Article

Comprehensive human proteome profiles across a

50-year lifespan reveal aging trajectories and
signatures
Graphical abstract Authors
Yingjie Ding, Yuesheng Zuo,
Bin Zhang, ..., Jing Qu, Weiqi Zhang,
Guang-Hui Liu

Correspondence
[email protected] (J.Y.),
[email protected] (J.Q.),
[email protected] (W.Z.),
[email protected] (G.-H.L.)

In brief
A 50-year multi-tissue proteomic atlas
reveals transcriptome-proteome
decoupling, proteostasis decline with
amyloid accumulation, and
asynchronous aging clocks, identifying
circulating senoproteins contributing to
vascular and systemic aging.

Highlights
• A proteomic blueprint of human organs across 50-year aging
stages

• Transcriptome-proteome decoupling and proteostasis

failure mark aged tissues

• Human organ proteomic clocks reveal aging inflection and

asynchrony

• Circulating senoproteins contribute to vascular and

systemic aging

Ding et al., 2025, Cell 188, 1–22

October 2, 2025 © 2025 Elsevier Inc. All rights are reserved, including those for
text and data mining, AI training, and similar technologies.
https://doi.org/10.1016/j.cell.2025.06.047 ll
 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll

Article
Comprehensive human proteome profiles across
a 50-year lifespan reveal aging trajectories
and signatures
Yingjie Ding,1,6,15 Yuesheng Zuo,1,6,15 Bin Zhang,2,6,15 Yanling Fan,1,15 Gang Xu,4,5,15,16 Zhongyi Cheng,12,15
Shuai Ma,2,3,6,13,14,15 Shuaiqi Fang,2,6,15 Ao Tian,2,6 Dandan Gao,8 Xi Xu,4,5 Qiaoran Wang,1,6 Yaobin Jing,7
Mengmeng Jiang,2,3 Muzhao Xiong,1,6 Jiaming Li,1,6 Zichu Han,1,6 Shuhui Sun,9 Si Wang,10,14,16 Fuchu He,11,16
Jiayin Yang,4,5,* Jing Qu,2,3,6,9,13,14,* Weiqi Zhang,1,3,6,13,14,* and Guang-Hui Liu2,3,6,10,13,14,17,*
1China National Center for Bioinformation and Beijing Institute of Genomics, Chinese Academy of Sciences, Beijing 100101, China
2State Key Laboratory of Organ Regeneration and Reconstruction, Institute of Zoology, Chinese Academy of Sciences, Beijing 100101, China
3Beijing Institute for Stem Cell and Regenerative Medicine, Beijing 100101, China
4Liver Transplant Center, Organ Transplant Center, West China Hospital of Sichuan University, Chengdu 610000, China
5Laboratory of Liver Transplantation, Key Laboratory of Transplant Engineering and Immunology, NHC, West China Hospital of Sichuan

University, Chengdu 610000, China

6University of Chinese Academy of Sciences, Beijing 100049, China
7International Center for Aging and Cancer, Hainan Academy of Medical Sciences, Hainan Medical University, Haikou 571199, China
8Department of Respiratory and Critical Care, Quzhou Affiliated Hospital of Wenzhou Medical University, Quzhou People’s Hospital, Quzhou

324000, China
9Beijing Anzhen Hospital, Capital Medical University, Beijing Institute of Heart, Lung, and Blood Vessel Diseases, Beijing 100029, China
10Advanced Innovation Center for Human Brain Protection, National Clinical Research Center for Geriatric Disorders, Aging Translational

Medicine Center, Beijing Municipal Geriatric Medical Research Center, Beijing Key Laboratory of Environment and Aging, Xuanwu Hospital
Capital Medical University, Beijing 100053, China
11State Key Laboratory of Medical Proteomics, Beijing Proteome Research Center, National Center for Protein Sciences (Beijing), Beijing

Institute of Lifeomics, Beijing 102200, China

12Jingjie PTM BioLab (Hangzhou) Co., Inc., HangZhou 310000, China
13Human Organ Physiopathology Emulation System, Institute of Zoology, Chinese Academy of Sciences, Beijing 100101, China
14Aging Biomarker Consortium (ABC), Beijing 100101, China
15These authors contributed equally
16Senior author
17Lead contact

*Correspondence: [email protected] (J.Y.), [email protected] (J.Q.), [email protected] (W.Z.), [email protected] (G.-H.L.)

https://doi.org/10.1016/j.cell.2025.06.047

SUMMARY

Proteins are the cornerstone of life. However, the proteomic blueprint of aging across human tissues remains
uncharted. Here, we present a comprehensive proteomic and histological analysis of 516 samples from 13
human tissues spanning five decades. This dynamic atlas reveals widespread transcriptome-proteome de-
coupling and proteostasis decline, characterized by amyloid accumulation. Based on aging-associated pro-
tein changes, we developed tissue-specific proteomic age clocks and characterized organ-level aging trajec-
tories. Temporal analysis revealed an aging inflection around age 50, with blood vessels being a tissue that
ages early and is markedly susceptible to aging. We further defined a plasma proteomic signature of aging
that matches its tissue origins and identified candidate senoproteins, including GAS6, driving vascular and
systemic aging. Together, our findings lay the groundwork for a systems-level understanding of human aging
through the lens of proteins.

