflammation influence each other.

Objective:

We explored the contributions of microglia and astrocytes in patients with symptomatic sporadic AD stratified according to AT(N) system and APOE genotype.

Methods:

We compared the cerebrospinal fluid (CSF) levels of sTREM-2 and markers of astrocytic activation (GFAP; β-S100) from 71 patients with AD (23 A+T–,48 A+T+; 38 APOE ɛ3, 33 APOE ɛ4) and 30 healthy controls (HC). With multivariate analyses we investigated associations between glial biomarkers, Aβ₄₂, and p-tau in all subgroups.

Results:

CSF sTREM-2 was higher in A+T+ [1.437 (0.264)] and A+T– [1.355 (0.213)] than in HC [1.042 (0.198); both p < 0.001]; GFAP and β-S100 were comparable across groups. Considering all patients, sTREM-2 positively associated with Aβ₄₂ (p = 0.04) and p-tau (=0.016), with the first being present only in the A+T– subgroup (p = 0.023). GFAP positively associated with Aβ₄₂ in all patients (p = 0.020) and in the A+T+ subgroup (p = 0.04). Stratifying by APOE, a positive association of sTREM-2 and p-tau was confirmed selectively in carriers of ɛ4 (p = 0.018). Finally, sTREM-2 positively correlated with β-S100 in all subgroups, and with GFAP in A+T+ (p = 0.042).

Conclusion:

Our results confirm the increase of CSF sTREM-2 in AD, which associates with reduced amyloidopathy in A+T– patients. Moreover, microglial activation seems to increase CSF tau levels in carriers of APOE ɛ4, is associated with astrocytic reactivity (GFAP) in A+T+, and likely leads the acquisition of a more neurotoxic astrocytic phenotype (β-S100).

Keywords

Alzheimer’s disease Alzheimer’s disease continuum astrocytes β-S100 biomarkers GFAP microglia soluble TREM-2

INTRODUCTION

Alzheimer’s disease (AD) is a clinically heterogeneous condition ranging from pre-clinical stages to overt dementia. Despite being the leading cause of cognitive decline worldwide [1], its biological substrate is still held as a complex phenomenon whose intricacy represents one of the most challenging enigmas of scientific research.

According to the amyloid hypothesis [2], the first hallmark of the disease is amyloid-β (Aβ) pathology, which eventually leads to tau-hyperphosphorylation [3] via a cascade of biological events that is triggered as much as 20 years before clinical manifestations [4], and that culminates in synaptic loss, neuronal death and ancillary events, such as atrophy and astrogliosis. A sort of hierarchical structure between these biological changes has been supported by both animal models and in vivo evidences stating that neurodegeneration and astrogliosis are final events in the cascade [5, 6]. However, many transversal molecular and cellular mechanism act synergistically at different time-points during the cascade, contributing to the pathophysiology of AD [7], including vascular and metabolic changes [8 –10].

More recently, attention has been drawn to the role of non-neuronal contributions, especially those of neuroinflammatory glial cells, since Aβ and tau-mediated mechanisms do induce a local neuroinflammatory response [11, 12], but conversely the activation of astrocytes and microglia plays a role in the regulation of Aβ and tau-related pathogenic events [13, 14].

In particular, hints on the importance of microglial activation also come from whole-genome analyses focusing on the Triggering Receptor Expressed on Myeloid cells 2 (TREM-2), a gene encoding for a transmembrane receptor present on brain microglial cells, whose loss of function increases the genetic risk of developing sporadic AD [15 –17]. Subsequent studies have proven an ambivalent role of TREM-2 and its cerebrospinal fluid (CSF) soluble form (sTREM-2) in AD [18 –20]. On one side it favors microglial viability, triggering a switch to disease-associated microglia (DAM) in response to pathogenic Aβ [21, 22] and enhancing microglia-mediated Aβ clearance [23 –25]. On the other, it induces the production of potentially damaging cytokines – namely IL-1β, TNFα and IL-6 – via its tonic overexpression [26, 27].

Likewise, changes in astrocytes’ secretomes in response to toxic stimuli have also been detected in AD [28, 29]. These changes include the release of biomarkers of astrocytic activation, such as Glial Fibrillary Acidic Protein (GFAP) and β-S100, a calcium-binding protein which has been linked to dystrophic neurites within Aβ plaques [30, 31] but also to neuroinflammatory astrocytic response [32].

Previous literature has focused on the longitudinal bond between Aβ pathology, tau-induced neurodegeneration and glial biomarkers, especially in the early stages of AD [33, 34], focusing on data from cognitively unimpaired pre-symptomatic individuals and patients with genetically inherited dementia [35, 36]. With the present work, we aimed at implementing these findings by evaluating a cohort of symptomatic patients diagnosed with sporadic AD stratified considering both the AT(N) system [37] and Apolipoprotein E (APOE) genotype. Indeed, much evidence links APOE to disease progression and to the modulation of neuroinflammation in AD [38, 39], and APOE ɛ4 carriers present an enhanced switching of microglial phenotype towards a neurodegenerative-associated phenotype which could further promote AD pathology [19]. In a previous work, we showed that human astrocytes cultures incubated with CSF samples from AD patients were vulnerable in terms of increased apoptosis only in the presence of high levels of tau protein and APOE ɛ4 [40]. More recently, we observed that individuals with isolated amyloidosis (A+T–) carrying APOE4, but not A+T+, had increased CSF levels of cytokines supporting neurotransmission and evidence of significant cognitive preservation [41]. Therefore, we expected different profiles of activation of neuroinflammatory cells in patients with different AT(N) status and APOE, which could be reflected by different expression of some of the proteins expressed by microglia and astrocytes.

First, we measured the CSF levels of all AD biomarkers, sTREM-2 and two markers of astrocytic activation, GFAP and β-S100. Then, we examined the correlations between these markers in all individual groups, to seek for potential differences and to investigate the dynamic relationship between neuronal and non-neuronal actors across the AD continuum.

MATERIALS AND METHODS

Subjects’ enrolment

Between September 2020 and December 2021, we evaluated a total of 100 outpatients that had been referred to the UOSD Centro Demenze of the University Hospital “Policlinico Tor Vergata” in Rome upon suspicion of AD [42]. After initial assessment and neurological examination, all patients underwent a complete diagnostic work-up including Mini Mental State Examination (MMSE), corrected by age and education, laboratory testing to rule out secondary cognitive decline, 3T brain MRI, and lumbar puncture. Eventually, 84 patients received a CSF biomarker-based diagnosis of AD, in conformity with current diagnostic criteria, and were further stratified according to the most recent NIA-AA research framework [37]. Patients were stratified as having either overt AD – with an increase of p-tau alongside Aβ₄₂ decrease (A+T+) – or sole amyloid pathology (A+T–). Genetic testing for APOE was also performed on all subjects.

We excluded patients with major comorbidities such as oncological history, inflammatory or autoimmune systemic conditions, decompensated diabetes, organ failure. Other exclusion criteria were a history of manifest acute stroke – Hachinski scale score >4 or radiological evidence of focal ischemic lesions – and clinical evidence or suspicion of other neurological disorders. Thus, our final sample included 71 patients within the AD continuum (ADc), namely 23 A+T– and 48 A+T+ [see Fig. 1].

Fig. 1

Patient selection flowchart summarizing enrolment procedures for the participants to the study.

