Microglia

The resident innate immune cells of the CNS (brain + spinal cord). Microglia are ontogenetically and functionally distinct from peripheral macrophages: they derive from yolk-sac primitive macrophages before embryonic day 8 and self-renew locally throughout life without significant contribution from circulating monocytes ¹. With aging, microglia undergo profound functional deterioration — becoming pro-inflammatory, morphologically dystrophic, and phagocytically impaired — and are central contributors to age-related neurodegeneration including alzheimers-disease, Parkinson’s disease, and ALS.

Identity

Ontogeny

Microglia are the only CNS-resident macrophages with a yolk-sac origin: they derive from primitive myeloid progenitors that colonize the brain rudiment before embryonic day 8 in mice, prior to the establishment of definitive hematopoiesis in the fetal liver ¹. This distinguishes them from all other CNS macrophages (perivascular, meningeal, and choroid plexus macrophages), which are partially replaced by monocyte-derived cells in adulthood.

Under homeostatic conditions, peripheral monocytes do not significantly replenish the microglial pool. Microglia maintain themselves by slow, local proliferation throughout the lifespan, with a turnover half-life estimated at ~100 days in adult mice needs-human-replication.

Dimension	Status
Ontogeny conserved in humans?	yes
Self-renewal from monocytes excluded in humans?	partial (human data are indirect; monocyte exclusion best established in mouse)
Replicated in humans?	partial — indirect evidence from CNS transplant studies

Canonical surface markers (mouse and human)

Marker	Notes
CD11b (ITGAM)	Pan-myeloid; microglia are CD11b+ CD45-low (distinguishes from infiltrating macrophages, which are CD45-high)
CX3CR1	Fractalkine receptor; high on homeostatic microglia; reduced in activated states
P2RY12	Purinergic receptor; homeostatic marker; lost in activated/disease-associated microglia
TMEM119	Highly specific homeostatic marker; lost upon activation
IBA1 (AIF1)	Pan-microglial; retained across activation states; standard IHC marker
TREM2	Triggering receptor expressed on myeloid cells; upregulated in disease-associated microglia (DAM)

Distribution and niche

Microglia constitute approximately 5–12% of all brain cells in adult mice (human estimates vary by region, ~0.5–16% of cells depending on method). Regional density is highest in areas vulnerable to neurodegeneration: substantia nigra, hippocampus, basal ganglia, and olfactory bulb. Cerebellar cortex and white matter have lower densities. unsourced — precise human regional density data across age are incomplete.

Each microglial cell surveys a discrete, largely non-overlapping territory via motile ramified processes, with the entire brain parenchyma covered by collective microglial surveillance at any given time.

Core functions

Phagocytosis and debris clearance

Microglia are the primary phagocytes of the CNS, clearing:

Apoptotic cells and cellular debris
Amyloid-beta (Aβ) oligomers and plaques
Myelin debris after axonal injury
Dead or dying neurons

Phagocytic capacity is critically dependent on TREM2 signaling: TREM2 promotes microglial metabolic reprogramming and migration toward amyloid deposits. TREM2 loss-of-function impairs Aβ clearance and accelerates plaque accumulation in mouse AD models ².

Synapse pruning

During development, microglia mediate complement-dependent synapse elimination via C1q/C3 opsonization and CR3-mediated phagocytosis (Beth Stevens lab work). This process normally refines neural circuits but is aberrantly reactivated in aging and neurodegeneration no-mechanism — the triggers for inappropriate synapse pruning in the aged brain are not fully established.

Cytokine secretion and immunosurveillance

Homeostatic microglia continuously survey the parenchyma and respond to PAMPs and DAMPs by secreting:

Pro-inflammatory: TNF-α, IL-1β, IL-6, IL-12
Anti-inflammatory/trophic: IL-10, TGF-β, BDNF, IGF-1

The balance between these programs is lost in aging (see Aging Features below).

Aging features

Morphological changes

Aged microglia adopt dystrophic morphology: shortened, de-ramified, beaded, or fragmented processes; cytoplasmic swelling; reduced process surveillance velocity. This contrasts with activated microglia (which show amoeboid retraction) and represents a distinct aged state. Dystrophic microglia are prominent in aged human brain post-mortem tissue, where they co-localize with Aβ plaques and tau tangles. unsourced — quantitative morphological studies across human lifespan are sparse.