INTRODUCTION and an elevated risk of chronic diseases, especially those of the

cardiovascular system.6–9 The onset of many chronic conditions
Among mammals, humans are notable for their extended life- is linked to organ aging, underscoring the significant potential for
spans, which, however, come with the burden of systemic ag- research in this field to inform strategies that could delay aging
ing.1–5 This aging is accompanied by a decline in organ function and the associated pathologies.10–17 Yet, despite the urgency,

Cell 188, 1–22, October 2, 2025 © 2025 Elsevier Inc. 1

All rights are reserved, including those for text and data mining, AI training, and similar technologies.
 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

B C D E

F G H I

(legend on next page)

2 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

our understanding of the aging patterns within individual human gan chronometry and uncovering the nexus between multi-or-
organs remains fragmented, and the molecular mechanisms that gan aging, biological markers, chronic disease risks, and ther-
govern the complex interplay of multi-organ aging are largely un- apeutic targets.
charted territories.18–21
Aging trajectories differ across organs, yet these organs RESULTS
exhibit close interactions and act synergistically to promote
organismal aging.22,23 Our understanding of both the shared A half-century proteomic compendium of human organ
and organ-specific characteristics of aging, as well as the sys- aging
temic changes they collectively undergo, remains limited. For To elucidate the intricate tapestry of human aging, we embarked
the development of innovative interventions targeting aging, pre- on a journey, compiling a proteomic encyclopedia that spans a
cise metrics that capture both organ-specific aging and sys- spectrum of tissues from individuals aged 14 to 68, inclusive of
temic aging heterogeneity are essential.24–32 The absence of both genders (Figure 1A; Table S1). Our meticulously curated
biomarkers capable of reflecting the aging pace of various or- collection, reflecting functional diversity and aging relevance,
gans complicates the establishment of organ-specific biological encompasses tissues from seven bodily systems: cardiovascu-
age.33–37 Furthermore, our understanding of the spatiotemporal lar (heart and aorta), digestive (liver, pancreas, and intestine), im-
dynamics in multi-organ aging is inadequate, and there is a crit- mune (spleen and lymph node), endocrine (adrenal gland and
ical need to identify tissues sensitive to aging at different life white adipose), respiratory (lung), integumentary (skin), and
stages and transition points.22,38–42 musculoskeletal (muscle), along with blood samples.
Proteins, the orchestrators of cellular function, face unceasing Using state-of-the-art mass spectrometry,52,53,58 we per-
disruptions to homeostasis as they age.4,26,43–49 It is crucial to formed quantitative proteomic analyses on each sample,
detail the decline of proteostasis across the entire spectrum of accompanied by parallel transcriptomic assays. This rigorous
protein quality control.50 However, our understanding of proteo- approach yielded a comprehensive catalog of over 12,771
mic changes within aged human tissues remains strikingly distinct proteins, covering the human protein-coding genome
limited, impeding a deep understanding of the fundamental across all chromosomes (Figures 1B, 1C, and S1A; Table S1).
mechanisms of aging and their progression to disease. Utilizing Beyond high-abundance proteins essential for translation and
state-of-the-art proteomic methods to quantify thousands of metabolic functions, we identified a substantial number of low-
plasma proteins with precision provides a systematic approach abundance proteins across tissues, including transporters and
to unraveling molecular responses to aging.40,51–57 Yet, research hormone regulators (Figure 1D). We also profiled the character-
has mainly focused on plasma proteomics, often overlooking or- istics of detected proteins based on their subcellular localization,
gan-specific proteostasis disruptions, protein exchanges be- cellular component, and molecular function. The majority were
tween the vasculature and various organs, and the organ origins intracellular, with additional detection of membrane-bound and
of these blood-based proteomic shifts. secreted proteins. A large number localized to nuclear and
Our study is poised to construct a comprehensive multi-tis- organelle compartments, particularly mitochondria. Function-
sue proteomic atlas spanning 50 years of the entire human ag- ally, these proteins span diverse roles, including transcriptional
ing process, elucidating the mechanisms behind proteostasis and epigenetic regulation, as well as receptor-ligand signaling,
imbalance in aged organs and revealing both universal and tis- with notable heterogeneity across tissues (Figure 1E). Uniform
sue-specific aging patterns. This work aims to broaden our manifold approximation and projection (UMAP) unveiled highly
comprehension of the proteomic signatures of aging and to en- tissue-specific proteomic signatures, with discernible similarities
gineer organ aging clocks that pinpoint the onset tissues and between heart and skeletal muscle, as well as between spleen
pivotal time points. Additionally, we seek to explore the dy- and lymph node tissues (Figures S1B and S1C). We identified tis-
namic interplay between circulating and organ proteins during sue-enriched and tissue-enhanced proteins, as well as those
aging, with the goal of identifying systemic biomarkers for or- common across tissues, which are vital for basic housekeeping

Figure 1. A multi-tissue proteomic atlas outlines age- and tissue-associated protein features
(A) Schematic overview of the study. Samples spanning > 50 years of age were collected from 76 individuals across 13 tissues (n = 516). Deep proteomic,
transcriptomic, and histological profiling enabled the construction of a tissue-specific aging atlas, incorporating aging clocks and biomarkers.
(B) Left: number of proteins identified per tissue. Right: density distribution of protein size, and darker colors indicate higher counts.
(C) Chromosomal distribution of detected proteins across tissues, with red indicating higher density.
(D) Left: protein abundance heatmap across tissues. Right: Gene Ontology (GO) enrichment for high/low-abundance proteins. Blue to red: low to high abundance.
(E) Dot plot of detected proteins by subcellular localization, cellular component, and function across tissues. Dot size: relative proportion; color bar: protein count.
(F) Ring chart of proteins per tissue: dark gray (matched, protein with mRNA), light gray (unmatched, protein without mRNA).
(G) Density plots of mRNA-protein Spearman’s correlations per tissue.
(H) Violin plots of age-stratified Spearman’s correlations between mRNA and protein levels (only positively correlated proteins are shown).
(I) Left: negatively aging-related proteins in proteostasis. Right: positively correlated proteins (amyloids, immunoglobulins, and complement components). The
diagram on the far right illustrates the molecular interactions among these proteins that are supported by the literature. Bar plot shows gene set enrichment
analysis (GSEA) scores for selected pathways.
(J) Immunofluorescence (IF) staining of SAA1/2, IGLC, IGHA, and C4A in lymph, heart, and intestine (n = 22–27). Each point: one individual. Correlation by
Spearman’s test. Scale bars, 50 μm.
See also Figures S1 and S2.