Furthermore, 30 healthy age-matched controls admitted to the Policlinico Tor Vergata Hospital Emergency Department between October 2019 and April 2022, whose CSF samples were collected in accordance with standard hospital practice, were also included in the study. Upon discharge all 24 patients had received a diagnosis of either tensive type headache or psychogenic symptoms, namely functional generalized weakness, pain or sensory symptoms, all with normal diagnostic tests. All control subjects underwent Head Computed Tomography or Brain MRI and showed no signs of cortico-subcortical atrophy. Active infections, incidental presence of cognitive impairment and other primary neurological conditions had been ruled out, including non-specific CSF changes such as increased CSF cell count (>4 cells/mmc), or altered biomarker profile, hence, all controls subjects were classified as A– T–. Demographics from the patients and control groups are reported in Table 1.

Table 1

Demographics and results of CSF dosages in groups expressed as median and interquartile range (IQR)

	Controls	A+T–	A+T+	APOEɛ3	APOEɛ4
	(n = 30)	(n = 23)	(n = 48)	(n = 38)	(n = 33)
Age	67.50 (11.00)	71.00 (10.00)	71.00 (9.00)	72.00 (7.75)	70.50 (11)
Sex (M:F)	16:14	12:11	19:29	16:22	15:18
MMSE	28.50 (1.00)	25.30 (2.25)	25.10 (2.90)	25.00 (3.00)	25.30 (2.50)
CSF/serum albumin	5.075 (0.74)	5.87 (4.87)	5.75 (3.21)	5.67 (2.79)	6.24 (5.39)
CSF Aβ₄₂ (pg/ml)	955.75 (590.33)	421.00 (155.93)	417.00 (111.35)	420.50 (118.153)	431.00 (116.00)
CSF p-tau (pg/ml)	30.60 (25.05)	40.00 (22.00)	117.00 (52.50)	89.00 (78.50)	97.00 (75.75)
CSF t-tau (pg/ml)	217.25 (83.18)	229.05 (98.95)	688.00 (418.65)	510.50 (560.62)	541.50 (541.60)
CSF sTREM-2 (ng/ml)	1.042 (0.198)	1.355 (0.213)	1.437 (0.264)	1.424 (0.260)	1.402 (0.275)
CSF GFAP (ng/ml)	0.047 (0.018)	0.048 (0.017)	0.050 (0.027)	0.048 (0.022)	0.050 (0.023)
CSF β-S100 (ng/ml)	761.60 (1795.75)	625.00 (2067.50)	866.50 (2246.25)	660.50 (2101.75)	953.00 (2452.00)

CSF, cerebrospinal fluid; A+T–, isolated amyloid pathology; A+T+, full-blown Alzheimer’s disease; F, female; M, male; MMSE, Mini-Mental State Examination; sTREM-2, soluble Triggering Receptor Expressed on Myeloid Cells 2; GFAP, Glial fibrillary acidic protein; IQR, interquartile range.

We obtained written consent from all participants and/or legally authorized representatives, the local ethical committee accounted for the study protocol as an observational retrospective design.

CSF sampling and laboratory analysis

A total amount of 10 ml CSF sample was collected for each patient. All lumbar punctures were performed with sterile technique between 8 and 10 am, samples were collected in polypropylene tubes. Around 2 ml were used for biochemical routine analysis, including cell and protein count. The other 6 ml were centrifuged at 2000 g at +4°C for 10 min, aliquoted in 1 ml portions and frozen at –80°C for further analysis of levels of Aβ₄₂, p-tau phosphorylated at Thr181, t-tau and glial biomarkers. Commercially available kits were used to carry out biochemical analysis (Flex reagent cartridge, Dimension Vista System, Siemens Healthcare Diagnostics GmbH, Munich, Germany); CSF Aβ₄₂, p-tau and t-tau, sTREM-2, GFAP and β-S100 concentrations were determined using a sandwich enzyme-linked immunosorbent assay (EUROIMMUN Aβ1-42 levels ELISA^©, EUROIMMUN Total tau ELISA^©, EUROIMMUN p-Tau (181) ELISA^©, FineTest Human-TREM2 ELISA^©, LUMINEX^© Multiple assays ELISA for GFAP and β-S100). Specific cut-off values provided by the laboratory were used to determine AT categories: A+ Aβ₄₂, <600 pg/ml, T + p-tau >65 pg/ml.

Blood samples were also drawn for complimentary analysis, i.e., CSF/serum albumin quotient (QAlb) accounting for blood-brain barrier (BBB) integrity [43, 44], and APOE genotyping which was conducted by allelic discrimination technology (TaqMan; Applied Biosystems).

Data management and statistical analysis

All continuous variables including levels of CSF biomarkers (i.e., Aβ₄₂, p-tau, t-tau, sTREM-2, GFAP, and β-S100), MMSE, age and the calculated CSF/serum albumin ratio were expressed in terms of median±interquartile (IQ) range, since data were not normally distributed per Shapiro-Wilk test (Table 1). Patients were stratified in the APOE ɛ4 subgroup when carrying either one (APOE ɛ3/ɛ4) or two (APOE ɛ4/ɛ4) alleles. No carriers of APOE ɛ2 were present in our cohort, and the remaining patients were all APOE ɛ3/ɛ3 (APOE ɛ3 subgroup).

The comparison between groups was carried out via the non-parametric Kruskal-Wallis test, using Dunn’s test with Holm correction for post hoc multiple comparisons. Binomial variables were compared via the Pearson’s Chi-squared test. Afterwards, we built a multivariate regression analysis computing CSF sTREM-2, GFAP, and β-S100 as individual dependent variables, to explore their relationship with CSF levels of Aβ₄₂ and p-tau, also adjusting for age and APOE status (APOE ɛ3 = 0; APOE ɛ4 = 1). Since CSF t-tau highly correlates with p-tau, we excluded it from the regression analyses. Finally, we performed a correlation analysis to explore the relationships between glial biomarkers in each condition, first stratifying patients according to the AT status and second according to APOE genotype.

Statistical analysis and data management were operated via JASP^© (Version 0.14 - Computer Software – JASP TEAM 2020) and GraphPad Prism^© version 9.3.1 for Windows (GraphPad Software, San Diego, CA, http://www.graphpad.com). All results were computed with two-tailed significativity tests; values of p < 0.05 were considered statistically significant.

RESULTS

CSF glial biomarkers’ levels comparison between groups

The study eventually included a total of 71 patients within the ADc (23 A+T– and 48 A+T+) and 30 healthy controls, as reported in the demographic table (Table 1). The BBB was intact in all groups [CSF/serum albumin, H (2) = 3.979, p = 0.137].

The Kruskal-Wallis test was significant for differences of CSF sTREM-2 between the AT(N)-stratified groups [H(2) = 38.099, p < 0.001], with A+T+ [1.437 (0.264)] and A+T– both showing higher mean values than healthy controls [1.042 (0.198); p_holm < 0.001 for both comparisons] and a difference was also found between A+T– and A+T+ (p_holm = 0.038). Neither CSF GFAP nor β-S100 levels resulted significatively different across groups (see Fig. 2A–C).

Fig. 2

CSF levels of sTREM-2 (A), GFAP (B), and β-S100 (C) in controls and patients with AD, stratified by AT status. Jitter elements represent single subjects (A+T– : isolated amyloid pathology; A+T+: full-blown Alzheimer’s disease; Ctrls: controls, ^*p < 0.05;^***p < 0.001).

Regression analyses between CSF glial biomarkers and biomarkers of AD

First, we performed multivariate regression analyses to evaluate the contribution of AD biomarkers (CSF Aβ₄₂ and p-tau levels), on the CSF levels of sTREM-2, GFAP, and β-S100 in the whole ADc, adjusting for APOE genotype and age (Table 2). The adjusted analyses highlighted that CSF sTREM-2 was positively associated with both Aβ₄₂ (β= 0.259, p = 0.040) and p-tau (β= 0.307, p = 0.016). CSF GFAP was also positively associated with Aβ₄₂ (β= 0.306, p = 0.020), while other covariates were not significant. The analysis of CSF β-S100 showed that neither Aβ₄₂ and p-tau nor other predictors were significantly associated with itslevels.

