Longevity & AgingReview ArticleOpen Access

Six Decades of Cellular Senescence Research Mapped in Landmark Review

A comprehensive review traces senescence science from Hayflick's 1961 discovery to today's senolytics and precision biomarkers.

Sunday, October 4, 2026 8 views
Published in EMBO J
A split microscope image showing normal dividing cells on one side and enlarged, flattened senescent cells stained blue with beta-galactosidase on the other side, on a lab bench

Summary

Senescent cells are cells that permanently stop dividing but stay alive. This review follows how scientists' understanding of them has changed since 1961. They were first seen as simple signs of cells wearing out. Researchers later found they are a stress response that blocks cancer. They also release a mix of inflammatory and signaling molecules, called the SASP, that reshapes surrounding tissue. Senescent cells help with wound healing, liver fibrosis control and embryonic development, but when they build up with age they drive chronic inflammation and disease. Mouse studies that removed them showed better health and longer life. The authors argue that senescent cells are diverse, not a single type. The next generation of treatments will need better single-cell tools and biomarkers to target the harmful ones precisely.

Detailed Summary

Cellular senescence has moved from a laboratory curiosity to one of the most intensively pursued targets in aging biology. This review by Klaver, Steeneken, Veeningen, Altulea and Demaria in The EMBO Journal traces six decades of work and the conceptual shifts that shaped it: replicative endpoint, stress-response program, regulator of tissue homeostasis, driver of chronic pathology, and clinically actionable process. Understanding these shifts matters because senolytic and other senotherapeutic strategies depend on knowing which senescent cells are helpful and which are harmful.

The story starts in 1961, when Hayflick and Moorhead showed that normal human fibroblasts divide only a finite number of times, overturning Alexis Carrel's dogma of unlimited cellular proliferation. In 1990 Harley and colleagues linked this limit to telomere shortening, which triggers a persistent DNA damage response. Biomarkers followed: SA-β-gal in 1995, p21 and p16 in the late 1990s, DNA damage foci such as γH2AX and 53BP1, and heterochromatin foci in 2003. The authors stress that no single marker defines senescence, and single-cell and spatial profiling now reveal substantial heterogeneity.

Senescence was then shown to occur without telomere erosion. In 1997 Serrano and colleagues found that oncogenic RAS induces senescence, establishing oncogene-induced senescence as a barrier to cancer. Rodent cells escaped senescence when either p53 or p16 was disrupted, whereas human cells required loss of both. Similar responses followed activation of Raf-MEK-ERK and PI3K-AKT signaling, and the authors describe an emerging OIS spectrum of cell states. Irradiation, chemotherapy drugs and oxidative stress, including mitochondrial reactive oxygen species, also trigger senescence, which led to the concept of stress-induced premature senescence.

The discovery of the senescence-associated secretory phenotype (SASP) reframed senescent cells as active communicators. Campisi's group showed in 2001 that senescent fibroblasts stimulate preneoplastic epithelial cells. The SASP is dynamic, with most factors appearing days after induction, and includes cytokines, chemokines, lipids and extracellular vesicles. There is no universal SASP. NF-κB, C/EBPβ, p38 MAPK, mTOR, cGAS-STING and mitochondrial RNA sensing drive inflammatory programs, while p53-driven programs promote immune clearance and NOTCH favors a TGF-β-dominated, less inflammatory secretome. This diversity helps explain antagonistic pleiotropy, in which a process that is beneficial early in life becomes harmful later.

Genetic mouse models supplied causal evidence. Clearing p16-positive cells with INK-ATTAC delayed pathology in progeroid mice in 2011 and improved healthspan and lifespan in naturally aged mice in 2016. Further models (p16-3MR, Glb1 and p21 reporters) expanded the toolkit. Senescence also has beneficial roles: it limited liver fibrosis in 2008, senescent fibroblasts accelerated wound closure through PDGF-AA in 2014, and in 2013 two groups reported p21-dependent, p16- and p53-independent developmental senescence, with senescent cells cleared by macrophages.

The authors conclude that the next era depends on single-cell technologies, precision biomarkers and targeted interventions that distinguish beneficial from detrimental senescent subtypes. Caveats include discrepancies between reporter models, such as the ongoing debate over p16-3MR, and limited evidence in some organisms such as planarians. This is a narrative review rather than a new dataset. Note also that the supplied text is truncated, so the later sections on senotherapeutics and clinical translation are not detailed here.

Key Findings

  • In 1961 Hayflick and Moorhead showed normal human fibroblasts divide only a finite number of times, overturning the Carrel dogma of unlimited proliferation held for roughly five decades
  • Harley and colleagues (1990) tied replicative senescence to progressive telomere shortening, and telomerase expression bypasses the arrest, which supports a causal role for telomere erosion
  • Serrano et al. (1997) showed oncogenic RAS induces senescence; rodent cells escaped via loss of p53 or p16 alone, while human cells needed both pathways inactivated
  • The SASP is dynamic and heterogeneous: most factors appear days after induction, and p53-driven, NF-κB-driven and NOTCH/TGF-β-driven secretomes have different immune and inflammatory effects
  • Clearing p16-positive cells in INK-ATTAC mice delayed age-related pathology in progeroid mice (2011) and improved healthspan and lifespan in naturally aged mice (2016)
  • Senescence has beneficial roles: it restrained liver fibrosis (2008), and senescent fibroblasts promoted wound closure via PDGF-AA, with their ablation delaying healing (2014)
  • Developmental senescence, reported in 2013, depends on p21 but not p16 or p53, and senescent embryonic cells are cleared by macrophages

Methodology

This is a narrative historical review, not a systematic review or meta-analysis, so it reports no pooled sample size, statistical tests or formal quality grading. The authors organize landmark in vitro, mouse and cross-species studies from 1961 onward into conceptual transitions. The supplied full text was truncated after the section on developmental senescence. The later sections on senotherapeutics, heterogeneity and precision targeting are therefore summarized only from the abstract.

Study Limitations

As a narrative review, it reflects the authors' selection and framing of the literature, and much of the causal evidence comes from mouse and cell-culture models whose translation to humans is not established. The authors acknowledge discrepancies between senescence reporter models, such as the debate over p16-3MR, and limited evidence in some organisms such as planarians. No conflict-of-interest information was included in the supplied text.

Enjoyed this summary?

Get the latest longevity research delivered to your inbox every week.

Enter your email to subscribe: