Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 68 of the 68 references it cites were found in a registry or an index; 36 of the 43 sources its claims cite were checked for support.

Distinct senotypes in p16- and p21-positive cells across human and mouse aging tissues

Open access

Abstract

Abstract Senescent cells drive age-related tissue dysfunction via the induction of a chronic senescence-associated secretory phenotype (SASP). The cyclin-dependent kinase inhibitors p21 Cip1 and p16 Ink4a have long served as markers of cellular senescence. However, their individual roles remain incompletely elucidated, particularly in vivo. Thus, we conducted a comprehensive examination of multiple single-cell RNA sequencing datasets spanning both murine and human tissues during aging. Our analysis revealed that p21 Cip1 and p16 Ink4a transcripts demonstrate significant heterogeneity across distinct cell types and tissues, frequently exhibiting a lack of co-expression. Moreover, we identified tissue-specific variations in SASP profiles linked to p21 Cip1 or p16 Ink4a expression. Using RNA velocity and pseudotime analyses, we discovered that p21+ and p16+ cells follow independent trajectory dynamics, with no evidence of direct transitions between these two states. Despite this heterogeneity, we identified a limited set of shared “core” SASP factors that may drive common senescence-related functions. Our study underscores the substantial diversity of cellular senescence and the SASP, emphasizing that these phenomena are inherently cell- and tissue-dependent.

Overview

CoChat Summary

## TL;DR

This paper analyzes multiple single-cell RNA-seq datasets from aging mouse and human tissues to compare p16- and p21-positive senescent cells. It finds that p16+ and p21+ cells are usually distinct, often non-overlapping subpopulations with different tissue-specific secretomes and independent trajectories.

## Key findings

  • In mouse hippocampus, the authors identified 5 primary cell populations and then found 3 senescence subpopulations: p21-only, p16-only, and p21/p16 double-positive cells; p16+ cells increased with age, while p21+ increases were restricted to specific clusters.
  • Across murine skeletal muscle, bone, and liver, they identified 9, 14, and 13 distinct cell populations, respectively, and again found mostly separate p16+ and p21+ subpopulations with rare co-expression; RNA velocity/pseudotime showed no common ancestor or direct transition between the two states.
  • In human eyelid skin (ages 18–76) they identified 16 cell populations and observed p21-exclusive, p16-exclusive, and p21-p16 co-expressing cells, with p21-exclusive cells predominant; in human lung (ages 21–78), most cells were p21+, with only a small number p16+ or double-positive.
  • A small “core” secretome emerged across tissues: in the SenMayo-based analysis, Icam1 was expressed across all 6 analyzed tissues, and Jun, Edn1, Igfbp6, Tnfrsf1a, Cxcl16, Ets2, Fas, Igfbp4, Plaur, and Ccl2 were expressed in 8 or more of the 14 analyzed tissues; broader analyses highlighted Cdkn1a as present in all analyzed tissues.

## Methods

  • The study reanalyzed multiple public and independent scRNA-seq datasets from aging mouse tissues (brain, skeletal muscle, bone, liver, plus spleen/kidney/lung in some analyses) and human skin and lung, using Cell Ranger, Seurat, SCTransform normalization, clustering, differential expression, RNA velocity (scVelo/PhyloVelo), pseudotime (Monocle3), CellChat, and SCENIC.
  • Cells with low gene counts, high mitochondrial content, Ki67 positivity, CD45high, or S-phase signatures were excluded before defining p16+ and p21+ populations and comparing SASP/secretome-associated genes using SenMayo, SenSig, CellAge, and Human Protein Atlas-derived secreted gene sets.
  • The authors validated transcriptomic findings with CyTOF in mouse bone, immunofluorescence in aged human skin and diseased lung, and MERFISH in human skin; they also used a p21-Cre mouse model to confirm enrichment of p21-associated signatures.

## Limitations

  • The paper states that scRNA-seq has technical biases, including dropouts and limited detection of low-abundance transcripts, which can miss key components; it says this could be improved with spatial transcriptomics or single-cell proteomics.
  • The paper notes that SASP heterogeneity was not yet addressed with spatial proteomics, so the functional picture remains incomplete.
  • Evident from the text: some analyses were constrained by low numbers of p16+ cells in certain tissues, so muscle was excluded from CellChat analyses and liver from some SCENIC analyses.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s44318-025-00601-2