Table 2

Multiple linear regressions in patients stratified according to the AT system

	AD continuum		A+T–		A+T+
CSF sTREM-2	β (std)	p	β (std)	p	β (std)	p
CSF Aβ₄₂	0.259	0.040 *	0.535	0.023 *	0.143	0.396
CSF p-tau	0.307	0.016 *	0.224	0.359	0.072	0.561
APOE (ɛ3 = 0; ɛ4 = 1)	0.065	0.605	–0.086	0.704	0.177	0.378
Age	–0.163	0.195	0.041	0.861	–0.301	0.077
	R² = 0.183		R² = 0.283		R² = 0.136
CSF GFAP
CSF Aβ₄₂	0.306	0.020 *	0.148	0.504	0.351	0.042 *
CSF p-tau	0.068	0.600	–0.410	0.108	0.010	0.952
APOE (ɛ3 = 0; ɛ4 = 1)	0.099	0.444	0.333	0.160	0.047	0.745
Age	0.020	0.875	0.025	0.917	0.061	0.713
	R² = 0.117		R² = 0.258		R² = 0.141
CSF β-S100
CSF Aβ₄₂	0.126	0.344	0.330	0.170	0.011	0.950
CSF p-tau	0.011	0.936	0.066	0.798	–0.050	0.771
APOE (ɛ3 = 0; ɛ4 = 1)	0.164	0.223	0.203	0.410	0.142	0.417
Age	–0.073	0.587	0.104	0.682	–0.164	0.350
	R² = 0.057		R² = 0.169		R² = 0.054

* = p < 0.05. CSF, cerebrospinal fluid; A+T–, isolated amyloid pathology; A+T+, full-blown Alzheimer’s disease; sTREM-2, soluble Triggering Receptor Expressed on Myeloid Cells 2; GFAP, Glial fibrillary acidic protein; std, standardized. Bold values represent statistical significance.

We then performed separate regression analyses in A+T– and A+T+ patients, to evaluate possible differences driven by p-tau status in the influence of covariates on glial biomarkers. Upon stratification, no subgroup showed persistence of the association between p-tau and sTREM-2. The positive association between CSF sTREM-2 and Aβ₄₂ that we had observed in the whole AD continuum was confirmed only in the A+T– group (β= 0.535, p = 0.023), while CSF GFAP resulted positively associated with Aβ₄₂ only in the A+T+ group (β= 0.351, p = 0.042). β-S100 did not associate with any covariate in neither A+T+ nor A+T–.

To account for potential APOE-specific dynamics, we repeated separate the regression analyses in the APOE ɛ3 and in the APOE ɛ4 subgroups. Selectively in carriers of APOE ɛ4 we identified a positive association of CSF sTREM-2 with p-tau (R² = 0.368), as observed in the whole AD continuum. No associations were found between β-S100 nor GFAP CSF levels and any of the predictors (Table 3).

Table 3

Multiple linear regressions in patients stratified according to the APOE genotype

	APOEɛ3		APOEɛ4
CSF sTREM-2	β (std)	p	β (std)	p
CSF Aβ₄₂	0.342	0.065	0.405	0.015 *
CSF p-tau	0.083	0.646	0.245	0.018 *
Age	–0.194	0.283	–0.258	0.122
	R² = 0.061		R² = 0.386
CSF GFAP
CSF Aβ₄₂	0.373	0.054	0.315	0.125
CSF p-tau	0.175	0.348	–0.044	0.818
Age	–0.099	0.519	0.194	0.315
	R² = 0.216		R² = 0.136
CSF β-S100
CSF Aβ₄₂	–0.023	0.907	0.248	0.171
CSF p-tau	0.069	0.727	–0.089	0.636
Age	–0.024	0.904	–0.203	0.287
	R² = 0.015		R² = 0.158

* = p < 0.05. CSF, cerebrospinal fluid; sTREM-2, soluble Triggering Receptor Expressed on Myeloid Cells 2; GFAP, Glial fibrillary acidic protein; std: standardized. Bold values represent statistical significance.

Correlations among CSF glial biomarkers

Considering the whole ADc group, we found a significant moderate positive correlation between CSF sTREM-2 and β-S100 (rho = 0.433, p < 0.001), that was retrievable even after stratification for both the AT(N) status (A+T+: rho = 0.414, p = 0.005; A+T– : rho = 0.460, p = 0.027) and APOE (APOE ɛ4: rho = 0.388, p = 0.039; APOE ɛ3: rho = 0.305, p = 0.023). Furthermore, in the ADc we identified a trend of correlation between CSF sTREM-2 and CSF GFAP (rho = –0.228, p = 0.055), that reached statistical significance in the A+T+ subgroup (rho = 0.305, p = 0.044) and in the APOE ɛ3 subgroup (rho = 0.400, p = 0.0336). Finally, levels of CSF GFAP were found to be inversely correlated with levels of CSF β-S100 selectively in the A+T– subgroup (rho = –0.417, p = 0.048). No other correlations were found between other biomarkers in all subgroups (Fig. 3).

Fig. 3

Heatmaps representing the Spearman’s rho of correlation among glial biomarkers in all subgroups. (A+T– : isolated amyloid pathology; A+T+: full-blown Alzheimer’s disease; ^*p < 0.05; ^**p < 0.01; ^***p < 0.001).

DISCUSSION

Emerging evidence depicts AD as a complex entity resulting from the perturbation of an equally complex equilibrium between many players, acting at once to guarantee the healthy functioning of the brain. We aimed at exploring this disruption in-vivo in a cohort of symptomatic patients with sporadic AD, focusing on the role of microglia, astrocytes, and APOE-related contributions to the disease.

Starting with microglia, previous studies identified an increase of CSF sTREM-2 in patients with AD with respect to controls, although these cohorts focused mainly on carriers of familiar AD mutations or very early-symptomatic subjects [36 , 46]. When accounting also for the AT(N) stratification [47, 48] it is possible to identify a further increase of CSF sTREM-2 in A+T+. Our results do confirm these findings, since we identified the same difference between AD and controls and a further slight increase of levels of CSF sTREM-2 in the A+T+ group with respect to the A+T– group. This adds interesting CSF documentation of microglial activation in a separate real-world cohort of clinically symptomatic patients, supporting the relevance of biological framing regardless of clinical status, since no difference was present between A+T+ and A+T– in terms of cognitive impairment.

In our regression analysis we identified a direct positive association between CSF levels of Aβ₄₂ and sTREM-2 in the whole ADc, and an even stronger one in the A+T– subgroup, which consolidate the hypothesis of an early activation of TREM-2-dependent microglial protective functions, since the earliest stages of AD [49, 50]. Indeed, this association is in line with recent findings that a high annual rate of increased sTREM-2 production longitudinally correlates with reduced rates of decreased Aβ₄₂ and hippocampal cortical shrinkage [34]. Plausible mechanisms behind this include the upregulation of TREM-2 enhancing DAM-mediated Aβ phagocytosis [51, 52] which amortizes Aβ’s toxic downstream effects on surrounding cells and possibly the onset of tau pathology and neurodegeneration [53]. Notably, our findings are referred to symptomatic subjects with sporadic AD, raising speculations that the persistence of elevated CSF sTREM-2 – triggered long before clinical onset – keeps playing a role even when cognitive decline sets in.