Cell 188, 1–22, October 2, 2025 3

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

A B

D E F

(legend on next page)

4 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

functions in biology. Tissue-enriched proteins, with their distinct and mitochondrial ribosomal proteins as well as aminoacyl-
expression profiles, underscore the specific physiological roles tRNA synthetases and peptidyl-tRNA hydrolase (Figure 1I).
of each tissue (Figures S1D–S1H; Table S1). Even within analo- This suggests a potential decline in translational capacity with
gous organs, our analysis precisely identified differentially ex- age. Post-translationally, proteins involved in folding, assembly,
pressed proteins (DEPs), as validated by western blotting for and transport, such as those in the chaperonin-containing
specific expressions, including SULT1B1, a liver xenobiotic- T-complex protein 1 subunit alpha (TCP1), Sec1/Munc18-family,
metabolizing enzyme59; PNLIPRP2, a pancreatic lipase that hy- endoplasmic reticulum membrane complex, and nuclear pore
drolyzes galactolipids60; aquaporin 4 (AQP4), a sarcolemma pro- complex, were found to decrease with aging (Figures 1I and
tein in muscle; and desmoglein 2 (DSG2), a desmosomal protein S2F). For example, histidyl-TRNA synthetase 1 (HARS1), a his-
in heart (Figure S1F).61,62 Additionally, we examined the impact tidyl-tRNA synthetase; chaperonin containing TCP1 subunit 5
of biological sex on protein expression across tissues and found (CCT5) and chaperonin containing TCP1 subunit 7 (CCT7), sub-
that only a few proteins exhibited sex-differentiated abundance, units of the ring-shaped chaperonin T-complex protein ring com-
such as the epigenetic regulator SGF29 in female hearts and the plex (TRiC); and proteasome subunit beta type-1 (PSMB1), a
matrix metalloprotease 3 (MMP3) in male white adipose tissue non-catalytic subunit of the 20S proteasome, were all reduced
(Figures S1I–S1K). at the protein level in aged tissues (Figure S2G). Consistent
with the decline in protein quality control, there is an accumula-
Aging disrupts mRNA-protein coupling and proteomic tion of amyloid proteins alongside immunoglobulin and comple-
homeostasis across tissues ment (Figure 1I). Proteomics data analysis and tissue immunoflu-
Protein homeostasis, a system encompassing the cellular mech- orescence analysis confirmed the accumulation of amyloid
anisms that govern protein synthesis, folding, modification, proteins, such as serum amyloid A1 (SAA1) and serum amyloid
translocation, and degradation, is a critical facet of cellular health A2 (SAA2), and the excessive activation of the immune pathway
(Figure S2A).43,44,63–65 To assess the transcriptional regulation of (such as immunoglobulins lambda constant 1 [IGLC], immuno-
proteomic patterns, we compared proteomic profiles with tran- globulin heavy constant alpha 1 [IGHA], and complement C4A)
scriptomic data. Over 95% of identified proteins showed a (Figures 1I and 1J). Considering that aberrant amyloid deposition
detectable transcriptional signal, yet the correlation between can trigger immunoglobulin activation, leading to the comple-
mRNA and protein levels was only moderately aligned (correla- ment pathway engagement,66–69 the amyloid-immunoglobulin-
tions ranged from 0.39 to 0.50) (Figures 1F and 1G). The positive complement axis may constitute a crucial component of the ag-
correlation diminishes with age, particularly pronounced in the ing tissue microenvironment.
spleen, muscle, and lymph nodes, underscoring an age-depen- We further observed increased variability in tissue-specific
dent decoupling between protein and transcription (Figures 1H protein expression in older individuals across organs, as indi-
and S2B–S2E; Table S2). Notably, these decoupled proteins cated by elevated standard deviation (SD) and coefficient of vari-
are pivotal for tissue functional maintenance, such as regulating ation (CV) in these proteins in aged samples (Figure S2H). In
B cell activation in the spleen (Figure S2D). These findings under- addition, we found a marked reduction in sex-biased proteomic
score the increasing disparity between mRNA and protein levels signatures in aged tissues (Figure S2I), suggesting a conver-
in aging organs, stressing the importance of understanding the gence of sex-specific expression patterns over time.
proteomic aspects of aging.
We next examined the key pathways across the entire protein Proteomic dynamics and biomarkers of aging in human
metabolism lifecycle, which may influence the protein quality organs
control capacity and subsequently cause mRNA-protein decou- To uncover the lifelong dynamics of the human proteome, we
pling during aging (Figure 1I). At the translational level, we focused on the proteins with differential abundance associated
observed a decrease in the abundance of key proteins involved with aging, identifying a substantial number of age-dependent
in protein synthesis across most tissues, including cytoplasmic that exhibited altered expression in at least one organ during

