p21 (CDKN1A)

Cyclin-dependent kinase inhibitor 1A — a 164-amino-acid, 18 kDa protein that enforces cell-cycle arrest downstream of p53 and is the dominant initiator of senescence-associated cell-cycle exit. Discovered in 1993 by two simultaneous reports as WAF1 (a p53-activated fragment) and CIP1 (a CDK-interacting protein) ¹². Central to aging biology as the gatekeeper of regenerative capacity: chronically elevated p21 in aged tissues limits stem-cell proliferation and tissue maintenance, while its acute induction protects against cancer.

Naming note: The gene is officially CDKN1A; the protein is p21 (colloquially). Aliases WAF1 (el-Deiry 1993) and CIP1 (Harper 1993) reflect the two independent discovery paths. This page uses “p21” as the primary protein name, consistent with the p53-pathway page.

Identity

Field	Value
UniProt	P38936 (CDN1A_HUMAN)
NCBI Gene	1026
HGNC symbol	CDKN1A
Ensembl	ENSG00000124762
Chromosomal location	6p21.2
Length	164 amino acids
MW	~18 kDa (runs anomalously high on SDS-PAGE at ~21 kDa, hence “p21”)
Mouse ortholog	Cdkn1a (one-to-one)
GenAge entry	284 (Homo sapiens)

Functional domains

N-terminal CDK-inhibitory domain (residues ~1–80) — binds the cyclin/CDK complex interface; two copies of this motif allow simultaneous engagement of the cyclin and CDK subunits. Responsible for inhibiting CDK2/cyclin-E (G1/S checkpoint) and CDK1/cyclin-B (G2/M checkpoint).
PIP-box (PCNA-binding, residues ~140–164) — binds the interdomain connecting loop of PCNA; blocks PCNA-stimulated DNA polymerase delta processivity while sparing PCNA-dependent nucleotide-excision repair ³. This is the mechanistic basis for the “anti-replication, not anti-repair” selectivity.
Nuclear localization signal (residues 141–156) — overlaps the PIP-box; cytoplasmic p21 (when phosphorylated at Thr145 by Akt) loses growth-suppressive activity.

Post-translational modifications

PTM	Site	Kinase	Functional consequence
Phosphorylation	Thr145	Akt, PKA, PIM1/2	Cytoplasmic retention; inactivates CDK-inhibitory and PCNA-binding functions
Phosphorylation	Ser146	PKC, NUAK1	Cytoplasmic accumulation; pro-survival in some contexts
Phosphorylation	Thr80	LKB1	Cytoplasmic localization
Ubiquitination	Multiple Lys	MKRN1, DCX(DTL), RNF114	Proteasomal degradation

Note (2026-05-04 verification): UniProt P38936 lists three additional phosphorylation sites not in this table: Ser-114 (GSK3β), Ser-130 (kinase unspecified), and Ser-160 (PKC). Functional consequences of these sites are not well-characterized; expand table on next curation pass. needs-replication

Cytoplasmic p21 can paradoxically function as an anti-apoptotic factor by inhibiting procaspase-3 and ASK1. This is context-dependent and represents a non-canonical function distinct from CDK inhibition. no-mechanism — the decision logic between nuclear (growth arrest) and cytoplasmic (pro-survival) p21 is incompletely understood.

Core mechanism: CDK inhibition and cell-cycle arrest

p21 is transcriptionally activated by p53 in response to DNA damage, oxidative stress, oncogene activation, and replicative stress ¹. Once induced:

G1/S checkpoint — p21 binds and inhibits CDK2/cyclin-E and CDK2/cyclin-A complexes, preventing Rb phosphorylation. Hypophosphorylated Rb sequesters E2F transcription factors, halting entry into S phase.
G2/M checkpoint — p21 inhibits CDK1/cyclin-B; this contribution to G2/M arrest is secondary to CDC25C regulation. needs-replication — the relative quantitative contribution of p21 vs other CDK inhibitors to G2/M arrest in different cell types is debated. Note: Brugarolas 1995 ⁴ characterizes p21’s role in the G1 checkpoint (irradiated p21-null MEFs show an intermediate G1 arrest defect, with S-phase fractions at 57.5–54.5% of untreated levels vs 39% for WT); that paper does not provide primary evidence for p21 at the G2/M checkpoint. The CDK1/cyclin-B inhibition by p21 is established by biochemical studies (including Harper 1993 ²) but the in vivo G2/M checkpoint role requires a separate citation. unsourced — primary citation needed for p21 requirement at G2/M in vivo.
PCNA inhibition — independent of CDK inhibition, the PIP-box directly blocks DNA polymerase delta processivity at PCNA, arresting replication forks ³.