We did not find the same association between CSF levels of Aβ₄₂ and sTREM-2 in the A+T+ subgroup. Embracing the concept that microglial TREM-2-related mechanisms might protect from Aβ-induced tau pathology, it is likely that in patients with A+T+ the effectiveness of such mechanisms has already been lost due to disease progression, which escapes microglial control [54].

Nevertheless, CSF sTREM-2 is higher in A+T+ than A+T–, and its relationship with tau-pathology is still debated [18]. Knapskog and colleagues identified a moderate positive association between CSF sTREM-2 and CSF p-tau in a cohort of sporadic AD patients ranging from MCI to dementia [48]. In our study we did detect an analogous positive association between sTREM-2 and CSF p-tau considering the whole ADc group, but not in the A+T+ and A+T– groups when considered separately, probably due to the small sample size after stratification. Moreover, an analogous correlation has been found also in other neurodegenerative disorders such as Parkinson’s disease [55], suggesting that the increase of sTREM-2 might be tied to tauopathy in a non-AD-specific manner, hence to the general progression of neurodegeneration.

Moving to astrocytes, we focused on the spectrum of glial activation in AD and its complex crosstalk with microglia. We did not find a significant increase of levels of CSF GFAP in A+ patients, despite raw values being higher in the latter than in healthy controls. This is consistent with previous studies which lead to inconsistent findings of increased CSF GFAP in different cohorts, depending on stratification strategies and cognitive status [56, 57].

Although levels of plasma GFAP seem to better reflect disease-specific Aβ-related astrocytic activation [58], likely due to its clearance mechanism via end-feet astrocytic projections draining directly into the bloodstream [59], the increase of CSF GFAP does predict the progression of cognitive decline, suggesting that it might account for non-specific glial inflammatory changes playing an active role in the progression of the disease [60]. In the A+T+ subgroup we retrieved a positive association between levels of CSF GFAP and sTREM-2, which could represent a microglia-induced astrocytic activation. Neuronal damage secondary to tauopathy might induce DAM-mediated rearrangements of astroglia that are not required if tauopathy is absent, as showed by the absence of this correlation in the A+T– subgroup. Intriguingly, we also found a positive correlation between levels of CSF GFAP and levels of CSF Aβ₄₂ in the A+T+ subgroup, suggesting that glial activation might be more efficient in the presence of less pathological levels of CSF Aβ₄₂. This could be because astrocytes can be themselves damaged by the progression of Aβ pathology.

Looking at β-S100, a recent meta-analysis reported the absence of changes in its CSF concentrations in AD [61], although a previous work identified a selective increase of β-S100 astrocytic expression in specific brain regions, such as the hippocampus and cortex layer I [62]. Accordingly, our results showed no significant increase of CSF β-S100 in patients with AD, but we did identify a consistent positive correlation with CSF sTREM-2 in the whole AD continuum, that was identifiable in all subgroups of stratification.

This novel finding supports previous evidence that, in response to the deposition of Aβ, activated microglia upregulates the expression of specific patterns of proinflammatory cytokines and chemokines in astrocytes – also known as alarmins [63] – including S100 proteins [64]. Specifically, the increment of β-S100 from nanomolar to micromolar concentrations has been linked to a passage from neurotrophic to neurotoxic pro-apoptotic effects [65], configuring it as a damage-associated extra-cellular acting molecule. Looking at sTREM-2 levels as a proxy for DAM’s function, its progressive increase seems to induce a parallel shift towards a neurotoxic-oriented pattern of secretion by astrocytes [66, 67]. Interestingly, this correlation is also present in the A+T– subgroup where sTREM-2 should be holding its most neuroprotective effects, pointing at a constant intertwining with toxicity all along the unravelling of AD. The slight inverse correlation between CSF β-S100 and CSF GFAP in the same A+T-subgroup might indirectly mirror this, as rising β-S100 levels depict a possible progressive shift towards specific disease-induced neurotoxicity rather than non-specific glial damage.

Finally, stratifying according to APOE genotype, we found that the association between CSF sTREM-2 and p-tau in the AD continuum seems to reach statistical significance exclusively in the APOE ɛ4 subgroup. This drives very interesting speculations, because although APOE ɛ4 is traditionally studied for its effects on astrocytes and Aβ accumulation [68], recent findings do show that it has specific effects also on upregulating the microglial transcriptome, inducing an imbalance from its homeostatic state towards the DAM state which has been shown to be strictly bound to the TREM-2 cascade [69, 70]. Moreover, APOE ɛ4 increases microglial cellular stress more than APOE ɛ3 does [71], and shifts microglial function towards pro-neurodegenerative behavior [72].

APOE ɛ4 has also been identified as an active contributor to the progression of tauopathy and neurodegeneration [73, 74]. In a mouse model of tauopathy DAM was demonstrated to be the driving-force of neurodegeneration, and it was shown that APOE seems to regulate this process mainly by modulating microglial activation [75]. Since functioning TREM-2 has been shown to block amyloid-induced tauopathy [53], speculations can be made that carriers of APOE ɛ4 might be equipped with an ineffective TREM-2 cascade, which could favor the progression of p-tau pathology [54]. Considering this, APOE ɛ4 could induce excessive DAM formation, represented in our results by the increase of CSF sTREM-2, which in turn would cause the spreading of tau pathology and thus a parallel increase of CSF levels of p-tau, as also supported by recent studies on animal models [76, 77].

Notably, no association was present between CSF sTREM-2 and either Aβ₄₂ or p-tau in the APOE ɛ3 subgroup, but we did identify the same correlation between CSF GFAP and sTREM-2 that we had retrieved in the A+T+ subgroup. In light of this, APOE ɛ3 might be more prone to induce a non-specific inflammatory response, depicted by the increase of CSF GFAP paralleling the increase of CSF sTREM-2, which could not be necessary in APOE ɛ4 thanks to the presence of other defense mechanisms that configure a better resilience to AD changes [41, 78]. This likely makes neuroinflammation in APOE ɛ4 more linked to AD-specific pathological changes than in APOE ɛ3.

We are aware that this study has some limitations including the limited sample size. Widening the study cohort could be useful both to confirm and to strengthen our results. Moreover, a bigger cohort would allow us to account for both ATN and APOE status at the same time – which was not possible in our study due to the limited sample size. Follow-up longitudinal data could also add meaningful information, in order to verify the burden of glial and microglial contributions on disease progression. Indeed, the cross-sectional design of our study does not allow us to explore the impact of our results on the rate of cognitive decline or the evolution of neuropsychological profiles. Besides, it would also be interesting to evaluate whether patterns of astrocytic and microglial response change alongside clinical progression.

Conclusions

In summary, we confirm the increase of CSF sTREM-2 in AD, which associates with reduced amyloidopathy in A+T–, but not when tauopathy sets in. This supports the evidence that the role of microglial activity could switch dynamically from neuroprotective to neurotoxic depending on disease stage. Moreover, in our cohort, CSF p-tau levels seem to be influenced by microglia in carriers of APOE ɛ4, meaning that APOE might play an active role in neurodegeneration. Finally, microglial-mediated inflammation is associated with astrocytic reactivity (GFAP) in A+T+, and with the acquisition of a more neurotoxic astrocytic phenotype (β-S100).