Figure 2. Molecular characterization of age-dependent DEPs

(A) Top: LOESS-fitted trajectories of scaled expression for upregulated (left) and downregulated (right) DEPs by tissue (number and proportion shown). Bottom:
heatmaps of scaled expression of upregulated and downregulated DEPs across tissues and ages, with subcellular localization distribution indicated.
(B) Heatmaps of mean Spearman’s correlation coefficients for upregulated (left) and downregulated (right) DEP modules across tissues. Blue to red: low to high
correlation coefficients.
(C) Age trajectories of 10 DEP modules with increased (top) or decreased (bottom) abundance. Each line: a tissue. Tissues exhibiting pronounced trends are
highlighted, with representative pathways annotated.
(D) Circos plots of shared and tissue-specific upregulated (left) and downregulated (right) DEPs. Arcs: shared proteins. Outer ring colors: tissues from (A).
(E) Heatmaps of DEPs upregulated (left) and downregulated (right) across tissues. Only DEPs with consistent changes across multiple tissues are shown. Proteins
in six or more tissues highlighted as UUPA and UDPA. Heatmap column colors: tissues from (A).
(F) Representative GO terms enriched in UUPA (top) and UDPA (bottom).
(G) SAP protein levels increase with age across tissues (proteomic analysis, Spearman’s correlation). Each point: one individual. Aorta (n = 24), heart (n = 29),
pancreas (n = 51), lung (n = 14), liver (n = 45), adrenal (n = 66), intestine (n = 28).
(H) Immunohistochemical (IHC) staining shows age-associated SAP upregulation across tissues. Each point: one individual. Aorta (n = 25), heart (n = 27),
pancreas (n = 34), lung (n = 26), liver (n = 50), adrenal (n = 57), intestine (n = 25). Spearman’s correlation was applied. Scale bars, 50 μm.
See also Figure S3.

Cell 188, 1–22, October 2, 2025 5

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

A B

C D

E F

(legend on next page)

6 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

human aging. Age-dependent DEPs were observed in various immune cell adhesion, and increases interleukin-6 (IL-6)
tissues, with some tissues showing pronounced sensitivity to ag- secretion (Figures S3G–S3K).
ing (Figures 2A, S3A, and S3B; Table S2). Through in-depth char-
acterization of the expression patterns of all DEPs across various Proteomic perturbation of biological pathways in aged
organs, we identified distinct modules for upregulated and human organs
downregulated DEPs. This analysis revealed organ-specific pro- To clarify proteomic shifts in aging organs, we established an ag-
tein signatures that reflect the dynamic changes associated with ing hallmark assessment system to track aging characteristics.
aging (Figures 2B, 2C, and S3C–S3F). Module 1 of upregulated This system includes markers for cell cycle arrest, genomic
DEPs prevalent across tissues and consistently rising with age instability, loss of heterochromatin, inflammation, and fibrosis
contains proteins involved in complement activation, immune (Figure 3A). In line with the proteomic data, histological staining
response, and amyloid fibril formation (Figure 2C), highlighting revealed elevated p21Cip1 expression, enhanced DNA damage
that these proteins form a ‘‘protein community’’ that synergisti- signals, reduced heterochromatin mark H3K9me3, increased
cally increases with aging across tissues. Other modules primar- pro-inflammatory factor IL-6, and alongside an elevation in tis-
ily consisted of tissue-specific DEPs that increased with age, sue fibrosis indicated by collagen IV in aged tissues, all of which
with corresponding pathways notably enriched in the aorta, are established biomarkers of aging (Figures 3B and S4A).
lung, heart, spleen, and lymph nodes, respectively (Figures 2C, Epigenetic instability is a key driver of aging, leading us to un-
S3C, and S3E). Regarding downregulated DEPs, we detected cover changes in epigenetic regulatory proteins associated
a steady decrease in proteins related to Golgi vesicle transport with aging, including those that regulate DNA methylation,
and protein maturation across distinct tissues. Other modules histone modification, chromatin remodeling, and RNA epige-
comprised tissue-specific DEPs that declined progressively netics (Figures 3C and S4B).9,26,70–80 Notably, YTHDC1 emerged
with age, with the downregulated pathways in various tissues as the most consistently downregulated epigenetic regulator
partially mirroring the aging-induced deterioration in tissue func- across tissues, and its reduced expression in aged liver and
tions (Figures 2C, S3D, and S3F). adrenal tissues was further validated by histological staining
Our analysis also revealed that while the majority of DEPs (Figure S4C). Among the aging hallmarks, we also observed
are tissue-specific, approximately 40% were found to be a widespread downregulation of mitochondrial proteins impli-
shared among at least two tissues (Figures 2D and 2E). cated in mitochondrial biogenesis, homeostasis, and respiration,
Notably, 31 proteins are uniformly downregulated across underscoring the importance of mitochondria in aging
over six tissues without upregulation in any organ. These (Figure 3D).
are termed ubiquitous downregulated proteins with aging Proteins often exert their functions within complexes, prompt-
(UDPAs), including pre-mRNA splicing tri-snRNP complex ing our quantitative dissection of the complexome to delineate
factor (PRPF4) involved in pre-mRNA splicing (Figures 2E aging-associated alterations. We analyzed a total of 816 protein
and 2F). Conversely, 29 proteins are consistently upregulated complexes and observed the downregulation of transcriptional
across more than six tissues without conflicting trends regulatory complexes (e.g., AR-NONO-SFPQ) and RNA splicing
(Figures 2E and 2F). These are designated ubiquitous upregu- complexes (e.g., spliceosome) across tissues (Figures 3E and
lated proteins with aging (UUPAs), associated with comple- S4D). By contrast, upregulated complexes were enriched in im-
ment activation, inflammatory response, and cell adhesion. mune and inflammatory responses (e.g., MAVS-TRADD) and
Among them, serum amyloid P-component (SAP, encoded extracellular matrix and cell adhesion (e.g., ITGA3-ITGB5) in
by APCS) is the most prominent UUPA, a canonical amyloid various tissues (Figure 3E). Gene set enrichment analysis
factor and a prime example of mRNA-protein decoupling (GSEA) corroborated these findings, revealing that upregulated
in the liver (Figures 2E, 2G, and S2E). Its elevation in multiple complexes were predominantly enriched in the inflammasome
aged tissues is confirmed by histological staining (Figure 2H). pathway, whereas downregulated complexes were linked to nu-
This protein’s upregulation with aging may promote vascular clear hormone receptors, nuclear transport, and spliceosomes
aging, damage, and inflammation, as treating human aortic (Figure 3F).
endothelial cells (hAECs) with recombinant SAP protein in- Finally, to elucidate the connection between tissue aging and
creases SA-β-Gal positivity, impairs tube formation, enhances chronic disease susceptibility, we constructed a protein-disease

