2025.08.28.672784v1.full-html.html

bioRxiv preprint

SUMMARY

Astrocytic calcium dysregulation and reactivity precede Aβ deposition in amyloid-β

deposition in Alzheimer’s disease (AD) but the neurotoxic mechanisms remain unclear.

We show that GSDME acts as a switch, linking MAM-mediated calcium release to

astrocyte-driven neurotoxicity. Specifically, Aβ-activated microglial signals activate

astrocytic GSDME, releasing its N-terminal fragment, which targets MAMs and

triggers ER calcium efflux. This induces biphasic CaMKIIα phosphorylation, initially

boosting NRF2 defenses, then activating NF-κB-driven inflammation, shifting

astrocytes from protective to toxic states. GSDME activation also drives astrocyte-

derived exosomes (ADEs) to carry neurotoxic tau, proinflammatory miRNAs, and toxic

lipids, propagating toxicity. GSDME deletion in AD mice reduces Aβ burden, restores

NF-κB/NRF2 balance, reprograms astrocytes and ADEs to protective states, and

rescues cognition. Multi-omics profiling of serum ADEs from AD patients reveals a

disease-specific signature with central neurotoxicity and peripheral immune regulation.

These findings position GSDME as a promising dual diagnostic and therapeutic target

for early AD invention.

Keywords

Alzheimer’s disease, astrocyte reactivity, GSDME; calcium signaling, mitochondrial-

associated ER membranes (MAMs), exosome, neuroinflammation; NF-κB/NRF2

balance.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

INTRODUCTION

Alzheimer’s disease (AD), the predominant form of dementia, is classically defined by

amyloid-β (Aβ) plaques and tau neurofibrillary tangles.

1-4

However, longitudinal

studies reveal that 20-30% of Aβ-positive individuals maintain cognitive stability over

decadal follow-ups,

5,6

underscoring the necessity to identify co-factors that convert Aβ

pathology into neurodegeneration.

Emerging preclinical evidence reveals a critical

temporal pattern, astrocyte reactivity-manifested by pro-inflammatory activation and

impaired glutamate uptake precedes detectable Aβ deposition and tau

hyperphosphorylation

7-10

This early astrocytic dysfunction is further supported by

human biomarker studies, where elevated plasma GFAP, a reactivity marker, predicts

subsequent tau propagation in presymptomatic stages.

9,11

Collectively, these findings

position astrocyte reactivity as a potential initiator of AD pathogenesis. Given the

limited efficacy of Aβ-targeting therapies in late-stage disease,

12,13

identifying the

molecular triggers and spatiotemporal progression of astrocyte reactivity

may offer a

strategic window for early interventions to halt synaptic degeneration before

irreversible damage occurs.

Astrocytic calcium dysregulation emerges as a pivotal early event in AD

pathogenesis

7,10,15

Under physiological conditions, calcium oscillations in astrocytes

maintain synaptic efficiency and neurovascular coupling through regulated

gliotransmitter release.

In early AD, however, these cells develop pathological

calcium hyperactivity, which

occurs prior to Aβ deposition and drives synaptic

hyperexcitability.

1,17

This aberrant signaling disrupts neurovascular coordination by

altering prostaglandin release and exacerbates neuroinflammation via uncontrolled

cytokine secretion

18,19

Crucially, restoring calcium homeostasis reverses cognitive

deficits in preclinical models,

while spatial transcriptomics identifies calcium-

dependent inflammatory pathways in plaque-associated astrocytes from human AD

brains

21,22

Despite these advancements, a significant knowledge gap remains regarding

the mechanisms by which early calcium signaling anomalies trigger the transition from

homeostatic astrocytes to a chronically reactive state.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Emerging evidence underscores the dualistic role of astrocyte reactivity in

neurodegeneration, governed by distinct activation states

19,23

Early classifications

proposed a binary paradigm that neurotoxic (A1) astrocytes induced by microglial

inflammatory signals, such as IL-1α and TNF-α, via JAK-STAT/NF-κB pathways, and

neuroprotective (A2) astrocytes activated in ischemic contexts.

7,19

Recently, single-cell

and spatial transcriptomics redefine this dichotomy, identifying disease-associated

astrocyte (DAA) subpopulations marked by inflammatory gene signatures, notably

exhibiting elevated expression of C3, alongside phagocytic dysfunction.

These DAAs

engage in bidirectional crosstalk with microglia through cytokines, including IL-6,

TGF-β,

and recruit peripheral immune cells via chemokine gradients such as

CXCL10,

thereby establishing self-reinforcing neuroinflammatory niches that

accelerate synaptic loss. Crucially, spatial mapping reveals DAAs forming peri-plaque

"hotspots" that expand with disease progression.

However, the fundamental

mechanism by which localized astrocytic inflammatory foci propagate across brain

networks remains elusive.

While gasdermin (GSDM)-gated pyroptosis has been reported to engaged in

microglial and neuronal dysfunction in neurodegeneration

, such as ALS

and MS

however, its role in reactive astrocytes and AD-related pathologies remains largely

unknown. Here, we identify a non-canonical GSDME activation pathway that

transforms astrocytes into neurodegenerative effectors via calcium-dependent

mechanisms. In response to

Aβ-activated microglial conditional culture medium,

GSDME cleavage targets its N-terminal fragment to mitochondria-associated ER

membranes (MAMs), inducing ER calcium efflux and activating the CaMKIIα-NF-κB

axis to drive reactive astrocyte-induced neurotoxicity. Notably, astrocytic neurotoxicity

can be propagated through astrocyte-derived exosomes (ADEs) enriched with

pathogenic cargo. In AD patient brains and 5×FAD mice, GSDME activation was

predominantly observed in reactive astrocytes, while genetic GSDME ablation reduces

Aβ deposition, restores NF-κB/NRF2 balance, and converts astrocytes and ADEs to

neuroprotective phenotypes. Multi-omics of serum ADEs from AD patients reveals a

neurotoxic signature, AD-ADEs activate microglia, induce neurotoxic astrocytes,

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

increase neuronal death, and also reduce platelet/neutrophil counts in mice. Our

findings collectively identify GSDME as a dual regulator of astrocyte reactivity and

ADE neurotoxicity, proposing therapeutic targeting of GSDME alongside serum ADE

monitoring for real-time disease tracking.

RESULTS

GSDME exhibits spatial and subcluster-specific heterogeneous expression in

human and mouse astrocytes.

To investigate the potential involvement of GSDM family members in the central CNS,

we analyzed publicly available RNA sequencing data from human

and mouse

brain

tissues. Notably, GSDME transcript levels were markedly higher in the brain compared

to other GSDM family members (

Figure S1

A and S1B). Further examination of single-

cell RNA-seq data

revealed pronounced GSDME expression in the

Slc1a3

high

astrocyte subpopulation (

Figure S1

), suggesting a cell-type-specific role in astrocytes.

Given the established heterogeneity of astrocytes, we next evaluated GSDME

expression patterns across astrocyte subtypes using high-resolution single-cell

transcriptomic data.

32,33

Our analysis showed that AST1 and AST4 were the main

subtypes with high GSDME expression under resting conditions (

Figure S1

D). These

subtypes are characterized by high

GFAP

and

Slc1a3

expression, consistent with prior

whole-brain single-cell studies

(

Figure S1

C). Intriguingly, GSDME expression

correlated strongly with

GFAP

, but less with other astrocytic markers, including

Slc1a3

Aldh1l1

, or

Slc1a2

(

Figure S

1E), underscoring varied GSDME expression among

astrocyte subtypes.

Based on the spatial position data concerning the

GFAP

high

AST1 and AST4

astrocyte subtypes,

we hypothesized that GSDME expression in astrocytes might

similarly favor the hippocampal region. To evaluate this hypothesis, we extracted

astrocytes from both the cerebral cortex and the hippocampus for immunoblot analysis.

Our results demonstrated that GSDME protein levels were higher in astrocytes isolated

from the hippocampus compared to those from the cerebral cortex (

Figure S1

F and

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

S1G). To verify the astrocyte-specific enrichment of GSDME in the hippocampus, we

isolated astrocytes, microglia, and neurons from this region and performed

immunoblotting. GSDME was detected exclusively in GFAP-positive astrocyte

fractions, while GSDMD was absent (

Figure S1

H and S1I). Consistently, mouse

hippocampal co-staining showed strong GSDME-GFAP colocalization, with no overlap

with NeuN or microglial markers (

Figure S1

J and S1K). Human hippocampal co-

staining confirmed GSDME-GFAP colocalization (

Figure S1

L and S1M), validating

single-cell RNA-seq findings.

Pathological GSDME activation in reactive astrocytes is shared by both AD

patients and 5×FAD mouse

To elucidate the pathogenic role of astrocytic GSDME in AD, we conducted an

integrative multi-omics analysis of human postmortem brain samples, complemented

by functional validation in 5×FAD mouse models. Single-nucleus RNA sequencing

comparing AD and other neurodegenerative patients to controls demonstrated a

significantly increase in

GSDME

transcript levels (

Figure 1

A), particularly within

reactive astrocyte clusters

34-40

(

Figure S1

N-S1Q). As GSDME activation is functionally

essential,

we analyzed hippocampal lysates from both AD patients and controls by

immunoblotting. The results showed elevated GSDME-NT fragments and increased

caspase-3 activation in AD samples,

correlating with decreased synaptic expression

compared to controls (

Figure 1

B and 1C).To substantiate the role of astrocytic GSDME

activation in AD we conducted co-immunostaining of human tissue sections using C3,

a marker for neurotoxic astrocytes, along with GFAP and GSDME-NT. The results

indicated a higher proportion of GSDME-NT-positive neurotoxic reactive astrocytes in

AD brain samples (

Figure 1

D and 1E). These observations imply that GSDME

activation in reactive astrocytes may contribute to AD pathogenesis.

To further validate the

in vivo

significance of GSDME activation in AD

pathogenesis, we examined cortical and hippocampal tissues from 5×FAD mice at

various aging timepoints. Immunoblotting revealed a significant linear correlation

between the intensity of the GSDME cleavage fragment band and mouse age,

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

suggesting progressive GSDME activation in parallel with AD pathology (

Figure 1

F-

1I). Immunofluorescence (IF) analysis in 6-month-old 5×FAD mice showed an increase

in GSDME-NT-positive astrocytes surrounding Aβ plaques, compared to age-matched

controls, with pronounced colocalization between GSDME-NT and C3 (

Figure 1

J and

1K). Notably, adult mouse brain cell fractionation assays confirmed that GSDME

cleavage products were predominantly localized within the astrocytic compartment

(

Figure 1

L and 1M). Collectively, these findings demonstrate that pathological

activation of GSDME in reactive astrocytes is a common feature observed in both AD

patients and the 5×FAD mouse model.

Non-pyroptotic GSDME activation is required for reactive astrocyte-mediated

neurotoxicity

Building upon the observed pathological correlation between GSDME activation and

AD progression, we subsequently explored its functional role in Aβ-induced glial

crosstalk, a recognized contributor to neurotoxicity in AD pathogenesis.

15,42

To emulate

disease-relevant interactions between microglia and astrocytes, we established an

vitro

model system. In this system, BV2 microglial cells were primed with pathogenic

Aβ

1-42

oligomers to produce activated microglia-conditioned medium (AMCM), and

primary astrocytes were subsequently exposed to this AMCM to observe phenotypic

transformation. Temporal transcriptomic analysis demonstrated that prolonged

exposure to AMCM induced a progressive transformation in astrocytes, characterized

by a stable neurotoxic transcriptional signature at 24, 48, 72 and 96 h post-stimulation

(hps) compared to microglia-conditioned medium (MCM) with PBS treatment. This

neurotoxic shift was characterized by three distinct molecular features: (1)

Upregulation of neurotoxic mediators, such as

, and

Srgn

, consistent with their

known roles in reactive astrogliosis and neuroinflammatory responses;

(2)

Dysregulation of protective pathways, evidenced by significant downregulation of

S100A10

Sphk1

and

Emp1

, thereby compromising their neuroprotective functions; and

(3) Deficits in neurotrophic support, as indicated by reduced levels of

Fgf1

Shh

and

Gpc4

, critically impairing synaptic maintenance and neuronal survival

(

Figure S2

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

and S2B).

These molecular changes collectively establish AMCM-induced astrocytic

transformation as a key mediator of AD-related neurodegeneration, with concurrent

activation of detrimental pathways and suppression of protective mechanisms.

To elucidate the functional role of GSDME activation in reactive astrocytes, we

investigated whether the caspase-3-GSDME signaling pathway (

Figure 2

A), which

executes pyroptosis through GSDME cleavage,

was operational in AMCM-treated

astrocytes. Temporal analysis following AMCM treatment demonstrated sustained

caspase-3 activation, leading to robust GSDME cleavage and production of GSDME-

NT fragments (

Figure 2

B and S2C). Contrary to the anticipated outcomes associated

with pyroptotic cell death, neither LDH release assays (

Figure 2

C and S2D) nor PI

staining (

Figure 2

D and S2E) detected compromised membrane integrity in AMCM-

treated groups. These findings contrast with the significant cell death increases

observed in GSDME-NT-EGFP overexpression controls. To further validate these

observations, we performed IF staining (

Figure 2

E, 2F and S2F-2H) and scanning

electron microscopy (SEM) on AMCM-treated astrocytes (

Figure 2

G and 2H).

Quantitative analysis revealed a marked increase in GSDME-NT-positive cells under

AMCM conditions,

while the cellular morphology remained indistinguishable from

controls, exhibiting no blebbing and intact membranes, and demonstrated significantly

lower PI-positive cell counts compared to GSDME-NT-EGFP controls (

Figure 2

E-2H).

These convergent data imply that while AMCM activates GSDME cleavage in

neurotoxic reactive astrocytes, this activation does not induce classical pyroptotic

features such as membrane pore formation or cell blebbing typically associated with

pyroptotic cell death.

Recent studies indicate that neurotoxic reactive astrocytes obtain aberrant

functions, including impaired phagocytosis and disrupted induction of

synaptogenesis.

To assess the role of GSDME in these abnormalities, we stimulated

both wild-type and GSDME KO astrocytes with MCM or AMCM for 24 h, evaluating

phagocytic activity by quantifying synaptosome uptake. The results demonstrated that

AMCM-treated GSDME KO astrocytes exhibited increased phagocytosis of pHrodo-

labeled synaptosomes compared to AMCM-treated WT astrocytes, aligning with

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

MCM-treated WT astrocytes (

Figure 2

I and 2J). This corresponds with upregulated

mRNA expression of phagocytic receptors,

Megf10, Mertk, Axl,

along with the

bridging molecule

Gas6

in AMCM-exposed GSDME KO astrocytes compared to

AMCM-treated WT astrocytes (

Figure 2

K).

Additionally, genes encoding synaptogenic

factors,

Thbs1, Gpc6, Sparcl1

were upregulated in AMCM-treated GSDME KO

astrocytes versus WT (

Figure 2

K). This suggests that the knockout of GSDME, to some

extent, alleviates the abnormalities in astrocytic function following AMCM treatment.

To assess neurotoxicity, we administered astrocyte-conditioned medium (ACM)

harvested from MCM-treated WT, MCM-treated GSDME KO, AMCM-treated WT,

and AMCM-treated GSDME KO astrocytes to cultures of mouse primary cortical

neurons. Our results indicated that, in comparison to the control group exposed to ACM

derived from AMCM-treated WT astrocytes, ACM from AMCM-treated GSDME KO

astrocytes significantly reduced neuronal deaths (

Figure 2

L and 2M). These findings

suggest that GSDME deficiency have the potential to attenuate neuronal death caused

by neurotoxic reactive astrocytes, thereby highlighting a possible therapeutic target for

neuroprotective strategies.

MAM-localized GSDME-NT drives ER-to-cytosol Ca²⁺ efflux to trigger neurotoxic

reactive astrogliosis

Given that GSDME was activated in reactive astrocytes without inducing pyroptosis,

we sought to identify the specific intracellular membranes targeted by the released

GSDME-NT. To determine the distribution of GSDME-NT across the intracellular

membranes of reactive astrocytes, we employed a standard protocol

for isolating

fractions of mitochondria, endoplasmic reticulum (ER), and mitochondria-associated

ER membranes (MAMs), constituting the primary intracellular membrane systems in

mammalian cells (

Figure 3

A). To assess the quality of subcellular fractionation,

immunoblotting was utilized to analyze marker proteins specific to different cellular

compartments. The results revealed that VDAC1, a mitochondrial marker, was

predominantly detected in the pure mitochondria fraction and was absent from the

cytosol fraction. Likewise, calnexin, acting as a marker for both ER and MAMs,

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

exhibited significant levels in both ER and MAM fractions. Consistent with these

findings, ACSL4, previously identified as an MAM-resident protein,

was enriched in

the MAM fraction. Conversely, β-actin, a cytosolic marker, was not detected in

mitochondria, MAMs, or ER fractions (

Figure 3

A). Collectively, these findings suggest

a high

quality of subcellular fractionation harvest.

Upon administering AMCM, we observed a notable elevation in GSDME-NT

fragments in the whole cell lysates of WT-AMCM astrocytes compared to MCM-

treated controls (

Figure 3

A). This finding implies the activation of GSDME in reactive

astrocytes. Further delving the activation of GSDME in various subcellular fractions,

we discovered a significant enrichment of GSDME-NT fragments specifically in the

MAM fraction, as opposed to the mitochondria fraction (

Figure 3

A). IF staining

followed by structured illumination microscopy (SIM) imaging revealing a significant

elevation in the colocalization of GSDME-FL with ACSL4 in WT-AMCM astrocytes

compared to

WT-MCM astrocytes (

Figure 3

B and 3C). Additional

SIM imaging

showcased a high degree of colocalization between GSDME-NT and ACSL4 (

Figure

D and 3E) and hinted at the targeting of GSDME-NT to MAMs. Furthermore,

transmission electron microscopy (TEM) observations confirmed frequent GSDME-

NT labeling at ER-mitochondria contact sites (

Figure 3

F). Stereological analysis further

revealed the distinct enrichment of GSDME-NT at MAMs (

Figure 3

G and 3H). These

results suggest that GSDME, activated by AMCM, is translocated to MAMs in reactive

astrocytes.

transfer is a primary function attributed to ER-mitochondrial contact sites.

Given that GSDME-NT targets and forms pores on the membrane, we investigated

whether the translocation of GSDME-NT to MAMs facilitates Ca

transfer. To test this

hypothesis, we utilized Fluo 8-AM, a fluorescent cytoplasmic Ca

indicator, to monitor

changes in cytoplasmic Ca

concentration in reactive astrocytes. Our findings reveal

that incubation with AMCM significantly increases cytoplasmic Ca

concentration in

these cells (

Figure 3

I and 3J).

However, this elevation was markedly impeded in the

absence of GSDME (

Figure 3

I and 3J). Additionally, we quantified the mRNA

expression of

Calm3

, an inherent calcium-binding protein capable of sensing cytosolic

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

levels,

as an indicator of cytoplasmic Ca

concentration changes in astrocytes.