Primed / exaggerated activation

Aged microglia become “primed” — displaying a baseline-elevated inflammatory tone and mounting an exaggerated, prolonged response to immune challenge ³:

Elevated resting levels of IL-1β, TNF-α, and IL-6 in aged vs. young brain
Following peripheral LPS challenge, aged microglia produce ~2–3× more IL-1β than young microglia (mouse; quantification from Norden 2013 review — primary sources cited therein not verified) no-fulltext-access — Norden 2013 full-text not accessible via archive (download failed); quantitative claim unverified against review source
Prolonged microglial activation after challenge correlates with intensified sickness behavior and cognitive decline in aged mice needs-human-replication

Impaired phagocytosis

Aged microglia show reduced phagocytic uptake of Aβ, myelin debris, and apoptotic cells relative to young microglia, despite increased pro-inflammatory output. This combination — heightened cytokine secretion + impaired clearance — is mechanistically central to plaque accumulation in Alzheimer’s disease.

Transcriptional heterogeneity revealed by scRNA-seq

Single-cell RNA-seq has resolved multiple functionally distinct microglial states across aging. Hammond et al. 2019 profiled 76,149 cells across development (E14.5–P540) and identified 9 transcriptionally distinct states ⁴; Keren-Shaul et al. 2017 identified disease-associated microglia (DAM) in an AD mouse model ⁵:

State	Key markers	Biology	Source
Homeostatic	P2RY12+, TMEM119+, CX3CR1-high	Surveillance; tissue maintenance	Both
DAM (Disease-Associated Microglia)	TREM2+, ApoE+, Lpl+, Cst7+, P2RY12-low	Activated around Aβ plaques; phagocytic; restricts plaque spread	⁵
Aging OA2	Ccl4+, Il1b+, Lpl+, Cst7+; 2–4× increase at P540 vs P100	Pro-inflammatory; increased in aged brain	⁴
Aging OA3 (interferon-response)	Ifitm3+, Irf7+, Rtp4+; small subset	Type-I IFN signaling; increased with age	⁴
Proliferating	Rrm2+, Ube2c+, Birc5+, Cenpa+	Reactive expansion; prominent in development and after injury	⁴
Injury-responsive	Apoe+, Cst7+, Lpl+, Axl+, P2ry12-low	Activated by LPC-demyelination; overlaps DAM transcriptionally	⁴

DAM emerge in a two-step process: a TREM2-independent first step (homeostatic → Stage 1 DAM, marked by downregulation of P2RY12 and Cx3cr1 and upregulation of Apoe) followed by a TREM2-dependent second step (Stage 1 → Stage 2 DAM, requiring TREM2 signaling for full upregulation of Lpl, Cst7, and Trem2 itself) ⁵. needs-replication — DAM two-step model from Keren-Shaul 2017 PDF not yet independently verified (download failed).

Dimension	Status
DAM state identified in humans?	yes — analogous TREM2-hi/P2RY12-low cluster in human AD post-mortem scRNA-seq
DAM protective or damaging?	contested — early DAM may be protective (clearance); chronic DAM activation may worsen inflammation contradictory-evidence
Replicated across labs?	yes (DAM replicated; additional substates vary by atlas)

Microglial senescence

Aged microglia share features with senescent cells (SA-β-gal activity, p21/p16 upregulation, SASP-like cytokine secretion) and have been classified as “senescent-like” in some studies. However, whether aged/dystrophic microglia meet the canonical definition of cellular senescence (permanent cell-cycle arrest + SASP) is debated — microglia are slowly proliferating, not post-mitotic, complicating the senescence designation. contradictory-evidence unsourced — dedicated studies quantifying p16+ microglia in aged human brain are lacking.

The functional overlap with chronic-inflammation is clear: aged/senescent-like microglia contribute substantially to CNS inflammaging — the chronic low-grade neuroinflammation that accelerates aging across many brain regions.