Figure 3. Shared aging proteomic signatures converge on core pathways across tissues
(A) Ridge plots of age-correlated proteins from GSEA. Spearman’s correlation was applied. Normalized enrichment score (NES) indicated by color.
(B) IHC or IF staining of aging markers in tissues. p21Cip1, γH2AX, H3K9me3, IL-6, and collagen IV were quantified and correlated with age. Each point: one
individual. Sample sizes vary by marker and tissue. Spearman’s test was applied. Scale bars, 50 μm.
(C) Dot plot of GSEA score and enriched p values for epigenetic pathways. Dot size: − log10(p value); color bar: NES. *p < 0.05, **p < 0.01, ***p < 0.001. NES
indicated by color.
(D) Dot plot of GSEA score and enriched p values for mitochondrial pathways. Dot size: − log10(p value); color bar: NES. *p < 0.05, **p < 0.01, ***p < 0.001. NES
indicated by color.
(E) Heatmaps of upregulated (top) and downregulated (bottom) protein complexes across tissues. The top 15 shared complexes are highlighted. Blue to red: low
to high correlation coefficients.
(F) Left: GSEA of protein complex groups. Right: representative complex abundance with age correlation across tissues. Each point: one individual. Skin (n = 40),
aorta (n = 24). Spearman’s correlation was applied. NES indicated by color.
See also Figure S4.

Cell 188, 1–22, October 2, 2025 7

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

A B

E F

G H I

(legend on next page)

8 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

interaction network by mapping aging-associated DEPs to Furthermore, sliding-window-based cumulative analysis re-
curated gene-disease annotations from the DisGeNET data- vealed that many tissues undergo substantial proteomic remod-
base.81 We revealed that the expression levels of 48 disease eling around age 50, particularly between 45 and 55 years
risk factors increased with age, encompassing a range of (Figures 4C, 4D, and S5I–S5K). Among all organs, the aorta ex-
chronic diseases, including cardiovascular conditions associ- hibited the most pronounced and continuous proteomic fluctua-
ated with APOE, CRP, ELN, and APOA1. Additionally, proteins tions across the lifespan. At age 30, the overall proportion of
like STAT3, TGFBR2, and MTOR are implicated in tissue fibrosis, DEPs remained low in most tissues, except for the aorta, spleen,
fatty liver, and hepatic tumors (Figure S4E).82–84 and adrenal gland, with the adrenal gland showing more signifi-
cant protein expression changes (Figures 4C and S5J). This sug-
Proteomic aging clocks for heterogeneous dynamics in gests that alterations in endocrine homeostasis may be one of
human organs the initiating events of systemic aging, consistent with our previ-
To quantitatively assess aging rate heterogeneity across human ous findings.24,89 At ages 45 and 55, the accumulation of DEPs
organs, we devised an innovative method for calculating the bio- intensified across organs, with the aorta showing the most prom-
logical age of each tissue (see STAR Methods) (Figure 4A). We inent changes, underscoring its exceptional sensitivity to aging.
established proteomic age clocks for 13 human tissues, expand- In parallel, the pancreas and spleen also exhibited sustained
ing and complementing previous clocks that solely depended on proteomic remodeling, likely reflecting aging-related deteriora-
plasma protein components.55,85,86 Employing the elastic net tion of metabolic and immune functions. Functionally, early alter-
regression algorithm, we identified proteins that precisely ations at age 30 were mainly associated with rRNA metabolic
predict each tissue’s chronological age, with Spearman’s corre- and translation, whereas transitions at ages 45 and 55 were en-
lation coefficients ranging from 0.74 to 0.95 (Figure 4B). We riched in RNA splicing and protein processing-related pathways
further dissected key proteins driving tissue-specific clocks (Figure S5K). These data collectively highlight the dynamic and
(Figures S5A–S5C; Table S3). Of particular interest, metallopro- heterogeneous nature of organ aging, with the vasculature
teinase inhibitor 3 (TIMP3), a regulator of metalloproteinase,87,88 emerging as one of the earliest and most affected systems.
was identified as a contributor to the tissue clocks across nine
distinct organs (Figures S5A–S5C). Our subsequent immuno- Tissue-specific profiling of secretory protein and inter-
staining confirmed that TIMP3 exhibits an age-associated in- tissue interactions
crease across various tissues (Figure S5D). Secretory proteins are key for extracellular functions and
This multi-time-point human organ protein navigator enabled mediate interactions among cells, organs, and systems.90–92
us to investigate critical transition points within the aging pro- We extracted secretory proteins from each organ to create an
cess. Through proteomic profiling across the human lifespan, age-related secretome dataset, comprising a total of 1,140 pro-
we pinpointed DEPs spanning the early, middle, and late teins. Our analysis included secretory proteins involved in immu-
stages of aging (Figures S5E–S5H). Despite their low preva- nity and inflammation (e.g., interleukins and chemokines), signal
lence, DEPs in early stages of aging associated with humoral transduction (e.g., neuropeptides and hormones), metabolism
immune responses and RNA splicing contribute to the aging and transport (e.g., apolipoproteins and enzymes), and struc-
DEPs across all age groups, with notable effects in the tural support (Figures S6A and S6B). We then categorized these
pancreas and aorta (Figures S5G and S5H). Middle-aged proteins into five modules based on their expression patterns,
DEPs constitute a substantial fraction of the lifespan-wide ag- which exhibited age-related changes (Figure S6C). The first
ing DEPs, underscoring the amplification of humoral immune module comprises secretory proteins that are universally
responses and complement activation across multiple tissues, increased across tissues, predominantly related to humoral im-
alongside the deregulation of RNA splicing in middle age mune responses and complement activation. The second mod-
(Figures S5G and S5H). ule includes proteins that are particularly upregulated in the small