Overall, our results add evidence that the integrity of the neural system is much more complex than neuronal integrity alone, by highlighting complex in vivo reciprocal influence between astrocytes, microglia, APOE genotype and both amyloid and tau pathology. Despite the rising number of data available, at the moment there is no consensus on the use of astrocytic and microglial biomarkers in AD, although efforts have been made to implement such profiling strategies [56]. Since a neuro-centric vision of AD – focused only on downstream neuronal damage – has led to several failures in clinical trials, key intercellular and molecular interactions should be held in higher consideration, as they could pave the way towards combined treatment strategies in the era of precision medicine.

Footnotes

ACKNOWLEDGMENTS

The authors have no acknowledgments to report.

FUNDING

The authors have no funding to report.

DISCLOSURE STATEMENT

The authors have no conflict of interest to report.

DATA AVAILABILITY

Data are available upon reasonable request.

References

Brayne

, Miller

(2017) Dementia and aging populations-A global priority for contextualized research and health policy. PLoS Med 14.

Cline

, Bicca

, Viola

, Klein

(2018) The amyloid-β oligomer hypothesis: Beginning of the third decade. J Alzheimers Dis 64, S567–S610.

Bloom

(2014) Amyloid-β and tau: The trigger and bullet in Alzheimer disease pathogenesis. JAMA Neurol 71, 505–508.

Barthélemy

, Li

, Joseph-Mathurin

, Gordon

, Hassenstab

, Benzinger

TLS

, Buckles

, Fagan

, Perrin

, Goate

, Morris

, Karch

, Xiong

, Allegri

, Mendez

, Berman

, Ikeuchi

, Mori

, Shimada

, Shoji

, Suzuki

, Noble

, Farlow

, Chhatwal

, Graff-Radford

, Salloway

, Schofield

, Masters

, Martins

, O’Connor

, Fox

, Levin

, Jucker

, Gabelle

, Lehmann

, Sato

, Bateman

, McDade

, Allegri

, Bateman

, Bechara

, Benzinger

, Berman

, Bodge

, Brandon

, Brooks

W (Bill)

, Buck

, Buckles

, Chea

, Chhatwal

, Chrem Mendez

, Chui

, Cinco

, Clifford

, Cruchaga

, Donahue

, Douglas

, Edigo

, Erekin-Taner

, Fagan

, Farlow

, Fitzpatrick

, Flynn

, Fox

, Franklin

, Fujii

, Gant

, Gardener

, Ghetti

, Goate

, Goldman

, Gordon

, Graff-Radford

, Gray

, Groves

, Hassenstab

, Hoechst-Swisher

, Holtzman

, Hornbeck

, DiBari

, Ikeuchi

, Ikonomovic

, Jerome

, Jucker

, Karch

, Kasuga

, Kawarabayashi

, Klunk

W (Bill)

, Koeppe

, Kuder-Buletta

, Laske

, Lee

, Levin

, Martins

, Mason

, Masters

, Maue-Dreyfus

, McDade

, Mori

, Morris

, Nagamatsu

, Neimeyer

, Noble

, Norton

, Perrin

, Raichle

, Renton

, Ringman

, Roh

, Salloway

, Schofield

, Shimada

, Sigurdson

, Sohrabi

, Sparks

, Suzuki

, Taddei

, Wang

, Xu

(2020) A soluble phosphorylated tau signature links tau, amyloid and the evolution of stages of dominantly inherited Alzheimer’s disease. Nat Med 26, 398–407.

Palmqvist

, Insel

, Stomrud

, Janelidze

, Zetterberg

, Brix

, Eichenlaub

, Dage

, Chai

, Blennow

, Mattsson

, Hansson

(2019) Cerebrospinal fluid and plasma biomarker trajectories with increasing amyloid deposition in Alzheimer’s disease. EMBO Mol Med 11, e11170.

Bateman

, Xiong

, Benzinger

TLS

, Fagan

, Goate

, Fox

, Marcus

, Cairns

, Xie

, Blazey

, Holtzman

, Santacruz

, Buckles

, Oliver

, Moulder

, Aisen

, Ghetti

, Klunk

, McDade

, Martins

, Masters

, Mayeux

, Ringman

, Rossor

, Schofield

, Sperling

, Salloway

, Morris

(2012) Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N Engl J Med 367, 795–804.

Guo

, Zhang

, Zeng

, Huang

, Xu

, Zhao

(2020) Molecular and cellular mechanisms underlying the pathogenesis of Alzheimer’s disease. Mol Neurodegener 15.

Cunnane

, Trushina

, Morland

, Prigione

, Casadesus

, Andrews

, Beal

, Bergersen

, Brinton

, de la Monte

, Eckert

, Harvey

, Jeggo

, Jhamandas

, Kann

, la Cour

, Martin

, Mithieux

, Moreira

, Murphy

, Nave

, Nuriel

, Oliet

SHR

, Saudou

, Mattson

, Swerdlow

, Millan

(2020) Brain energy rescue: An emerging therapeutic concept for neurodegenerative disorders of ageing. Nat Rev Drug Discov 19, 609–633.

Bonomi

, De Lucia

, Mascolo

, Assogna

, Motta

, Scaricamazza

, Sallustio

, Mercuri

, Koch

, Martorana

(2021) Brain energy metabolism and neurodegeneration: Hints from CSF lactate levels in dementias. Neurobiol Aging 105, 333–339.

10.

Diomedi

, Rocco

, Bonomi

, Mascolo

, De Lucia

, Marrama

, Sallustio

, Koch

, Martorana

(2021) Haemodynamic impairment along the Alzheimer’s disease continuum. Eur J Neurol 28, 2168–2173.

11.

Zhong

, Wang

, Martens

, Wu

, Xu

, Wang

, Li

, Huang

, Can

, Xu

, Bu

, Chen

(2018) Amyloid-beta modulates microglial responses by binding to the triggering receptor expressed on myeloid cells 2 (TREM2). Mol Neurodegener 13.

12.

Wyssenbach

, Quintela

, Llavero

, Zugaza

, Matute

, Alberdi

(2016) Amyloid β-induced astrogliosis is mediated by β1-integrin via NADPH oxidase 2 in Alzheimer’s disease. Aging Cell 15, 1140–1152.

13.

Heneka

, Carson

, Khoury

J El

, Landreth

, Brosseron

, Feinstein

, Jacobs

, Wyss-Coray

, Vitorica

, Ransohoff

, Herrup

, Frautschy

, Finsen

, Brown

, Verkhratsky

, Yamanaka

, Koistinaho

, Latz

, Halle

, Petzold

, Town

, Morgan

, Shinohara

, Perry

, Holmes

, Bazan

, Brooks

, Hunot

, Joseph

, Deigendesch

, Garaschuk

, Boddeke

, Dinarello

, Breitner

, Cole

, Golenbock

, Kummer

(2015) Neuroinflammation in Alzheimer’s disease. Lancet Neurol 14, 388–405.

14.

Henstridge

, Hyman

, Spires-Jones

(2019) Beyond the neuron-cellular interactions early in Alzheimer disease pathogenesis. Nat Rev Neurosci 20, 94–108.

15.

Gratuze

, Leyns

CEG

, Holtzman

(2018) New insights into the role of TREM2 in Alzheimer’s disease. Mol Neurodegener 13.

16.

Jonsson

, Stefansson

, Steinberg

, Jonsdottir

, Jonsson

, Snaedal

, Bjornsson

, Huttenlocher

, Levey

, Lah

, Rujescu

, Hampel

, Giegling

, Andreassen

, Engedal

, Ulstein

, Djurovic

, Ibrahim-Verbaas

, Hofman

, Ikram

, van Duijn

, Thorsteinsdottir

, Kong

, Stefansson

(2013) Variant of TREM2 associated with the risk of Alzheimer’s disease. N Engl J Med 368, 107–116.

17.

