cst7

CST7 (cystatin F / leukocystatin)

CST7 encodes cystatin F (also leukocystatin), a secreted and lysosomal glycoprotein of the cystatin superfamily that inhibits cysteine endopeptidases — primarily cathepsins L, F, K, and V — in haematopoietic cells and their tissue-resident derivatives including microglia. It is a canonical transcriptomic marker of disease-associated microglia (DAM) and is strongly induced in neurodegeneration and demyelination. In the aging field, Cst7 has emerged as one of the universal conserved mammalian genes upregulated with both chronological age and expected mortality, naming alongside GPNMB and CDKN1A as a core feature of the multi-species transcriptomic mortality clock.

Identity

Field	Value
UniProt	O76096 (CST7_HUMAN)
NCBI Gene	8530
HGNC symbol	CST7
HGNC ID	HGNC:2479
Ensembl	ENSG00000077984
Chromosomal location	20p11.21
Mouse ortholog	Cst7 (NCBI Gene 13011; one-to-one)
GenAge	not listed

Biochemistry and structure

Cystatin F is a secreted glycoprotein with two N-linked glycosylation sites (Asn-62, Asn-115) and a 19-amino-acid signal peptide (residues 1–19); the mature protein is 126 amino acid residues (~15.5 kDa unglycosylated per UniProt O76096; ~17 kDa glycosylated monomer by gel; extracellular dimer ~35 kDa by non-denaturing gel).¹ Unlike most cystatins (which circulate as monomers), the secreted form of cystatin F exists predominantly as a disulfide-linked covalent dimer that is inhibitory-inactive; the dimer is internalized by neighbouring cells via M6P-mediated endocytosis and trafficked to the endolysosomal system, where the lysosomal protease cathepsin V cleaves 15 N-terminal amino acids from the full-length monomer to generate the fully active inhibitory monomeric form.¹ This proform-activation switch is unusual within the cystatin superfamily and provides a mechanism for spatial and temporal control of lysosomal cysteine-protease activity.

Canonical inhibitory targets per Wang 2025 (the most comprehensive summary review to date):

The fully processed active monomer tightly inhibits cysteine endopeptidases — cathepsins L, F, K, and V — with weaker interactions with cathepsins S and H. Critically, cathepsins B, C, and X (exopeptidases) are not inhibited by the active monomeric form.¹ UniProt O76096 independently confirms inhibition of papain and cathepsin L, with lower affinity than other cystatins, noting a potential role in immune regulation via “inhibition of a unique target in the hematopoietic system.”

The connection to cathepsin C activity is indirect and context-dependent: the extracellular dimer can physically interact with cathepsin C in secretory contexts (e.g., neutrophil degranulation), and indirect effects on cathepsin C-dependent pathways (granzyme activation in CTL/NK cells) are documented in several reports — but direct intracellular inhibition of cathepsin C by the active monomer is not supported by Wang 2025 or by functional KO data in Daniels 2023 (which found no change in intracellular cathepsin C activity on Cst7 deletion).²¹

• Legumain (asparaginyl endopeptidase / AEP) — some reviews list legumain as a target; not confirmed by Wang 2025’s target enumeration for the active monomer. needs-verification

Subcellular distribution (per UniProt O76096): secreted to the extracellular space, but also found in lysosomes, endosomes, ER, Golgi, and multivesicular bodies — reflecting the proform biosynthetic pathway and active-monomer sequestration to the endolysosomal system.

Expression pattern

CST7/Cst7 is expressed predominantly in haematopoietic tissues (peripheral blood cells, spleen) and their tissue-resident derivatives — particularly NK cells, cytotoxic T lymphocytes (CTL), and myeloid-lineage cells including microglia.¹ Expression in non-immune tissues is low under homeostatic conditions. This immune-restricted expression distinguishes it from the housekeeping cystatins (CST3 cystatin C, CST1 cystatin SN) and underlies its roles in regulating cytotoxic immune responses. needs-verification

In the central nervous system, microglial expression is low in homeostasis and strongly induced during disease — CST7 is one of the defining transcriptomic markers of the disease-associated microglia (DAM) state first described in Alzheimer’s disease models.¹