Figure 4. Dynamic profiling of aging-associated proteins and secreted factors

(A) Schematic overview of proteomic aging clock construction using elastic net regression.
(B) Performance of aging clocks across tissues. x axis: chronological age. y axis: predicted age. Each point: one individual. Pearson’s correlation and mean
absolute error (MAE) values are shown.
(C) Sliding-window analysis of dynamic DEPs across age. Left: time-resolved DEPs within each sliding window, normalized to detected proteins per tissue. Right:
cumulative DEPs from successive sliding windows, similarly normalized.
(D) Digital body map of time-resolved DEPs across tissues. Red borders: tissue inflection points. Color intensity: relative number of DEPs.
(E) Network of age-dependent upregulated secretory proteins overlapping with known SASP gene sets. Node size: number of related tissues. Node color: mean
Spearman’s correlation coefficient.
(F) IHC validation of age-associated CXCL12 increase in multiple tissues. Each point: one individual. Aorta (n = 23), lymph (n = 21), liver (n = 49), adrenal (n = 62).
Spearman’s correlation was used. Scale bars, 50 μm.
(G) Tissue interaction networks in young (14–30 years, left) vs. old (45–68 years, right) groups. Edge thickness and color: interaction counts.
(H) Sankey plot of plasma-mediated ligand-receptor interactions across tissues. Left: ligand-expressing tissues. Right: receptor-expressing tissues. Colors:
tissues; link colors: age association.
(I) Dot plot of age-related changes in plasma-mediated ligand-receptor interactions. Color: differential interactions during aging. GAS6 structure: predicted using
AlphaFold.
See also Figures S5 and S6.

Cell 188, 1–22, October 2, 2025 9

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

A B C D E

F G H I J

K L M N

O P

Q S

(legend on next page)

10 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

intestine, mainly associated with immunoglobulin-mediated im- level. We identified 24 ligand-receptor pairs that mediate tissue
mune response and leukocyte-mediated immunity. The third communication via plasma and show age-related changes,
module consists of secretory proteins that are markedly upregu- mainly originating from the aorta and spleen. This highlights
lated in the aorta, regulating chemotaxis and blood coagulation. the key roles of vascular and immune tissues in orchestrating
The fourth module encompasses secretory proteins that are multi-organ aging (Figures 4H and 4I). Given that arteries age
specifically downregulated in the aorta, functioning in endothelial early and dramatically, as previously shown, these findings
cell migration and collagen fibril organization (Figure S6C). further support that arteries may be a key origin of systemic ag-
To decipher the intricacies of the senescence-associated ing (referred to as ‘‘senohub’’). They are not only highly sensitive
secretory phenotype (SASP), we have meticulously crafted tis- to aging-related changes but also serve as a major source of
sue-specific ‘‘SASP roadmaps’’ through an integrated analysis secretory senescence-promoting proteins (termed senopro-
of the aging-related secretome and the well-established teins), amplifying systemic aging signals.
SASP collection.42,93 Overall, we identified 579 distinct SASP Notably, growth arrest specific 6 (GAS6)—a senokine elevated
proteins, about half of them shared across at least two tissues in both plasma and its source tissue, the aorta—showed height-
(Figure S6D). CXCL12, a chemokine,94 accumulated in nine tis- ened interaction with the broadly expressed TAM receptors
sues (Figures 4E and S6D) and was validated by immunostaining (TYRO3, AXL, and MERTK) (Figures 4I and S6G). This interaction
in aged human tissues (Figure 4F). Henceforth, cytokines may activate pro-thrombotic signaling pathways, contributing
released by senescent tissues or cells are collectively termed to endothelial dysfunction and arteriopathic conditions
senokines. (Figures S6G and S6H).96–98
Secretory proteins are key mediators of inter-organ communi- To test this hypothesis, we first validated the age-related upre-
cation, yet our understanding of how these interactions change gulation of GAS6 in human vascular tissues and plasma using
with aging in human organs is limited. We profiled aging-associ- immunohistochemical staining and ELISA (Figures 5A and 5B).
ated secretory protein dynamics and mapped ligand-receptor in- We then treated hAECs and human vascular smooth muscle
teractions across tissues via CellPhoneDB (Figures 4G and cells (hVSMCs) with recombinant human GAS6 protein
S6E).95 Using this approach, we uncovered potential inter-organ (Figure 5C). GAS6 treatment in hAECs increased SA-β-Gal pos-
communication networks, with the aorta, immune tissues such itivity and p21Cip1 expression, impaired tube formation and
as the spleen and lymph nodes, and endocrine organs like the ad- migratory capacity, elevated IL-6 secretion, and enhanced im-
renal gland showing extensive interactions with other tissues. mune cell adhesion (Figures 5D–5I). Similarly, in hVSMCs, seno-
Notably, these inter-organ interactions intensify with age, primarily kine GAS6 accelerated cellular senescence and elevated inflam-
driven by receptor-ligand pairs involving the inflammatory matory cytokine levels (Figures 5J–5L).
chemokines CXCL12 and CXCL14 (Figure S6F; Table S4). By To evaluate the in vivo effects of circulating GAS6, we intrave-
contrast, aging-related reductions in inter-organ interactions nously administered recombinant mouse GAS6 to 8-month-old
were observed for ephrin A1 (EFNA1), ephrin B1 (EFNB1), and mice every 5 days for a total of four doses (Figure 5M). At the
ephrin B2 (EFNB2), members of the Ephrin family, which mediate physiological level, GAS6-treated mice exhibited impaired phys-
bidirectional signaling through Eph receptors and are involved in ical performance, including reduced grip strength, decreased
regulating cell migration, tissue boundary formation, and vascular wire-hang endurance, and compromised balance and coordina-
development (Figure S6F; Table S4). tion, as evidenced by increased foot slips and prolonged
Considering proteins mediate organ-organ interactions via the crossing times in balance beam tests (Figure 5N). Histological
bloodstream, we integrated plasma proteomics to elucidate analyses showed prominent markers of vascular aging, including
these long-range, plasma-mediated interactions at the protein elevated p21Cip1 expression, increased oxidative DNA damage

Figure 5. GAS6 promotes human endothelial senescence and mouse multi-organ aging
(A) IHC staining of vascular wall GAS6 in individuals of different ages (n = 23). Spearman’s correlation was applied. Scale bars, 50 μm.
(B) ELISA showing plasma GAS6 levels across ages (n = 62). Each point: one individual. Pearson’s correlation was applied.
(C) Schematic of GAS6 treatment in hAECs and hVSMCs.
(D–I) SA-β-Gal staining (D), western blot (E), tube formation (F), migration assay (G), IL-6 ELISA (H), and monocyte adhesion assay (I) in hAECs treated with vehicle
(n = 6 for D and F–I; n = 3 for E) or GAS6 (n = 6 for D and F–I; n = 3 for E). Scale bars, 100 μm.
(J) SA-β-Gal staining in hVSMCs treated with vehicle (n = 6) or GAS6 (n = 6). Scale bars, 200 μm.
(K) IF staining of H3K9me3 and LAP2 in hVSMCs treated with vehicle (n = 300 nuclei) or GAS6 (n = 300 nuclei). Scale bars, 20 μm.
(L) ELISA of IL-6 in hVSMCs treated with vehicle (n = 6) or GAS6 (n = 6).
(M) Schematic of in vivo GAS6 injection in 8-month-old mice (every 5 days, 4 doses), with behavioral and histological assays after the final dose. Image of
AlphaFold-predicted mouse GAS6 structure.
(N) Functional tests (grip strength, wire hang, beam-crossing time, and foot slips) in mice treated with vehicle (n = 7) or GAS6 (n = 10).
(O) Aortic staining for p21Cip1, 8-OHdG, MMTV/CD31, oil red O, S100A8, and vWF/CD31 (vehicle: n = 6–8, GAS6: n = 9–10). Scale bars, 20 and 5 μm (zoomed-in
image).
(P) Liver staining for p21Cip1, H3K9me3, H&E, and oil red O (vehicle: n = 7–8, GAS6: n = 10). Scale bars, 25 μm.
(Q) Spleen staining for H&E, p21Cip1, and IL-6 (vehicle: n = 7, GAS6: n = 10). Scale bars, 50 μm.
(R) Lymph staining for IL-6, S100A8, and γH2AX (vehicle: n = 7, GAS6: n = 6). Scale bars, 25 μm.
(S) Adrenal staining for IL-6 and S100A8 (vehicle: n = 7, GAS6: n = 9). Scale bars, 20 μm.
Data are shown as mean ± SEM. Two-tailed unpaired t test or Mann-Whitney test were applied (D–L and N–S).