Guerreiro

, Wojtas

, Bras

, Carrasquillo

, Rogaeva

, Majounie

, Cruchaga

, Sassi

, Kauwe

JSK

, Younkin

, Hazrati

, Collinge

, Pocock

, Lashley

, Williams

, Lambert

J-C

, Amouyel

, Goate

, Rademakers

, Morgan

, Powell

, St.George-Hyslop

, Singleton

, Hardy

(2013) TREM2 variants in Alzheimer’s disease. N Engl J Med 368, 117–127.

18.

Yang

, Fu

, Zhang

, Xiong

, Meng

, Zhang

(2020) TREM2 ectodomain and its soluble form in Alzheimer’s disease. J Neuroinflammation 17.

19.

Carmona

, Zahs

, Wu

, Dakin

, Bras

, Guerreiro

(2018) The role of TREM2 in Alzheimer’s disease and other neurodegenerative disorders. Lancet Neurol 17, 721–730.

20.

Udeochu

, Sayed

, Gan

(2018) TREM2 and amyloid beta: A love-hate relationship. Neuron 97, 991–993.

21.

Zhong

, Chen

, Wang

, Liao

, Wang

, Huang

, Wang

, Li

, Wu

, Jia

, Zheng

, Painter

, Atagi

, Liu

, Zhang

, Fryer

, Xu

, Bu

(2017) Soluble TREM2 induces inflammatory responses and enhances microglial survival. J Exp Med 214, 597–607.

22.

Mazaheri

, Snaidero

, Kleinberger

, Madore

, Daria

, Werner

, Krasemann

, Capell

, Trümbach

, Wurst

, Brunner

, Bultmann

, Tahirovic

, Kerschensteiner

, Misgeld

, Butovsky

, Haass

(2017) TREM2 deficiency impairs chemotaxis and microglial responses to neuronal injury. EMBO Rep 18, 1186–1198.

23.

Lee

CYD

, Daggett

, Gu

, Jiang

, Langfelder

, Li

, Wang

, Zhao

, Park

, Cooper

, Ferando

, Mody

, Coppola

, Xu

, Yang

(2018) Elevated TREM2 gene dosage reprograms microglia responsivity and ameliorates pathological phenotypes in Alzheimer’s disease models. Neuron 97, 1032–1048.e5.

24.

Del-Aguila

, Benitez

, Li

, Dube

, Mihindukulasuriya

, Budde

, Farias

FHG

, Fernández

, Ibanez

, Jiang

, Perrin

, Cairns

, Morris

, Harari

, Cruchaga

(2019) TREM2 brain transcript-specific studies in AD and TREM2 mutation carriers. Mol Neurodegener 14, 18.

25.

Condello

, Yuan

, Schain

, Grutzendler

(2015) Microglia constitute a barrier that prevents neurotoxic protofibrillar Aβ42 hotspots around plaques. Nat Commun 6, 6176.

26.

Nordengen

, Kirsebom

, Henjum

, Selnes

, Gísladóttir

, Wettergreen

, Torsetnes

, Grøntvedt

, Waterloo

, Aarsland

, Nilsson

LNG

, Fladby

(2019) Glial activation and inflammation along the Alzheimer’s disease continuum. J Neuroinflammation 16, 46.

27.

Mosher

, Wyss-Coray

(2014) Microglial dysfunction in brain aging and Alzheimer’s disease. Biochem Pharmacol 88, 594–604.

28.

Sadick

, O’Dea

, Hasel

, Dykstra

, Faustin

, Liddelow

(2022) Astrocytes and oligodendrocytes undergo subtype-specific transcriptional changes in Alzheimer’s disease. Neuron 110, 1788-1805.e10.

29.

Perez-Nievas

, Serrano-Pozo

(2018) Deciphering the astrocyte reaction in Alzheimer’s disease. Front Aging Neurosci 10, 114.

30.

Mrak

, Sheng

, Griffin

WST

(1996) Correlation of astrocytic S100 beta expression with dystrophic neurites in amyloid plaques of Alzheimer’s disease. J Neuropathol Exp Neurol 55, 273–279.

31.

Nishiyama

, Knöpfel

, Endo

, Itohara

(2002) Glial protein S100B modulates long-term neuronal synaptic plasticity. Proc Natl Acad Sci U S A 99, 4037–4042.

32.

Cristóvaõ

, Gomes

(2019) S100 proteins in Alzheimer’s disease. Front Neurosci 13, 463.

33.

Elahi

, Casaletto

, La Joie

, Walters

, Harvey

, Wolf

, Edwards

, Rivera-Contreras

, Karydas

, Cobigo

, Rosen

, DeCarli

, Miller

, Rabinovici

, Kramer

(2020) Plasma biomarkers of astrocytic and neuronal dysfunction in early- and late-onset Alzheimer’s disease. Alzheimers Dement 16, 681–695.

34.

Morenas-Rodríguez

, Li

, Nuscher

, Franzmeier

, Xiong

, Suárez-Calvet

, Fagan

, Schultz

, Gordon

, Benzinger

TLS

, Hassenstab

, McDade

, Feederle

, Karch

, Schlepckow

, Morris

, Kleinberger

, Nellgard

, Vöglein

, Blennow

, Zetterberg

, Ewers

, Jucker

, Levin

, Bateman

, Haass

, Adams

, Allegri

, Araki

, Barthelemy

, Bechara

, Berman

, Bodge

, Brandon

, Brooks

W (Bill)

, Brosch

, Buck

, Buckles

, Carter

, Cash

, Chen

, Chhatwal

, Chrem

, Chua

, Chui

, Cruchaga

, Day

, Cruz

CDLa

, Denner

, Diffenbacher

, Dincer

, Donahue

, Douglas

, Duong

, Egido

, Esposito

, Farlow

, Feldman

, Fitzpatrick

, Flores

, Fox

, Franklin

, Friedrichsen

, Fujii

, Gardener

, Ghetti

, Goate

, Goldberg

, Goldman

, Gonzalez

, Gräber-Sultan

, Graff-Radford

, Graham

, Gray

, Gremminger

, Grilo

, Groves

, Häsler

, Hellm

, Herries

, Hoechst-Swisher

, Hofmann

, Holtzman

, Hornbeck

, Igor

, Ihara

, Ikeuchi

, Ikonomovic

, Ishii

, Jack

, Jerome

, Johnson

, Käser

, Kasuga

, Keefe

, Klunk

W (Bill)

, Koeppe

, Koudelis

, Kuder-Buletta

, Laske

, Levey

, Lopez

, Marsh

, Martinez

, Martins

, Mason

, Masters

, Mawuenyega

, McCullough

, Mejia

, MountzMD

, Mummery

, Nadkarni

, Nagamatsu

, Neimeyer

, Niimi

, Noble

, Norton

, Nuscher

, O’Connor

, Obermüller

, Patira

, Perrin

, Ping

, Preische

, Renton

, Ringman

, Salloway

, Schofield

, Senda

, Seyfried

, Shady

, Shimada

, Sigurdson

, Smith

, Snitz

, Sohrabi

, Stephens

, Taddei

, Thompson

, Wang

, Weamer

, Xu

(2022) Soluble TREM2 in CSF and its association with other biomarkers and cognition in autosomal-dominant Alzheimer’s disease: A longitudinal observational study. Lancet Neurol 21, 329–341.

35.