TREM2 and Alzheimer’s disease risk

TREM2 R47H is among the highest-impact single-gene risk variants for late-onset Alzheimer’s disease. Jonsson et al. found OR 2.90 (95% CI 2.16–3.91, P=2.1×10⁻¹², combined Icelandic + replication cohorts) for rs75932628-T ². Guerreiro et al. confirmed this in an independent European case-control series (1092 AD cases + 1107 controls discovery set; 1887 cases + 4061 controls replication), finding OR 4.5 (95% CI 1.7–11.9) for R47H specifically ⁶. The two independent concurrent papers were published simultaneously in the same issue of NEJM (January 10, 2013, Vol. 368:2).

Dimension	Status
TREM2 R47H effect size replicated?	yes — multiple independent cohorts
Mechanism (how R47H impairs microglia)?	partial — reduced TREM2 ligand binding and downstream PI3K/AKT signaling; impaired phagocytosis and metabolic fitness in microglia; precise human mechanism still under study
Therapeutic target status?	active — anti-TREM2 agonist antibodies and TREM2 mimetics in Phase 1–2 trials

TREM2 is expressed almost exclusively by microglia in the brain; TREM2 risk variants thus directly implicate microglial dysfunction in AD pathogenesis. See alzheimers-disease for broader disease context.

Disease relevance

Alzheimer’s disease

Microglia surround and partially phagocytose Aβ plaques in early AD; insufficient clearance allows plaque growth.
Tau pathology triggers a distinct microglial activation program, potentially amplifying neuronal loss.
GWAS hits enriched in myeloid-expressed genes (TREM2, CR1, CLU, BIN1) indicate microglial biology is genetically central to AD risk.
DAM state may be initially protective (Keren-Shaul 2017) but chronic activation contributes to synaptic damage and neuronal loss.

Parkinson’s disease

Substantia nigra microglia are activated in PD; α-synuclein aggregates act as DAMPs to trigger TLR2/4-dependent microglial activation.
Activated microglia produce NO and ROS that accelerate dopaminergic neuron loss. unsourced — human in vivo evidence limited; mechanism established in vitro and in rodent models.

ALS

Spinal cord microglia shift from an early neuroprotective to a late-stage neurotoxic profile during ALS progression in SOD1 mouse models. needs-human-replication

Therapeutic relevance: microglial repopulation

CSF1R inhibitors (e.g., PLX3397, PLX5622) deplete ~99% of brain microglia within ~3 weeks of dietary administration (PLX3397 at 290 mg/kg chow) by blocking the survival signal from CSF1R (c-Fms) ⁷. Upon inhibitor withdrawal, repopulation begins within 48–72 hours via proliferation of nestin-expressing progenitor cells found throughout the CNS; microglial numbers return to control levels within ~7 days ⁷.

Rationale for repopulation therapy in aging:

Replace aged, dysfunctional microglia with a “young” functional population
Eliminate accumulated epigenetic damage in the long-lived microglial pool

Current evidence:

Elmore 2014 (PLX3397) showed that microglia depletion in young adult mice causes no cognitive deficits and mildly enhanced Barnes maze learning; cognitive benefits from repopulation in aged mice are from subsequent work by the Green lab and others, not from Elmore 2014 itself. needs-replication unsourced — aged-mouse repopulation benefit claims require dedicated citation
Long-term cognitive effects and optimal repopulation protocols under active investigation. long-term-unknown
No human trials of CSF1R inhibition for microglial repopulation as of 2026-05-04 (CSF1R inhibitors are in trials for other CNS indications, e.g., ALSP). needs-human-replication

Dimension	Status
Repopulation demonstrated in mouse?	yes ⁷
Cognitive benefit in aged mouse?	partial — some studies positive; heterogeneous results across labs
Human trial data?	none (as of 2026-05-04)

Limitations and open questions

Human data lag: most mechanistic microglial aging data are from mouse models. Human post-mortem scRNA-seq studies are growing but are confounded by agonal state and post-mortem interval. needs-human-replication
Senescence vs. priming vs. dystrophy: no consensus on whether these represent distinct states or a spectrum; lack of uniform markers hampers comparison across studies. contradictory-evidence
Sex differences: microglial density, transcriptome, and aging trajectory differ between male and female mice; extent in humans unclear. unsourced
DAM functional interpretation: whether DAM is protective (restrains plaque expansion, Keren-Shaul 2017) or damaging (drives neuroinflammation) remains contested and likely context-dependent. contradictory-evidence
Repopulation therapy durability: repopulated microglia age at the same rate as native microglia; whether iterative repopulation can extend CNS healthspan is unknown. long-term-unknown

