🧬 Aging Wiki

homo-sapiens

Properties1
aliaseshuman, Homo sapiens sapiens, modern human

11 min read

Homo sapiens (human)

Homo sapiens is the target organism for all aging research in this wiki. Unlike other entries in model-organisms/, humans are the subject being studied, not a proxy for another species. This page documents human-specific aging biology, lifespan statistics, major research cohorts, and the features that are poorly captured by model organisms — providing the baseline reference against which all model→human extrapolations (see _extrapolation-guide) must be measured.

Lifespan distribution

MetricValueNotes
Average life expectancy (OECD)~80 yearsVaries 72–86 yr by country; female > male ~5 yr
Median adult healthspan~60–70 yearsOnset of first chronic disease or disability
Supercentenarian threshold110+ yearsRare: estimated ~300–500 alive globally at any time
Maximum verified lifespan122.5 yearsJeanne Calment (France, 1875–1997) 1

The gap between average (~80 yr) and maximum (~122.5 yr) lifespan is far larger than in any laboratory model organism, reflecting the dominant role of lifestyle, environment, and stochastic damage accumulation in human aging outcomes. unsourced — precise healthspan estimates vary substantially by definition (disability-free, disease-free, functional independence); no single widely-adopted operational definition exists.

Per-organism extrapolation profile

(Per _extrapolation-guide § Per-organism extrapolation profiles)

1. Genome similarity to human: 100% (by definition; humans are the reference)

2. Lifespan ratio: 1.0 (reference); mouse:human ratio ~1:30; worm:human ratio ~1:2000

3. Conserved systems (well-modeled by other organisms)

Aging mechanisms that are robustly conserved across model organisms and demonstrated in humans:

  • Insulin/IGF-1 signaling (IIS) — insulin-igf1 pathway downregulation extends lifespan in worm, fly, and mouse; human longevity associations confirmed for FOXO3 and IGF1R 2 3
  • mTOR inhibition — mtor pathway is conserved; rapamycin extends lifespan in mice; observational human data supportive but no human RCT lifespan endpoint
  • AMPK activation — ampk pathway activates with energy stress; metformin (indirect AMPK activator) is associated with longevity in humans 4; see metformin
  • Cellular senescence — p16+/p21+ senescent cell accumulation with age documented in human tissue; sasp composition partly overlapping with mouse SASP data contradictory-evidence (SASP factor lists differ between mouse and human senescent cells)
  • Telomere attrition — Unlike standard lab mice (which have very long telomeres and active telomerase), humans rely on telomere shortening as a replication counter in most somatic tissues; attrition-associated replicative senescence is a major human aging driver
  • Mitochondrial dysfunction — progressive respiratory-chain defects accumulate in human post-mitotic tissues (heart, brain, skeletal muscle); well-modeled by mouse heteroplasmy models
  • Sarcopenia — age-related muscle loss (see sarcopenia) is universal in humans; onset detectable ~40-50 yr; clinically significant by 70+ yr
  • Epigenetic drift — DNA methylation age (see Horvath clock 5) accelerates with chronological age; clock-slowing by caloric restriction demonstrated in humans (CALERIE trial) 6
  • Chronic inflammation (inflammaging) — rising IL-6, TNF-alpha, CRP with age documented in large human cohorts; see immunosenescence