Consistently,

we noted a reduction in

Calm3

transcription in GSDME KO-AMCM

astrocytes compared to their WT-AMCM counterparts (

Figure 3

K). These results

suggest that GSDME-NT targeting MAMs contributes to the elevation of cytoplasmic

concentration in reactive astrocytes.

Next, we expressed genetically encoded Ca

indicators specifically targeting

cytosol (GCaMP6s), ER (CEPIA1-ER), and mitochondria (CEPIA4-Mito) to monitor

dynamics

48,49

(

Figure S3

A-S3E). Notably, the depletion of extracellular Ca

did

not change the intracellular Ca

elevation through GCaMP6s fluorescence intensity

observation in astrocytes treated with AMCM (

Figure 3

L). This suggests that the

cytosolic Ca

rise primarily originates from GSDME pores located at the endosomal

membrane, rather than the plasma membrane. Meanwhile, the fluorescence intensity

enhancement of GCaMP6s in WT-AMCM astrocytes was suppressed in the absence of

GSDME (

Figure 3

L). These results reinforce the hypothesis that GSDME-NT targeting

MAMs contributes to the cytoplasmic Ca

concentration elevation. To pinpoint the

source of this cytoplasmic Ca

surge, we

dissected Ca

dynamics within mitochondria

and ER, the principal Ca

reservoirs in astrocytes. The analysis reveals that AMCM

triggers a decrease in ER Ca

dynamics, as evidenced by CEPIA1-ER fluorescence

(

Figure 3

M), while simultaneously inducing an increase in mitochondrial Ca

dynamics, demonstrated by CEPIA4-Mito fluorescence (

Figure 3

N), in reactive

astrocytes compared to WT-MCM astrocytes. Therefore, we believe that GSDME-NT

localizes to MAMs, facilitating the release of Ca

from ER into cytosol.

To mechanistically interrogate the causal linkage between GSDME-NT-driven ER

calcium mobilization and neurotoxic astrocyte transformation, we designed two

molecular probes

，

namely full-length EMC10-GSDME-NT-EGFP incorporating the

ER-localizing EMC10 domain

and its N-terminal 56-amino acid

truncated

counterpart EMC10-GSDME-NT(Δ56)-EGFP. Lentiviral delivery of EMC10-

GSDME-NT-EGFP in primary astrocytes preserved viability equivalent to EMC10-

EGFP controls, consistent with prior observations of pyroptosis-independent GSDME

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

activation (

Figure S3

F). IF quantification demonstrated a notable elevation in C3

expression in cells expressing EMC10-GSDME-NT, a finding further substantiated by

qPCR, which indicated a coordinated transcriptional upregulation of genes specific to

neurotoxicity (

Figure 3

O-3P). In contrast, the Δ56 truncation mutant not only inhibited

C3 induction but also repressed the expression of neurotoxic-specific genes (

Figure 3

O-

3P).These evidences collectively suggest that MAM-anchored GSDME-NT mediates

compartment-specific calcium shutting from ER to cytosol, which functions as a

molecular switch initiating the neurotoxic astrogliosis program.

Phase-Specific CaMKⅡα phosphorylation by GSDME governs the

neuroprotective-to-neurotoxic transition

To elucidate how GSDME-NT induces MAM-dependent calcium dysregulation and

drives astrocytes to a neurotoxic state, we combined transcriptomic profiling with

functional validation. Unexpected, bulk RNA sequencing demonstrated that GSDME

knockout suppressed transcription of neurotoxic genes such as

and

Lcn2

while

upregulating protective markers including

S100A10

Emp1

, and

Sphk1

(

Figure 4

A).

These data indicate that GSDME is essential for

reprogramming astrocytes toward a

neuroprotective phenotype. KEGG pathway enrichment analysis identified calcium

signaling as the most significantly affected pathway (

Figure 4

B), aligning with

GSDME’s role in modulating calcium dynamics at the ER-mitochondria interface.

Further differential expression analysis within the calcium signaling category

highlighted

Camk2a

(the gene encoding CaMKIIα) as a central regulator of the

GSDME-driven neurotoxic transition (

Figure 4

C). Functionally validation via siRNA-

mediated knockdown of

Cam

k2a demonstrated that suppression of CaMKIIα abrogates

the induction of protective gene (

Figure 4

D), confirming CaMKIIα as a key effector in

the GSDME-calcium signaling axis.

Time-resolved immunoblot analysis revealed that astrocytes adopt a

neuroprotective phenotype within 12 hps, as indicated by increased expression

S100A10,

SPHK1, and

GPC4,

along with suppressed levels of

and

LCN2

(

Figure 4

E and 4F). Strikingly, after 12 hps, astrocytes transition to a neurotoxic state:

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

and

LCN2 are robustly upregulated, whereas protective markers S100A10,

GPC4,

and

SPHK1 decline (

Figure 4

E and 4F). This biphasic pattern demonstrates that

induction of a neuroprotective phase precedes toxic conversion, consistent with

previous observation.

Concurrently, CaMKⅡα phosphorylation (p-CaMKⅡα) exhibits

a dynamic biphasic profile (

Figure 4

G and 4H), implicating posttranslational regulation

of astrocyte state. Detailed analysis of NF-κB and NRF2 signaling revealed that, during

the early phase (0-12 hps), phosphorylated NRF2 (p-NRF2) predominates while

phosphorylated NF-κB subunit p65 (p-p65) remains low. In contrast, the late phase (12-

48 hps) shows diminished p-NRF2 and increased p-p65 (

Figure 4

G and 4H).

Importantly, the ratio of p-NRF2 to p-p65 correlates positively with p-CaMKIIα activity

(

Figure 4

H), suggesting that CaMKⅡα dynamically balances NF-κB and NRF2

signaling to dictate astrocyte fate.

To validate the spatiotemporal regulation of CaMKⅡα phosphorylation by

GSDME, we conducted time-resolved immunoblotting in astrocytes with GSDME

silenced. During the early phase (0-12 hps), GSDME silencing did not significantly

alter protective marker expression, likely reflecting minimal GSDME activation at this

stage (

Figure 4

I and 4J). However, in the late phase (12-48 hps), GSDME knockdown

prevented induction of neurotoxic markers and maintained protective marker

expression, effectively blocking the transition to a neurotoxic phenotype (

Figure 4

I and

4J). Correspondingly, GSDME silencing attenuated late-phase CaMKⅡα

phosphorylation and preserved the

p-NRF2/p-p65 ratio (

Figure 4

K and 4L). These

findings suggest that early-phase CaMKⅡα phosphorylation is required for

neuroprotection, whereas excessive phosphorylation at later stages disrupts the NF-

κB/NRF2 balance and promotes toxicity.

Pharmacological modulation corroborated the phase-specific requirement for

CaMKⅡα in astrocyte fate transition. Continuous inhibition at the beginning of stimulus

with CaMKIIα-IN-1

failed to prevent the neurotoxic conversion (Data not shown). In

contrast, a sequential treatment recapitulated the protective effects observed with

GSDME ablation. This regimen elevated S100A10, GPC4, and SPHK1 and reduced C3

and LCN2 levels (

Figure 4

M and 4N). Mechanistically, early-phase activation of

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

CaMKⅡα enhanced the p-NRF2/p-p65 ratio, while subsequent inhibition prevented

late-phase CaMKIIα hyperphosphorylation and suppressed neurotoxic marker

induction (

Figure 4

O and 4P). Together, these findings underscore the necessity of

precise temporal regulation of CaMKⅡα activity: early activation potentiates NRF2-

mediated neuroprotection, whereas unchecked hyperphosphorylation in later phases

shifts the balance toward NF-κB-driven neurotoxicity. Overall, GSDME functions a

molecular switch that dynamically modulates CaMKIIα phosphorylation to maintain

NF-κB/NRF2 signaling equilibrium and govern the transition between neuroprotective

and neurotoxic states in astrocytes.

GSDME reprograms astrocyte-derived exosomes to propagate neurotoxic signals

Given that GSDME activation elevates cytosolic Ca²⁺ levels in reactive astrocytes, we

examined whether altered calcium signaling regulates exosome production in these

cells. Using a standardized exosome isolation protocol, we obtained astrocyte-derived

exosomes (ADEs) from ACM and aACM, respectively (

Figure 5

A-5C). TEM imaging

confirmed intact vesicular structures (

Figure S5

A), while nanoparticle tracking analysis

(NTA) revealed size distributions between 30-200 nm, consistent with canonical

exosome dimensions (

Figure S5

B).

Notably, quantitative comparison showed

equivalent average diameters between ACM-ADEs (ADEs) and aACM-ADEs (aADEs)

(

Figure S5

C), while the particle counts of aADEs were much higher than those of ADEs

(

Figure S5

D) Functional validation through GSDME knockout revealed markedly

reduced aADEs production in GSDME-deficient astrocytes cultured in MCM (

Figure

B), suggesting GSDME as a critical regulator of ADEs biogenesis in reactive

astrocytes.

Exosomes have emerged as pivotal mediators of neuroinflammation and cognitive

decline,

53-55

prompting investigation into their role in aACM-mediated neurotoxicity.

Fractionation of neurotoxic aACM revealed that its exosomal component recapitulated

the full neurotoxicity of equivalent aACM volumes (

Figure 5

D), with effects scaling

dose-dependently (

Figure S5

E). Removal of aADEs from aACM abolished its

neurotoxic activity, restoring neuronal viability to ACM levels (

Figure 5

D), confirming

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

exosomes as primary neurotoxic effectors. To interrogate the role of GSDME, we

isolated ADEs from WT-ACM, GSDME KO-ACM, WT-aACM, and GSDME KO-

aACM, and confirmed equivalent neuronal uptake across all groups (

Figure 5

G and

S5F). While WT-aADEs reduced neuronal survival compared to WT-ADEs

counterparts (

Figure 5

E), GSDME ablation completely rescued this neurotoxicity.

Synaptic analysis through MAP2/SYP immunostaining further demonstrated that WT-

aADEs disrupted synaptogenesis more severely than WT-ADEs, an effect entirely

reversed by GSDME deficiency (

Figure 5

F and 5G). These findings indicate GSDME-

controlled ADEs as indispensable mediators of pathological neurotoxicity, directly

linking GSDME activation in reactive astrocytes to synaptic pathology and neuronal

demise.

Exosomes serve as critical carriers of neurotoxic cargoes, including misfolded

proteins and inflammatory mediators, directly contributing to the pathogenesis of

neuronal injury.

To systematically dissect the neurotoxic molecular components in

aADEs, we performed multi-dimensional molecular profiling analyses.

Immunoblotting further confirmed differential loading of neurotoxic proteins,

including C3, LCN2, HMGB1, iNOS and neuroprotective proteins, GPC4 and BDNF,

in WT-aADEs (

Figure 5

H and 5I). Lipidomic profiling by LC-MS revealed

significantly elevated levels of neurotoxic long-chain saturated lipids

in WT-aADEs

compared to WT-ADEs (

Figure 5

J). Concurrently, miRNA profiling demonstrated

marked upregulation of neurotoxic miRNAs,

55,57

such as miR-141-3p, miR-155-5p,

miR-135a-5p, miR-125a-5p, and downregulation of neuroprotective miRNAs

such as

miR-9-5p, miR-124-3p, miR-132-3p, miR-29b in WT-aADEs (

Figure 5

K). Notably,

transcriptomic analysis of ADEs revealed that WT-aADEs recapitulated the

transcriptional signature of their parental cells, characterized by activation of

neurotoxic transcripts and suppression of neuroprotective transcripts (

Figure S5

H).

This finding suggests that ADEs may function as biological sentinels reflecting

astrocyte reactivity states. Importantly, GSDME deficiency not only inhibited the

incorporation of neurotoxic proteins, lipids and miRNAs or mRNAs, but also enhanced

the packaging efficiency of neuroprotective miRNAs, mRNAs and selectively enriched

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

neurotrophic factors (

Figure 5

H-5L). Collectively, these findings reveal that GSDME

ablation profoundly reprogrammed the molecular composition of aADEs

converting

these exosomes from neurotoxic effectors to neuroprotective signaling entities.

Next, we investigated whether the aADEs mediate propagating

neuroinflammation. We isolated WT-ADEs, WT-aADEs and GSDME-KO-aADEs and

assessed their effects on BV2 microglia and primary astrocytes (

Figure 5

M). Compared

to WT-ADEs, WT-aADEs upregulated pro-inflammatory genes (

Figure 5

N) and signals

in BV2 cells (

Figure 5

O), while inducing neurotoxic transcripts (

Figure 5

P) and

neurotoxicity (

Figure 5

Q) of primary astrocytes, suggesting their role as

neuroinflammatory propagators. Notably, GSDME ablation suppressed pro-

inflammatory gene expression in microglia (

Figure 5

N) and inhibited neurotoxic

phenotype conversion (

Figure 5

P and 5Q) in astrocytes. These findings position

GSDME as a key regulator of aADEs-mediated neuroinflammatory cascades, operating

through coordinated control of cytotoxic cargo sorting and intercellular signaling

amplification.

Blocking astrocytic GSDME activation alleviates pathology in 5×FAD mice

Given our prior observation of pathological GSDME activation in astrocytes from

both AD patients and 5×FAD transgenic mice (

Figure 1

), we hypothesized that

astrocyte-specific knockdown of GSDME might attenuate Aβ-induced

neurodegeneration in 5×FAD mice. To achieve targeted knockdown in astrocytes

throughout the developing brain, we employed a PHP.eB-serotype AAV vector, which

is capable of crossing the blood–brain barrier to express GSDME-specific shRNA

(AAV-shGE) or a non-targeting control shRNA (AAV-shNT) under the GFAP

promoter.

58,59

Neonatal (postnatal day 1–3) mice received stereotaxic

intracerebroventricular injections, and all subsequent analyses were performed at 6

months of age (

Figure 6

A). Whole-brain mapping of viral expression confirmed robust

distribution of both constructs and high colocalization with GFAP-positive astrocytes

(

Figure 6

B and 6C), validating our approach for cell-type-specific intervention.

Quantification by immunoblotting revealed that total GSDME protein levels were

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

significantly reduced in the hippocampi of AAV-shGE–treated 5×FAD mice compared

to the AAV-shNT controls (

Figure 6

D and 6E). Importantly, the abundance of the

cleaved GSDME N-terminal fragment (GSDME-NT) was likewise diminished (

Figure

D and 6E), suggesting effective inhibition of GSDME activation in astrocytes.

To evaluate therapeutic efficacy, we performed neuropathological assessments and

behavioral testing six months after injection. Immunoblot analysis showed that

GSDME silencing not only reduced Aβ

1-42

levels but also prevented the downregulation

of synaptic markers (

Figure 6

D and 6E). Consistent with these findings, QM-FN-SO₃

staining, which specifically targets Aβ plaques, revealed a marked reduction in plaque

burden in AAV-shGE-treated mice (

Figure 6

F and 6G), suggesting that astrocyte-

specific GSDME silencing suppresses Aβ pathology. NeuN immunostaining further

revealed preservation of layer V cortical pyramidal neurons in the AAV-shGE group

(

Figure 6

H and 6I), implying a neuroprotective effect mediated by astrocytic GSDME

knockdown.

In the open-field test, no significant differences in locomotor activity were

observed among treatment groups (

Figure 6

J and 6K), ruling out motor dysfunction as

a confounding variable for cognitive readouts. In contrast, AAV-shGE-treated 5×FAD

mice exhibited significantly improved spatial memory in the Y-maze spontaneous

alternation test (

Figure 6

L and 6M) and displayed greater discrimination in the novel

object recognition assay (

Figure 6

N and 6O). These data collectively demonstrate that

astrocyte-specific inhibition of GSDME activation rescues both neuropathology and

cognitive deficits in the 5×FAD model, highlighting GSDME as a promising

therapeutic target for AD.

GSDME ablation orchestrates dual reprogramming of astrocyte and exosome

reactivity to halt neurotoxic signaling

in vivo

To examine the potential modulation of astrocyte state transitions by astrocyte-specific

GSDME knockdown, we analyzed brain tissues from 5×FAD mice through

immunoblotting and transcriptomic sequencing. Immunoblotting revealed that

GSDME ablation suppressed neurotoxic factor C3 expression, while upregulating

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

neuroprotective genes, such as S100A10, SHH, GPC4, and BDNF, accompanied by

reduced Iba1, and GFAP levels (

Figure 7

A and 7B), indicating attenuated

neuroinflammatory signaling. Mechanistically, astrocytic GSDME knockout restored

aberrantly low p-CaMKIIα phosphorylation to wild-type levels (

Figure 7

C and 7D), a

molecular signature linked to improved synaptic plasticity markers

and elevated p-

NRF2/p-p65 (

Figure 7

C and 7D), suggesting its contribution to neuronal protection.

Transcriptomics and GSEA further demonstrated suppression of pro-inflammatory

mediators such as

Lcn2

Nlrc5

, and

Irf7

, alongside upregulation of neuroprotective

factors, such as

Sphk1

and

B3GNT5

, as well as neurotrophic factors,

BMP6

and

BMP7

(

Figure 7

E and 7F). Pathway analysis revealed inhibition of type I interferon and

oxidative stress signaling, coupled with activation of neurotransmitter biosynthesis and

glial-derived neurotrophic factor synthesis (

Figure 7

F). qPCR also confirmed

downregulated

and

Lcn2

, and upregulated

EMP1

and

Sphk1

mRNA levels in

hippocampus (

Figure 7

G). IF validation confirmed reduced proportions of C3-positive

toxic astrocytes (

Figure 7

H and 7I), increased S100A10 (

Figure 7

J and 7K) or BDNF-

positive (

Figure 7

L and 7M) protective astrocytes, and diminished microglial

infiltration (

Figure 7

N and 7O), establishing GSDME deletion as a suppressor of

astrocyte neurotoxic phenotypic switching

in vivo

To investigate whether GSDME regulates the reactive state of ADEs

in vivo

, we

isolated ADEs from mouse serum for functional validation (

Figure 7

P and S6).

Immunoblot analysis revealed that astrocytic GSDME knockdown suppressed C3

loading in ADEs, while promoting the enrichment of neurotrophic factors, GPC4 and

BDNF (

Figure 7

Q and 7R). The qPCR results demonstrated that astrocytic GSDME

ablation significantly reduced transcriptional levels of

and

Lcn2

and upregulated the

neuroprotective factor

Gdnf, Tgm, Sphk1

(

Figure 7

S). Lipidomic analysis further

confirmed that GSDME deletion attenuated the production of toxic lipids (

Figure 7

T).

Collectively,

astrocyte-specific

GSDME

knockdown

dually

suppresses

neuroinflammation and restores protective phenotypes intracellularly, while blocking

exosomal toxins transfer and boosting reparative signals extracellularly, systemically

resetting CNS and peripheral pathology. These findings establish GSDME as a

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

molecular switch that regulates astrocyte-bystander cell communication through ADEs

to limit disease progression.