The senescence axis is distinct from transient checkpoint arrest:

p53 → p21 establishes the initial cell-cycle arrest (hours–days).
Sustained p21 expression → Rb hypophosphorylation → E2F repression → reinforced arrest.
Parallel activation of p16ink4a (via independent signaling) → CDK4/6 inhibition → additional Rb hypophosphorylation. p16 maintains senescence even if p21 declines.
Together, p21-initiated + p16-maintained arrest creates the irreversible senescent state.

Checkpoint	CDK complex inhibited	Rb axis effect	p21 required?
G1/S	CDK2/cyclin-E, CDK2/cyclin-A	Rb hypophosphorylated	yes (primary effector)
S-phase	PCNA (direct)	—	yes (replication fork stalling)
G2/M	CDK1/cyclin-B	—	partial/unsourced in vivo unsourced
Senescence induction	CDK2/cyclin-E + CDK4/6 (indirect via p16)	permanent Rb hypophosphorylation	yes (initiation); p16 takes over for maintenance

Discovery

Two independent groups published back-to-back in Cell, November 1993:

el-Deiry et al. 1993 — identified WAF1 via subtractive hybridization in colorectal cancer cells as a transcript induced by wild-type p53. Showed a p53-response element upstream of WAF1, and demonstrated growth suppression upon introduction into tumor cell lines ¹.
Harper et al. 1993 — identified CIP1 biochemically as a 21 kDa protein co-immunoprecipitating with multiple cyclin/CDK complexes. Showed potent inhibition of CDK kinase activity and block of Rb phosphorylation ².

Both groups were simultaneously identifying the same protein; the genes were confirmed identical shortly after publication.

Role in aging

p21 as a senescence initiator

p21 is the canonical effector linking upstream damage signals to permanent cell-cycle exit. Chronically activated p53 (from telomere attrition, accumulated DNA damage, oxidative stress, or oncogenic signaling) drives sustained p21 expression in aged tissues. This positions p21 as the proximate cause of the senescent cell-cycle block that underlies loss of tissue regeneration ⁵.

Evidence from p21-deficient mice with telomere dysfunction:

Deletion of p21 in mice with short, dysfunctional telomeres (late-generation Terc−/− mice) partially rescued tissue stem-cell function and extended lifespan, establishing that p21 is causally — not merely correlatively — required for telomere-damage-driven stem-cell depletion ⁶. needs-replication — this finding is from a single genetic background; independent replication in other contexts would strengthen the causal claim.

Dimension	Status	Notes
Pathway conserved in humans?	yes	CDK2/cyclin-E axis, PCNA interaction, and p53 → p21 induction are conserved; same transcriptional response elements
Phenotype conserved in humans?	partial	Senescent cells accumulate in human skin and organs with age, evidenced by SA-β-Gal staining ⁷; p21 itself is induced in aged human tissues unsourced — the specific claim about p21 protein levels in aged human tissues needs a primary citation distinct from Dimri 1995 (which establishes SA-β-Gal as a senescence marker, not p21 specifically); causal rescue experiments (as in mice) have not been done in humans
Replicated in humans?	no	needs-human-replication

Universal transcriptomic mortality marker (multi-species)

Beyond its mechanistic role, Cdkn1a/p21 is empirically one of the most consistent up-regulated markers of mammalian ageing and mortality across species and tissues. In the multi-species transcriptomic-clock meta-analysis of Tyshkovskiy et al. 2026 (>11,000 transcriptomes; mouse, rat, macaque, human), Cdkn1a was a top conserved gene positively associated with chronological age, expected mortality and (negatively) maximum lifespan, and one of three headline universal mortality genes alongside GPNMB and LGALS3 ⁸. Specifically, p21:

was up-regulated in ≥5 of 9 rodent chronic-disease models and a strong contributor to disease-associated mortality-tAge ⁸;
rose in replicative senescence (top contributor to in-vitro tAge), was up in Klotho-KO progeria (both kidney and muscle), and fell under rejuvenating perturbations — heterochronic parabiosis and partial reprogramming;
was down-regulated up to embryonic day ~E10 then re-activated, tracking the U-shaped “ground-zero” tAge rejuvenation trajectory (information-theory-of-aging) ⁸;
showed plasma protein (CDKN1A) positively associated with all-cause mortality in UK Biobank (n=51,276, adjusted age+sex) — extending the senescence-marker role to a circulating human mortality biomarker ⁸.

See transcriptomic-clock-tage for how p21 enters the transcriptomic mortality clock.

Regenerative capacity loss

Elevated p21 in aged tissues restricts proliferative responses of tissue stem cells:

Skeletal muscle satellite cells in aged animals show increased p21 expression and reduced self-renewal capacity (consistent with the PRMT7/DNMT3b/p21 axis in muscle stem cells ⁹).
Hematopoietic stem cells and intestinal epithelial stem cells show similar p21-associated age-related proliferative decline. unsourced — hematopoietic and intestinal data need primary citations; added as a direction for further curation.

Antagonistic pleiotropy context

p21 is a downstream effector of the same p53 antagonistic-pleiotropy trade-off described on p53. Hyperactive p53 → excess p21 → excess senescence → tissue aging. The p53+/m model (see p53) in which accelerated aging phenotypes appear is mediated in part through p21 over-induction, though the specific quantitative contribution of p21 vs other p53 targets (PUMA, BAX) in that model has not been resolved ¹⁰. no-mechanism

Senescent cell burden and senolytics

p21-expressing cells are a component of the senescent cell pool that accumulates with age and drives SASP-mediated tissue dysfunction. Senescent-cell clearance studies (using p16Ink4a as the biomarker) demonstrate that reducing this pool delays age-associated pathologies ¹¹¹². p21 and p16 are partially overlapping but not identical markers of senescence; the relationship between p21+ and p16+ cell subsets in different tissues and their respective contributions to tissue aging is an active research area.

needs-replication — the Baker 2011 and Baker 2016 clearance experiments used p16-driven systems; equivalent p21-targeted clearance experiments across wild-type aging mice have not been widely published as of 2026.

Pathway membership

p53-pathway — p21 is the primary transcriptional effector of p53 for cell-cycle arrest; directly listed as a target on that page
dna-damage-response — ATM/ATR → p53 Ser15 phosphorylation → p21 induction; p21 is the distal effector
cell-cycle-regulation — core G1/S and G2/M checkpoint effector alongside p16ink4a, CDC25 phosphatases, and Rb
apoptosis-pathway — cytoplasmic p21 can inhibit caspase-3 and apoptosis (anti-apoptotic, non-canonical role); nuclear p21 participates in PUMA/NOXA induction context

Key interactors

Interactor	Interaction type	Functional consequence
p53	transcriptional regulator → p21	p53 binds WAF1 promoter; induces p21 expression under stress
CDK2/cyclin-e	direct inhibition	blocks G1/S transition; prevents Rb phosphorylation
CDK1/cyclin-b	direct inhibition	blocks G2/M transition
pcna	PIP-box binding	blocks DNA polymerase delta processivity; spares NER
mdm2	indirect (MDM2 degrades p53 which induces p21)	MDM2-p53 feedback controls p21 levels
atm	upstream kinase (phosphorylates p53 → activates p21 induction)	genotoxic stress → p21 induction
Akt	phosphorylates p21 at Thr145	cytoplasmic sequestration; inactivation

Aging interventions that modulate p21

senolytics — clear p21-expressing senescent cells downstream; do not directly target p21 itself
caloric-restriction — reduces senescent cell burden in some tissues; mechanism includes attenuation of p53/p21 activation, possibly via mtor and sirtuin pathways no-mechanism
rapamycin — mTOR inhibition reduces p21-driven senescence in some contexts; mechanistic link proposed via S6K1 → p53 stability, but human evidence absent needs-human-replication
exercise — aerobic exercise reduces senescent cell markers including p21 in muscle; mechanistic basis unclear unsourced

Limitations and open questions

Gap	Tag	Notes
Human causal rescue experiments lacking	needs-human-replication	p21 deletion rescues telomere-damage-driven aging in mice; no equivalent in humans
Cytoplasmic vs nuclear decision logic	no-mechanism	What determines whether p21 acts as CDK inhibitor vs anti-apoptotic effector?
p21+ vs p16+ senescent cell subsets	needs-replication	The two markers identify overlapping but non-identical senescent populations; tissue-specific relationship unclear
Hematopoietic/intestinal stem-cell data	unsourced	Claims about p21 elevation in HSC and intestinal stem cells need primary citations
p21 contribution to Tyner 2002 phenotype	no-mechanism	Quantitative share of p21 vs other p53 targets in accelerated aging unclear
Optimal therapeutic window	dose-response-unclear	Transient p21 suppression (to rescue regeneration) vs permanent suppression (cancer risk): no clinical framework
GenAge classification	—	GenAge entry 284 notes “role in human ageing remains unknown” — the evidence is model-organism-heavy