4. Divergent systems (poorly captured by model organisms)

SystemHow humans differImplication for extrapolation
Telomere biologyLab mice (C57BL/6) have telomeres 5–10× longer than humans; constitutive somatic telomeraseMouse telomere-manipulation findings do not translate cleanly; use wild-derived or Terc-KO mice for telomere models
Cognitive aging trajectoryHumans show a decades-long cortical thinning and synaptic decline; rodent models poorly capture frontotemporal agingNeurodegeneration findings from rodents need primate confirmation
Atherosclerosis kineticsHumans accumulate atherosclerotic plaques over decades; mice require genetic manipulation (ApoE-/-, LDLR-/-) to develop equivalent diseaseCardiovascular aging timeline is fundamentally different
Gut microbiome compositionHuman microbiome is substantially more diverse and diet-dependent than standard inbred mouse microbiomeMicrobiome-longevity findings from mice require human replication
Drug metabolism (CYP)Human CYP3A4 substrate specificity differs from mouse CYP3A; some compounds metabolized to active form in mice are not in humans (and vice versa)Senolytic dosing extrapolation is particularly unreliable
Immune aging patternsHuman Th17/Treg balance, NK cell decline kinetics, and vaccine-response decline differ from mouse patterns; thymic involution timeline differsImmunosenescence interventions in mice need human immunological validation
Psychosocial agingChronic psychological stress, social isolation, loneliness, and socioeconomic factors causally modulate human aging rate (epigenetic clock data); rodents lack equivalent exposure complexityHuman aging cannot be fully modeled without accounting for psychosocial covariates
Caloric restriction effect sizeCR extends lifespan ~30–40% in mice; CALERIE (2-year RCT targeting 25% CR) achieved only ~11.7% CR on average, with significant cardiometabolic biomarker improvements (TNF-α, CRP, T3, lipids) and DunedinPACE slowing, but no lifespan endpoint 7 8 no-fulltext-access (Kraus 2019)Achieved CR was well below target; biomarker effects were significant but modest; lifespan translation unknown (see caloric-restriction)

5. Strengths as the target organism

  • No extrapolation required — findings in humans are directly applicable to humans
  • Large cohort availability — Framingham (~15,000 person-years), BLSA (ongoing since 1958), LLFS, NECS provide longitudinal aging data at scale
  • Genetic association studies — GWAS with millions of participants enable discovery of longevity loci (FOXO3, APOE, CETP, etc.)
  • Epigenetic clocks validated in human tissue — biological age can be estimated in living subjects 5
  • Intervention trials possible — CALERIE, TAME, senolytic Phase 2 trials; surrogate endpoints established

6. Failure modes / limitations as a research subject

  • No mortality endpoint — RCTs with human lifespan as a primary outcome are ethically and practically impossible (decades-long, prohibitive cost); must rely on surrogate endpoints (epigenetic age, biomarkers of aging, healthspan metrics)
  • Extreme individual variation — genetic background heterogeneity means effect sizes are smaller and harder to detect than in inbred mouse strains; larger samples required
  • Ethical constraints on interventions — genetic interventions, extreme dietary restriction, and gerotherapeutics with unknown long-term safety cannot be tested as freely as in animal models
  • Confounding by lifestyle / SES — diet, exercise, smoking, socioeconomic status, access to healthcare all affect aging rate; disentangling these from genetic effects is a central challenge
  • Survivor bias in centenarian studies — individuals studied at 100+ yr are survivors of decades of selection pressure; their genetic and physiological profiles are not representative of the general population trajectory

Human longevity genetics

The following longevity-associated loci have the strongest and most-replicated human evidence:

FOXO3 (FOXO3A)

The rs2802292 GG genotype in FOXO3 is associated with ~2.75-fold increased odds of reaching 95+ years in the Honolulu Heart Program / Kuakini cohort 2, with later replications across European, Okinawan, German, Danish, and Chinese cohorts. FOXO3 encodes a forkhead transcription factor in the insulin-igf1 pathway; longevity alleles appear to upregulate autophagy, stress resistance, and ampk-dependent pathways. See foxo3 for mechanistic detail. needs-replication — individual replication study effect sizes vary; meta-analysis support is strong but OR estimates differ by study.

IGF1R (IGF-1 receptor)

Suh et al. (2008) identified heterozygous loss-of-function mutations in IGF1R in Ashkenazi Jewish centenarians (n=384 centenarians vs 312 controls); 9 centenarians (2.3%) carried either of two nonsynonymous mutations (Ala-37-Thr or Arg-407-His) vs 1 control (0.3%), p=0.02. Carriers showed reduced IGF-1 receptor (IGFIR) signaling activity in immortalized lymphocytes, and female offspring of centenarians overall had 35% higher serum IGF-1 levels (p<0.01) — interpreted as a compensatory response to reduced IGFIR sensitivity — and were 2.5 cm shorter than controls (p<0.001) 3. Consistent with the evolutionarily conserved IIS longevity axis (worm daf-2, fly InR, mouse Igf1r+/-).

APOE

  • APOE ε2 allele: longevity-associated; lower Alzheimer’s risk, lower cardiovascular risk
  • APOE ε4 allele: anti-longevity; 3–4× increased Alzheimer’s risk; common in short-lived individuals; strongly depleted in centenarian populations 9

Other replicated longevity loci

CETP, HMGA2, CDKN2A/B, and multiple GWAS loci with small individual effects; polygenic longevity score work ongoing. unsourced for individual loci beyond FOXO3/IGF1R/APOE — see sebastiani-2012-centenarian-genetics for NECS overview.

Major human aging research cohorts

CohortStartN (approx)Focus
Framingham Heart Study1948~15,000 original + offspring + third genCardiovascular aging; now broad multi-system
Baltimore Longitudinal Study of Aging (BLSA)1958~3,000 cumulativeMulti-system longitudinal phenotyping; oldest US aging study
New England Centenarian Study (NECS)1994~1,600 centenariansExtreme longevity genetics; family studies
Okinawan Centenarian Study1975~900 centenariansDietary and genetic factors in exceptional longevity
Long Life Family Study (LLFS)2006~5,000 (clustered families)Family-based genetic architecture of longevity
CALERIE (Phase 2)2007220 randomized (218 started)Only human RCT of sustained caloric restriction (targeting 25% CR × 2 yr; achieved ~11.7%); epigenetic + cardiometabolic outcomes 7 6
TAME trialdesign 2016 (Barzilai)3,000 (planned)Metformin vs placebo; geroscience-based trial; primary endpoint: composite aging-related events. Not currently ClinicalTrials.gov-registered as of 2026-05 (NCT04977829 is invalid; NCT02432287 is unrelated MILES n=16; NCT03138915 is unrelated radiomics study). See metformin.

Surrogate endpoints and biomarkers of aging

Because mortality cannot serve as a primary endpoint in human aging trials, validated surrogate endpoints are critical. The most established:

  • Epigenetic clocks — Horvath’s pan-tissue clock (353 CpG sites) predicts chronological age with ~3.6 yr MAE in test data across human tissues 5; accelerated clock predicts all-cause mortality, frailty, and disease risk. Second-generation clocks (GrimAge, PhenoAge, DunedinPACE) are more predictive of aging-related outcomes. CALERIE showed that 2-year CR intervention (mean ~11.9% achieved, not the 25% target) slowed the pace of aging as measured by DunedinPACE (24-month d=−0.25, 95% CI −0.41 to −0.09, p<0.003), corresponding to a ~2–3% slowing of biological aging rate per year; however, CR did not significantly change PhenoAge or GrimAge 6.
  • Frailty indices — Fried phenotype (5-item) and Rockwood frailty index (30–70 items) quantify functional aging; predictive of hospitalization, disability, mortality
  • Biological age algorithms — composite scores (phenotypic age, metabolomic age) from clinical laboratory values
  • Telomere length — average leukocyte telomere length (LTL) declines ~20–50 bp/yr; shorter LTL associated with aging diseases, but causal directionality unclear for most outcomes no-mechanism

Hallmarks in human context

Hallmarks of aging documented in humans (with human-specific caveats):

HallmarkHuman Evidence LevelKey Caveats
genomic-instabilityStrongSomatic mutation accumulation in normal human tissues well-documented by clonal hematopoiesis studies
telomere-attritionStrongReplicative senescence is a real human aging driver; unlike lab mice, humans have short telomeres and limited somatic telomerase
epigenetic-alterationsStrongHorvath clock; methylation drift well-documented in human longitudinal cohorts
loss-of-proteostasisStrongProtein aggregates (Aβ, tau, TDP-43, α-syn) accumulate in human brain aging
deregulated-nutrient-sensingModerateIIS axis longevity associations confirmed; CR biomarker effects modest in humans vs mice
mitochondrial-dysfunctionStrongRespiratory-chain mutations accumulate in human muscle and brain with age
cellular-senescenceStrongp16+ senescent cells accumulate in human tissues; ex vivo senolytic clearance demonstrated
stem-cell-exhaustionStrongHSC clonal hematopoiesis; satellite cell decline in sarcopenia; intestinal stem cell dysfunction
altered-intercellular-communicationModerateInflammaging, SASP factors measurable in human plasma; causal role under investigation
chronic-inflammationStrongInflammaging CRP/IL-6/TNF-alpha trends documented in Framingham, BLSA and others
dysbiosisLimitedGut microbiome diversity changes with age in humans; causality unclear
disabled-macroautophagyModerateAutophagy flux decline with age in human peripheral blood cells and muscle; biopsy data limited