Why CST7 matters for aging

Universal mortality gene (Tyshkovskiy 2026)

In the Tyshkovskiy et al. 2026 multi-species transcriptomic-clock study, Cst7 was identified as one of the conserved genes positively associated with both chronological ageing and expected mortality across mouse, rat, crab-eating macaque, and human, contributing positive clock coefficients to the elastic-net mortality-clock model.³ The paper places Cst7 alongside Gpnmb and Cdkn1a/p21 as the headline universal upregulated markers — genes reliably rising with both calendar age and age-adjusted death probability in the multi-species dataset.⁴

This positions Cst7 as part of the chronic-inflammation and immune-dysregulation co-expression module whose upregulation tracks mortality more faithfully than chronological-age clocks alone. See transcriptomic-clock-tage for the clock architecture and module-level context.

Unlike gpnmb and CDKN1A, circulating CST7 protein was not independently validated against UK Biobank mortality outcome data in this study — the plasma proteomics validation was specifically for GPNMB, CDKN1A, and LGALS3 (Olink panel coverage). Whether plasma cystatin F is an independent human mortality biomarker remains to be tested. needs-human-replication

Disease-associated microglia (DAM) and neuroinflammation

CST7/Cst7 is a canonical DAM marker: single-cell RNA-seq studies of murine Alzheimer’s disease models systematically identify it as one of the most upregulated genes in the shift from homeostatic to disease-associated microglial states.¹ Its induction is interpreted as part of a lysosomal-activation programme enabling microglia to handle increased phagocytic load (amyloid-beta, myelin debris, apoptotic cells).

However, functional studies reveal that the net effect of cystatin F upregulation on neuroinflammation is context- and sex-dependent rather than uniformly protective or pathological:

• In a 5xFAD amyloid-driven Alzheimer’s model (App^NL-G-F^ knock-in mice, 12 months, both sexes), Cst7 deletion produced sexually dimorphic outcomes: in females, loss of Cst7 enhanced endolysosomal gene expression, increased lysosomal burden, and increased microglial Aβ uptake in vivo — paradoxically increasing amyloid-beta burden in the subiculum specifically (significant; cortex and hippocampus: non-significant trend); in males, Cst7 deletion reduced inflammatory signalling (Il1b, Il1a, Tnf, Cxcl2 downregulated) without any change in Aβ burden in any region.² Mechanistically, the increased Aβ burden in females appears to be due to increased phagocytosis of Aβ “seeding” plaques rather than impaired degradation — intracellular cathepsin L and C activities were unchanged by Cst7 KO.² The Nature Reviews Immunology commentary by Baleviciute and Keane 2023 highlighted these data as evidence that microglial CST7 function is sex-specific.⁵
• In a murine coronavirus CNS infection model (JHMV, C57BL/6 mice, intracerebral), Cst7-KO mice controlled viral replication normally (no difference in viral titers at day 7 p.i. vs. controls). However, Cst7-deficient mice developed more severe demyelination and impaired remyelination during chronic disease (days 14–21 p.i.), correlating with increased T-cell infiltration and elevated T-cell chemoattractant chemokines Cxcl9 and Cxcl10, and elevated IFN-γ and perforin in CD8+ T cells — suggesting Cst7 normally dampens maladaptive T-cell amplification in demyelinating contexts without being essential for anti-viral host defence.⁶
• Monocyte-derived cystatin F (as a secreted dimer) was shown to physically associate with amyloid-beta, impairing monocyte phagocytosis and worsening amyloid pathology and cognitive deficits in mouse AD models.⁷ This monocyte-sourced pool of cystatin F may act differently from the microglial-intrinsic pool.

The review by Wang et al. 2025 summarises the current picture: “cystatin F is significantly upregulated in several CNS diseases” and “its role differs across disease contexts — serving a neuroprotective function while promoting pathological progression” depending on which cell type expresses it and which disease process is active.¹

Aging neuroinflammation

The combination of (a) its universal upregulation with chronological age/mortality in multi-tissue transcriptomic data and (b) its status as a DAM marker implicates cystatin F as a molecular node linking aging-associated chronic neuroinflammation to declining microglial lysosomal efficiency. The working model is that age-dependent upregulation of Cst7 in microglia suppresses cathepsin C-dependent protease cascades, progressively impairing phagolysosomal digestion capacity and contributing to chronic-inflammation module activation. This is mechanistically plausible but not yet demonstrated by a causal experiment with age-specific Cst7 manipulation (independently of disease context). no-mechanism

Druggability (aging-context)

Tier 4 — undruggable in the current aging-intervention landscape.