Suárez-Calvet

, Kleinberger

, Araque Caballero

MÁ

, Brendel

, Rominger

, Alcolea

, Fortea

, Lleó

, Blesa

, Gispert

, Sánchez-Valle

, Antonell

, Rami

, Molinuevo

, Brosseron

, Traschütz

, Heneka

, Struyfs

, Engelborghs

, Sleegers

, Van Broeckhoven

, Zetterberg

, Nellgård

, Blennow

, Crispin

, Ewers

, Haass

(2016) sTREM2 cerebrospinal fluid levels are a potential biomarker for microglia activity in early-stage Alzheimer’s disease and associate with neuronal injury markers. EMBO Mol Med 8, 466–476.

36.

Suárez-Calvet

, Morenas-Rodríguez

, Kleinberger

, Schlepckow

, Caballero

MÁA

, Franzmeier

, Capell

, Fellerer

, Nuscher

, Eren

, Levin

, Deming

, Piccio

, Karch

, Cruchaga

, Shaw

, Trojanowski

, Weiner

, Ewers

, Haass

(2019) Early increase of CSF sTREM2 in Alzheimer’s disease is associated with tau related-neurodegeneration but not with amyloid-β pathology. Mol Neurodegener 14, 1.

37.

Jack

, Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

, Elliott

, Masliah

, Ryan

, Silverberg

(2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562.

38.

Shi

, Yamada

, Liddelow

, Smith

, Zhao

, Luo

, Tsai

, Spina

, Grinberg

, Rojas

, Gallardo

, Wang

, Roh

, Robinson

, Finn

, Jiang

, Sullivan

, Baufeld

, Wood

, Sutphen

, McCue

, Xiong

, Del-Aguila

, Morris

, Cruchaga

, Fagan

, Miller

, Boxer

, Seeley

, Butovsky

, Barres

, Paul

, Holtzman

(2017) ApoE4 markedly exacerbates tau-mediated neurodegeneration in a mouse model of tauopathy. Nature 549, 523–527.

39.

Hudry

, Dashkoff

, Roe

, Takeda

, Koffie

, Hashimoto

, Scheel

, Spires-Jones

, Arbel-Ornath

, Betensky

, Davidson

, Hyman

(2013) Gene transfer of human Apoe isoforms results in differential modulation of amyloid deposition and neurotoxicity in mouse brain. Sci Transl Med 5, 212ra161.

40.

Koch

, Di Lorenzo

, Loizzo

, Motta

, Travaglione

, Baiula

, Rimondini

, Ponzo

, Bonnì

, Toniolo

, Sallustio

, Bozzali

, Caltagirone

, Campana

, Martorana

(2017) CSF tau is associated with impaired cortical plasticity, cognitive decline and astrocyte survival only in APOE4-positive Alzheimer’s disease. Sci Rep 7, 13728.

41.

Motta

, Finardi

, Toniolo

, Di Lorenzo

, Scaricamazza

, Loizzo

, Mercuri

, Furlan

, Koch

, Martorana

(2020) Protective role of cerebrospinal fluid inflammatory cytokines in patients with amnestic mild cognitive impairment and early Alzheimer’s disease carrying Apolipoprotein E4 genotype. Alzheimers Dis 76, 681–689.

42.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

, Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

43.

Skillbäck

, Delsing

, Synnergren

, Mattsson

, Janelidze

, Nägga

, Kilander

, Hicks

, Wimo

, Winblad

, Hansson

, Blennow

, Eriksdotter

, Zetterberg

(2017) CSF/serum albumin ratio in dementias: A cross-sectional study on 1861 patients. Neurobiol Aging 59, 1–9.

44.

Reiber

, Peter

(2001) Cerebrospinal fluid analysis: Disease-related data patterns and evaluation programs. J Neurol Sci 184, 101–122.

45.

Piccio

, Deming

, Del-Águila

, Ghezzi

, Holtzman

, Fagan

, Fenoglio

, Galimberti

, Borroni

, Cruchaga

(2016) Cerebrospinal fluid soluble TREM2 is higher in Alzheimer disease and associated with mutation status. Acta Neuropathol 131, 925–933.

46.

Suárez-Calvet

, Caballero

MÁA

, Kleinberger

, Bateman

, Fagan

, Morris

, Levin

, Danek

, Ewers

, Haass

(2016) Early changes in CSF sTREM2 in dominantly inherited Alzheimer’s disease occur after amyloid deposition and neuronal injury. Sci Transl Med 8, 369ra178.

47.

Milà-Alomà

, Salvadó

, Gispert

, Vilor-Tejedor

, Grau-Rivera

, Sala-Vila

, Sánchez-Benavides

, Arenaza-Urquijo

, Crous-Bou

, González-de-Echávarri

, Minguillon

, Fauria

, Simon

, Kollmorgen

, Zetterberg

, Blennow

, Suárez-Calvet

, Molinuevo

(2020) Amyloid beta, tau, synaptic, neurodegeneration, and glial biomarkers in the preclinical stage of the Alzheimer’s continuum. Alzheimers Dement 16, 1358–1371.

48.

Knapskog

, Henjum

, Idland

, Eldholm

, Persson

, Saltvedt

, Watne

, Engedal

, Nilsson

LNG

(2020) Cerebrospinal fluid sTREM2 in Alzheimer’s disease: Comparisons between clinical presentation and AT classification. Sci Rep 10, 15886.

49.

Parhizkar

, Arzberger

, Brendel

, Kleinberger

, Deussing

, Focke

, Nuscher

, Xiong

, Ghasemigharagoz

, Katzmarski

, Krasemann

, Lichtenthaler

, Müller

, Colombo

, Monasor

, Tahirovic

, Herms

, Willem

, Pettkus

, Butovsky

, Bartenstein

, Edbauer

, Rominger

, Ertürk

, Grathwohl

, Neher

, Holtzman

, Meyer-Luehmann

, Haass

(2019) Loss of TREM2 function increases amyloid seeding but reduces plaque-associated ApoE. Nat Neurosci 22, 191–204.

50.

Ewers

, Franzmeier

, Suárez-Calvet

, Morenas-Rodriguez

, Caballero

MAA

, Kleinberger

, Piccio

, Cruchaga

, Deming

, Dichgans

, Trojanowski

, Shaw

, Weiner

, Haass

(2019) Increased soluble TREM2 in cerebrospinal fluid is associated with reduced cognitive and clinical decline in Alzheimer’s disease. Sci Transl Med 11, eaav6221.

51.

Meilandt

, Ngu

, Gogineni

, Lalehzadeh

, Lee

, Srinivasan

, Imperio

, Wu

, Weber

, Kruse

, Stark

, Chan

, Kwong

, Modrusan

, Friedman

, Elstrott

, Foreman

, Easton

, Sheng

, Hansen

XDV.

(2020) TREM2 deletion reduces late-stage amyloid plaque accumulation, elevates the Aβ42: Aβ40 ratio, and exacerbates axonal dystrophy and dendritic spine loss in the PS2ApP Alzheimer’s mouse model. J Neurosci 40, 1956–1974.

52.

Kim

, Mun

, Lee

, Joh

, Lee

, Ji

, Choi

, Lee

, Kim

, Jang

, Song

, Mook-Jung

, Choi

, Kang

(2017) TREM2 promotes Aβ phagocytosis by upregulating C/EBPα-dependent CD36 expression in microglia. Sci Rep 7, 11118.

53.

Lee

, Meilandt

, Xie

, Gandham

, Ngu

, Barck

, Rezzonico

, Imperio

, Lalehzadeh

, Huntley

, Stark

, Foreman

, Carano

RAD

, Friedman

, Sheng

, Easton

, Bohlen

, Hansen

(2021) Trem2 restrains the enhancement of tau accumulation andneurodegeneration by β-amyloid pathology, Neuron 109, 1283–1301.e6.

54.

Hardy

, Salih

(2021) TREM2-mediated activation of microglia breaks link between amyloid and tau. Lancet Neurol 20, 416–417.

55.

Wilson

, Swarovski

, Linortner

, Shahid

, Zuckerman

, Wang

, Channappa

, Minhas

, Mhatre

, Plowey

, Quinn

, Zabetian

, Tian

, Longo

, Cholerton

, Montine

, Poston

, Andreasson

(2020) Soluble TREM2 is elevated in Parkinson’s disease subgroups with increased CSF tau. Brain 143, 932–943.

56.

Van Hulle

, Jonaitis

, Betthauser

, Batrla

, Wild

, Kollmorgen

, Andreasson

, Okonkwo

, Bendlin

, Asthana

, Carlsson

, Johnson

, Zetterberg

, Blennow

(2021) An examination of a novel multipanel of CSF biomarkers in the Alzheimer’s disease clinical and pathological continuum. Alzheimers Dement 17, 431–445.

57.

Benedet

, Milà-Alomà

, Vrillon

, Ashton

, Pascoal

, Lussier

, Karikari

, Hourregue

, Cognat

, Dumurgier

, Stevenson

, Rahmouni

, Pallen

, Poltronetti

, Salvadó

, Shekari

, Operto

, Gispert

, Minguillon

, Fauria

, Kollmorgen

, Suridjan

, Zimmer

, Zetterberg

, Molinuevo

, Paquet

, Rosa-Neto

, Blennow

, Suárez-Calvet

(2021) Differences between plasma and cerebrospinal fluid glial fibrillary acidic protein levels across the Alzheimer disease continuum. JAMA Neurol 78, 1471–1483.

58.

Pereira

, Janelidze

, Smith

, Mattsson-Carlgren

, Palmqvist

, Teunissen

, Zetterberg

, Stomrud

, Ashton

, Blennow

, Hansson

(2021) Plasma GFAP is an early marker of amyloid-β but not tau pathology in Alzheimer’s disease. Brain 144, 3505–3516.

59.

Giannoni

, Badaut

, Dargazanli

, De Maudave

AFH

, Klement

, Costalat

, Marchi

(2018) The pericyte-glia interface at the blood-brain barrier. Clin Sci (Lond) 132, 361–374.

60.

Verberk

IMW

, Laarhuis

, van den Bosch

, Ebenau

, van Leeuwenstijn

, Prins

, Scheltens

, Teunissen

, van der Flier

(2021) Serum markers glial fibrillary acidic protein and neurofilament light for prognosis and monitoring in cognitively normal older people: A prospective memory clinic-based cohort study. Lancet Heal Longev 2, e87–e95.

61.

Bellaver

, Ferrari-Souza

, Uglione da Ros

, Carter

, Rodriguez-Vieitez

, Nordberg

, Pellerin

, Rosa-Neto

, Leffa

, Zimmer

(2021) Astrocyte biomarkers in Alzheimer disease: A systematic review and meta-analysis, Neurology 96, e2944–e2955.

62.

Simpson

, Ince

, Lace

, Forster

, Shaw

, Matthews

, Savva

, Brayne

, Wharton

(2010) Astrocyte phenotype in relation to Alzheimer-type pathology in the ageing brain. Neurobiol Aging 31, 578–590.

63.

González-Reyes

, Nava-Mesa

, Vargas-Sánchez

, Ariza-Salamanca

, Mora-Muñoz

(2017) Involvement of astrocytes in Alzheimer’s disease from a neuroinflammatory and oxidative stress perspective. Front Mol Neurosci 10, 427.

64.

Venegas

, Heneka

(2017) Danger-associated molecular patterns in Alzheimer’s disease. J Leukoc Biol 101, 87–98.

65.

Mrak

, Griffin

WST

(2001) The role of activated astrocytes and of the neurotrophic cytokine S100B in the pathogenesis of Alzheimer’s disease. Neurobiol Aging 22, 915–922.

66.

Ridler

(2017) Neuroimmunology: Microglia-induced reactive astrocytes-toxic players in neurological disease? Nat Rev Neurol 13, 127.

67.

Liddelow

, Guttenplan

, Clarke

, Bennett

, Bohlen

, Schirmer

, Bennett

, Münch

, Chung

, Peterson

, Wilton

, Frouin

, Napier

, Panicker

, Kumar

, Buckwalter

, Rowitch

, Dawson

, Stevens

, Barres

(2017) Neurotoxic reactive astrocytes are induced by activated microglia. Nature 541, 481–487.

68.

Liu

, Zhao

, Fu

, Wang

, Linares

, Tsai

, Bu

(2017) ApoE4 accelerates early seeding of amyloid pathology. Neuron 96, 1024–1032.e3.

69.

Pimenova

, Marcora

, Goate

(2017) A tale of two genes: Microglial Apoe and Trem2. Immunity 47, 398–400.

70.

Krasemann

, Madore

, Cialic

, Baufeld

, Calcagno

, El Fatimy

, Beckers

, O’Loughlin

, Xu

, Fanek

, Greco

, Smith

, Tweet

, Humulock

, Zrzavy

, Conde-Sanroman

, Gacias

, Weng

, Chen

, Tjon

, Mazaheri

, Hartmann

, Madi

, Ulrich

, Glatzel

, Worthmann

, Heeren

, Budnik

, Lemere

, Ikezu

, Heppner

, Litvak

, Holtzman

, Lassmann

, Weiner

, Ochando

, Haass

, Butovsky

(2017) The TREM2-APOE pathway drives the transcriptional phenotype of dysfunctional microglia in neurodegenerative diseases. Immunity 47, 566–581.e9.

71.

Machlovi

, Neuner

, Hemmer

, Khan

, Liu

, Huang

, Zhu

, Castellano

, Cai

, Marcora

, Goate

(2022) APOE4 confers transcriptomic and functional alterations to primary mouse microglia., Neurobiol Dis 164.

72.

Young

, Jayadev

(2022) Neighborhood matters: Altered lipid metabolism in APOE4 microglia causes problems for neurons., Cell Stem Cell 29, 1159–1160.

73.

Martens

, Zhao

, Liu

C-C

, Kanekiyo

, Yang

, Goate

, Holtzman

, Bu

(2022) ApoE Cascade Hypothesis in the pathogenesis of Alzheimer’s disease and related dementias. Neuron 110, 1304–1317.

74.

Chakrabarty

, Borchelt

(2021) Astrocytic apoE4 and tau: Deadly combination for neurons. Cell Rep Med 2, 100316.

75.

Shi

, Manis

, Long

, Wang

, Sullivan

, Serrano

, Hoyle

, Holtzman

(2019) Microglia drive APOE-dependent neurodegeneration in a tauopathy mouse model. J Exp Med 216, 2546–2561.

76.

Gratuze

, Schlachetzki

JCM

, D’Oliveira Albanus

, Jain

, Novotny

, Brase

, Rodriguez

, Mansel

, Kipnis

, O’Brien

, Pasillas

, Lee

, Manis

, Colonna

, Harari

, Glass

, Ulrich

, Holtzman

(2023) TREM2-independent microgliosis promotes tau-mediated neurodegeneration in the presence of ApoE4. Neuron 111, 202–219.e7.

77.

Heneka

(2023) ApoE4 makes microglia trem2bling. Neuron 111, 142–144.

78.

Assogna

, Motta

, Bonnì

, Borghi

, Casula

, Martorana

, Koch

(2022) Isolated amyloid-β pathology is associated with preserved cortical plasticity in APOE4 Alzheimer’s disease patients. J Alzheimers Dis 86, 773–778.