Cross-references

chronic-inflammation — CNS inflammaging is largely microglial-mediated
cellular-senescence — aged microglia share senescent-like features; relationship debated
alzheimers-disease — TREM2 genetics, DAM, Aβ clearance
neurodegeneration — shared microglial contributions across AD, PD, ALS
trem2 — the key microglial risk gene for AD (protein page; implicit stub)
csf1r — survival receptor; target for microglial depletion/repopulation (implicit stub)
inflammaging — systemic and CNS chronic inflammation in aging

Footnotes

doi:10.1126/science.1194637 · Ginhoux F et al. · in-vivo (mouse fate mapping) · model: Runx1-MERCre-MER × Rosa26^R26R-eYFP lineage tracing; progenitors arise before E8.0, seed brain via blood vessels E8.5–E9.5 · foundational paper establishing yolk-sac origin of adult microglia · cited_by: ~4919 ↩ ↩²
doi:10.1056/NEJMoa1211103 · Jonsson T et al. · observational (genome sequencing + case-control) · model: Icelandic population cohort + replication in US, Norway, Netherlands, Germany · TREM2 R47H (rs75932628) OR 2.92 (95% CI 2.09–4.09) discovery; OR 2.90 (95% CI 2.16–3.91, P=2.1×10⁻¹²) combined · cited_by: ~2511 · note: OR confirmed via abstract (full-text PDF download stalled) ↩ ↩²
doi:10.1111/j.1365-2990.2012.01306.x · Norden DM, Godbout JP · review (published 2012/2013, Neuropathology and Applied Neurobiology) · model: mouse + review of aging neuroinflammation literature · cited_by: ~744 · no-fulltext-access — PDF download failed; quantitative IL-1β claims not verified against full text ↩
doi:10.1016/j.immuni.2018.11.004 · Hammond TR et al. · in-vivo (mouse scRNA-seq across lifespan) · model: C57BL/6 mice ages E14.5, P4/5, P30, P100, P540 + LPC-injury; n=76,149 cells from 41 mice · identified 9 transcriptionally distinct microglia states; aging enriched OA2 (Ccl4+, Il1b+) and OA3 (interferon-response: Ifitm3+, Irf7+) clusters · cited_by: ~2166 ↩ ↩² ↩³ ↩⁴ ↩⁵
doi:10.1016/j.cell.2017.05.018 · Keren-Shaul H et al. · in-vivo (mouse scRNA-seq) · model: 5XFAD Alzheimer mouse model · identified DAM two-step transition (Stage 1 TREM2-independent; Stage 2 TREM2-dependent) · cited_by: ~5115 · no-fulltext-access — PDF download failed (Cell.com 403); n-cells figure and precise DAM marker list not verified against primary source ↩ ↩² ↩³
doi:10.1056/NEJMoa1211851 · Guerreiro R et al. · observational (case-control; exome + Sanger sequencing) · model: European case-control; 1092 AD cases + 1107 controls (discovery); 1887 AD cases + 4061 controls (replication) · TREM2 R47H OR 4.5 (95% CI 1.7–11.9); all-variant combined OR 4.59 (95% CI 2.49–8.46, P=9×10⁻⁹) · confirmed TREM2 risk in independent population · cited_by: ~2949 ↩
doi:10.1016/j.neuron.2014.02.040 · Elmore MR et al. · in-vivo (mouse) · model: C57BL/6 adult mice (2–18 months); PLX3397 (290 mg/kg chow) as primary in vivo CSF1R inhibitor · ~99% microglial elimination by 21 days; repopulation from nestin+ progenitors begins 48–72 hr after drug withdrawal; full numerical recovery by ~7 days · cited_by: ~1920 ↩ ↩² ↩³

🧬 Aging Wiki

Explorer

microglia