Cites

68 of 68 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Image 1
CC BY · PubMed Central
Image 2
CC BY · PubMed Central
Figure 1 Distinct p16+ and p21+ cell populations with differential secretory profiles in the murine hippocampus. (A) A t-SNE plot depicting the five main populations of young and old murine hippocampus (Ogrodnik et al, 2021, GSE161340). (B) From the original population, CD45− cells are selected, followed by Ki67-negativity and cells not in the S phase. From these cells, just p16(Cdkn2a)+ cells and p21(Cdkn1a)+ cells are selected. From the p16+ cells, microglia and oligodendrocytes depict the main populations, while in the p21+ population, microglia is predominant. (C) t-SNE visualization of p16 + (red) and p21+ (blue) cells, along with a few double-positive cells (green), shown in the Cd45-Ki67-S- population. (D) The dPo (double positive) cells are high in both Cdkn2a and Cdkn1a, while p16+ cells just express Cdkn2a, but no Cdkn1a and vice versa. (E) Velocity depicts that there is no common ancestor for p16+ cells or p21+ cells, but that these are rather distinct entities. (F) The standard error for SASP factors is substantially higher in the p16 compared to the p21 population. (G) Utilizing the SenMayo gene set, SASP factors exclusively expressed by p16+ cells are fewer than those expressed in p21+ cells, with some (Cxcl16, Plaur) being expressed by both. Very few secretory genes are expressed by p16-negative and p21-negative cells. The size of the dots represents the fold change compared to all other populations shown in (C).
Figure 1 Distinct p16+ and p21+ cell populations with differential secretory profiles in the murine hippocampus. (A) A t-SNE plot depicting the five main populations of young and old murine hippocampus (Ogrodnik et al, 2021, GSE161340). (B) From the original population, CD45− cells are selected, followed by Ki67-negativity and cells not in the S phase. From these cells, just p16(Cdkn2a)+ cells and p21(Cdkn1a)+ cells are selected. From the p16+ cells, microglia and oligodendrocytes depict the main populations, while in the p21+ population, microglia is predominant. (C) t-SNE visualization of p16 + (red) and p21+ (blue) cells, along with a few double-positive cells (green), shown in the Cd45-Ki67-S- population. (D) The dPo (double positive) cells are high in both Cdkn2a and Cdkn1a, while p16+ cells just express Cdkn2a, but no Cdkn1a and vice versa. (E) Velocity depicts that there is no common ancestor for p16+ cells or p21+ cells, but that these are rather distinct entities. (F) The standard error for SASP factors is substantially higher in the p16 compared to the p21 population. (G) Utilizing the SenMayo gene set, SASP factors exclusively expressed by p16+ cells are fewer than those expressed in p21+ cells, with some (Cxcl16, Plaur) being expressed by both. Very few secretory genes are expressed by p16-negative and p21-negative cells. The size of the dots represents the fold change compared to all other populations shown in (C).CC BY · PubMed Central
Figure EV1 Comprehensive examination of the mRNA expression profiles for the whole secretome and SenMayo in aging murine hippocampus. (A) p16+ cells are significantly more frequent in the old brain. (B) Microglia has more p21+ cells in the old brain. (C) In the murine hippocampus, there is minimal overlap between the mRNA expression of the whole secretome + SenMayo in p21+ and p16+ cells. (D) SenMayo composition in p16+ vs. p21+ microglia shows some common secretory genes like Plaur and Cxcl16, while in (E) oligodendrocytes, there is no common secretory phenotype. (F) The velocity trajectory of p16+ cells, p21+ cells and the remaining brain cells shows no overlap or common origin (G) Monocle3 applied to the backbone from f shows that there is no common “ancestor” for these cell types. A: unpaired t-test, n = 2 per condition, error bars: sd, B: multiple paired t-tests, n = 2 per condition, error bars: sd.
Figure EV1 Comprehensive examination of the mRNA expression profiles for the whole secretome and SenMayo in aging murine hippocampus. (A) p16+ cells are significantly more frequent in the old brain. (B) Microglia has more p21+ cells in the old brain. (C) In the murine hippocampus, there is minimal overlap between the mRNA expression of the whole secretome + SenMayo in p21+ and p16+ cells. (D) SenMayo composition in p16+ vs. p21+ microglia shows some common secretory genes like Plaur and Cxcl16, while in (E) oligodendrocytes, there is no common secretory phenotype. (F) The velocity trajectory of p16+ cells, p21+ cells and the remaining brain cells shows no overlap or common origin (G) Monocle3 applied to the backbone from f shows that there is no common “ancestor” for these cell types. A: unpaired t-test, n = 2 per condition, error bars: sd, B: multiple paired t-tests, n = 2 per condition, error bars: sd.CC BY · PubMed Central
Figure EV2 A second brain sample (Ximerakis et al, 2019) was used to confirm the findings from the first brain sample. (A) A t-SNE plot indicating all cell types. (B) The selection of p16+ cells focuses on oligodendrocyte precursors, while p21+ cells were mostly endothelial cells. (C) The t-SNE plot demonstrates that most cells are either p16+ or p21+, but double-positive cells are rare. (D) The expression of p16+, p21+ cells, and the rare dPo population. (E) There is only a very small overlap between the p16- and p21-positive secretory phenotype, namely just one gene (Igfbp3). (F) When combining both brain datasets, we observed no consistent overlap between the secretory profiles of p16+ and p21+ cells, underscoring that they represent distinct populations. Shared genes across both datasets are shown in solid color, whereas genes unique to a single dataset are displayed transparently. (G) p16+ cells are significantly more frequent in the old brain. (H) Microglia (MG) has more p21+ cells in the old brain. (G): unpaired t-test, n = 8 per condition, error bars: sd, (H) multiple paired t-tests, n = 8 per condition, error bars: sd.
Figure EV2 A second brain sample (Ximerakis et al, 2019) was used to confirm the findings from the first brain sample. (A) A t-SNE plot indicating all cell types. (B) The selection of p16+ cells focuses on oligodendrocyte precursors, while p21+ cells were mostly endothelial cells. (C) The t-SNE plot demonstrates that most cells are either p16+ or p21+, but double-positive cells are rare. (D) The expression of p16+, p21+ cells, and the rare dPo population. (E) There is only a very small overlap between the p16- and p21-positive secretory phenotype, namely just one gene (Igfbp3). (F) When combining both brain datasets, we observed no consistent overlap between the secretory profiles of p16+ and p21+ cells, underscoring that they represent distinct populations. Shared genes across both datasets are shown in solid color, whereas genes unique to a single dataset are displayed transparently. (G) p16+ cells are significantly more frequent in the old brain. (H) Microglia (MG) has more p21+ cells in the old brain. (G): unpaired t-test, n = 8 per condition, error bars: sd, (H) multiple paired t-tests, n = 8 per condition, error bars: sd.CC BY · PubMed Central
Figure 2 Comparative analysis of p16+ and p21+ cells and their secretory phenotype in different murine tissues during aging. (A) In murine skeletal muscle (Zhang et al, 2022, GSE172410), nine different cell types can be distinguished. (B) p21+ cells constitute the majority, while double positive (dPo) cells are infrequent. (C) The developmental trajectory points to different points of origin for p16+- and p21+ cells. (D) The secretory phenotype is heterogeneous, with p21+ cells expressing a vast array of SASP factors, exhibiting minimal overlap with p16-associated secretory factors. (E) In murine bone (Baryawno et al, 2019, GSE128423), 17 cell types can be distinguished. (F) p21+ cells once again dominate the senescent cell population. (G) There is no common ancestor for p16+- and p21+ cells, and the (H) secretory profile in murine bone remains diverse, with p21+ cells expressing a larger number of secretory factors compared to p16+ cells. (I) In the murine liver (Su et al, 2021, GSE166504), 13 different cell types are identified. (J) The liver has the lowest proportion of p21+ cells from all tissues analyzed, although they still form the majority of senescent cells, with a few dPo cells. (K) There are different origins for p16+ and p21+ cells in the liver. (L) The secretory phenotype linked to p21 is more extensive than that associated with p16, with minimal overlap between the two.
Figure 2 Comparative analysis of p16+ and p21+ cells and their secretory phenotype in different murine tissues during aging. (A) In murine skeletal muscle (Zhang et al, 2022, GSE172410), nine different cell types can be distinguished. (B) p21+ cells constitute the majority, while double positive (dPo) cells are infrequent. (C) The developmental trajectory points to different points of origin for p16+- and p21+ cells. (D) The secretory phenotype is heterogeneous, with p21+ cells expressing a vast array of SASP factors, exhibiting minimal overlap with p16-associated secretory factors. (E) In murine bone (Baryawno et al, 2019, GSE128423), 17 cell types can be distinguished. (F) p21+ cells once again dominate the senescent cell population. (G) There is no common ancestor for p16+- and p21+ cells, and the (H) secretory profile in murine bone remains diverse, with p21+ cells expressing a larger number of secretory factors compared to p16+ cells. (I) In the murine liver (Su et al, 2021, GSE166504), 13 different cell types are identified. (J) The liver has the lowest proportion of p21+ cells from all tissues analyzed, although they still form the majority of senescent cells, with a few dPo cells. (K) There are different origins for p16+ and p21+ cells in the liver. (L) The secretory phenotype linked to p21 is more extensive than that associated with p16, with minimal overlap between the two.CC BY · PubMed Central
Figure EV3 Comprehensive analysis of mRNA expression profiles in aging murine muscle. (A) In the muscle, there is some overlap with a noticeably larger p21-associated secretome compared to p16. (B) The selection criteria for p21+ and p16+ cells. While most of the p16+ cells are FAPs (black), in the p21+ cells there are also some endothelial cells (red). (C) Key markers for the abovementioned populations. (D) Secretory phenotype composition in p16+ and p21 + FAPs. (E) A second muscle dataset (Tabula_Muris_Consortium, 2020) showed nine key cell populations, with (F) most of the p16+ and p21+ cells originating from the mesenchymal stem cell compartment. (G) The majority of cells is p21+. (H) Key markers for these three populations. (I) The secreted factors between these two populations are heterogeneous, and (J) Integration of both datasets shows no overlap in secreted factors between these two populations.
Figure EV3 Comprehensive analysis of mRNA expression profiles in aging murine muscle. (A) In the muscle, there is some overlap with a noticeably larger p21-associated secretome compared to p16. (B) The selection criteria for p21+ and p16+ cells. While most of the p16+ cells are FAPs (black), in the p21+ cells there are also some endothelial cells (red). (C) Key markers for the abovementioned populations. (D) Secretory phenotype composition in p16+ and p21 + FAPs. (E) A second muscle dataset (Tabula_Muris_Consortium, 2020) showed nine key cell populations, with (F) most of the p16+ and p21+ cells originating from the mesenchymal stem cell compartment. (G) The majority of cells is p21+. (H) Key markers for these three populations. (I) The secreted factors between these two populations are heterogeneous, and (J) Integration of both datasets shows no overlap in secreted factors between these two populations.CC BY · PubMed Central
Figure EV5 Liver analysis demonstrates differences between p21+ cells and p16+ cells in aging. (A) Within the liver, there is minimal overlap between p21+ cells and p16+ cells, with the p21+ cells predominantly expressing the majority of factors. (B) The selection of p16+ cells shows mostly endothelial cells, while in the p21+ cells, there are also some hepatocytes. (C) The key markers for the three key populations. (D) p21- and p16-associated SenMayo secretory phenotype in ECs. (E) A second hepatic dataset (Tabula_Muris_Consortium, 2020) was analyzed, while (F) endothelial cells were the most abundant in p16+ cells, and hepatocytes the most prominent in the p21+ cells. (G, H) p21+ cells were the majority compared to p16+ cells, and (I) there was a substantial overlap between p21- and p16-associated secretory factors, although p21+ cells expressed a greater number of factors than p16+ cells. (J) Combining both datasets, there was just one overlapping factor (Cxcl16).
Figure EV5 Liver analysis demonstrates differences between p21+ cells and p16+ cells in aging. (A) Within the liver, there is minimal overlap between p21+ cells and p16+ cells, with the p21+ cells predominantly expressing the majority of factors. (B) The selection of p16+ cells shows mostly endothelial cells, while in the p21+ cells, there are also some hepatocytes. (C) The key markers for the three key populations. (D) p21- and p16-associated SenMayo secretory phenotype in ECs. (E) A second hepatic dataset (Tabula_Muris_Consortium, 2020) was analyzed, while (F) endothelial cells were the most abundant in p16+ cells, and hepatocytes the most prominent in the p21+ cells. (G, H) p21+ cells were the majority compared to p16+ cells, and (I) there was a substantial overlap between p21- and p16-associated secretory factors, although p21+ cells expressed a greater number of factors than p16+ cells. (J) Combining both datasets, there was just one overlapping factor (Cxcl16).CC BY · PubMed Central
Figure EV4 Analysis of the secretome in bone during aging. (A) Similar to the muscle, in the bone, there is some overlap but a more pronounced p21-associated secretome. (B) The selection criteria and key populations (C) for p16+ and p21+ cells. (D) Just Lepr+ MSCs and p16- and p21-associated SenMayo phenotype, respectively. (E) A second bone dataset (Tabula_Muris_Consortium, 2020) shows a (F) high amount of B cells in p16+ cells and monocytes in p21+ cells. (G, H) There are a few p21+ cells, and almost no dPo cells. (I) There is some overlap between p16+ and p21+ cells, while a summary of both bone datasets (J) shows no single gene overlapping for p16- and p21-associated secretory factors.
Figure EV4 Analysis of the secretome in bone during aging. (A) Similar to the muscle, in the bone, there is some overlap but a more pronounced p21-associated secretome. (B) The selection criteria and key populations (C) for p16+ and p21+ cells. (D) Just Lepr+ MSCs and p16- and p21-associated SenMayo phenotype, respectively. (E) A second bone dataset (Tabula_Muris_Consortium, 2020) shows a (F) high amount of B cells in p16+ cells and monocytes in p21+ cells. (G, H) There are a few p21+ cells, and almost no dPo cells. (I) There is some overlap between p16+ and p21+ cells, while a summary of both bone datasets (J) shows no single gene overlapping for p16- and p21-associated secretory factors.CC BY · PubMed Central
Figure EV6 CyTOF analyses of p16, p21 and associated secreted factors in murine bone. (A) CyTOF dataset from murine bone (Doolittle et al, 2023). (B) p16+ and p21+ cells are largely distinct, with minimal overlap. (C) The associated secretory phenotype of p16+ and p21+ cells shows some overlap but is limited to only a few factors, likely reflecting the limited antibody panel available.
Figure EV6 CyTOF analyses of p16, p21 and associated secreted factors in murine bone. (A) CyTOF dataset from murine bone (Doolittle et al, 2023). (B) p16+ and p21+ cells are largely distinct, with minimal overlap. (C) The associated secretory phenotype of p16+ and p21+ cells shows some overlap but is limited to only a few factors, likely reflecting the limited antibody panel available.CC BY · PubMed Central
Figure EV7 Spleen and kidney display p16- and p21-associated differences consistent with previous murine datasets. (A) In the spleen calico dataset (Kimmel et al, 2019, GSE 132901), eight different populations were identified and (B) analyzed according to their Cd45-negativity, Ki67-negativity and S-phase exclusion, followed by p16- (mostly B cells) and p21-positivity (mostly Cd8 macrophages). (C) The majority of cells were p21+, although (D) only a small fraction of total cells were positive for either marker. (E) Secreted factors showed limited overlap, with most associated with p21+ cells. (F) The kidney dataset consisted of 11 different cell types, out of which (G) T cells were the most abundant in p16+ cells and capillary endothelial cells were the most abundant in p21+ cells. (H, I) The majority of cells was p21+. (j) A vast majority of secreted factors was associated with p21+ cells.
Figure EV7 Spleen and kidney display p16- and p21-associated differences consistent with previous murine datasets. (A) In the spleen calico dataset (Kimmel et al, 2019, GSE 132901), eight different populations were identified and (B) analyzed according to their Cd45-negativity, Ki67-negativity and S-phase exclusion, followed by p16- (mostly B cells) and p21-positivity (mostly Cd8 macrophages). (C) The majority of cells were p21+, although (D) only a small fraction of total cells were positive for either marker. (E) Secreted factors showed limited overlap, with most associated with p21+ cells. (F) The kidney dataset consisted of 11 different cell types, out of which (G) T cells were the most abundant in p16+ cells and capillary endothelial cells were the most abundant in p21+ cells. (H, I) The majority of cells was p21+. (j) A vast majority of secreted factors was associated with p21+ cells.CC BY · PubMed Central
Figure EV8 Distinct secretory phenotypes of p16+ and p21+ cells in MASH and aging liver. (A) Analysis of a diseased liver (MASH, GSE218300, (Bendixen et al, 2024)), (B) a western-style diet (wd) compared to chow resulted in a moderate increase of p21+ cells, while (C) aging resulted in a profound increase of p21+ cells. (D) The secretory phenotype in p21+ cells was different to a large degree in aging vs. MASH, with a higher amount of genes upregulated in aging. Similarly, the (E) secretory phenotype in p16+ cells was markedly different in aging vs. MASH, but with more genes upregulated in the western diet compared to aging.
Figure EV8 Distinct secretory phenotypes of p16+ and p21+ cells in MASH and aging liver. (A) Analysis of a diseased liver (MASH, GSE218300, (Bendixen et al, 2024)), (B) a western-style diet (wd) compared to chow resulted in a moderate increase of p21+ cells, while (C) aging resulted in a profound increase of p21+ cells. (D) The secretory phenotype in p21+ cells was different to a large degree in aging vs. MASH, with a higher amount of genes upregulated in aging. Similarly, the (E) secretory phenotype in p16+ cells was markedly different in aging vs. MASH, but with more genes upregulated in the western diet compared to aging.CC BY · PubMed Central
Figure EV9 Overexpression of p16 and p21 reveals distinct secretory profiles (GSE117278). (A) A PCA plot illustrating control-MEFs on day 4 and 10, as well as p16-overexpressing cells (adeno-Cre-EGFP virus Ai14;L-p16 injection into the tail) on day 4 and 10, compared to p21-overexpressing cells (Ai14;L-p21) on day 4 and 10. (B) A comparable number of factors are expressed in p16+ (red)- vs. p21+ (blue) cells, with overlap indicated in yellow.
Figure EV9 Overexpression of p16 and p21 reveals distinct secretory profiles (GSE117278). (A) A PCA plot illustrating control-MEFs on day 4 and 10, as well as p16-overexpressing cells (adeno-Cre-EGFP virus Ai14;L-p16 injection into the tail) on day 4 and 10, compared to p21-overexpressing cells (Ai14;L-p21) on day 4 and 10. (B) A comparable number of factors are expressed in p16+ (red)- vs. p21+ (blue) cells, with overlap indicated in yellow.CC BY · PubMed Central
Figure 3 Comparative analysis of p16+ and p21+ cells and their secretory phenotype in human skin and lung during aging. (A) In human skin samples (Zou et al, 2021, HRA000395), 16 distinguishable cell populations are identified. (B) Among these populations, p21+ cells are notably abundant, while double-positive (dPo) cells are rare. (C) RNA-Velocity reveals a distinct pattern for p16+ and p21+ cells, and a higher heterogeneity regarding the secretory factor expression in p16+ cells. (D) The secretory profile expressed by p21+ cells is significantly larger compared to the secreted factors associated with p16+ cells. (E) Immunofluorescence analysis on aged human skin depicting p16+- (red) and p21+ (green) cells display distinct entities. Scale bar: 20 µm. (F) In the human lung (Jia et al, 2023), GSE122960, GSE128033, GSE130148, and GSE212109), 18 discernible cell types are identified, and (G) within these cell types, p21+ cells constitute the majority of senescent cells. (H) RNA-Velocity reveals no common ancestor of p16+ and p21+ cells, while p16+ cells exhibit a more diverse secretory expression profile. (I) The secretory profile of p21+ cells is extensive compared to that of p16+ cells, with minimal overlap. (J) In human lung tissue (COPD and pulmonary fibrosis/IPF), similarly to aged skin tissue, p21+ (green) and p16+ (red) cells display distinct entities. Scale bar: 20 µm. Source data are available online for this figure.
Figure 3 Comparative analysis of p16+ and p21+ cells and their secretory phenotype in human skin and lung during aging. (A) In human skin samples (Zou et al, 2021, HRA000395), 16 distinguishable cell populations are identified. (B) Among these populations, p21+ cells are notably abundant, while double-positive (dPo) cells are rare. (C) RNA-Velocity reveals a distinct pattern for p16+ and p21+ cells, and a higher heterogeneity regarding the secretory factor expression in p16+ cells. (D) The secretory profile expressed by p21+ cells is significantly larger compared to the secreted factors associated with p16+ cells. (E) Immunofluorescence analysis on aged human skin depicting p16+- (red) and p21+ (green) cells display distinct entities. Scale bar: 20 µm. (F) In the human lung (Jia et al, 2023), GSE122960, GSE128033, GSE130148, and GSE212109), 18 discernible cell types are identified, and (G) within these cell types, p21+ cells constitute the majority of senescent cells. (H) RNA-Velocity reveals no common ancestor of p16+ and p21+ cells, while p16+ cells exhibit a more diverse secretory expression profile. (I) The secretory profile of p21+ cells is extensive compared to that of p16+ cells, with minimal overlap. (J) In human lung tissue (COPD and pulmonary fibrosis/IPF), similarly to aged skin tissue, p21+ (green) and p16+ (red) cells display distinct entities. Scale bar: 20 µm. Source data are available online for this figure.CC BY · PubMed Central
Figure EV10 Human skin secretome, associated with p21 and p16, incorporating datasets from Zou et al, (2021) and Solé-Boldo et al, (2020). (A) In the skin, there is minimal overlap between p16+ and p21+ cells in terms of mRNA expression of the whole secretome. Unlike the majority of examined tissues, the mRNA expression of the entire secretome associated with p16 is as extensive as that associated with p21. (B) CD45-negativity, followed by Ki67-negativity and exclusion of S-phase cells were followed by p21- or p16-positivity for selection of proposedly senescent cells, which key markers are depicted in (C). (D) SenMayo in p16+ and p21+ keratinocytes, respectively. (E) Five human skin samples (GSE130973) were likewise (F) applied these selection criteria, but p16+ cells were lacking, resulting in an abundance of (G) p21+ cells and (H) no p16+ cells. (I) Subsequently, there is just a p21-associated SASP, and no overlap (J) in both datasets for p16+ and p21+ cells. (K) Multiplexed error-robust fluorescence in situ hybridization (MERFISH) analysis on human skin samples, with different cell types within a human skin sample. Within these cell types, (L) p16+ (red) and p21+ (blue) cells are largely distinct, with minimal overlap between these markers. (M) p16+ and p21+-associated secretory profiles within the MERFISH dataset.
Figure EV10 Human skin secretome, associated with p21 and p16, incorporating datasets from Zou et al, (2021) and Solé-Boldo et al, (2020). (A) In the skin, there is minimal overlap between p16+ and p21+ cells in terms of mRNA expression of the whole secretome. Unlike the majority of examined tissues, the mRNA expression of the entire secretome associated with p16 is as extensive as that associated with p21. (B) CD45-negativity, followed by Ki67-negativity and exclusion of S-phase cells were followed by p21- or p16-positivity for selection of proposedly senescent cells, which key markers are depicted in (C). (D) SenMayo in p16+ and p21+ keratinocytes, respectively. (E) Five human skin samples (GSE130973) were likewise (F) applied these selection criteria, but p16+ cells were lacking, resulting in an abundance of (G) p21+ cells and (H) no p16+ cells. (I) Subsequently, there is just a p21-associated SASP, and no overlap (J) in both datasets for p16+ and p21+ cells. (K) Multiplexed error-robust fluorescence in situ hybridization (MERFISH) analysis on human skin samples, with different cell types within a human skin sample. Within these cell types, (L) p16+ (red) and p21+ (blue) cells are largely distinct, with minimal overlap between these markers. (M) p16+ and p21+-associated secretory profiles within the MERFISH dataset.CC BY · PubMed Central
Figure EV11 Analysis of lung reveals limited overlap in the secretory phenotype and distinct developmental profiles of p16- and p21-positive cells. (A) In the human lung, the p21+ cells express a large number of secreted factors, with a very small overlap with p16+ cells. (B) Selection criteria for human lung tissue include CD45-negativity, followed by Ki67-negativity and exclusion of S-phase cells, before the cells are distributed between p16+ or p21+. (C) The markers for these three key populations. (D) p16- and p21-SenMayo associated secretory phenotype in p16+ and p21 + AT2 cells, respectively. (E) The murine Calico lung dataset (Kimmel et al, 2019) provides 13 different cell types, which are selected (F) as abovementioned. (G, H) The vast majority of cells are p21+. (I) The secretory phenotype itself is more abundant in the p21+ cells and there is (I) a small overlap of p16+ and p21+ associated factors between both datasets.
Figure EV11 Analysis of lung reveals limited overlap in the secretory phenotype and distinct developmental profiles of p16- and p21-positive cells. (A) In the human lung, the p21+ cells express a large number of secreted factors, with a very small overlap with p16+ cells. (B) Selection criteria for human lung tissue include CD45-negativity, followed by Ki67-negativity and exclusion of S-phase cells, before the cells are distributed between p16+ or p21+. (C) The markers for these three key populations. (D) p16- and p21-SenMayo associated secretory phenotype in p16+ and p21 + AT2 cells, respectively. (E) The murine Calico lung dataset (Kimmel et al, 2019) provides 13 different cell types, which are selected (F) as abovementioned. (G, H) The vast majority of cells are p21+. (I) The secretory phenotype itself is more abundant in the p21+ cells and there is (I) a small overlap of p16+ and p21+ associated factors between both datasets.CC BY · PubMed Central
Figure 4 Common p16- and p21-associated secretory phenotype, SenSig, and CellAge across murine and human tissues. (A) The p16-associated secretory phenotype is characterized by the predominant presence of Ccl2, Ccl4, Ccl7, Mmp12, and Ccl24, with some markers, such as Tnfrsf11b, expressed in multiple tissues. (B) The p21-associated secretory phenotype displays greater diversity, with genes like Ccl2, Cxcl12, Cxcl16, and Jun commonly expressed across tissues. (C) From the SenMayo panel, the p16-core-secreted factors include Il10, Mif, Timp2, and Tnfrsf11b. (D) The p21-associated secreted factors are dominated by Icam1, Jun, and Edn1 along with Igfbp6 and Tnfrsf1a that are expressed consistently by p21+ cells in at least 11 of the 14 analyzed tissues. (E) A “core”-secreted phenotype for senescent cells comprises Icam1, Jun, Igfbp6, Tnfrsf1a, and Edn1. (F–H) The significantly associated genes with p16 and p21, respectively, from SenSig (Cherry et al, 2023) and (I–K) CellAge (Avelar et al, 2020) gene sets. The depicted genes are those significantly overexpressed (p < 0.05, FC >0), with cell type colors corresponding to those in Figs. 1–3. In (A, B), the dot size reflects the log2FC, the circular y-axis is in log10- scale. For (C) only markers expressed in one or more tissues are shown; in (D) those expressed in eight or more tissues; in (F) those expressed in four tissues; and in (G), those expressed in more than 11 tissues are displayed.
Figure 4 Common p16- and p21-associated secretory phenotype, SenSig, and CellAge across murine and human tissues. (A) The p16-associated secretory phenotype is characterized by the predominant presence of Ccl2, Ccl4, Ccl7, Mmp12, and Ccl24, with some markers, such as Tnfrsf11b, expressed in multiple tissues. (B) The p21-associated secretory phenotype displays greater diversity, with genes like Ccl2, Cxcl12, Cxcl16, and Jun commonly expressed across tissues. (C) From the SenMayo panel, the p16-core-secreted factors include Il10, Mif, Timp2, and Tnfrsf11b. (D) The p21-associated secreted factors are dominated by Icam1, Jun, and Edn1 along with Igfbp6 and Tnfrsf1a that are expressed consistently by p21+ cells in at least 11 of the 14 analyzed tissues. (E) A “core”-secreted phenotype for senescent cells comprises Icam1, Jun, Igfbp6, Tnfrsf1a, and Edn1. (F–H) The significantly associated genes with p16 and p21, respectively, from SenSig (Cherry et al, 2023) and (I–K) CellAge (Avelar et al, 2020) gene sets. The depicted genes are those significantly overexpressed (p < 0.05, FC >0), with cell type colors corresponding to those in Figs. 1–3. In (A, B), the dot size reflects the log2FC, the circular y-axis is in log10- scale. For (C) only markers expressed in one or more tissues are shown; in (D) those expressed in eight or more tissues; in (F) those expressed in four tissues; and in (G), those expressed in more than 11 tissues are displayed.CC BY · PubMed Central
Figure EV12 In an adipose tissue dataset (Wang et al, 2024a, GSE269660), the GFP/Cdkn1a-positive cells exhibited an enrichment of the p21-associated secretory phenotype. (A) In adipose tissue, ten different populations were identified, out of which the (B) macrophages were the highest in p21/GFP. (C) The p21-associated SenMayo, (D) SenSig, and (E) CellAge were all (F) significantly higher in the GFP-positive cells compared to the GFP-negative cells. (G) Likewise, a samplewise comparison (n = 10) between GFP-positive (n = 5) and GFP-negative (n = 5) samples shows an increase of p21-associated SenMayo in the positive samples. ****p value <0.0001, ***p value <0.001, **p value <0.01, *p value <0.05. (C–E) Exact p values = 2 × 10−16, Wilcoxon test, (G) Exact p value = 0.004, t-test, error bars: 2 SE. n = 5 per condition
Figure EV12 In an adipose tissue dataset (Wang et al, 2024a, GSE269660), the GFP/Cdkn1a-positive cells exhibited an enrichment of the p21-associated secretory phenotype. (A) In adipose tissue, ten different populations were identified, out of which the (B) macrophages were the highest in p21/GFP. (C) The p21-associated SenMayo, (D) SenSig, and (E) CellAge were all (F) significantly higher in the GFP-positive cells compared to the GFP-negative cells. (G) Likewise, a samplewise comparison (n = 10) between GFP-positive (n = 5) and GFP-negative (n = 5) samples shows an increase of p21-associated SenMayo in the positive samples. ****p value <0.0001, ***p value <0.001, **p value <0.01, *p value <0.05. (C–E) Exact p values = 2 × 10−16, Wilcoxon test, (G) Exact p value = 0.004, t-test, error bars: 2 SE. n = 5 per conditionCC BY · PubMed Central
Figure 5 Unequal communicational patterns between p21+ and p16+ cells across tissues. (A) The network centrality scores depict low (blue) and high (red) communication networks for seventeen significant interactions, with p21+ and p16+ cells shown pairwise in each tissue. (B) In the brain, p21+ cells exhibit higher interaction strength in CCL communication, primarily to microglia cells, while p16+ cells demonstrate increased secretory activity via MAG to oligodendrocytes. (C) In bone, p21+ cells exclusively utilize the THBS and FN1 pathways to communicate with hypertrophic chondrocytes, osteoblasts, and lymphocytes. (D) In the skin, p16+ cells show higher interaction strength in the CD99-pathway to fibroblasts and melanocytes, while p21+ cells predominantly signal through the DESMOSOME-pathway to granular and spinous cells, including sebocytes. (E) In muscle, p21+ cells mainly signal through the COLLAGEN and LAMININ pathways to fibro-adipogenic progenitors (FAPs) and tendon cells, as p16+ cells are too few for calculating communicational patterns. (F) In the liver, the JAM pathway is predominantly used by p21+ cells to communicate with myofibroblasts and hepatic progenitors, whereas p16+ cells primarily employ the CDH5-pathway to interact with endothelial and Kupffer cells. (G) In the lung, p16+ cells exhibit the highest signaling strength in the CD45 pathway, mainly used to communicate with NK cells and T lymphocytes, while p21+ cells primarily use the MIF pathway for contact with lymphatic endothelial and NK cells. Displayed are only significant interactions (padj < 0.05), with cell type colors corresponding to those in Figs. 1–3. The color scale (blue-white-red) of the interactions in (B–G) corresponds to the centrality score in (A).
Figure 5 Unequal communicational patterns between p21+ and p16+ cells across tissues. (A) The network centrality scores depict low (blue) and high (red) communication networks for seventeen significant interactions, with p21+ and p16+ cells shown pairwise in each tissue. (B) In the brain, p21+ cells exhibit higher interaction strength in CCL communication, primarily to microglia cells, while p16+ cells demonstrate increased secretory activity via MAG to oligodendrocytes. (C) In bone, p21+ cells exclusively utilize the THBS and FN1 pathways to communicate with hypertrophic chondrocytes, osteoblasts, and lymphocytes. (D) In the skin, p16+ cells show higher interaction strength in the CD99-pathway to fibroblasts and melanocytes, while p21+ cells predominantly signal through the DESMOSOME-pathway to granular and spinous cells, including sebocytes. (E) In muscle, p21+ cells mainly signal through the COLLAGEN and LAMININ pathways to fibro-adipogenic progenitors (FAPs) and tendon cells, as p16+ cells are too few for calculating communicational patterns. (F) In the liver, the JAM pathway is predominantly used by p21+ cells to communicate with myofibroblasts and hepatic progenitors, whereas p16+ cells primarily employ the CDH5-pathway to interact with endothelial and Kupffer cells. (G) In the lung, p16+ cells exhibit the highest signaling strength in the CD45 pathway, mainly used to communicate with NK cells and T lymphocytes, while p21+ cells primarily use the MIF pathway for contact with lymphatic endothelial and NK cells. Displayed are only significant interactions (padj < 0.05), with cell type colors corresponding to those in Figs. 1–3. The color scale (blue-white-red) of the interactions in (B–G) corresponds to the centrality score in (A).CC BY · PubMed Central
Figure EV13 SCENIC analysis of the regulating transcription factors of p16+ and p21+ cells. (A) When analyzing five distinct tissues for regulatory factors in p16+ cells, each tissue exhibits a substantial array of transcription factors governing the behavior of p16+ cells. (Since the liver has very few p16-positive cells for proper calculation with SCENIC, these were excluded). (B) The x-axis represents the number of genes, plotted against the y-axis, illustrating the relative activity of the respective transcription factor, highlighting the significance of each factor. Interestingly, only four factors (Sox8, Creb3l2, Sox10, and Relb) exhibit consistency across multiple tissues in the context of p16+ cells. (C) In p21+ cells, the regulating transcription factors show a high heterogeneity across tissues. (D) The transcription factors associated with p21 are predominantly specific to a single organ, but a few transcription factors, such as Erg, Sox17, Klf4, Jun, Klf2, and others, regulate p21+ cells in two tissues.
Figure EV13 SCENIC analysis of the regulating transcription factors of p16+ and p21+ cells. (A) When analyzing five distinct tissues for regulatory factors in p16+ cells, each tissue exhibits a substantial array of transcription factors governing the behavior of p16+ cells. (Since the liver has very few p16-positive cells for proper calculation with SCENIC, these were excluded). (B) The x-axis represents the number of genes, plotted against the y-axis, illustrating the relative activity of the respective transcription factor, highlighting the significance of each factor. Interestingly, only four factors (Sox8, Creb3l2, Sox10, and Relb) exhibit consistency across multiple tissues in the context of p16+ cells. (C) In p21+ cells, the regulating transcription factors show a high heterogeneity across tissues. (D) The transcription factors associated with p21 are predominantly specific to a single organ, but a few transcription factors, such as Erg, Sox17, Klf4, Jun, Klf2, and others, regulate p21+ cells in two tissues.CC BY · PubMed Central

Verification

Claims the paper makes

28 Valid · 2 Not valid · 5 Not checked · 102 Own results and interpretations

66 more claims · 3 with their cited sources checked

Create a free account to see all 137 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 25
  • It may state two separate claims as one. 46

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Self-check results appear here once an independent check has confirmed them.

Cited works that were retracted

Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.

1 of 1

Citations

Cites

1 retracted

Show all 66 references

Research works printed but not linked: 1

  1. Tabula_Muris_Consortium (2020) A single-cell transcriptomic atlas characterizes ageing tissues in the mouse. Nature 583:590–595 · DOI · Paper page

Cited by

See all 45

Metadata

Authors
Saul, D., Jurk, D., Doolittle, M. L., Kosinsky, R. L., Han, Y., Zhang, X., Franco, A. C., Kim, S. Y., Wyles, S. P., Prakash, Y. S., Monroe, D. G., Ferrucci, L., LeBrasseur, N. K., Robbins, P. D., Niedernhofer, L. J., Khosla, S., & Passos, J. F.
Affiliations
  • Mayo Clinic (US)
  • Robert Bosch Hospital (DE)
  • Konkuk University (KR)
  • National Institute on Aging (US)
  • University of Minnesota (US)
Corresponding author
  • Saul, Dominik
  • Khosla, Sundeep
  • Passos, João F
Venue
The EMBO Journal
Published
2025-10-29
Volume
44
Issue
23
Pages
7295-7325
ISSN
  • 0261-4189 (ISSN-L)
  • 1460-2075
OpenAlex
W4415668439
Cited by
45
References
66
Citations by year
2026: 44 · 2025: 1
FWCI
28.68
Type
journal article
Language
en
Open access
Licence
Locations
  • The EMBO Journal (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • GoeScholar The Publication Server of the Georg-August-Universität Göttingen (Georg-August-Universität Göttingen) (submittedVersion, cc-by)
Funders
  • HHS | National Institutes of Health (P01 AG062413, R01AG068048; R01AG827; UG3CA26810308; UH3CA268103, R01 AG068182, R01 AG086085, R01 AG055529, R01AG82708, R01 AG063543, R01AG07651, U54 AG079754, U54 AG076041, U19 AG056278)
  • Hevolution (HR-GRO-23-1199144-8)
  • National Research Foundation of Korea (2020M3A9D8038014)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cell Biology › Telomeres, Telomerase, and Senescence
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Aging › Genetics, Aging, and Longevity in Model Organisms
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
Keywords
cellular senescence, senescence-associated secretory phenotype, p16INK4a, p21 Cip1, single-cell transcriptomics, RNA velocity, pseudotime
MeSH terms
Senescence-Associated Secretory Phenotype, Aging, Animals, Humans, Cellular Senescence, Cyclin-Dependent Kinase Inhibitor p16, Cyclin-Dependent Kinase Inhibitor p21, Mice, Single-Cell Analysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-29, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-29 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Distinct senotypes in p16- and p21-positive cells across human and mouse aging tissues. https://app.cochat.ai/papers/10.1038/s44318-025-00601-2
The paper
Saul, D., Jurk, D., Doolittle, M. L., Kosinsky, R. L., Han, Y., Zhang, X., Franco, A. C., Kim, S. Y., Wyles, S. P., Prakash, Y. S., Monroe, D. G., Ferrucci, L., LeBrasseur, N. K., Robbins, P. D., Niedernhofer, L. J., Khosla, S., & Passos, J. F. (2025). Distinct senotypes in p16- and p21-positive cells across human and mouse aging tissues. The EMBO Journal. https://doi.org/10.1038/s44318-025-00601-2

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.