CST7 is a secreted inhibitory protein rather than an enzymatic target. There are no clinical drugs, approved probes, or advanced-stage biologics targeting cystatin F for any indication as of 2026-05-29. Open Targets Platform was not directly queried (API 400/500 errors); tier 4 is assessed on absence of any clinical compound and the structural basis (endogenous inhibitor, no binding pocket for small molecules).

Wang 2025 catalogues several preclinical-only research strategies: RIPK1 pathway inhibition (necrostatin-1s), miRNA targeting (anti-miR-29a), mRNA stabilization (targeting ELAVL1), necroptosis inhibition (NSA for PD models), and A2AR modulation — all animal/in-vitro stage, no clinical agents.¹ The note about brensocatib (a cathepsin C inhibitor in bronchiectasis) is removed — cathepsin C is not a direct monomeric substrate of cystatin F (see Biochemistry section above), making this connection speculative. needs-verification on Open Targets Platform direct query.

The sexual dimorphism and context-dependence described above make the druggability path particularly unclear: the protective vs. pathological effects of cystatin F may be cell-type and disease-state specific, such that a blanket inhibitor or inducer would carry unpredictable effects. dose-response-unclear

Relationship to other aging markers

Marker	Relationship to Cst7 / CST7
gpnmb	Co-top universal mortality gene (Tyshkovskiy 2026); both are DAM-expressed; GPNMB has UK Biobank plasma validation, CST7 does not yet
p21 (CDKN1A)	Co-top universal mortality gene; mechanistically upstream (p53 → p21 → cell-cycle arrest in senescent cells); not co-expressed with Cst7 in DAM specifically
lgals3	Third headline mortality gene; galectin-3 is also myeloid-enriched and has plasma mortality validation
chronic-inflammation	Cst7/CST7 is a positive contributor to inflammation + immune modules in the Tyshkovskiy mortality clock; DAM induction is part of the aging neuroinflammatory programme
disabled-adaptive-immunity	Via indirect effects on cathepsin C-dependent pathways: cystatin F in extracellular / secreted contexts can suppress cathepsin C activity and thereby dampen granzyme activation in NK cells / CTL (Wang 2025); direct intracellular monomer-mediated cathepsin C inhibition is not supported by current evidence (active monomer targets cathepsins L/F/K/V, not C)

Evidence extrapolation

Dimension	Status
Pathway conserved in humans?	yes — cystatin family proteins and cathepsin targets are conserved; human CST7 and mouse Cst7 are one-to-one orthologs
Age-upregulation conserved in humans?	yes — Tyshkovskiy 2026 multi-species transcriptomic clock shows conservation across rodents and primates including human tissues
Plasma/protein validated in humans?	no — UK Biobank proteomics validation in Tyshkovskiy 2026 covered GPNMB/CDKN1A/LGALS3 but not CST7; needs-human-replication
Causal for mortality?	not tested — transcriptomic associations are observational; no MR study exists

Limitations and gaps

• No plasma protein validation in humans — the Tyshkovskiy 2026 transcriptomic data are human-tissue RNA-seq; circulating CST7 protein as a mortality biomarker is untested. needs-human-replication
• No aging-specific causal experiment — existing Cst7 KO studies are in disease models (AD, coronavirus infection), not aged WT mice. Whether reducing or eliminating Cst7 in aged-only animals affects neuroinflammation or survival is unknown. no-mechanism
• Sexual dimorphism unresolved at the mechanistic level — why Cst7 has opposite effects in male and female microglia is not known; sex-specific studies dominate the recent literature but no unifying mechanism is established. no-mechanism
• gtex-aging-correlation not populated — GTEx tissue-by-age expression summary not retrieved; for verifier pass. needs-verification
• mr-causal-evidence: not-tested — no instruments from germline GWAS at the CST7 locus have been used in a Mendelian randomization study for an aging endpoint to date.
• druggability-tier: 4 assigned based on absence of clinical drug or probe; Open Targets Platform not independently queried — verify on verifier pass. needs-verification