Multi-omics analysis of human serum astrocyte-derived exosomes uncovers

central-to-peripheral immune-regulatory signaling

To determine the clinical roles of ADEs in mediating neuroinflammatory signaling in

AD, we performed a comprehensive multi-omics analysis of serum-derived ADEs from

AD patients and age-matched healthy controls (

Figure 8

A). Our transcriptomic analysis

revealed that ADEs from AD patients exhibit a distinct neurotoxic molecular signature,

characterized by the significant upregulation of neurotoxic factors, including

Lcn2

and

Hmgb1

, alongside a marked downregulation of protective factors like

S100A10

Sphk1

, and

Tgm1

, as well as neurotrophic factors such as

Fgf1

Nptx1

, and

Ntf3

(

Figure

B). This gene expression pattern closely mirrors the profile of neurotoxic reactive

astrocytes observed

in vitro

. Further analysis using the KEGG pathway framework

confirmed that ADEs from AD patients are enriched in pro-inflammatory pathways,

such as TNF, mTOR, and PI3K-AKT signaling (

Figure 8

C), while concurrently

showing suppression of protective pathways, including JAK-STAT and neurotrophin

signaling (

Figure 8

D). Mass spectrometry analysis corroborated these findings by

identifying elevated levels of neurotoxic proteins like LCN2 and B2M, alongside

reduced levels of neurotrophic factors, such as THBS1(

Figure 8

E-8G). Gene Set

Enrichment Analysis (GSEA) highlighted significant enrichment in pro-inflammatory

pathways, including NF-κB and IL-17 signaling (

Figure 8

H and 8I). Notably,

immunoblotting and single-molecule array quantification indicated elevated levels of

LCN2 (

Figure 8

J), p-Tau181 (

Figure 8

K), and p-Tau217

(Figure 8

L) in AD-ADEs.

Moreover, full-length GSDME protein was detected exclusively in AD-ADEs for the

first time (

Figure 8

M). Lipidomics analysis further showed an enrichment of long-chain

saturated fatty acids, such as C16:0 and C18:0, in AD-ADEs (

Figure 8

N). Taken

together, these cross-omics data outline a neurotoxic molecular signature in AD-ADEs,

characterized by the synergistic presence of pathogenic proteins, pro-inflammatory

signaling, and disrupted lipid metabolism, which may contribute to the propagation of

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

neurodegeneration via neurotoxic and inflammatory amplification.

To functionally validate the impact of AD-ADEs on neuroinflammatory cascades,

we performed a series of in vitro and in vivo assays.

In vitro

stimulation of human

microglial cell lines with AD-ADEs induced the transcriptional upregulation of pro-

inflammatory mediators

, including

C1q

Tnfα

, and

Cxcl12

, alongside the suppression

of anti-inflammatory regulators such as

NFKBIE

Sirt2

, and

HDAC1

, suggesting the

activation of microglia (

Figure 8

O). In parallel, AD-ADEs triggered a neurotoxic

reactive state in human primary astrocytes, illustrating their ability to transmit

neuroinflammatory signals across glial populations (

Figure 8

P). Moreover, AD-ADEs

significantly increased neuronal mortality compared to control exosomes (NC-ADEs),

confirming their direct neurotoxic effects (

Figure 8

Q and 8R). Considering the

peripheral origin of serum ADEs, we also investigated their systemic

immunomodulatory effects. Intravenous administration of AD-ADEs in mice resulted

in a specific reduction in platelet and neutrophil counts (

Figure 8

S-8U), highlighting

their potential to modulate peripheral immune responses. Collectively, our findings

suggest that AD-ADEs orchestrate a dual pathological mechanism, propagating

neuroinflammation and neurotoxicity within the central nervous system, while

exacerbating systemic pathology through peripheral immune dysregulation.

DISCUSSION

Our study redefines GSDME as a calcium-dependent molecular rheostat that

orchestrates astrocyte and ADE state transitions through spatiotemporal regulation of

ER calcium flux at MAMs. Unlike its canonical roles in anti-infection and tumor

immunity

, GSDME-NTs localize to MAMs, where they induce ER calcium leakage

without plasma membrane pore formation. This subcellular activity drives biphasic

CaMKIIα phosphorylation dynamics

that early-phase activation enhances NRF2-

mediated defenses (neuroprotective state), while sustained phosphorylation

paradoxically amplifies NF-

κB

-driven neuroinflammation (neurotoxic state). These

findings align with the prevailing consensus that astrocytes can adopt diverse substates

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

in response to various pathological stimuli, extending beyond the simplistic A1/A2 or

neurotoxic/neuroprotective framework.

Notably, genetic ablation of GSDME

stabilizes the neuroprotective phenotype, as evidenced by reduced expression of

neurotoxic markers, such as C3 and LCN2, elevated expression of neurotoxic markers,

such as EMP1, SPHK1 and GPC4 levels, while rescuing neuronal viability and

cognitive deficits in 5×FAD mice. This mechanistic cascade positions GSDME as a

druggable regulator of astrocyte plasticity, which provides a mechanistic foundation for

precision therapeutics targeting glial state transitions in AD.

Recent studies have demonstrated that GSDM-NTs target not only the plasma

membranes but also diverse organelle membranes.

64-66

Here, our work demonstrates

that GSDME-NT is preferentially localized at MAMs in neurotoxic astrocytes. MAMs,

lipid rafts in the endoplasmic reticulum near mitochondria, are crucial for cell health by

regulating intracellular Ca

transfer

67,68

They regulate Ca

release from the ER via

IP3Rs and its entry into mitochondria through VDACs, while also balancing cytosolic

with the MCU complex for uptake and NCLX for release.

Although dysfunctions

in MAMs are linked to various pathological conditions and neurodegenerative

diseases,

69,70

their precise role in AD pathogenesis remains inadequately understood.

The targeted relocation of GSDME-NTs to MAMs results in ER calcium leakage, as

opposed to forming pores in the plasma membrane. Thus, we hypothesize that GSDME-

NTs permeabilize MAMs in neurotoxic astrocytes, thereby facilitating the transfer of

from the ER to both the cytosol and mitochondria.

This specific localization

mechanism elucidates why the activation of GSDME as a cytosolic calcium channel

does not trigger pyroptosis but instead promotes transcriptomic alterations by

modulating calcium signaling pathways. This subcellular targeting may constitute a

common characteristic within the GSDM family, offering a novel framework for

understanding the diversity of their pathological roles.

Astrocytic calcium dysregulation, a hallmark of early AD, leads to synaptic issues

and cognitive decline through reactive astrogliosis

17,71

Although normalizing Ca²⁺

transients can reverse cognitive deficits in AD mice,

72,73

the specific mechanisms

connecting altered calcium signaling to astrocyte-induced neurotoxicity are not yet

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

understood. Here, we elucidate a temporal rheostat mechanism wherein calcium-

dependent phosphorylation dynamics of CaMKIIα govern the transition from

neuroprotective to neurotoxic through phase-specific transcriptional reprogramming.

During the early phase, CaMKIIα phosphorylation enhances NRF2-mediated

antioxidant defenses, thereby promoting neuroprotection. Conversely, prolonged

phosphorylation paradoxically intensifies NF-κB-driven neuroinflammation,

converting astrocytes into a neurotoxic state. This biphasic regulation challenges the

traditional view of CaMKIIα as a solely neuroprotective kinase

and underscores the

importance of timing in therapeutic interventions.

GSDME is identified as a crucial

temporal checkpoint, its activation triggers ER Ca²⁺ leakage, initiating NF-κB signaling.

Interestingly, GSDME knockdown selectively inhibits late-phase CaMKIIα

phosphorylation, thereby preventing NF-κB hyperactivation and subsequent

neurotoxicity.

These findings align with broader calcium signaling dynamics, where

transient vs. sustained calcium signals differentially regulate synaptic plasticity and

cellular homeostasis. Consistently, pharmacological testing with antagonists shows a

time-dependent dual effect that early CaMKIIα activation is protective, whereas late

inactivation hampers toxicity, indicating metabolic feedback with distinct

neuroprotective outcomes. Taken together, our work reconceptualizes astrogliosis as a

calcium-regulated process orchestrated by the dual functions of CaMKIIα. By

modulating its activity in a phase-specific manner, we propose a “late inhibition”

strategy aimed at mitigating pathogenic astrogliosis through targeted p-CaMKIIα. This

methodology integrates AD pathomechanisms with therapeutic development,

presenting a precision neuroprotection framework for early intervention in

neurodegenerative disorders.

A growing body of evidence positions exosomes as critical mediators of

neuropathological progression, with ADE serving as potent coordinators of

multicellular neurodegeneration.

74-77

Unlike prior models attributing astrocyte-

mediated neuronal damage to direct cytokine, toxic lipid release or oxidative

stress,

56,78,79

our findings reveal that ADEs function as context-sensitive pathogenic

platforms whose cargo dynamically adapts to neurotoxic microenvironments. Multi-

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

omics profiling demonstrates that ADEs from reactive astrocytes exhibit enrichment in

neurotoxic lipids, pro-apoptotic miRNAs, and pathogenic tau isoforms relative to

homeostatic astrocytes, while co-encapsulating mRNA cargo that coordinately disrupts

synaptic architecture, astrocyte to pro-inflammatory microglial activation, and

astrocyte-to-astrocyte reactive phenotype propagation. This suggests a tripartite

mechanism unprecedented in classical single-pathway neurotoxicity models. Central to

this paradigm shift is the identification of GSDME as a molecular switch regulating

ADE pathogenicity. Genetic ablation of GSDME prevents reactive astrocyte transition

from neuroprotective to neurotoxic state, while reprogramming ADE cargo toward

neuroprotective mediators, increasing BDNF,

GPC4 levels

and enhancing anti-

inflammatory miRNAs. These findings redefine GSDME as a master regulator of

astrocyte exosome biology, further extending its canonical role in pyroptosis to include

ADE-mediated intercellular communication, and position ADEs as critical mediators

of neuroinflammation propagation.

While early intervention remains pivotal for AD therapeutics,

current limited

glial markers, including serum GFAP, lack resolution to decode the functional

heterogeneity of reactive astrocytes driving AD progression.

11,24

Here, we pioneer a

multi-dimensional serum ADE biomarker platform that integrates transcriptomic,

lipidomic, and proteomic profiling to transcend static tissue paradigms. This strategy

identifies an AD-specific molecular fingerprint, including coordinated depletion of

neuroprotective transcripts, amplification of pro-degenerative effectors, and lipidomic

dysregulation marked by C16:0, C18:0 elevation. Unlike conventional methods, this

cross-omics framework captures dynamic astrocyte state transitions through serum

ADE-mediated intercellular networks, enabling presymptomatic stratification with

mechanistic granularity. By bridging real-time molecular pathology and therapeutic

targeting, ADE multi-omics repositions biomarker discovery as a mechanism-to-

therapeutics pipeline, where early diagnosis directly informs astrocyte-modulating

precision strategies.

In summary, our work redefines astrocyte activation as a calcium-regulated

continuum. The discovery of GSDME-NTs localized at MAM and its dual regulation

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

of cell-autonomous signaling via CaMKIIα and non-cell-autonomous communication

through ADEs provides a unified framework for understanding glial-driven

neurodegeneration. This paradigm shift integrates calcium dynamics, transcriptional

reprogramming, and intercellular communication, offering novel insights into AD

pathogenesis. Future studies should explore the conservation of GSDME-MAM

interactions in human AD progression and investigate the therapeutic potential of

combinatorial strategies targeting both calcium signaling and ADE-mediated

inflammation.

Limitations of the study

Although our study establishes GSDME as a critical regulator of astrocyte state

transitions, several limitations warrant further investigation. First, the molecular

mechanisms driving GSDME-NT’s preferential localization to MAMs remain

incompletely defined; whether specific lipid modifications or chaperones mediate this

subcellular targeting requires systematic interrogation. Second, the therapeutic

validation is restricted to Aβ pathology models, lacking examination in Tau-inclusive

models like 3×Tg, which may limit generalizability to AD with mixed pathologies.

Third, our ADE multi-omics approach, while identifying disease-specific signatures,

does not resolve heterogeneity among ADE subpopulations, which may obscure

functionally distinct cargo subsets. Single-vesicle profiling and spatial transcriptomics

may address this limitation. Fourth, the long-term consequences of GSDME ablation

on physiological processes remain unexplored, raising potential safety concerns for

therapeutic targeting. Finally, human tissue validation of GSDME associations is

limited to postmortem AD samples; future studies should employ PET-MRI imaging to

correlate GSDME activity with real-time neuroinflammation in living patients.

Addressing these gaps will refine our understanding of the spatiotemporal regulation of

GSDME and accelerate translation of astrocyte-targeted therapies.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Acknowledgments:

We thank Dr. M. Xu from the Center for Brain-Inspired

Intelligence at the Chinese Academy of Sciences for the comments.

We thank Dr. J.

Xiang and Dr. Y.Z. Wang from the Air Force Medical University for the helpful advices.

We thank the analysis & testing laboratory for life sciences and medicine of the Air

Force Medical University for their assistance with electron microscopy imaging, as

approved by Y. Zhao, and structured illumination microscopy imaging, as approved by

D.L. Shi. We thank the

Chinese Brain Bank Center for providing the human brain

samples.

Funding:

Funding was obtained from the National Key Research and Development

Plan (No.2022YFC3400103 to Q.L. and X.B.H.), the ECUST-OPM Open Fund (No.

20220701 to Q.L.).

Author contributions:

Q. L. and S.W. C., X.B. H. conceived the project; Q. L., S.W.

C., X.M. X designed the experiments and wrote the manuscript. Q. L., S.W. C., X.B.

H., X.M. X , J.M. X., X.Q. L., X.Y. L., X.M. W. and Y.X. Z. revised the manuscript.

S.W. C., J.M. X., X.Q. L., X.Y. L. and X.M. W., S.X. L. and Y.X. G. performed the

experiments with the assistance from Y.Z. C., G.G. G.; and Q. L.,

S.W. C. and X.M. X,

J.M. X., X.Q. L., X.Y. L., X.M. W. analyzed the data and created the figures. All

authors discussed the results and commented on the manuscript.

Competing interests:

The authors declare that they have no competing interests.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

FIGURES

Figure 1. Pathological GSDME activation in reactive astrocytes is preserved across

Alzheimer’s disease patients and 5×FAD mouse models

(A) Comparison of GSDME expression in astrocytic clusters between disease (Dg) and

healthy (Hg) groups was conducted based on open single-nucleus RNA sequencing

(snRNA-seq) datasets of AD (Case 1

= 8,

= 8; Case 2

= 7,

= 27;

Case 3

= 8,

= 8; Case 4

= 16,

= 16).

(B and C) Immunoblot analysis of the activation of GSDME, caspase-3, as well as the

expression of NeuN and Synaptophysin (SYN) in hippocampus tissue lysates collected

from AD patients and healthy controls (hippocampus,

= 3,

= 3), measured by

Image J based on gray scale.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

(D and E) Representative IF images of hippocampus region from AD patients and

healthy controls (hippocampus,

= 3,

= 3) co-stained with anti-GSDME-NT, anti-

GFAP and anti-C3 antibody. Scale bars, 100 μm. Percentage of GSDME-NT and C3-

double positive cells within GFAP-positive cell populations was quantified using Image

J software. Each point means two sections.

(F-I) Immunoblot analysis of GSDME activation in cortex (F and G) and hippocampus

(G and H) tissue lysates collected from 5×FAD and wild type (WT) mice. Regression

analysis of GSDME-NT expression dynamics during aging in 5×FAD mice. GSDME-

NT levels were quantified by ImageJ (

= 3 for each group).

(J and K) Representative IF images of cortex and hippocampus region from 5×FAD (

= 3) and WT mice (

= 3) co-stained with anti-GSDME-NT, anti-GFAP and anti-6E10

antibody (J). Scale bars, 30 μm. Percentage of GSDME-NT within GFAP-positive cell

populations (K) was quantified by Image J.

(L-M) Immunoblot analysis of GSDME-NT, GFAP, as well as the expression of NeuN

and Iba1 in cortex tissue homogenates, astrocyte-enriched fractions, and non-astrocytic

fractions isolated from 6-month-old 5×FAD mice, measured by Image J (

= 3 for each

fraction).

Statistical significance determined using a two-tailed Student’s

test (C, E, G, I, K and

M), or two-sided Wilcoxon signed-rank test (A). All values are represented as mean ±

SEM, *

< 0.05; **

< 0.01; ***

< 0.001; ****

< 0.0001. See also

Figure S1

and

Table S1

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Image J (E) and representative IF images were shown in (F). GSDME-NT (anti

GSDME-NT; green) in primary astrocytes (anti-GFAP; red) 48 h post MCM or AMCM

treatment. Scale bars, 20 μm. Arrows indicate the GSDME-NT

astrocytes (

= 6

replicates per group).

(G and H) Quantification of holes (G) on the primary astrocytic cell membrane via SEM

at 48 h post MCM or AMCM treatment. Astrocytes overexpressing GSDME-NT were

used as positive controls. Representative SEM images are shown in (H). Scale bars, 3

μm. Arrows indicate the holes.

(I and J) Representative images of WT or GSDME KO (knock-out) primary astrocytes

treated with MCM or AMCM as indicated, showing the engulfment of pHrodo

conjugated synaptosomes (I) and the extent of engulfing pHrodo-conjugated

synaptosomes was calculated with Image J (J). The inserts show the engulfed particles

(red). Scale bars, 10 μm.

(K) Relative mRNA expression of phagocytosis-related proteins (

Axl

Gas6

Megf10

Mertk

), synaptogenic factors (

Sparc

Sparcl1

GPC4

GPC6

Thbs1

), quantified by qRT-

PCR (

= 3 replicates per group).

(L and M) Representative live images of primary cortical neuronal death induced by

ACM, analyzed with the Calcein AM/PI staining (L) and the ratio of neuronal death

was calculated by Image J (M) (

= 3 replicates per group). Scale bars, 20

Statistical significance determined using a two-tailed Student’ s

test. All values are

represented as mean ± SEM, *

< 0.05; **

< 0.01; ****

< 0.0001. See also

Figure

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Enlarged images of the boxed area are shown (

= 2 replicates, at least 8 cells were

calculated each group per replicate). Scale bars, 5 μm.

(D and E) SIM images stained for GSDME-NT and ACSL-4 at 48 hps (D).

Quantification of Pearson’s correlation between GSDME-NT and ACSL-4 (E).

Enlarged images of the boxed area are shown (

= 2 replicates, at least 8 cells were

calculated each group per replicate). Scale bars, 10 μm.

(F-H) TEM images stained with GSDME-NT-immunogold (F). Yellow dashed lines

outline mitochondria, arrows indicate the GSDME-NT-immunogold. Scale bars, 200

nm. Quantification of gold particles in WT-MCM astrocytes versus WT-AMCM

astrocytes was shown in (G). The relative subcellular distribution of gold particles in

WT-AMCM astrocytes, including cytosol, mitochondria, MAMs were further

quantified in (H) (

= 3 replicates, 10-20 cells each group per replicate).

(I-K) Representative IF image stained with Fluo8-AM was shown in (I). Quantification

of Fluo8 AM fluorescence intensity to indicate the cytosolic Ca

was shown in (J) (

= 8 replicates per group). Scale bars, 50 μm. Relative mRNA expression of endogenous

cytosolic Ca

sensor, Calm3, quantified by qRT-PCR was shown in (K) (

= 3

replicates per group).

(L-N) Quantification of relative fluorescence intensity over time in WT-MCM, WT-

AMCM or GSDME KO-AMCM astrocytes by the genetically encoded Ca

indicators,

GCaMP6s (L) for the cytosol, CEP1AE-ER (M) for the ER, and CEPIA4-Mito (N) for

the mitochondria.

(O-Q) Representative IF image stained with C3 (grey) in primary mouse astrocytes

(anti-GFAP; red) was shown in (O), quantified by qRT-PCR was shown in (P) (

= 4

replicates per group). Relative mRNA expression of reactive astrocyte marker

(

Serping1

H2-D1

Srgn

H2-T23

Ggta1

Ligp1

Psmb8

LCN2

), quantified by qRT-

PCR (

= 3 replicates per group).

Statistical significance determined using a two-tailed Student’ s

test. All values are

represented as mean ± SEM, **

< 0.01; ***

< 0.001; ****

< 0.0001. See also

Figure

and

Movie S1

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Significantly upregulated (red, |log

FC| > 1, FDR < 0.05) and downregulated (blue)

genes are annotated.

Camk2α

(asterisk) is highlighted as a top DEG.

(D) Heatmap analysis of neuroprotective and neurotrophic gene expression by qPCR

in astrocytes transfected with CaMKIIα-targeting siRNA (siCamk2α) or non-targeting

scrambled siRNA (siNC), followed by 24-h AMCM stimulation. Data are normalized

Gapdh

and presented as fold change relative to siNC.

(E-G) Time-dependent immunoblot analysis of GSDME-FL and GSDME-NT,

neurotoxic (C3, LCN2), neuroprotective (S100A10, SPHK1), and neurotrophic (GPC4)

markers in astrocytes transfected with non-targeting scrambled siRNA (siNT) following

AMCM treatment for 48 h (E). Quantification of grayscale values (ImageJ) and

temporal trends (GraphPad Prism fit spline analysis) are shown in (F). Phosphorylation

kinetics (G) of CaMKIIα, p65, and NRF2 in (E).

Quantification of grayscale values

(ImageJ) and temporal trends (GraphPad Prism fit spline analysis) are shown in (H).

(I-L) Time-dependent immunoblot of GSDME-FL/NT, neurotoxic/neuroprotective

markers, and GPC4 (I and J) and phosphorylation kinetics (L and M) in astrocytes

transfected with siRNA targeted GSDME (siGE) under identical conditions in (E).

Quantification and trend fitting are shown in (J and L).

(M-P

)

Time-dependent immunoblot of GSDME-FL/NT, neurotoxic/neuroprotective

markers, and GPC4 (M and N) and phosphorylation kinetics (O and P) in astrocytes

treated with CaMKⅡα-IN-1at 8 hps with AMCM. Quantification and trend fitting are

shown in (O and Q).

Statistical significance determined using a two-tailed Student’s

test. All values are

represented as mean ± SEM with at least 3 biological replicates. Each point represents

one replicate. See also

Figure S4.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

the treatment with WT-ACM, aACM (Whole), aACM-isolated exosomes (Exo) and

Exo-free aACM (Exo-free) (

= 3 replicates per group).

(E) Quantification of the neuronal cell survival rate of mouse cortical neurons 24 h after

the treatment of ACM-isolated exosomes from WT-ACM, GSDME KO-ACM, WT-

aACM and GSDME KO-aACM astrocytes supernatant (

= 3 replicates per group).

(F and G) Representative IF images of neurons, as treated in (E), stained for both MAP2

and SYN (F). Enlarged images of the boxed area are shown (

= 2 replicates, at least

10 cells were calculated each group per replicate). The number of puncta per 200

as in (D) is calculated(G). Scale bars, 10 μm.

(H and I) Immunoblotting of equal amounts of exosomes with antibodies, C3, LCN2,

HMGB1, iNOS, GPC4 and BDNF, controlled by HSP70 (H), measured by ImageJ

based on gray scale (I).

(J) Quantification of saturated lipids in equal amounts of exosomes, as treated in (E),

by liquid chromatography-mass spectrometry (LC-MS). FFA, free fatty acid (

= 4

replicates per group).

(K) Quantification of pro- and anti-inflammatory miRNAs in exosomes, as treated in

(E), by RT-qPCR (

= 3 replicates per group).

(L) Relative expression of neurotoxic, neuroprotective and neurotrophic genes in

exosomes, as treated in (E), by RT-qPCR (

= 3 replicates per group).

(M)

Schematic model of ADE-mediated neuroinflammatory propagation across

neuronal-glial networks.

(N and O) Heatmap depicting relative expression of pro-inflammatory genes in BV2

cells treated with different exosomes by transcriptional sequencing (

= 3 replicates

per group) (N) . Analysis of upregulated transcriptome signal pathway by GO pathway

between BV2 cells treated with aADEs vs ADEs (O).

(P) Relative expression of neurotoxic, neuroprotective and neurotrophic genes in

primary astrocytes treated with different exosomes by RT-qPCR (

= 3 replicates per

group).

(Q) Quantification of neuronal death induce by

supernatant from WT-ADEs, WT-

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Figure 6.

Inhibition of Astrocytic GSDME Activation Attenuates Pathological

Progression in 5×FAD Mice

(A)

Experimental timeline and group design. Neonatal WT and 5×FAD mice were

injected with AAV.PHP.eB-hGFAPS-GFP expressing either a scrambled control shRNA

(shNT) or GSDME-targeting shRNA (shGSDME, shGE).

Behavioral and pathological

assessments were performed at indicated time points. (

＞

10 mice per group).

(B and C) Representative confocal images showing GFP

cells (green) colocalized with

NeuN, Iba1 or GFAP positive cells (Red) in the cortex of shNT-WT mice (B).

Quantification of the ratio of GFAP

⁺

GFP

⁺

, Iba1

⁺

GFP

⁺

, and NeuN

⁺

GFP

⁺

double

positive cells within total GFP

⁺

cells in the cortex (

= 3 replicates per group) (C).

Scale

bar, 50 μm.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

(D and E) Representative immunoblots of cortical tissues from AAV-treated mice

showing protein expression levels of GSDME-FL, GSDME-N, Aβ

1-42

, synaptic

markers (SYP and PSD-95), neuronal nuclei marker (NeuN), with GAPDH as loading

control (D). Quantitative analysis of protein bands by ImageJ: GSDME isoforms

(FL/NT), Aβ

1-42

, synaptic proteins (SYP/PSD-95), and NeuN. Data normalized to

GAPDH with background subtraction (

= 4 mice per group) (E).

(D and E) Immunoblot analysis of the activation of GSDME, A

1-42

, SYP, PSD-95 and

NeuN in cortical tissue lysates collected from AAV-treated mice (

= 4 replicates per

group) (D) , (E) measured by Image J based on gray scale.

(F and G) Representative images of QM-FN-SO3 staining of Aβ plaques in coronal

whole-brain sections of shGE and shNT-5

FAD mice (F) and quantification of plaque

burden (G). Scale bar, 1000 μm (

= 4 replicates per group) .

(H and I) Representative IF images of sections collected from shGE and shNT-5

FAD

mice co-stained with anti-NeuN, anti-GFAP antibodies (H). Number of NeuN-positive

cells was quantified using Image J (I).Scale bar, 1000 μm (

= 3 replicates per group).

(J and K) Open-field test for assessing anxiety-like behavior and locomotor activity.

Times in the centre zone (J) and tolal distance traveled (K) (

＞

10 mice per group).

(L and M) Y maze test on the same groups of mice. Spontaneous alternation rate,

calculated using the standard method, represent working memory capacity and

cognitive flexibility (L).Total entries into all three arms reflect baseline locomotor

activity (M) (

＞

10 mice per group).

(N and O) Novel object recognition (NOR) test on the same groups of mice. Object

recognition index (ORI), reflecting short-term memory performance (N).Total distance

traveled during the test session, indicating baseline locomotor activity (O) (

＞

10 mice

per group).

Statistical significance determined using a two-tailed Student’s

test. All values are

represented as mean ± SEM, *

< 0.05; **

< 0.01; ***

< 0.001; ****

< 0.0001.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

p65 in cortex tissue lysates collected from shGE and shNT 5

FAD mice (C),

(D)measured by Image J based on gray scale (

= 3 replicates per group).

(E) Representative RNA-seq volcano plot of cortical differentially expressed genes

(DEGs) between shGE-5

FAD and shNT-5

FAD groups; Red and blue dots indicate

significantly upregulated and downregulated genes respectively (|log2FC| > 0.585,

FDR < 0.05). Vertical dashed lines mark fold-change thresholds.

(F) Cortical pathway enrichment analysis of shGE-5

FAD vs shNT-5

FAD groups by

GSEA; Significantly altered pathways are displayed according to normalized

enrichment score (NES) absolute values. Horizontal dashed line indicates significance

threshold.

(G) Hippocampal qPCR validation of dysregulated genes; Heatmap showing relative

mRNA expression levels of

Lcn2

Emp1

, an

d Sphk1

in shGE-5

FAD, shNT-5

FAD, and shNT-WT groups (

= 4). Data normalized to

Gapdh

and expressed as log

fold changes relative to shNT-WT controls (2^

ΔΔ

method).

(H and I) Representative IF image of sections collected from shGE- and shNT-5

FAD

mice cortex co-stained with anti-GFAP , anti-C3 antibodies (H). Percentage of C3-

positive cells within GFAP

cell populations were quantified by Image J (n

replicates per group) (I).

(J and K) Representative IF image of sections collected from shGE- and shNT-5

FAD

mice cortex co-stained with anti-GFAP , anti-S100A10 antibodies (J). Percentage of

S100A10-positive cells within GFAP

cell populations were quantified by Image J (

= 4 replicates per group) (K).

(L and M) Representative IF image of sections collected from shGE- and shNT-5

FAD mice cortex co-stained with anti-BDNF antibody (L). Relative expression level of

Iba1 was quantified by Image J (

= 4 replicates per group) (M).

(N and O) Representative IF image of sections collected from shGE- and shNT-5

FAD

mice cortex co-stained with anti-Iba1 antibody (N). Relative expression level of Iba1

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

was quantified by Image J (

= 4 replicates per group) (O).

(P) Schematic workflow of astrocyte-derived exosomes (ADEs) isolation from mouse

serum via orbital venous plexus blood collection.

(Q and R) Immunoblot analysis of C3, GPC4, BDNF and CD81 in C3 in ADEs

collected from shNT-WT, shGE- and shNT-5

FAD mice (Q), (R) measured by Image

J based on gray scale (

= 3 replicates per group).

(S) Transcriptomic heatmap of ADE-derived differentially expressed genes (DEGs)

across shNT-WT, shGE and shNT-5

FAD groups (

= 3 replicates per group).

(T) Lipidomic heatmap of ADE-losded long-chain fatty acid profiles from the same

experimental groups (

= 3 replicates per group).

Statistical significance determined using a two-tailed Student’s

test. All values are

represented as mean ± SEM, *

< 0.05; **

< 0.01; ***

< 0.001. See also

Figure S6

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

factors from AD- ADEs compared to NC

ADEs (

= 6 replicates per group).

(F and G) Mass spectrometry reveals co-upregulation and co-downregulation of

differentially expressed genes (DEGs) and proteins (DEPs) between NC and AD.

(H and I) Gene Set Enrichment Analysis (GESA) showing enrichment in pro-

inflammatory pathways of AD-ADEs.

(J-M) Immunoblotting and simoa quantification showing elevated LCN2 (J), p-Tau181

(K), p-Tau217 (L) and GSDME-FL (M) in AD-ADEs

(

13 replicates per group).

(N) Lipidomic analysis showing enrichment of long-chain saturated fatty acids in AD-

ADEs (

6 replicates per group).

(O and P) Transcriptome heatmap of human glial cells simulated by serum AD-ADEs

compared to NC-ADEs showing activated of microglia (O) and activated astrocytes (P)

in vitro

(Q and R) Representative images of primary neuron stimulated with AD-ADEs (Q) and

neuronal mortality (R) were detected with Calcein-AM and PI stain, quantified by

Image J (

= 6 replicates per group)

(S) Schematic diagram of the peripheral pro-inflammation role of AD-ADEs in a mouse

model.

(T and U) The count of blood cells in mice treated with AD-ADEs showing decreased

platelets (T) and increased neutrophils (U) 2 h post tail vein injected.

Statistical significance determined using a two-tailed Student’s

test. All values are

represented as mean ± SEM, *

< 0.05; **

< 0.01; ***

< 0.001.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

STAR Methods

Detailed methods are provided in the online version and include the following:

⚫

KEY RESOURCES TABLE

⚫

RESOURCE AVAILABILITY



Lead contact



Materials availability



Data and code availability

⚫

EXPERIMENTAL MODEL AND SUBJECT DETAILS



Animals



Cell culture and treatment

⚫

METHOD DETAILS



Single-cell transcriptome analysis using open snRNA-seq data



RNA extraction and RNA-seq analysis



Isolation of astrocytic exosomes



Nanoparticle tracking analysis (NTA)



Synaptosome engulfment assay



Neurotoxicity assay



Scanning electron microscope (SEM)



Transmission electron microscopy (TEM)



Immuno-electron microscope (IEM)



Western blotting



Immunostaining



Characterization of subcellular Ca



Saturated lipid analysis



Behavioral test



siRNA transfection



Lentivirus preparation and infection



Peripheral effect of ADVs

in vivo

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Supplementary information

Figure S1.

Spatiotemporal and disease-associated heterogeneity of GSDME expression

in astrocytes

related to

Figure 1

Figure S2.

AMCM induces the neurotoxic phenotype and GSDME activation in

primary astrocytes, related to

Figure 2

Figure S3.

Detection of Ca

dynamics, related to

Figure 3

Figure S4.

Temporal dynamics of GSDME cleavage and signaling pathways in PBS-

treated astrocytes , related to

Figure 4

Figure S5.

Uptake of astrocytic exosomes by neurons, related to

Figure 5

Figure S6

Isolation and identification of serum-derived astrocytic exosomes

related to

Figure 7

and

Figure 8

Table S1.

The characteristics of human tissue donors, related to

Figure 1

Table S2.

Primers used in this study

Movie 1.

Detection of cytosolic Ca

dynamics

in WT-MCM, WT-AMCM and

GSDME KO-AMCM astrocytes following the stimulation within 8 h,

related to

Figure

Movie 2.

Detection of ER Ca

dynamics

in WT-MCM and WT-AMCM astrocytes

following the stimulation within 8 h,

related to

Figure 3

Movie 3.

Detection of mitochondrial Ca

dynamics

in WT-MCM and WT-AMCM

astrocytes following the stimulation within 8 h,

related to

Figure 3

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

REFERENCES

Busche, M.A., and Hyman, B.T. (2020). Synergy between amyloid-beta and tau in Alzheimer's
disease. Nat Neurosci

, 1183-1193. 10.1038/s41593-020-0687-6.

Long, J.M., and Holtzman, D.M. (2019). Alzheimer Disease: An Update on Pathobiology and
Treatment Strategies. Cell

179

, 312-339. 10.1016/j.cell.2019.09.001.

Hardy, J.A., and Higgins, G.A. (1992). Alzheimer's disease: the amyloid cascade hypothesis.
Science

256

, 184-185. 10.1126/science.1566067.

Braak, H., and Del Tredici, K. (2011). Alzheimer's pathogenesis: is there neuron-to-neuron
propagation? Acta Neuropathol

121

, 589-595. 10.1007/s00401-011-0825-z.

Chetelat,  G.,  La  Joie,  R.,  Villain,  N.,  Perrotin,  A.,  de  La  Sayette,  V.,  Eustache,  F.,  and
Vandenberghe,  R.  (2013).  Amyloid  imaging  in  cognitively  normal  individuals,  at-risk
populations  and  preclinical  Alzheimer's  disease.  Neuroimage  Clin

, 356-365.

10.1016/j.nicl.2013.02.006.

Mattsson-Carlgren,  N.,  Salvado,  G., Ashton,  N.J., Tideman,  P.,  Stomrud,  E.,  Zetterberg,  H.,
Ossenkoppele,  R.,  Betthauser,  T.J.,  Cody,  K.A.,  Jonaitis,  E.M.,  et  al.  (2023).  Prediction  of
Longitudinal  Cognitive  Decline  in Preclinical Alzheimer Disease  Using  Plasma  Biomarkers.
JAMA Neurol

, 360-369. 10.1001/jamaneurol.2022.5272.

Patani, R., Hardingham, G.E., and Liddelow, S.A. (2023). Functional roles of reactive astrocytes
in neuroinflammation and neurodegeneration. Nat Rev Neurol

, 395-409. 10.1038/s41582-

023-00822-1.

Rothstein, J.D., Dykes-Hoberg, M., Pardo, C.A., Bristol, L.A., Jin, L., Kuncl, R.W., Kanai, Y.,
Hediger, M.A., Wang, Y., Schielke, J.P., and Welty, D.F. (1996). Knockout of glutamate
transporters reveals a major role for astroglial transport in excitotoxicity and clearance of
glutamate. Neuron

, 675-686. 10.1016/s0896-6273(00)80086-0.

Bellaver, B., Povala, G., Ferreira, P.C.L., Ferrari-Souza, J.P., Leffa, D.T., Lussier, F.Z., Benedet,
A.L., Ashton, N.J., Triana-Baltzer, G., Kolb, H.C., et al. (2023). Astrocyte reactivity influences
amyloid-beta effects on tau pathology in preclinical Alzheimer's disease. Nat Med

, 1775-

1781. 10.1038/s41591-023-02380-x.

10.

Escartin, C., Galea, E., Lakatos, A., O'Callaghan, J.P., Petzold, G.C., Serrano-Pozo, A.,
Steinhauser, C., Volterra, A., Carmignoto, G., Agarwal, A., et al. (2021). Reactive astrocyte
nomenclature, definitions, and future directions. Nat Neurosci

, 312-325. 10.1038/s41593-

020-00783-4.

11.

Pereira, J.B., Janelidze, S., Smith, R., Mattsson-Carlgren, N., Palmqvist, S., Teunissen, C.E.,
Zetterberg, H., Stomrud, E., Ashton, N.J., Blennow, K., and Hansson, O. (2021). Plasma GFAP
is an early marker of amyloid-beta but not tau pathology in Alzheimer's disease. Brain

144

3505-3516. 10.1093/brain/awab223.

12.

De Strooper, B., and Karran, E. (2016). The Cellular Phase of Alzheimer's Disease. Cell

164

603-615. 10.1016/j.cell.2015.12.056.

13.

van Dyck, C.H., Sabbagh, M., and Cohen, S. (2023). Lecanemab in Early Alzheimer's Disease.
Reply. N Engl J Med

388

, 1631-1632. 10.1056/NEJMc2301380.

14.

Henstridge, C.M., Hyman, B.T., and Spires-Jones, T.L. (2019). Beyond the neuron-cellular

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

interactions early in Alzheimer disease pathogenesis. Nat Rev Neurosci

, 94-108.

10.1038/s41583-018-0113-1.

15.

Brandebura, A.N., Paumier, A., Onur, T.S., and Allen, N.J. (2023). Astrocyte contribution to
dysfunction, risk and progression in neurodegenerative disorders. Nat Rev Neurosci

, 23-39.

10.1038/s41583-022-00641-1.

16.

Verkhratsky, A., and Nedergaard, M. (2018). Physiology of Astroglia. Physiol Rev

, 239-389.

10.1152/physrev.00042.2016.

17.

Shah, D., Gsell, W., Wahis, J., Luckett, E.S., Jamoulle, T., Vermaercke, B., Preman, P., Moechars,

D., Hendrickx, V., Jaspers, T., et al. (2022). Astrocyte calcium dysfunction causes early network
hyperactivity in Alzheimer's disease. Cell Rep

, 111280. 10.1016/j.celrep.2022.111280.

18.

Takano, T., Tian, G.F., Peng, W., Lou, N., Libionka, W., Han, X., and Nedergaard, M. (2006).
Astrocyte-mediated control of cerebral blood flow. Nat Neurosci

, 260-267. 10.1038/nn1623.

19.

Liddelow, S.A., Guttenplan, K.A., Clarke, L.E., Bennett, F.C., Bohlen, C.J., Schirmer, L.,
Bennett, M.L., Munch, A.E., Chung, W.S., Peterson, T.C., et al. (2017). Neurotoxic reactive
astrocytes are induced by activated microglia. Nature

541

, 481-487. 10.1038/nature21029.

20.

Reichenbach, N., Delekate, A., Breithausen, B., Keppler, K., Poll, S., Schulte, T., Peter, J.,
Plescher, M., Hansen, J.N., Blank, N., et al. (2018). P2Y1 receptor blockade normalizes network
dysfunction and cognition in an Alzheimer's disease model. J Exp Med

215

, 1649-1663.

10.1084/jem.20171487.

21.

Hasel, P., Rose, I.V.L., Sadick, J.S., Kim, R.D., and Liddelow, S.A. (2021). Neuroinflammatory
astrocyte subtypes in the mouse brain. Nat Neurosci

, 1475-1487. 10.1038/s41593-021-

00905-6.

22.

Chen, W.T., Lu, A., Craessaerts, K., Pavie, B., Sala Frigerio, C., Corthout, N., Qian, X.,
Lalakova, J., Kuhnemund, M., Voytyuk, I., et al. (2020). Spatial Transcriptomics and In Situ
Sequencing to Study Alzheimer's Disease. Cell

182

, 976-991 e919. 10.1016/j.cell.2020.06.038.

23.

Zamanian, J.L., Xu, L., Foo, L.C., Nouri, N., Zhou, L., Giffard, R.G., and Barres, B.A. (2012).
Genomic

analysis

reactive

astrogliosis.

Neurosci

6391-6410.

10.1523/JNEUROSCI.6221-11.2012.

24.

Habib, N., McCabe, C., Medina, S., Varshavsky, M., Kitsberg, D., Dvir-Szternfeld, R., Green,
G., Dionne, D., Nguyen, L., Marshall, J.L., et al. (2020). Disease-associated astrocytes in
Alzheimer's disease and aging. Nat Neurosci

, 701-706. 10.1038/s41593-020-0624-8.

25.

Rothhammer, V., Borucki, D.M., Tjon, E.C., Takenaka, M.C., Chao, C.C., Ardura-Fabregat, A.,
de Lima, K.A., Gutierrez-Vazquez, C., Hewson, P., Staszewski, O., et al. (2018). Microglial
control of astrocytes in response to microbial metabolites. Nature

557

, 724-728.

10.1038/s41586-018-0119-x.

26.

Wheeler, M.A., Clark, I.C., Tjon, E.C., Li, Z., Zandee, S.E.J., Couturier, C.P., Watson, B.R.,
Scalisi, G., Alkwai, S., Rothhammer, V., et al. (2020). MAFG-driven astrocytes promote CNS
inflammation. Nature

578

, 593-599. 10.1038/s41586-020-1999-0.

27.

Fu, J., Schroder, K., and Wu, H. (2024). Mechanistic insights from inflammasome structures.
Nat Rev Immunol

, 518-535. 10.1038/s41577-024-00995-w.

28.

Neel, D.V., Basu, H., Gunner, G., Bergstresser, M.D., Giadone, R.M., Chung, H., Miao, R.,
Chou, V., Brody, E., Jiang, X., et al. (2023). Gasdermin-E mediates mitochondrial damage in
axons and neurodegeneration. Neuron

111

, 1222-1240 e1229. 10.1016/j.neuron.2023.02.019.

29.

Pollock, N.M., Fernandes, J.P., Woodfield, J., Moussa, E., Hlavay, B., Branton, W.G., Wuest,

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

M., Mohammadzadeh, N., Schmitt, L., Plemel, J.R., et al. (2024). Gasdermin D activation in
oligodendrocytes and microglia drives inflammatory demyelination in progressive multiple
sclerosis. Brain Behav Immun

115

, 374-393. 10.1016/j.bbi.2023.10.022.

30.

Fagerberg, L., Hallstrom, B.M., Oksvold, P., Kampf, C., Djureinovic, D., Odeberg, J., Habuka,
M., Tahmasebpoor, S., Danielsson, A., Edlund, K., et al. (2014). Analysis of the human tissue-
specific expression by genome-wide integration of transcriptomics and antibody-based
proteomics. Mol Cell Proteomics

, 397-406. 10.1074/mcp.M113.035600.

31.

Yue, F., Cheng, Y., Breschi, A., Vierstra, J., Wu, W., Ryba, T., Sandstrom, R., Ma, Z., Davis, C.,
Pope, B.D., et al. (2014). A comparative encyclopedia of DNA elements in the mouse genome.
Nature

515

, 355-364. 10.1038/nature13992.

32.

Han, X., Wang, R., Zhou, Y., Fei, L., Sun, H., Lai, S., Saadatpour, A., Zhou, Z., Chen, H., Ye,
F., et al. (2018). Mapping the Mouse Cell Atlas by Microwell-Seq. Cell

173

, 1307.

10.1016/j.cell.2018.05.012.

33.

Batiuk, M.Y., Martirosyan, A., Wahis, J., de Vin, F., Marneffe, C., Kusserow, C., Koeppen, J.,
Viana, J.F., Oliveira, J.F., Voet, T., et al. (2020). Identification of region-specific astrocyte
subtypes at single cell resolution. Nat Commun

, 1220. 10.1038/s41467-019-14198-8.

34.

Yang, A.C., Vest, R.T., Kern, F., Lee, D.P., Agam, M., Maat, C.A., Losada, P.M., Chen, M.B.,
Schaum, N., Khoury, N., et al. (2022). A human brain vascular atlas reveals diverse mediators
of Alzheimer's risk. Nature

603

, 885-892. 10.1038/s41586-021-04369-3.

35.

Morabito, S., Miyoshi, E., Michael, N., Shahin, S., Martini, A.C., Head, E., Silva, J., Leavy, K.,
Perez-Rosendahl, M., and Swarup, V. (2021). Single-nucleus chromatin accessibility and
transcriptomic characterization of Alzheimer's disease. Nat Genet

, 1143-1155.

10.1038/s41588-021-00894-z.

36.

Lau, S.F., Cao, H., Fu, A.K.Y., and Ip, N.Y. (2020). Single-nucleus transcriptome analysis
reveals dysregulation of angiogenic endothelial cells and neuroprotective glia in Alzheimer's
disease. Proc Natl Acad Sci U S A

117

, 25800-25809. 10.1073/pnas.2008762117.

37.

Garcia, F.J., Sun, N., Lee, H., Godlewski, B., Mathys, H., Galani, K., Zhou, B., Jiang, X., Ng,
A.P., Mantero, J., et al. (2022). Single-cell dissection of the human brain vasculature. Nature

603

, 893-899. 10.1038/s41586-022-04521-7.

38.

Gerrits,  E.,  Brouwer,  N.,  Kooistra,  S.M.,  Woodbury,  M.E.,  Vermeiren,  Y.,  Lambourne,  M.,
Mulder, J., Kummer, M., Moller, T., Biber, K., et al. (2021). Distinct  amyloid-beta  and tau-
associated  microglia  profiles  in  Alzheimer's  disease.  Acta  Neuropathol

141

, 681-696.

10.1007/s00401-021-02263-w.

39.

Pineda, S.S., Lee, H., Ulloa-Navas, M.J., Linville, R.M., Garcia, F.J., Galani, K., Engelberg-
Cook, E., Castanedes, M.C., Fitzwalter, B.E., Pregent, L.J., et al. (2024). Single-cell dissection
of the human motor and prefrontal cortices in ALS and FTLD. Cell

187

, 1971-1989 e1916.

10.1016/j.cell.2024.02.031.

40.

Wang, Q., Wang, M., Choi, I., Sarrafha, L., Liang, M., Ho, L., Farrell, K., Beaumont, K.G.,
Sebra, R., De Sanctis, C., et al. (2024). Molecular profiling of human substantia nigra identifies
diverse neuron types associated with vulnerability in Parkinson's disease. Sci Adv

, eadi8287.

10.1126/sciadv.adi8287.

41.

Rogers, C., Fernandes-Alnemri, T., Mayes, L., Alnemri, D., Cingolani, G., and Alnemri, E.S.
(2017). Cleavage of DFNA5 by caspase-3 during apoptosis mediates progression to secondary
necrotic/pyroptotic cell death. Nat Commun

, 14128. 10.1038/ncomms14128.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

42.

Deng, Q., Wu, C., Parker, E., Liu, T.C., Duan, R., and Yang, L. (2024). Microglia and Astrocytes
in Alzheimer's Disease: Significance and Summary of Recent Advances. Aging Dis

, 1537-

1564. 10.14336/AD.2023.0907.

43.

Kim, J., Yoo, I.D., Lim, J., and Moon, J.S. (2024). Pathological phenotypes of astrocytes in
Alzheimer's disease. Exp Mol Med

, 95-99. 10.1038/s12276-023-01148-0.

44.

Wieckowski, M.R., Giorgi, C., Lebiedzinska, M., Duszynski, J., and Pinton, P. (2009). Isolation
of mitochondria-associated membranes and mitochondria from animal tissues and cells. Nat
Protoc

, 1582-1590. 10.1038/nprot.2009.151.

45.

de Brito, O.M., and Scorrano, L. (2008). Mitofusin 2 tethers endoplasmic reticulum to
mitochondria. Nature

456

, 605-610. 10.1038/nature07534.

46.

Brahimi-Horn, M.C., and Mazure, N.M. (2014). Hypoxic VDAC1: a potential mitochondrial
marker for cancer therapy. Adv Exp Med Biol

772

, 101-110. 10.1007/978-1-4614-5915-6_5.

47.

Toutenhoofd, S.L., Foletti, D., Wicki, R., Rhyner, J.A., Garcia, F., Tolon, R., and Strehler, E.E.
(1998). Characterization of the human CALM2 calmodulin gene and comparison of the
transcriptional activity of CALM1, CALM2 and CALM3. Cell Calcium

, 323-338.

10.1016/s0143-4160(98)90028-8.

48.

Suzuki, J., Kanemaru, K., Ishii, K., Ohkura, M., Okubo, Y., and Iino, M. (2014). Imaging
intraorganellar Ca2+ at subcellular resolution using CEPIA. Nat Commun

, 4153.

10.1038/ncomms5153.

49.

Chang-Graham, A.L., Perry, J.L., Strtak, A.C., Ramachandran, N.K., Criglar, J.M., Philip, A.A.,
Patton, J.T., Estes, M.K., and Hyser, J.M. (2019). Rotavirus Calcium Dysregulation Manifests
as Dynamic Calcium Signaling in the Cytoplasm and Endoplasmic Reticulum. Sci Rep

, 10822.

10.1038/s41598-019-46856-8.

50.

Guna, A., Volkmar, N., Christianson, J.C., and Hegde, R.S. (2018). The ER membrane protein
complex is a transmembrane domain insertase. Science

359

, 470-473. 10.1126/science.aao3099.

51.

Zhang, L., Xu, Z., Jia, Z., Cai, S., Wu, Q., Liu, X., Hu, X., Bai, T., Chen, Y., Li, T., et al. (2025).
Modulating mTOR-dependent astrocyte substate transitions to alleviate neurodegeneration. Nat
Aging

, 468-485. 10.1038/s43587-024-00792-z.

52.

Tian, Y., Shehata, M.A., Gauger, S.J., Ng, C.K.L., Solbak, S., Thiesen, L., Bruus-Jensen, J.,
Krall, J., Bundgaard, C., Gibson, K.M., et al. (2022). Discovery and Optimization of 5-
Hydroxy-Diclofenac toward a New Class of Ligands with Nanomolar Affinity for the
CaMKIIalpha Hub Domain. J Med Chem

, 6656-6676. 10.1021/acs.jmedchem.1c02177.

53.

Li, K.L., Huang, H.Y., Ren, H., and Yang, X.L. (2022). Role of exosomes in the pathogenesis
of inflammation in Parkinson's disease. Neural Regen Res

, 1898-1906. 10.4103/1673-

5374.335143.

54.

Wang,  J.,  Li,  L.,  Zhang,  Z., Zhang,  X.,  Zhu, Y.,  Zhang,  C.,  and  Bi, Y.  (2022).  Extracellular
vesicles  mediate  the  communication  of  adipose  tissue  with  brain  and  promote  cognitive
impairment  associated  with  insulin  resistance.  Cell  Metab

, 1264-1279 e1268.

10.1016/j.cmet.2022.08.004.

55.

Upadhya, R., Zingg, W., Shetty, S., and Shetty, A.K. (2020). Astrocyte-derived extracellular
vesicles: Neuroreparative properties and role in the pathogenesis of neurodegenerative disorders.
J Control Release

323

, 225-239. 10.1016/j.jconrel.2020.04.017.

56.

Guttenplan, K.A., Weigel, M.K., Prakash, P., Wijewardhane, P.R., Hasel, P., Rufen-Blanchette,
U., Munch, A.E., Blum, J.A., Fine, J., Neal, M.C., et al. (2021). Neurotoxic reactive astrocytes

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

induce cell death via saturated lipids. Nature

599

, 102-107. 10.1038/s41586-021-03960-y.

57.

Liu, Z., Zhang, H., Liu, S., Hou, Y., and Chi, G. (2023). The Dual Role of Astrocyte-Derived
Exosomes and Their Contents in the Process of Alzheimer's Disease. J Alzheimers Dis

, 33-

42. 10.3233/JAD-220698.

58.

Nagai, J., Bellafard, A., Qu, Z., Yu, X., Ollivier, M., Gangwani, M.R., Diaz-Castro, B., Coppola,
G., Schumacher, S.M., Golshani, P., et al. (2021). Specific and behaviorally consequential
astrocyte G(q) GPCR signaling attenuation in vivo with ibetaARK. Neuron

109

, 2256-2274

e2259. 10.1016/j.neuron.2021.05.023.

59.

Li, M., Liu, Z., Wu, Y., Zheng, N., Liu, X., Cai, A., Zheng, D., Zhu, J., Wu, J., Xu, L., et al.
(2024). In vivo imaging of astrocytes in the whole brain with engineered AAVs and diffusion-
weighted magnetic resonance imaging. Mol Psychiatry

, 545-552. 10.1038/s41380-022-

01580-0.

60.

Yasuda, R., Hayashi, Y., and Hell, J.W. (2022). CaMKII: a central molecular organizer of
synaptic plasticity, learning and memory. Nat Rev Neurosci

, 666-682. 10.1038/s41583-022-

00624-2.

61.

Yun, J., Shin, D., Lee, E.H., Kim, J.P., Ham, H., Gu, Y., Chun, M.Y., Kang, S.H., Kim, H.J., Na,
D.L., et al. (2025). Temporal Dynamics and Biological Variability of Alzheimer Biomarkers.
JAMA Neurol

, 384-396. 10.1001/jamaneurol.2024.5263.

62.

Lee, H.G., Lee, J.H., Flausino, L.E., and Quintana, F.J. (2023). Neuroinflammation: An
astrocyte perspective. Sci Transl Med

, eadi7828. 10.1126/scitranslmed.adi7828.

63.

De Schutter, E., Roelandt, R., Riquet, F.B., Van Camp, G., Wullaert, A., and Vandenabeele, P.
(2021). Punching Holes in Cellular Membranes: Biology and Evolution of Gasdermins. Trends
Cell Biol

, 500-513. 10.1016/j.tcb.2021.03.004.

64.

Zhu, C., Xu, S., Jiang, R., Yu, Y., Bian, J., and Zou, Z. (2024). The gasdermin family: emerging
therapeutic targets in diseases. Signal Transduct Target Ther

, 87. 10.1038/s41392-024-01801-

65.

Hu, J.J., Liu, X., Xia, S., Zhang, Z., Zhang, Y., Zhao, J., Ruan, J., Luo, X., Lou, X., Bai, Y., et
al. (2020). FDA-approved disulfiram inhibits pyroptosis by blocking gasdermin D pore
formation. Nat Immunol

, 736-745. 10.1038/s41590-020-0669-6.

66.

Miao, R., Jiang, C., Chang, W.Y., Zhang, H., An, J., Ho, F., Chen, P., Zhang, H., Junqueira, C.,
Amgalan, D., et al. (2023). Gasdermin D permeabilization of mitochondrial inner and outer
membranes accelerates and enhances pyroptosis. Immunity

, 2523-2541 e2528.

10.1016/j.immuni.2023.10.004.

67.

Rizzuto, R., Marchi, S., Bonora, M., Aguiari, P., Bononi, A., De Stefani, D., Giorgi, C., Leo, S.,
Rimessi, A., Siviero, R., et al. (2009). Ca(2+) transfer from the ER to mitochondria: when, how
and why. Biochim Biophys Acta

1787

, 1342-1351. 10.1016/j.bbabio.2009.03.015.

68.

Belosludtsev, K.N., Dubinin, M.V., Belosludtseva, N.V., and Mironova, G.D. (2019).
Mitochondrial Ca2+ Transport: Mechanisms, Molecular Structures, and Role in Cells.
Biochemistry (Mosc)

, 593-607. 10.1134/S0006297919060026.

69.

Liu, J., and Yang, J. (2022). Mitochondria-associated membranes: A hub for neurodegenerative
diseases. Biomed Pharmacother

149

, 112890. 10.1016/j.biopha.2022.112890.

70.

Barazzuol, L., Giamogante, F., and Cali, T. (2021). Mitochondria Associated Membranes
(MAMs): Architecture and physiopathological role. Cell Calcium

, 102343.

10.1016/j.ceca.2020.102343.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

71.

Kuchibhotla, K.V., Lattarulo, C.R., Hyman, B.T., and Bacskai, B.J. (2009). Synchronous
hyperactivity and intercellular calcium waves in astrocytes in Alzheimer mice. Science

323

1211-1215. 10.1126/science.1169096.

72.

Korte, N., Barkaway, A., Wells, J., Freitas, F., Sethi, H., Andrews, S.P., Skidmore, J., Stevens,
B., and Attwell, D. (2024). Inhibiting Ca(2+) channels in Alzheimer's disease model mice
relaxes pericytes, improves cerebral blood flow and reduces immune cell stalling and hypoxia.
Nat Neurosci

, 2086-2100. 10.1038/s41593-024-01753-w.

73.

Princen, K., Van Dooren, T., van Gorsel, M., Louros, N., Yang, X., Dumbacher, M., Bastiaens,
I., Coupet, K., Dupont, S., Cuveliers, E., et al. (2024). Pharmacological modulation of septins
restores calcium homeostasis and is neuroprotective in models of Alzheimer's disease. Science

384

, eadd6260. 10.1126/science.add6260.

74.

Asai, H., Ikezu, S., Tsunoda, S., Medalla, M., Luebke, J., Haydar, T., Wolozin, B., Butovsky, O.,
Kugler, S., and Ikezu, T. (2015). Depletion of microglia and inhibition of exosome synthesis
halt tau propagation. Nat Neurosci

, 1584-1593. 10.1038/nn.4132.

75.

Cheng, L., Vella, L.J., Barnham, K.J., McLean, C., Masters, C.L., and Hill, A.F. (2020). Small
RNA fingerprinting of Alzheimer's disease frontal cortex extracellular vesicles and their
comparison with peripheral extracellular vesicles. J Extracell Vesicles

, 1766822.

10.1080/20013078.2020.1766822.

76.

Liu, X., Shen, L., Wan, M., Xie, H., and Wang, Z. (2024). Peripheral extracellular vesicles in
neurodegeneration: pathogenic influencers and therapeutic vehicles. J Nanobiotechnology

170. 10.1186/s12951-024-02428-1.

77.

Sardar  Sinha,  M., Ansell-Schultz, A.,  Civitelli,  L.,  Hildesjo,  C.,  Larsson,  M.,  Lannfelt,  L.,
Ingelsson,  M.,  and  Hallbeck,  M.  (2018).  Alzheimer's  disease  pathology  propagation  by
exosomes  containing  toxic  amyloid-beta  oligomers.  Acta  Neuropathol

136

, 41-56.

10.1007/s00401-018-1868-1.

78.

Rabah, Y., Berwick, J.P., Sagar, N., Pasquer, L., Placais, P.Y., and Preat, T. (2025). Astrocyte-
to-neuron H(2)O(2) signalling supports long-term memory formation in Drosophila and is
impaired in an Alzheimer's disease model. Nat Metab

, 321-335. 10.1038/s42255-024-01189-

79.

Pons-Espinal, M., Blasco-Agell, L., Fernandez-Carasa, I., Andres-Benito, P., di Domenico, A.,
Richaud-Patin, Y., Baruffi, V., Marruecos, L., Espinosa, L., Garrido, A., et al. (2024). Blocking
IL-6 signaling prevents astrocyte-induced neurodegeneration in an iPSC-based model of
Parkinson's disease. JCI Insight

. 10.1172/jci.insight.163359.

80.

Ameroso, D., Meng, A., Chen, S., Felsted, J., Dulla, C.G., and Rios, M. (2022). Astrocytic
BDNF signaling within the ventromedial hypothalamus regulates energy homeostasis. Nat
Metab

, 627-643. 10.1038/s42255-022-00566-0.

81.

Allen, N.J., Bennett, M.L., Foo, L.C., Wang, G.X., Chakraborty, C., Smith, S.J., and Barres,
B.A. (2012). Astrocyte glypicans 4 and 6 promote formation of excitatory synapses via GluA1
AMPA receptors. Nature

486

, 410-414. 10.1038/nature11059.

82.

Jack, C.R., Jr., Bennett, D.A., Blennow, K., Carrillo, M.C., Dunn, B., Haeberlein, S.B.,
Holtzman, D.M., Jagust, W., Jessen, F., Karlawish, J., et al. (2018). NIA-AA Research
Framework: Toward a biological definition of Alzheimer's disease. Alzheimers Dement

, 535-

562. 10.1016/j.jalz.2018.02.018.

83.

Wolfes, A.C., and Dean, C. (2018). Culturing In Vivo-like Murine Astrocytes Using the Fast,

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies

Chicken polyclonal anti-GFAP

Abcam

Cat# ab4674; RRID:
AB_304558

Mouse monoclonal anti-GFAP

Cell Signaling
Technologies

Cat# 3670S; RRID:
AB_561049

Rabbit monoclonal anti-GSDME

Abcam

Cat# ab215191;
RRID:

AB_2737000

Rabbit monoclonal anti-cleaved N-terminal GSDME

Cell Signaling
Technologies

Cat# E8G4U; RRID:
AB_2923216

Rabbit monoclonal anti-GSDMD

Abcam

Cat# ab209845;
RRID: AB_2783550

Goat polyclonal anti-complement C3

Thermo Fisher
Scientific

Cat# PA1-29715;
RRID:

AB_2066730

Mouse monoclonal anti-complement C3

Santa Cruz
Biotechnology

Cat#

sc-28294;

RRID: AB_627277

Rabbit monoclonal anti-LCN2

Abcam

Cat# ab63929; RRID:

AB_1140965

Rabbit polyclonal anti-HMGB1

Abcam

Cat# ab18256;
RRID: AB_444360

Rabbit polyclonal anti-iNOS

Abclonal Technology

Cat# A14031;
RRID: AB_2760886

Rabbit monoclonal anti-ACSL4

Abcam

Cat# ab155282;
RRID: AB_2714020

Mouse polyclonal anti-IP3R

Boster Biological
Technology

Cat# PB9225,
RRID: AB_3082024

Rabbit polyclonal anti-VDAC1

Proteintech

Cat# 55259-1-AP;

RRID: AB_10837225

Rabbit polyclonal anti-GAPDH

HuaBio

Cat# ER1706-83;
RRID: AB_3069094

Mouse monoclonal anti-β-actin

HuaBio

Cat# EM21002;
RRID: AB_2819164

Rabbit monoclonal anti-caspase-3

Abcam

Cat# ab184787;
RRID:

AB_2827742

Rabbit polyclonal anti-Iba1

FUJIFILM Wako

Cat# 019-19741;
RRID:

AB_839504

Rabbit monoclonal anti-Iba1

Cell Signaling
Technologies

Cat# 17198T; RRID:

AB_2820254

Rabbit monoclonal anti-NeuN

Abcam

Cat# ab177487;
RRID:

AB_2532109

Mouse monoclonal anti-PSD-95

Santa Cruz
Biotechnology

Cat# sc-71934;
RRID:

AB_1128590

Rabbit polyclonal anti-synaptophysin

Proteintech

Cat# 17785-1-AP;
RRID:

AB_2271365

Chicken polyclonal anti-MAP2

Abcam

Cat#ab5392; RRID:

AB_2138153

Rabbit monoclonal anti-calnexin, CD9, CD63, CD81, Hsp70,
TSG101

Abcam

Cat#ab275018

Rabbit polyclonal anti-SHH

Cusabio

Cat# CSB-
PA623000LA01HU

Rabbit polyclonal anti-BDNF

Cusabio

Cat# CSB-
PA05775A0Rb

Rabbit polyclonal anti-EMP1

Cusabio

Cat# CSB-
PA007648LA01HU

Rabbit polyclonal anti-SPHK1

Cusabio

Cat# CSB-
PA022564LA01HU

Rabbit polyclonal anti-Phospho-NRF2 (p-NRF2)

ABclonal

Cat# AP1133 RRID:
AB_2864001

Rabbit monoclonal anti-LCN2

ABclonal

Cat# A11207 RRID:
AB_2861520

Rabbit polyclonal anti-S100A10

ABclonal

Cat# A14658 RRID:
AB_2761534

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Mouse monoclonal anti-CaMKII alpha

Proteintech

Cat# 66843-1-Ig
RRID: AB_2882184

Rabbit recombinant anti-Phospho-NF-kB p65 (p-p65)

ABclonal

Cat# AP1294 RRID:
AB_3099756

Rabbit Recombinant anti-NRF2

Proteintech

Cat# 80593-1-RR
RRID: AB_2918904

Rabbit anti-Phospho-CaMKII alpha (T286) (p- CaMKIIα)

ABclonal

Cat# AP1386

Rabbit polyclonal anti-GPC4

Cusabio

Cat# CSB-
PA009706LA01HU

Rabbit recombinant anti-GDNF

Proteintech

Cat# 83087-2-RR
RRID: AB_3670803

Rabbit polyclonal anti-SPHK1

Proteintech

Cat# 10670-1-AP
RRID: AB_2195809

Rabbit monoclonal anti-GSDME

abcam

Cat# ab215191
RRID: AB_2737000

Anti-mouse IgG secondary antibody, FITC conjugated

Boster Biological
Technology

Cat#BA1101

Anti-mouse IgG secondary antibody, Alexa Fluor® 568
conjugated

Abcam

Cat#ab175473;
RRID:

AB_2895153

Anti-mouse IgG secondary antibody, Alexa Fluor® 647
conjugated

Abcam

Cat#ab150115;
RRID:

AB_2687948

Anti-Rabbit IgG Secondary Antibody, Cy3 Conjugated

Boster Biological
Technology

Cat#BA1032; RRID:

AB_2716305

Anti-rabbit IgG secondary antibody, Alexa Fluor® 647
conjugated

Abcam

Cat#ab150075;
RRID:

AB_2752244

Anti-chicken IgY secondary antibody, Alexa Fluor® 594
conjugated

Abcam

Cat#ab150176;
RRID:

AB_2716250

Anti-chicken IgY secondary antibody, Alexa Fluor® 647
conjugated

Abcam

Cat#ab150171;
RRID:

AB_2921318

Anti-goat IgG secondary antibody, Alexa Fluor® 647 conjugated

Abcam

Cat#ab150135;
RRID:

AB_2687955

Anti-rabbit IgG secondary antibody, 10 nm gold conjugated

Abcam

Cat#ab39601;
RRID:

AB_954434

Biological samples

Human brain tissue

Chinese Brain Bank
Center

N/A

Chemicals, peptides, and recombinant proteins

PEI

Proteintech

Cat#PR40001

Lipopolysaccharides from

E. coli

O 111: B4

Sigma-Aldrich

Cat#L2630

Nigericin

InvivoGen

Cat#tlrl-nig

E. coli

OMVs

InvivoGen

Cat#tlrl-omv-1

poly-L-lysine

Boster Biological
Technology

Cat#AR0003

Fetal bovine serum

Thermo Fisher
Scientific

Cat#10091148

DMEM, high glucose, pyruvate

Thermo Fisher
Scientific

Cat#C11995500BT

0.25% trypsin-EDTA

Beyotime
Biotechnology

Cat#C0201

N-Acetyl-L-cysteine

Sigma-Aldrich

Cat#A8199-10G

Glutamax

Gibco

Cat#35050061

apo-Transferrin human

Sigma-Aldrich

Cat#T1147-100MG

HB-EGF protein, human

MedChemExpress

Cat#HY-P7017

Recombinant human HB-EGF protein

R&D Systems

Cat#259-HE

B27

Gibco

Cat#17504044

4% Paraformaldehyde (PFA) solution in PBS

Boster Biological
Technology

Cat#AR1068

ATP

New England
BioLabs (NEB)

Cat#P0756S

Critical commercial assays

CytoTox 96®non-radioactive cytotoxicity assay kit

Promega

Cat#G1780

Calcein/PI cell viability/cytotoxicity assay kit

Beyotime
Biotechnology

Cat#C2015M

NF-κB pathway sampler kit

Cell Signaling

Cat#9936T

The copyright holder for this

this version posted September 2, 2025.

https://satijalab.org/s
eurat/

bioRxiv preprint

Experimental models: Cell lines

Human: HEK293T

ATCC

Cat# CRL-3216

Mouse: BV-2

Procell Life
Science&Technology

Cat#CL-0493A

Experimental models: Animals

Mouse: C57BL/6J; 5×FAD

Animal Resources
Center of the Fourth
Military Medical
University

N/A

Mouse: C57BL/6J GSDME

-/-

NIBS

N/A

Recombinant DNA

pCMV-VSV-G

N/A

Addgene: #8454

pCMV-dR8.91

YouBio

Cat: #VT1441

Plasmid: pGP-CMV-GCaMP6s

HonorGene

Addgene: #40753

Plasmid:

pLVX-puro_G-CEPIA1er

HonorGene

Addgene: #164590

Plasmid: pCMV-CEPIA4mt

HonorGene

Addgene:

#58220

Software and algorithms

ImageJ

National Institutes of
Health

https://imagej.nih.go
v/ij/

Graphpad Prism

GraphPad Software

https://www.graphpa
d.com

CellMarker

Harbin Medical
University

http://xteam.xbio.top
/CellMarker/

Human Protein Atlas

N/A

https://www.proteina
tlas.org/

Seurat (v.4.3.0)

Satija Lab

RStudio (1.1.442)

N/A

https://posit.co/down
loads/

R (v.4.1.1)

N/A

https://www.r-
project.org/

ggplot2 (v.3.4.4)

GitHub

https://github.com/er
ikgahner/awesome-
ggplot2

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

All experimental procedures involving mice were conducted under a protocol approved
by  the  Fourth  Military  Medical  University  Institutional  Animal  Care  and  Use
Committee, adhering to the guidelines outlined by the Animal Welfare Act and relevant
Chinese regulations concerning animal experimentation. Wild-type C57BL/6 mice and
5×FAD  transgenic  mice  (C57BL/6  background)  were  supplied  by  the  Animal
Resources  Center  of  the  Fourth  Military  Medical  University  (Xi’an,  China),  while
GSDME

-/-

transgenic mice (C57BL/6) were kindly provided by Professor Feng Shao.

All strains were maintained under specific pathogen-free conditions by breeding with
C57BL/6 mice, with animals housed in controlled environments (temperature: 22 ± 1°C,
humidity: 50 ± 10%, 12-h light/dark cycle) and provided ad libitum access to food and
water. To eliminate bias, animals were randomly assigned identification numbers and
subsequently evaluated in a blinded manner throughout the experimental procedures.

Cell culture and treatment

Astrocytes

Mouse primary astrocytes were cultured following an established protocol with minor
modifications.

Briefly, astrocytes were isolated from the cortices or hippocampus

obtained from mouse pups of mixed sexes at postnatal day 0 (P0). P0 mouse pups were
decapitated, and brains were isolated and kept in ice-cold Dulbecco’s modified Eagle
medium (DMEM).

After removing the meninges and capillaries, the cortices or

hippocampus were minced and digested with 0.125% Trypsin-EDTA

at 37°C for 10

min. Digestion was halted with DMEM containing 10% fetal bovine serum (FBS).
Large tissue aggregates were removed by filtering through a 40-μm mesh. The cells
were pelleted by centrifugation at 300



for 10 min and resuspended in DMEM with

10% FBS. Following a differential adhesion period of 20 min, the culture medium was
replaced with a serum-free base medium.

The culture medium consisted of a mixture

of 50% Neurobasal and 50% DMEM, supplemented with 100 U/mL penicillin, 100
μg/mL streptomycin, 1 mM sodium pyruvate, 292 μg/mL glutamine, 5 ng/mL HBEGF,
and 5 μg/mL N-acetylcysteine. Cells were seeded on T75 culture flasks pre-coated with
poly-L-lysine. After 4 d, the medium was replaced with DMEM supplemented with 10%
FBS.

For the collection of astrocyte-conditioned medium (ACM), astrocyte cultures were

randomly chosen as control or reactive. Following 8 d in culture, the flask was agitated
at 260 rpm for 8 h at 37°C to eliminate microglial cells and then harvested by
trypsinization using 0.25% Trypsin-EDTA. Neurotoxic astrocytes were generated

vitro

by culturing purified astrocytes for 8 d, followed by a 48-h exposure to activated

microglia conditioned medium (AMCM). The conditioned medium from both control
(ACM) and neurotoxic (aACM) astrocytes was harvested and concentrated using

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

ultrafiltration concentration tubes with a 10 kDa retention capacity until it reached
approximately a tenfold volume concentration for subsequent neuronal toxicity testing
or exosome isolation.

Primary human astrocytes (HA, ScienCell Research Laboratories, Cat# HA) were

cultured in poly-L-lysine-coated (2 μg/cm²) 24-well plates. Prior to cell seeding, plates
were pre-coated with poly-L-lysine (10 mg/ml, Cat# 0413) diluted in sterile water (2
μg/cm²)  and  incubated  overnight  at  37°C.  Cells  were  thawed  and  expanded  in T-75
flasks following the manufacturer’s protocol, then subcultured into 24-well plates at a
density  of  5,000  cells/cm²  in  complete  Astrocyte  Medium  (AM,  Cat#  1801)
supplemented  with AGC  and  FBS.  Cultures  were  maintained  at  37°C  in  a  5%  CO₂
incubator  until  reaching  70%  confluency  (typically  3-4  days  post-seeding).  For
exosome stimulation, cells were gently washed twice with pre-warmed DPBS (Cat#
0303) to remove serum and growth factors. Serum- and AGC-free Astrocyte Medium
(500 μL/well) was added, followed by exosomes resuspended in the same medium at a
final  concentration  of  100  μg/mL.  Control  wells  received  vehicle  solution  (equal
volume  of  PBS  or  exosome-free  medium). The  plate  was  incubated  under  standard
conditions (37°C, 5% CO₂) for 24 h.

Microglia

Mouse primary microglia were cultured following an established protocol with minor
modifications.

Briefly, microglia were isolated from the cortices obtained from mouse

pups of mixed sexes at postnatal day 0 (P0). P0 mouse pups were decapitated, and
brains were isolated and kept in ice-cold DMEM. After removing the meninges and
capillaries, the cortices were minced and digested with 0.125% trypsin-EDTA at 37°C
for a duration of 15 min. The digestion process was halted with DMEM containing 10%
FBS. Large tissue aggregates were removed by filtering through a 40 μm mesh. The
cells were pelleted at 600



for 10 min

and resuspended in DMEM with 10% FBS.

Following a differential adhesion period of 1 h, the culture medium was replaced with
a serum-free base medium. The medium was changed the next day and changed every
4 d for the next 12 d. When mixed glial cultures were completely confluent, microglia
were harvested by shake-off at 200 rpm at 37°C for 6 h. Microglia were counted and
plated in 24-well culture plates at a density of 1.5×10

cells/ml. The culture purity

was >99%, verified by immunofluorescence with GFAP (for astrocytes), NeuN (for
neurons) and CD11b (for microglia).

Neurons

Primary cortical neurons from mouse embryos at embryonic day 14-18 were isolated
and cultured in neurobasal medium following a previously established protocol with
minor modifications.

Briefly, following a differential adhesion period of 4 h in

DMEM containing 10% FBS, the culture medium was replaced with neurobasal
medium supplemented with 1% B-27, 0.5 mM glutamine and 1% penicillin-
streptomycin.

Following 7 d in culture, neurotoxicity assay was conducted

in vitro

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

METHOD DETAILS

Single-cell transcriptome analysis using open snRNA-seq data

Basic processing and image plotting of the snRNA-seq data were performed with Seurat
(v.4.3.0), ggplot2 (v.3.4.4) and ggpurb (v.0.4.0), run on RStudio (v.1.1.442), using R (v.
4.1.1).

Cells with the number of genes (nFeature_RNA) less than 200 and more than

10,000, as well as the percentage of mitochondrial genes larger than 5%were discarded.
The  expression  matrix  of  snRNA-seq  were  converted  into  a  standard  input  file.
Following  clustering,  astrocytes  were  distinguished  from  different  clusters  using
CellMarker  (

http://xteam.xbio.top/CellMarker/

) and the Human Protein Atlas

(

https://www.proteinatlas.org/

). Subsequently, the comparation of astrocytic

GSDME

expression (disease group versus healthy group) was identified using the FindMarkers
function. Analyses of datasets from various databases were conducted separately.
Identification of astrocytes using conventional snRNA seq analysis methods. Further
subgroup analysis of astrocytes with a resolution ≥ 0.1. Within each cluster, we assessed
the average log

(fold change) of

Gfap

. Clusters with an avg_log

(fold change) > 0 were

categorized as the reactive astrocyte group, while the remaining clusters were classified
as the non-reactive astrocyte group. The cells were discarded with zero expression
levels of GSDME. The analysis of astrocytic

GSDME

expression across various

subtypes in Figure 6A was conducted post-clustering. Each subtype of astrocytes (ASTs)
was subsequently paired with its counterpart to identify ASTs based on marker genes

The correlations between

GSDME

and different astrocytic marker genes across

different ASTs were analyzed using the Spearman method.

RNA extraction and RNA-seq analysis

Total RNA was isolated from mouse cortical tissues using Trizol Reagent (Invitrogen
Life Technologies), followed by direct lysis of cell samples in 24-well culture plates.
RNA  concentration,  purity,  and  integrity  were  assessed  via  NanoDrop
spectrophotometry  (Thermo  Scientific)  and  agarose  gel  electrophoresis.  For  library
construction,  3  μg  of  RNA  was  subjected  to  mRNA  enrichment  with  poly-T  oligo-
attached magnetic beads, followed by fragmentation in Illumina buffer using divalent
cations (94°C, 5 min). First-strand cDNA was synthesized with SuperScript II reverse
transcriptase  and  random  hexamers,  and  second-strand  cDNA  was  generated  using
DNA  Polymerase  I/RNase  H.  Blunt-end  repair,  3′-adenylation,  and  Illumina  adapter
ligation  were  sequentially  performed.  Library  fragments  (400–500  bp)  were  size-
selected with AMPure XP beads (Beckman Coulter, USA), amplified by 15-cycle PCR,
and  quantified  via Agilent  Bioanalyzer  2100  (Agilent  Technologies).  Final  libraries
were  sequenced  on  an  Illumina  NovaSeq  6000  platform  (Shanghai  Personal
Biotechnology Co., Ltd).

RNA sequencing was performed as follows: Raw sequencing data (FASTQ files)

were quality-controlled using Cutadapt (v1.15) to remove adapters and low-quality

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

reads, generating clean data. HISAT2 (v2.0.5) aligned filtered reads to the reference
genome. Gene expression quantification was calculated via HTSeq (0.9.1) with FPKM
normalization. Differential expression analysis employed DESeq (v1.39.0) with
thresholds set at |log

FoldChange| >1.5 (tissue samples) or >2 (cellular samples) and

adjusted

-value <0.05. Bidirectional clustering of differentially expressed genes

(DEGs) was visualized using the pheatmap R package (v1.0.8), with Euclidean distance
and complete linkage algorithms.

Functional enrichment analysis encompassed GO

term enrichment via topGO (v2.40.0) using a hypergeometric test (

<0.05), KEGG

pathway annotation using clusterProfiler (v3.16.1) (

<0.05), and multi-database Gene

Set Enrichment Analysis (GSEA) performed with clusterProfiler (v3.16.1). For GSEA,
pre-ranked gene lists (sorted by log

FoldChange) were tested against gene sets from

MSigDB Hallmark, GO Biological Process, and KEGG pathways. Significance for
GSEA was defined as FDR <0.25 and |Normalized Enrichment Score (NES)| >1.

Isolation of astrocytic exosomes

The ACM of WT-MCM, GSDME KO-MCM, WT-AMCM, or GSDME KO-AMCM
astrocytes were harvested 48 h post-stimulation. The astrocytic exosomes were then
isolated through differential ultracentrifugation, following the method previously
described.

Briefly, the collected medium underwent centrifugation at 300×

for 10

min and 2,000×

for 10 min at 4°C to eliminate dead cells and cellular debris.

Subsequently, the supernatant was centrifuged at 10,000×

for 30 min at 4°C to remove

larger microvesicles and apoptotic bodies. The remaining supernatant was subjected to
ultracentrifugation at 100,000×

for 70 min at 4°C to yield the astrocytic exosome

pellet. This pellet was washed once with PBS to ensure the removal of non-specific
proteins and debris.

Nanoparticle tracking analysis (NTA)

The size and quantity of isolated astrocytic exosomes were assessed using a ZetaView
Nanoparticle Tracker and Particle Metrix software. Exosome samples were diluted with
1× PBS buffer for size and concentration measurements. The NTA measurements were
taken at 11 different positions, and the ZetaView system was calibrated with 250 nm
polystyrene particles. The temperature during the measurements was controlled at
approximately 23°C and 30°C.

Synaptosome engulfment assay

Synaptosomes were isolated from mouse brain according to the procedure described in
a previous study

and labeled with pHrodo Red succinimidyl ester. Excess pHrodo Red

was removed by centrifugation, and the pHrodo-labeled synaptosomes were then
resuspended in PBS containing 5% dimethyl sulfoxide (DMSO) before being stored at
-80°C. Primary mouse astrocytes were seeded on Poly-L-lysine precoated 24-well
plates at a density of 1×10

cells per well for 24 h and exposed to MCM or aMCM for

an additional 48 h. Subsequently, the cells were incubated with pHrodo-labeled

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

synaptosomes (5 μL) for 24 h, followed by co-staining with Calcein AM to visualize
the uptake of synaptosomes by astrocytes. Images were captured with a confocal
microscope (IXplore SpinSR).

Neurotoxicity assay

Mouse primary cortical neurons were isolated and cultured as mentioned before. The
cultured cortical neurons were incubated with the concentrated ACM or purified
exosomes collected from astrocytes in neurobasal medium for 24 h. Neuronal viability
was measured using the Calcein AM/PI method with a viability/cytotoxicity assay kit
(Beyotime, Cat#C2015M).

Scanning electron microscope (SEM)

Astrocytes growing on cover slips were stimulated by MCM, AMCM or overexpressed
EGFP-tagged GSDME-NT. Subsequently, the cells were rinsed thrice with PBS,
followed by fixation with 2.5% glutaraldehyde at 4°C for 24 h. The slips were rinsed
twice with PBS after glutaraldehyde removed. The morphology of astrocytes on glass
slides was examined using a scanning electron microscope (ThermoFisher Quattro S)
following treatment with osmium tetroxide, dehydration, and gold coating. For the
exosome samples,

20 μL of exosome suspension droplets were pipetted into the copper

mesh grid and retained on the copper mesh for a few moments (more than 1 min).
Negative staining was performed by dripping 2% phosphotungstic acid onto the grid
for 1-10 min, and dry naturally at room temperature after the filter paper is blotted dry.
Samples were observed and photographed under a biological transmission electron
microscope at 120 kV accelerating voltage.

Transmission electron microscopy (TEM)

Astrocytes were treated with 0.25% trypsin at 24 h post-stimulation. Subsequently, the
cells were pelleted by centrifugation at 600 rpm for 3 min, and resuspended in 3%
glutaraldehyde. The resuspended cells were then centrifuged at 5000 rpm for 10 min to
form tight cell clusters, after which the supernatant was carefully removed. Fresh
glutaraldehyde was slowly added to the tubes, which were then stored at 4°C for 24 h.
The cell clusters underwent osmium acid treatment, dehydration, and embedding, and
were sectioned to a thickness of 70 nm. Finally, the sections were stained with lead
citrate and uranyl acetate for observation under an electron microscope (HITACHI, HT-
7800).

Immuno-electron microscope (IEM)

Astrocytes stimulated with microglia conditioned medium for 24 h were collected for
TEM samples, avoiding osmium-acid treatment. The cell clusters were cut into ultrathin
70 nm sections and placed on a nickel mesh with 200-300 mesh holes. Sections were
soaked in 1% H

for 10 min and then rinsed in double-distilled water for 10 min.

Blocking was performed with 1% PBSA (1% BSA in PBS buffer) for 30-60 min at

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

room  temperature.  1:50  Rabbit  cleaved  GSDME  primary  antibody  (Cell  Signaling
Technologies,  E8G4U)  was  prepared  in  1%  PBSA,  preincubated  for  1  h  at  room
temperature, and then incubated overnight at 4°C. The next day, sections were rinsed
with PBS for 5 min, three times. Sections were then incubated with 1% PBSA for 5 min,
followed by incubation with 1:40 colloidal gold-labeled secondary antibody (Abcam,
ab39601) solution  in  1% PBSA for 1 h at  room  temperature. Sections were washed
three  times  with  double-distilled  water  for  8  min  each,  stained  with  uranyl  acetate
(prepared in double-distilled water) for 12 min, and then washed with double-distilled
water for 10 min. Finally, sections were stained with lead citrate for 6 min and washed
with double-distilled water for 10 min. Residual water around the sections was carefully
drained using filter paper, and images were taken using an electron microscope.

RT-qPCR

Total RNA of both cell and exosomes (collected from 4×10

cells) was extracted with

Trizol (Invitrogen) using a standard protocol. The synthesis of cDNA was accomplished
using  either  the  TransScript®  One-Step  gDNA  Removal  and  cDNA  Synthesis
SuperMix (TransGen Biotech) for standard RNA reverse transcription, or the miRNA
1st  Strand  cDNA  Synthesis  Kit  (Vazyme  Biotech)  for  the  reverse  transcription  of
microRNA.  Quantitative  RT-PCR  was  run  using  1 μL  cDNA  and  SYBR  green
chemistry  (Monad  Biotech)  using  the  supplier’s  protocol  and  a  cycling  program  of
2 min at 95°C followed by 40 cycles of 95°C for 3 s and 60°C for 30 s on QuantStudio
3 Real-Time PCR Systems (Applied Biosystems). The primer sequences were shown
in supplementary table 2. Either GAPDH or U6 snRNA was served as internal controls.
The 2

−ΔΔ

method was used for the calculation of relative fold expression of target

genes.

Western blotting

Protein  samples  were  collected  in  PBS  at  4°C  and  lysed  using  RIPA  buffer
supplemented  with  1  mM  phenylmethylsulfonyl  fluoride  (PMSF).  The  total  protein
concentration  was  determined  using  the  bicinchoninic  acid  (BCA)  assay,  and  equal
amounts  of  total  protein  were  loaded  onto  12.5%  SDS-polyacrylamide  gel
electrophoresis (PAGE) gels from Epizyme Biotech. After electrophoresis at 100 V for
60 min, the proteins were transferred to polyvinylidene difluoride (PVDF) membranes
from Millipore. The membranes were blocked with 5% milk for 2 h at room temperature,
followed by overnight incubation at 4°C with primary antibodies in the blocking buffer.
Subsequently, the blots were washed three times with phosphate-buffered saline with
Tween  (PBST)  and  then  incubated  with  horseradish  peroxidase  (HRP)-conjugated
secondary antibodies for 2 h at room temperature. The protein bands were visualized
using the Omni-ECL Femto Light Chemiluminescence Kit, and imaging was performed
with a Chemi-Image System from Tanon.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Immunostaining

Mice  were  anesthetized  using  50  mg/kg  Pentobarbital  sodium  and  subsequently
perfused transcardially with PBS following behavioural testing. The left hemisphere of
the brain was excised and preserved in 4% paraformaldehyde at 4°C for a duration of
48 h, prior to dehydration with 20% and 30% sucrose solutions. The brains were then
sectioned coronally at 30 μm using a frozen section machine (Leica Camera, Inc.). The
sections  underwent  an  overnight  incubation  with  primary  anti-mouse  antibodies.
Fluorescent-conjugated  secondary  antibodies  were  used  to  reveal  the  staining.  The
sections were then imaged using a confocal microscope (Zeiss LSM800).

For primary cell staining, 1.5×10

astrocytes or 1×10

neurons were cultured on poly-

D-lysine coated glass slides in each well of a 24-well plate, using serum-free base
medium for 24 h. The slides were then fixed in 4% paraformaldehyde for 30 min at
room temperature, followed by permeabilization with PBS containing 0.1% Triton X-
100 for another 30 min at room temperature. Subsequently, the cells were blocked with
3% BSA (PBS) mixture for 2 h at room temperature. An overnight incubation at 4°C
was then conducted using primary anti-mouse antibodies. Fluorescein-conjugated
secondary antibodies were applied to the slides for 1 h in light-protected conditions at
room temperature. Washing with PB (0.1M; Na

HPO

: NaH

= 1:1, mol/mol) was

performed three times between each step. Images were captured using a fluorescence
microscope after a 5 min incubation with DAPI at room temperature.

Characterization of subcellular Ca

To obtain Ca

relative concentration of interested subcellular area, we constructed

lentiviral expression vectors expressing genetically encoded Ca

indicators (GECIs).

The plasmid pLVX-puro_G-CEPIA1er was purchased from HonorGene (Plasmid
#164590). Plasmids for cytoplasm and mitochondria Ca

detection were obtained by

replacing G-CEPIA1er with CEPIA4mt of pCMV CEPIA4mt (FENGHUISHENGWU,
Plasmid #58220) and GCaMP6s of pGP-CMV-GCaMP6s (FENGHUISHENGWU,
Plasmid #40753) respectively.

48,49

The functionality of these indicators was validated

by capturing subcellular Ca

dynamics using a confocal microscope (IXplore SpinSR).

This allowed for the characterization of real-time fluorescence intensity changes by
Image J observed 10 min post the application of 50 µM ATP (NEB, Cat# P0756S) or
PBS. To explore astrocyte subcellular Ca

dynamics induced by aMCM or MCM

，

the

real-time fluorescence intensity changes were recorded for 8 h after the stimulation
using a

High-content Imaging Analysis System (ImageXpress Micro Confocal).

Saturated lipid analysis

Lipids were extracted from astrocytic exosomes according to the modified Bligh-Dyer
method. Briefly, the same quantity of exosomes (80-100 μg) from indicated groups
added with a deuterated lipid internal standard (Stearic acid 18,18,18-D3, #61163-39-
7, Macklin Inc.) were incubated at 4°C and 1500 rpm for 1 h in a chloroform extract
with 1% butylated hydroxytoluene (BHT) in methanol (1:2, v/v). Following the

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

incubation, 350 µL of ice-cold Milli-Q water and 250 µL of ice-cold chloroform was
added to induce phase separation. The mixture was then centrifuged at 16, 260

for 5

min at 4°C to transfer the lower organic phase containing lipids and oxidized sterols to
a new tube. Subsequently, another 450 µL of ice-cold chloroform was added to the
remaining aqueous phase for a second extraction. The extracts were combined and dried
under organic mode in a SpeedVac. The lipid extract was re-suspended in 200 µL of
chloroform:methanol (1:1, v/v). The exosomes lipid extract was utilized for targeted
lipidomics analysis by a high-performance liquid chromatography-mass spectrometry
(LC-MS, Thermo Scientific™ Q Exactive Plus™).

Intracerebroventricular viral delivery in neonatal mice

Neonatal mice (postnatal days 0–3, P0–P3) were anesthetized by hypothermia on a pre-
chilled  ice  pad  for  2–3  min  until  the  loss  of  reflex  response.  The  pup’s  head  was
disinfected  with  70%  ethanol  and  stabilized  in  a  lateral  recumbent  position  under  a
stereotaxic apparatus (Model details, Manufacturer). The skull was rotated to align the
target injection site (lateral ventricle: anteroposterior −0.5 mm, mediolateral ±1.0 mm
relative  to  bregma)  upward,  and  the  coordinates  were  microscopically  verified
(Microscope model, Manufacturer). A 34-gauge microinjection needle (Hamilton, USA)
preloaded with 10 µL of AAV suspension (1×10¹² vg/mL) was vertically inserted into
the lateral ventricle at a depth of 1.35 mm. Viral solution (1-2 µL) was infused at a rate
of 0.2 µL/min using a microprocessor-controlled pump (Model details, Manufacturer).
The  needle  was  retained  in  situ  for  2  min  post-injection  to  prevent  backflow  and
gradually withdrawn, followed by gentle pressure application with a sterile cotton swab
to  seal  the  puncture  site. After  allowing  5  min  for  wound  closure,  the  contralateral
hemisphere was injected using identical parameters. Pups were transferred to a heating
pad (37°C) for 30 min to restore normothermia, monitored for respiratory stability and
motor activity, and returned to the dam’s nest. Postoperative survival and maternal care
were assessed for 24 h prior to subsequent experiments.

AAV viral vectors were generated by PackGene Biotech (Guangzhou, China). Briefly,

HEK293  cells  were  transiently  transfected  with  a  GOI  plasmid  (pAAV-
mir30shRNA.WPRE.SV40p), an adenovirus helper plasmid (pHelper), and an AAV RC
plasmid (pAAV-php.eb). 72 h post-transfection, recombinant AAV were collected by
lysing the cells to release the virus particles into the supernatant. The crude viral lysate
was then purified via iodixanol-gradient centrifugation. Viral titers, reported in genome
copies  per  milliliter  (gc/mL),  were  measured  using  real-time  PCR  with  Universal
SYBR  Green  Mix  (BIO-RAD,  USA)  and  primers  designed  to  target  the  inverted
terminal  repeat  (ITR)  sequence  of  the AAV  vector. Vectors  that  met  quality  control
standards were aliquoted and stored at -80°C until further use.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Behavior test

Open field test

The open-field arena consisted of a white polypropylene chamber (50 × 50 × 40 cm).
The central zone was defined as a 20 × 20 cm square area in the center of the arena,
with the remaining area designated as the periphery. Illumination was adjusted to 60
lux  using  a  calibrated  light  meter.  Behavioral  sessions  were  recorded  and  analyzed
using  Noldus  Ethovision  XT  11.5  software  (Noldus  Information  Technology,
Netherlands). Mice were transported to the behavioral testing room in covered cages
(to minimize visual stress) and allowed to acclimate to the environment for 1 hour prior
to testing. At the start of each trial, mice were gently grasped by the base of the tail and
positioned head-first toward the corner of the arena, ensuring abdominal contact with
the wall. Mice were allowed to freely explore the arena for 10 min while being video-
recorded. After the trial, mice were returned to their home cages, and individual animals
were identified by tail markings corresponding to their assigned arena and video codes.
The arena and surrounding surfaces were thoroughly cleaned with 70% ethanol between
trials to eliminate residual odors. The tracking software automatically quantified two
parameters. First, center zone exploration, defined as the percentage of time spent in
the  central  zone  (20  ×  20  cm).  Second,  total  distance  traveled  (cm),  representing
cumulative locomotor activity across the entire arena.

Y-Maze

The Y-maze comprised three identical opaque acrylic arms (length 50 cm × width 18
cm  ×  height  35  cm)  arranged  at  120°  angles.  Distinct  geometric  patterns  (triangle,
square, and circle) were affixed to each arm's distal end as visual-spatial cues. Ambient
illumination  was  maintained  at  40  lux,  verified  using  a  calibrated  light  meter.
Behavioral  sessions  were  recorded  and  analyzed  with  Noldus  Ethovision  XT  11.5
software (Noldus Information Technology, Netherlands). Mice were transported to the
behavioral room in covered cages and acclimated for 1 h. Each mouse was placed at
the distal end of a randomly selected arm, facing away from the center, and allowed 8
minutes of free exploration. An arm entry required all four paws to enter an arm. Three
parameters  were  quantified:  total  arm  entries  (complete  entries  during  the  session),
spontaneous alternation (sequential entries into three distinct arms without repetition),
and alternation score (calculated as [number of alternations / maximum alternations] ×
100%).

Novel object recognition test

The test was conducted in the same open-field arena (50 cm × 50 cm × 40 cm) used for
baseline behavioral assessments. Two identical objects (e.g., cylindrical batteries; 5 cm

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

height × 2 cm diameter) were placed symmetrically in the arena during training phases.
A novel object (distinct in shape/texture but matched for size) replaced one familiar
object in the final test phase. Behavioral data were recorded using Noldus Ethovision
XT 11.5 software (Noldus Information Technology, Netherlands).

Prior to testing, mice were transported to the behavioral testing room in covered

cages and allowed to acclimate for 1 hour. Twenty-four hours before formal testing,
each mouse was placed in the empty arena for a 5-min habituation session to reduce
environmental novelty. After 24 hours, two identical objects were symmetrically
positioned 15 cm from the arena walls, and mice were allowed to freely explore the
objects for 10 min. Twelve hours later, the same two objects were reintroduced, and
exploration behavior was recorded for 10 min to reinforce object familiarity. Six hours
after the second training session, one familiar object was replaced with a novel object
(position counterbalanced across subjects), and mice were tested for 10 min. Between
trials, the arena and objects were cleaned with 70% ethanol to eliminate residual odors.

Object exploration was operationally defined as direct sniffing or tactile contact with

the object, measured when the snout was within 2 cm of it. Novel object preference
index (PI) was calculated according to two distinct metrics. Specifically, time-based PI
represented the ratio of time spent exploring the novel object to total exploration time
(both objects), expressed as a percentage [(Novel object exploration time / Total
exploration time) × 100%]. Frequency-based PI reflected the proportion of contacts
made with the novel object relative to total contacts with both objects, calculated as
[(Number of novel object contacts / Total contacts) × 100%].

siRNA transfection

Primary  astrocytes  were  transfected  with  siRNA  using  Lipofectamine  LTX&PLUS
Reagent  (Thermo  Fisher  Scientific,  Cat#15338100)  following  the  manufacturer’s
protocol. Briefly, 1 hour prior to transfection, the culture medium was replaced with
Opti-MEM  Reduced-Serum  Medium  (Gibco,  Cat#31985062)  to  minimize  serum
interference. For each well of a 24-well plate, a transfection complex was prepared by
combining 50 nM siRNA (targeting

GSDME

Camk2a

; scrambled siRNA served as a

negative control), 1 μL PLUS Reagent, and 1 μL Lipofectamine LTX Reagent in 50 μL
Opti-MEM. The mixture was incubated at room temperature for 5 minutes to allow
complex formation and then added dropwise to the cells.

Lentivirus preparation and infection

Lentiviral  stocks  were  generated  by  co-transfecting  HEK293T  cells  with  the  target
plasmid  (pLV-GSDME-NTD-G4S-eGFP,  pLV-EMC10-GSDME-NTD-G4S-eGFP,
pLV-EMC10-GSDME-NTD(Δ56)-G4S-eGFP,  pLV-Emc10-eGFP,  pLKO.1-shRNA-
Scramble and pLKO.1-shRNA-Gsdme) and helper plasmids pCMV-Δ8.91, and pCMV-
VSV-G using CarpTrans (OPM, Cat#AC501302) as recommended by the manufacturer.
After 48 h post-transfection, lentivirus supernatant was filtered through 0.45 μm filter

The copyright holder for this

this version posted September 2, 2025.

http://bis.zju.edu.cn/MCA/

bioRxiv preprint

ENCODE transcriptome data.

brain were shown according to the published single-cell database

(

(D) Spatial distribution of GSDME in different astrocyte subtypes was analyzed by

open single-cell database.

(E) Correlation between

GSDME

and

GFAP

Slc1a3

Slc1a2

and

Adh1l1

is calculated

as Spearman correlation based on the single-cell RNA-seq expression data.

(F and G) Immunoblot analysis of GSDME expression in primary astrocytes isolated

from cerebral cortex (CC) or hippocampus (HC) of P1-3 mouse pups measured by

ImageJ based on gray scale (

= 3 replicates per group).

(H and I) Immunoblot analysis of GSDME expression in primary astrocytes (GFAP

microglia (Iba1

) and neurons (NeuN

) isolated from hippocampus DG region of P1-

3 mouse pups measured by ImageJ based on gray scale (

= 3 mice per group).

(J and K) Representative IF images of mouse hippocampus DG region co-stained with

GSDME, GFAP and either Iba1 or NeuN antibodies (J). Enlarged images of the boxed

area are shown. For colocalization, each dot represents two coronal sections of one

mouse (K). Scale bars, 40 μm.

(L and M) Representative immunofluorescence (IF) images of hippocampus region

from three human donors co-stained with anti-GSDME antibody and anti-GFAP

antibody. The yellow arrows indicate GFAP

GSDME

cells, while the blue arrows

indicate GFAP

GSDME

cells. Scale bars, 100 μm. Percentage of GSDME-positive

cells within GFAP

or GFAP

cell populations were quantified using Image J software.

(N-R) Comparison of GSDME mRNA levels between non-reactive (NR) and reactive

(R) astrocytes in snRNA-seq datasets of (N), AD (AD1

= 165,

= 51; AD2

= 177,

= 82; AD3

= 103,

= 61; AD4

= 367,

= 128) (D), PD

(

= 367,

= 31) (E), HD

(

= 367,

= 17) (F), ALS

(

= 106,

= 19) (G),

FTLD

(

= 82,

= 253). Significance levels are determined using a two-sided

Wilcoxon signed-rank test. All values are represented as mean ± SEM, ***

< 0.001.

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Table S1. The characteristics of human donors, related to

Figure 1

Non-AD

Non-PD

ALS

Non-ALS

Number (

)

Age (years)

90.67

3.61

77.67 ± 2.11

Brain region

Hippocampus

Temporal

lobes

Temporal

lobes

Sex

(male/female)

1/2

2/1

1/0

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Table S2. Primers used in this study

Oligonucleotides

SOURCE

IDENTIFIER

U6 snRNA

Forward:

ACGTTGGTGGATACGTGCAT

Reverse: CTCTTCCCACTTCTCGGCAG

This paper

N/A

hsa-miR-9-3p

Forward: TTTTGGTTATCTAGCTTATGA

Reverse: AGTGCAGGGTCCGAGGTATT

This paper

N/A

hsa-miR-200a-3p

Forward: TAACACTGTCTGGTAACGATGT

Reverse: AGTGCAGGGTCCGAGGTATT

This paper

N/A

hsa-miR-155-5p

Forward: TTAATGCTAATTGTGATAGGGGT

Reverse: AGTGCAGGGTCCGAGGTATT

This paper

N/A

hsa-miR-146a-5p

Forward: TGAGAACTGAATTCCATGGGTT

Reverse: AGTGCAGGGTCCGAGGTATT

This paper

N/A

hsa-miR-141-3p

Forward: TAACACTGTCTGGTAAAGATGG

Reverse: AGTGCAGGGTCCGAGGTATT

This paper

N/A

GSDME

Forward: ACGGACACCAATGTAGTGCTGG

Reverse: CTCTCATGCTCGAAGCCACCAT

This paper

N/A

Il-1α

Forward: CGCTTGAGTCGGCAAAGAAATC

Reverse: GAGAGAGATGGTCAATGGCAGAAC

This paper

N/A

tnf-α

Forward: TTCCCAAATGGCCTCCCTCTCATC

Reverse: TCCTCCACTTGGTGGTTTGCTAC

This paper

N/A

c1q

Forward: CCATATCGCGGCCTCTATGG

Reverse: GTCTTCCTGCCTCCCCTTTC

This paper

N/A

c1q

Forward: GGCCACCGACAAGAACTCACTAC

Reverse:

CCATATCTGGAAAGAGCAGGAACC

This paper

N/A

Gbp2

Forward: GGGGTCACTGTCTGACCACT

Reverse: GGGAAACCTGGGATGAGATT

This paper

N/A

Ggta1

Forward: GTGAACAGCATGAGGGGTTT

Reverse: GTTTTGTTGCCTCTGGGTGT

This paper

N/A

Ligp1

Forward: GGGGCAATAGCTCATTGGTA

Reverse: ACCTCGAAGACATCCCCTTT

This paper

N/A

H2-D1

Forward: TCCGAGATTGTAAAGCGTGAAGA

Reverse: ACAGGGCAGTGCAGGGATAG

This paper

N/A

H2-T23

Forward: GGACCGCGAATGACATAGC

Reverse: GCACCTCAGGGTGACTTCAT

This paper

N/A

Srgn

Forward: GCAAGGTTATCCTGCTCGGA

Reverse: TGGGAGGGCCGATGTTATTG

This paper

N/A

Psmb8

Forward: CAGTCCTGAAGAGGCCTACG

Reverse: CACTTTCACCCAACCGTCTT

This paper

N/A

Amigo2

Forward: GAGGCGACCATAATGTCGTT

Reverse: GCATCCAACAGTCCGATTCT

This paper

N/A

Ugt1a

Forward: CCTATGGGTCACTTGCCACT

Reverse: AAAACCATGTTGGGCATGAT

This paper

N/A

Serping1

Forward: ACAGCCCCCTCTGAATTCTT

This paper

N/A

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Reverse: GGATGCTCTCCAAGTTGCTC

Fkbp5

Forward: TATGCTTATGGCTCGGCTGG

Reverse: CAGCCTTCCAGGTGGACTTT

This paper

N/A

Fbln5

Forward: TATGCTTATGGCTCGGCTGG

Reverse: CAGCCTTCCAGGTGGACTTT

This paper

N/A

Gfap

Forward: AGAAAGGTTGAATCGCTGGA

Reverse: CGGCGATAGTCGTTAGCTTC

This paper

N/A

Forward: TGTGATGGGAATGGGCAATGA

Reverse: AGTGTATAGAGGATGTTCCAGGTCA

This paper

N/A

S1pr3

Forward: AAGCCTAGCGGGAGAGAAAC

Reverse: TCAGGGAACAATTGGGAGAG

This paper

N/A

Cd44

Forward: ACCTTGGCCACCACTCCTAA

Reverse: GCAGTAGGCTGAAGGGTTGT

This paper

N/A

Lcn2

Forward: CCAGTTCGCCATGGTATTTT

Reverse: CACACTCACCACCCATTCAG

This paper

N/A

Timp1

Forward: AGTGATTTCCCCGCCAACTC

Reverse: GGGGCCATCATGGTATCTGC

This paper

N/A

Vim

Forward: AGACCAGAGATGGACAGGTGA

Reverse: TTGCGCTCCTGAAAAACTGC

This paper

N/A

Aspg

Forward: GCTGCTGGCCATTTACACTG

Reverse: GTGGGCCTGTGCATACTCTT

This paper

N/A

Serpina3n

Forward: CCTGGAGGATGTCCTTTCAA

Reverse: TTATCAGGAAAGGCCGATTG

This paper

N/A

Hspb1

Forward: GACATGAGCAGTCGGATTGA

Reverse: GGATGGGGTGTAGGGGTACT

This paper

N/A

Steap4

Forward: CCCGAATCGTGTCTTTCCTA

Reverse: GGCCTGAGTAATGGTTGCAT

This paper

N/A

Osmr

Forward: GTGAAGGACCCAAAGCATGT

Reverse: GCCTAATACCTGGTGCGTGT

This paper

N/A

Cd14

Forward: GGACTGATCTCAGCCCTCTG

Reverse: GCTTCAGCCCAGTGAAAGAC

This paper

N/A

Cd109

Forward: CACAGTCGGGAGCCCTAAAG

Reverse: GCAGCGATTTCGATGTCCAC

This paper

N/A

Clcf1

Forward: CTTCAATCCTCCTCGACTGG

Reverse: TACGTCGGAGTTCAGCTGTG

This paper

N/A

Ptgs2

Forward: GCTGTACAAGCAGTGGCAAA

Reverse: CCCCAAAGATAGCATCTGGA

This paper

N/A

Sphk1

Forward: GATGCATGAGGTGGTGAATG

Reverse: TGCTCGTACCCAGCATAGTG

This paper

N/A

Tm4sf1

Forward: GCCCAAGCATATTGTGGAGT

Reverse: AGGGTAGGATGTGGCACAAG

This paper

N/A

Tgm1

Forward: CTGTTGGTCCCGTCCCAAA

Reverse: GGACCTTCCATTGTGCCTGG

This paper

N/A

Mertk

Forward: ACGTTGGTGGATACGTGCAT

This paper

N/A

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Reverse: CTCTTCCCACTTCTCGGCAG

Megf10

Forward: AGCGCCTATGATGGGGAATG

Reverse: ATAGCCGTATGTTCCCGCAG

This paper

N/A

Gas6

Forward: ATGAAGATCGCGGTAGCTGG

Reverse: CCAACTCCTCATGCACCCAT

This paper

N/A

Axl

Forward: TTCAACTGTGCTACGTCCCC

Reverse: GGGTCCCTCTAGGTAAGCCA

This paper

N/A

Gpc4

Forward: CTGCTTTCCATCGGGTCTCA

Reverse: CGCACTTCCGAGCAACTTTT

This paper

N/A

Gpc6

Forward: GAGGACTGTGTGGAGGACCAT

Reverse: GTTCCTGCTGTCACACGTTC

This paper

N/A

Sparc

Forward: ACCTGGACTACATCGGACCA

Reverse: GTGGGACAGGTACCCATCAA

This paper

N/A

Thbs1

Forward: TGACAATTTTCAGGGGGTGCT

Reverse: ACCACGTTGTTGTCAAGGGT

This paper

N/A

Sparcl1

Forward: GCAGACAACCAAGAGGCCAA

Reverse: GGTTTCCCTTGTGGATCGGT

This paper

N/A

Thbs1

Forward: CAATTTTCAGGGGGTGCTGC

Reverse: CCGTTCACCACGTTGTTGTC

This paper

N/A

Crel

Forward: ACCAGAACGCAGACCTTTGT

Reverse: AGTATTTGGGGCACGGTTGT

This paper

N/A

Ikb

Forward: ACCTCAATAAACCGGAGCCT

Reverse: ATCGTCCTCGTCCTCAGGTT

This paper

N/A

Ikk

Forward: CCCAAAGAACAGAGACCGCT

Reverse: TCTAAGAGCCGATGCGATGTC

This paper

N/A

p50

Forward: CCTGGACGACTCTTGGGAGA

Reverse: TCCCGGAGTTCATCTCATAGT

This paper

N/A

p65

Forward: CACACCCCACCATCAAGATCA

Reverse: CTCTATAGGAACTATGGATACTGCG

This paper

N/A

IL-6

Forward: CCACTTCACAAGTCGGAGGCTTA

Reverse: TGCAAGTGCATCATCGTTGTTC

This paper

N/A

Forward: AGCTTCAGGGTCCCAGCTAC

Reverse: CTGGAATCTTGATGGAGACGC

This paper

N/A

NOS2

Forward: GGAGCATCCCAAGTACGAGT

Reverse: ATTGATCTCCGTGACAGCCC

This paper

N/A

Calm3

Forward: GACAAGGATGGAGATGGCACCA

Reverse: GAACTCTGGGAAGTCAATGGTCC

This paper

N/A

il-1β

Forward: TGTGTAATGAAAGACGGCACACC

Reverse: GTATTGCTTGGGATCCACACTCTC

This paper

N/A

HMGB1

Forward: ACTGACTGAGCTCATCACGT

Reverse: ACATCTCGATGATCGACGGA

This paper

N/A

Elovl1

Forward: CCTCGAATCATGGCTAATCG

Reverse: CGAACCATCCGAAGTGCTTC

This paper

N/A

ApoE

Forward:

GAACCGCTTCTGGGATTACCTG

This paper

N/A

The copyright holder for this

this version posted September 2, 2025.

bioRxiv preprint

Reverse:

GCCTTTACTTCCGTCATAGTGTC

ApoJ

Forward:

GATGATCCACCAGGCTCAACAG

Reverse:

ACACAGTGCGGTCATCTTCACC

This paper

N/A

ifn-γ

Forward: GAGGAACTGGCAAAAGGATG

Reverse: TGAGCTCATTGAATGCTTGG

This paper

N/A

il-4

Forward: ATCATCGGCATTTTGAACGAGG

Reverse: TGCAGCTCCATGAGAACACTA

This paper

N/A

il-7

Forward: TCTAATGGTCAGCATCGATCA

Reverse: GTGGAGATCAAAATCACCAG

This paper

N/A

il-8

Forward: CGGCAATGAAGCTTCTGTAT

Reverse: CCTTGAAACTCTTTGCCTCA

This paper

N/A

il-9

Forward: AATGCCACACAGAAATCAAGAC

Reverse: ACACGTGATGTTCTTTAGGACT

This paper

N/A

il-10

Forward: CTAACGGAAACAACTCCTTG

Reverse: GAAAGGACACCATAGCAAAG

This paper

N/A

il-15

Forward: ATCGCCATAGCCAGCTC

Reverse: ATGAATGCCAGCCTCAGT

This paper

N/A

g-csf

Forward: AGCCCAGATCACCCAGAATC

Reverse: TCTTCCTCACTTGCTCCAGG

This paper

N/A

App

Forward:

AGAAGGACAGACAGCACACC

Reverse:

TCATAACCTGGGACCGGATCT

This paper

N/A

Tau

Forward: GAACCACCAAAATCCGGAGA

Reverse: CTCTTACTAGCTGATGGTGAC

This paper

N/A

GAPDH

Forward: CATCACTGCCACCCAGAAGACTG

Reverse: ATGCCAGTGAGCTTCCCGTTCAG

This paper

N/A

The copyright holder for this

this version posted September 2, 2025.