Limitations and gaps

  • needs-replication — TAME trial (metformin vs aging composite endpoint) is ongoing; no primary result yet as of 2026
  • long-term-unknown — long-term effects of most candidate geroprotectors (rapamycin, senolytics, NAD+ precursors) remain unknown in humans
  • no-mechanism — mechanisms by which psychosocial factors (loneliness, SES, stress) modulate epigenetic aging rate in humans are incompletely understood
  • unsourced — population-level healthspan estimates vary widely by definition; a single operational definition has not been adopted across cohort studies
  • contradictory-evidence — SASP composition in human vs mouse senescent cells overlaps only partially; cross-species SASP-based therapeutic targets need human-side confirmation
  • No dedicated page yet exists for the frailty-index methodology or the DunedinPACE clock; these would be useful downstream stubs

See also

Footnotes

  1. coles-2004-supercentenarian-demography · observational · model: human supercentenarians · doi:10.1093/gerona/59.6.b579 · no-fulltext-access (download failed; not_oa) ↩

  2. willcox-2008-foxo3a-longevity · n=615 (213 longevity cases ≥95 yr + 402 average-lived controls; drawn from HHP/HAAS cohort of 3,741 Japanese-American men) · nested case-control · OR=2.75 (GG vs TT, 95% CI 1.51–5.02, p=0.0007) for survival to 95+ · model: Honolulu Heart Program / Kuakini cohort · doi:10.1073/pnas.0801030105 · locally available ↩ ↩2

  3. suh-2008-igf1r-centenarian · n=384 centenarians / 312 controls · observational (case-control) · model: Ashkenazi Jewish centenarians · doi:10.1073/pnas.0705467105 · locally available ↩ ↩2

  4. barzilai-tame-trial · TAME design paper (Barzilai et al. 2016, Cell Metab 23:1060–1065) · NOT currently ClinicalTrials.gov-registered as of 2026-05; previously cited NCT04977829, NCT02432287, NCT03138915 are all incorrect (verified via CT.gov v2 API 2026-05-08) · model: adults 65–79 yr (planned ~3,000) · long-term-unknown no-mechanism ↩

  5. horvath-2013-epigenetic-clock · observational (multi-tissue) · n=7,844 non-cancer samples from 82 datasets (51 tissues/cell types) · test-set MAE=3.6 yr · model: human tissues across age · doi:10.1186/gb-2013-14-10-r115 ↩ ↩2 ↩3

  6. waziry-2023-calerie-epigenetic-clock · n=197 with baseline + follow-up DNAm data (128 CR, 69 AL; from 220 randomized) · rct · model: healthy non-obese adults (CALERIE) · doi:10.1038/s43587-022-00357-y ↩ ↩2 ↩3

  7. ravussin-2015-calerie · n=220 randomized (218 ITT; 145 CR, 75 AL) · rct · achieved 11.7±0.7% CR over 2 yr (target: 25%) · model: healthy non-obese adults aged 21–51 yr · doi:10.1093/gerona/glv057 ↩ ↩2

  8. kraus-2019-calerie-cardiometabolic · n=218 · rct · model: healthy non-obese adults · doi:10.1016/s2213-8587(19)30151-2 · no-fulltext-access (download failed; OA URL present but unreachable) ↩

  9. sebastiani-2012-centenarian-genetics · review (NECS overview 1994–2012; GWAS sub-study: 801 centenarians + 914 controls) · model: New England Centenarian Study · doi:10.3389/fgene.2012.00277 ↩

Graph View

Backlinks