Related pages

gpnmb · p21 · lgals3 · transcriptomic-clock-tage · tyshkovskiy-2026-universal-transcriptomic-hallmarks · chronic-inflammation · disabled-adaptive-immunity · microglia

Footnotes

1. doi:10.1186/s12974-025-03526-z · Wang YT et al. · J Neuroinflammation 2025 · 22(1):203 · review · comprehensive review of cystatin F structure, function and CNS disease roles; active monomer inhibits cathepsins L/F/K/V (not B/C/X); dimer inactive until endolysosomally processed by cathepsin V; upregulated in AD, MS, PD, ALS, stroke, AGS, prion disease, GBM; role context-dependent (neuroprotective vs. pathological depending on disease context and cell type) · model: review · verified 2026-05-29 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

2. doi:10.7554/eLife.85279 · Daniels MJD et al. · eLife 2023 · n=5–12/group · in vivo App^NL-G-F^ 12-month-old mice, both sexes · Cst7 KO female mice showed increased endolysosomal genes, increased lysosomal burden, increased microglial Aβ uptake in vivo, and increased subicular amyloid plaque number (significant) but not cortex/hippocampus; male KO reduced inflammatory genes (Il1b, Il1a, Tnf, Cxcl2) without amyloid change in any region; intracellular cathepsin L and C activity unchanged in both sexes · model: mouse (App^NL-G-F^ amyloid knock-in, sex-stratified) · verified 2026-05-29 ↩ ↩² ↩³

3. tyshkovskiy-2026-universal-transcriptomic-hallmarks · n=11,165 transcriptomes, 4 species (mouse/rat/macaque/human), >25 tissues · meta-analysis + new RNA-seq · elastic-net + Bayesian-ridge mortality clocks · model: multi-species ↩

4. tyshkovskiy-2026-universal-transcriptomic-hallmarks · linear mixed-effects gene-trait associations, P_adj<0.05 · multi-tissue rodent meta-dataset + multi-species cross-validation · Cst7 named as a positive contributor to chronological-age and mortality clocks alongside Gpnmb and Cdkn1a ↩

5. doi:10.1038/s41577-022-00830-0 · Baleviciute A, Keane L · Nature Reviews Immunology 2023 · 23:73 · commentary · sex-specific profiles of microglial CST7 in AD highlighted · model: mouse/review · NOT yet verified against full text ↩

6. doi:10.1186/s12974-024-03153-0 · Syage AR, Lane TE et al. · J Neuroinflammation 2024 · 21(1):157 · n=6/group for scRNA-seq; n=5–8 for histology · C57BL/6 Cst7-KO vs WT, JHMV i.c. infection · viral titers equivalent at day 7; chronic disease (days 14–21): Cst7-KO had increased T-cell infiltration, more axonal damage and demyelination, impaired remyelination, elevated Cxcl9/Cxcl10 T-cell chemoattractants and IFN-γ/perforin in CD8+ T cells · model: mouse (JHMV coronavirus, C57BL/6) · verified 2026-05-29 ↩

7. doi:10.1186/s12974-024-03119-2 · Li Q et al. · J Neuroinflammation 2024 · 21(1):125 · AD patients + APP/PS1 mice + 5XFAD injection model · monocyte-specific cystatin F overexpression increased Aβ deposition in APP/PS1 mice; secreted cystatin F dimer in plasma exists in dimeric form and physically interacts with Aβ via specific amino acid interactions, inhibiting monocyte internalization of Aβ; tail-vein injection of recombinant cystatin F dimer (200 µg/kg every 3 days × 1 month) in 3-month-old 5XFAD mice worsened amyloid pathology and cognitive deficits; monocytic cystatin F mRNA specifically elevated in sporadic AD patients · model: mouse (APP/PS1 + 5XFAD) + human monocytes · verified 2026-05-29 ↩