Footnotes

eldeiry-1993-waf1-p53 · doi:10.1016/0092-8674(93)90500-p · n=N/A · in-vitro + in-vivo · model: SW480 colorectal carcinoma cells, nude mice · 8,360 citations; not locally downloaded (closed access) ↩ ↩² ↩³
harper-1993-cip1-cdk-inhibitor · doi:10.1016/0092-8674(93)90499-g · n=N/A · in-vitro · model: human cell extracts and recombinant protein · 5,642 citations; not locally downloaded (closed access) ↩ ↩² ↩³
waga-1994-p21-pcna-replication-repair · doi:10.1038/371534a0 · n=N/A · in-vitro (biochemical) · model: SV40 DNA replication system · 652 citations; not locally downloaded (closed access) ↩ ↩²
doi:10.1038/377552a0 · in-vitro (mouse embryo fibroblasts, MEFs) · model: p21−/− MEFs from B6↔129/Sv chimeras vs WT · 5.5 Gy γ-irradiation · key finding: p21-null MEFs show intermediate G1 arrest defect (S-phase fraction 57.5–54.5% of untreated vs 39% for WT at 18 h post-irradiation); DNA-damage-induced CDK2 inhibition also requires p21 · paper does NOT address G2/M checkpoint · locally downloaded and verified ↩
doi:10.1016/j.cell.2013.05.039 · review (Cell 2013) · n=N/A · model: review of human and mouse data · 14,200 citations; locally downloaded ↩
GenAge entry 284 (CDKN1A, Homo sapiens) · https://genomics.senescence.info/genes/entry.php?hgnc=CDKN1A · cites multiple primary sources; see GenAge for individual references ↩
doi:10.1073/pnas.92.20.9363 · in-vitro + observational · model: human fibroblasts (WI-38, HCA2, AG06234 and others) + skin biopsies from 20 donors aged 20–90 yr · key finding: SA-β-Gal activity at pH 6 marks senescent cells in culture and accumulates in dermal fibroblasts and epidermal keratinocytes in aged (>69 yr) human skin; SA-β-Gal not expressed in quiescent or terminally differentiated cells · paper establishes SA-β-Gal as a senescence biomarker; does NOT directly study p21 protein — do not cite for p21 induction in aged human skin · locally downloaded and verified ↩
tyshkovskiy-2026-universal-transcriptomic-hallmarks · doi:10.1038/s41586-026-10542-3 · Nature 2026 · n=11,165 transcriptomes across 4 species (meta-analysis) + UK Biobank plasma (CDKN1A n=51,276) · elastic-net/mixed-effects clock coefficients; Cox PH for plasma · model: mouse/rat/macaque/human · Cdkn1a a top conserved up-with-mortality / down-with-lifespan gene; plasma CDKN1A positively associated with all-cause mortality (adj. age+sex) ↩ ↩² ↩³ ↩⁴
doi:10.1016/j.celrep.2016.01.022 · in-vivo (mouse, PRMT7-KO whole-body and Pax7-CreERT2 conditional) · model: 8-month-old PRMT7−/− and PRMT7FL/FL;Pax7CrEERT2/+ mice · key finding: PRMT7 depletion in satellite cells elevates p21 expression and induces premature senescence (~30% SA-β-Gal positive vs ~22% in conditional KO); PRMT7 regulates p21 epigenetically via DNMT3b/CpG methylation at the Cdkn1a promoter; restoration of DNMT3b rescues senescence by ~65% · locally downloaded and verified ↩
tyner-2002-p53-mutant-aging · doi:10.1038/415045a · n=35 (p53+/m) + 56 (p53+/+) · in-vivo (mouse, transgenic) · P<0.0001 (survival curve) · model: p53+/m on mixed C57BL/6 × 129/Sv background · locally downloaded ↩
doi:10.1038/nature10600 · in-vivo (progeroid mouse, INK-ATTAC transgene) · model: BubR1 progeroid mice · OA (green) but download failed — PDF not locally available; claims not verified against full text no-fulltext-access ↩
doi:10.1038/nature16932 · in-vivo (naturally aged mouse, INK-ATTAC) · model: wild-type ATTAC mice; two cohorts: 129Sv×C57BL/6J×FVB mixed background + congenic C57BL/6J · AP20187 treatment from 12 months, twice weekly · median lifespan extended 27% (mixed background) and 24% (C57BL/6J); ranges 17–35% by sex/background · clearance partial and tissue-selective (colon and liver refractory) · uses p16Ink4a-driven system, not p21-targeted · locally downloaded and verified ↩

🧬 Aging Wiki

Explorer

p21