Cell 188, 1–22, October 2, 2025 11

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

(8-OHdG), and heightened expression of endogenous retrovirus algorithms, we developed a plasma protein proxy tissue aging
(MMTV) (Figure 5O).70 GAS6 treatment additionally promoted clock that achieves comparable age prediction accuracy to the
lipid deposition while upregulating expression of inflammatory tissue protein matrix-derived clock and differentiates aging
marker S100A8 and vascular damage marker von Willebrand across various tissues (see STAR Methods) (Figures S7D–S7H).
factor (vWF) within the aorta (Figure 5O).
Beyond vascular tissues, GAS6 promoted accelerated senes- Humoral senoproteins driving vascular and systemic
cence phenotypes in multiple target organs (Figures 5P–5S). In aging
the liver, GAS6 treatment increased p21Cip1 expression, reduced We next explored whether specific senoproteins accumulating in
H3K9me3 levels, stimulated immune infiltration, and lipid accu- plasma could drive vascular aging, given their direct interaction
mulation (Figure 5P). In the spleen, GAS6 disrupted the structural with the vascular wall, especially endothelial cells (Figures 7A
integrity of the white pulp and elevated p21Cip1 and IL-6 expres- and 7B). To test this hypothesis, we treated hAECs with recom-
sion (Figure 5Q). GAS6 administration elevated expression of in- binant forms of plasma proteins that increase with age
flammatory markers IL-6 and S100A8 in lymph nodes and adre- (Figure 7B). Seven proteins—GPNMB, COMP, HTRA1, SLPI,
nal glands, while increasing DNA damage marker γH2AX in IGFBP7, NEGR1, and NOTCH3—induced hallmark features of
lymph nodes (Figures 5R and 5S). Collectively, these findings vascular endothelial senescence. These features included
demonstrate that elevated GAS6 contributes to accelerated increased SA-β-Gal activity, elevated expression of p21Cip1,
vascular and systemic aging. and reduced levels of H3K9me3 and LAP2 (Figures 7C–7F,
S8A, and S8B). Elevated expression of pro-inflammatory cyto-
Circulating protein biomarkers predicting organ aging kines (IL-6 and IL-8) and endothelial dysfunction marker
Unveiling blood-based biomarkers for organ aging is essential for (VCAM1), impaired migration and tube formation, enhanced
non-invasive aging assessments. We integrated plasma and or- monocyte adhesion, and increased IL-6 secretion were also
gan proteomic data to identify proteins with concordant changes observed (Figures 7G–7J and S8C–S8E).
in both, thereby characterizing potential fluid biomarkers for organ Transcriptomic analysis further revealed that treatment with
aging (see STAR Methods). Employing this approach, we identi- these proteins activated canonical senescence-associated
fied a total of 211 plasma-tissue paired DEPs that vary with aging genes and upregulated pathways related to cellular senescence,
in plasma and show age-dependent changes in at least one tissue oxidative stress, and inflammation (Figure S8F). To determine
in the same direction as in plasma (Figures 6A, S7A, and S7B). The whether these effects extend beyond endothelial cells, we
highest number of plasma-tissue paired DEPs was observed in treated hVSMCs with GPNMB, HTRA1, and IGFBP7, which simi-
the aorta, adrenal glands, and spleen, highlighting their influence larly induced senescence phenotypes, including elevated SA-
on aging-related alterations in plasma protein composition β-Gal activity, reduced H3K9me3, compromised LAP2, and
(Figure S7A). Furthermore, we focused primarily on the upregu- increased IL-6 expression (Figures S8G–S8P). As an in vivo
lated DEPs. Approximately 40% of the plasma-tissue paired example to examine whether aging-associated plasma proteins
DEPs were uniquely present in a single tissue within our examined can drive vascular aging and subsequent systemic decline, we
range (Figures 6A, S7A, and S7B). Notably, several plasma-tissue tested GPNMB via intravenous injection in middle-aged mice.
paired DEPs were identified that correspond between plasma and GPNMB-treated mice exhibited reduced grip strength and
at least one tissue, and these proteins exhibit remarkable accu- impaired motor coordination and balance (Figure 7K). Histologi-
mulation in aging plasma. For example, GPNMB, a reported cal analysis revealed that GPNMB accelerated vascular aging,
senescence-associated antigen,99 has been confirmed by us marked by increased DNA oxidation, reactivated retrotranspo-
and other studies to accumulate in aging plasma.100,101 Addition- sons, cell cycle arrest in vascular wall cells, endothelial damage
ally, we identified a series of plasma-tissue paired aging bio- with upregulated vWF, and increased vascular inflammation and
markers, including COMP, CHI3L1, NOTCH3, ITLN1, EFEMP1, lipid accumulation (Figures 7L–7R). Together, these findings
HTRA1, and LTBP2, whose elevation in human plasma was further demonstrate that senoproteins secreted by aging tissues can
validated by ELISA (Figure 6B). We also confirmed the aging- amplify senescence-promoting signals, mediating vascular and
related upregulation of GPNMB, COMP, NOTCH3, HTRA1, and multi-organ degeneration.
LTBP2 proteins across multiple tissues through histological im-
munostaining, with the most significant increases observed in DISCUSSION
the aorta (Figure 6C).
In addition, we identified age-dependent downregulation of Human organ aging is a primary driver of numerous chronic dis-
plasma-tissue paired proteins, including HARS1, CCT5, and eases, yet the molecular mechanisms that govern its complexity
PSMB1 (Figure S7B). Notably, HARS1, a protein associated and heterogeneity across organs remain largely elusive.11,106–108
with neurodegenerative and autoimmune diseases,102,103 ex- Our study harnesses advanced quantitative proteomics to reveal
hibited the most widespread decline across tissues (Figure the protein underpinnings of organ aging, creating a pioneering
S7B). ELISA and immunohistochemistry further confirmed the atlas across human tissues. This work offers innovative and in-
age-related decrease of HARS1 in plasma and its reduction in depth perspectives for understanding, anticipating, and ad-
the aorta (Figures S2G and S7C). Drawing from our analyses, dressing aging with a protein-centric approach.
we posited that a plasma protein-based clock could act as a We introduce protein-based aging clocks and dynamic aging
surrogate for the organ clocks constructed from actual tissue trajectories for human organs, illuminating their biological age
protein matrices.24,104,105 Accordingly, using machine learning and disease risks. Notably, tissues generally exhibit an increase

12 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Figure 6. Proteins with coordinated upregulation in human plasma and tissues

(A) Dot plot of DEPs co-upregulated with age in plasma and tissues (Spearman’s correlation > 0, p < 0.05). Dots: plasma-tissue pairs; line graph: plasma
abundance. Dot colors: low to high correlation coefficients.
(B) ELISA analysis of plasma protein levels for indicated DEPs. Each point: one individual. Sample sizes: COMP (n = 72), CHI3L1 (n = 71), NOTCH3 (n = 72), ITLN1
(n = 72), EFEMP1 (n = 72), HTRA1 (n = 63), and LTBP2 (n = 64). Individuals were randomly selected. Pearson’s correlation was applied.
(C) IHC staining of age-upregulated proteins in age-stratified tissues. Each point: one individual. Sample sizes vary by protein and tissue. Spearman’s correlation
was applied. Scale bars: 50 μm.
See also Figure S7.

in aging rate around 50 years, aligning with our established non- components. Integrating plasma and tissue data, we identified
invasive multidimensional aging clocks.24,27 Our study highlights age-accumulating humoral senoproteins—pro-inflammatory se-
the early and pronounced aging of the aorta, underscoring the nokines—that accelerate vascular and systemic decline. This re-
critical role of vascular senescence in initiating systemic aging veals that blood vessels, as the primary senohub, are not only
through its extensive interactions with other organs and blood extremely aging-sensing units and reservoirs for receiving

Cell 188, 1–22, October 2, 2025 13

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

A B C

E F

G H

I J

K L M N

O P Q R

(legend on next page)

14 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

aging-related factors from peripheral tissues but are also signif- for future research. Demographic limitations in tissue sampling
icantly affected by organ aging itself. may also impact the generalizability of our findings. Despite
Our study also probed proteostasis decline, a key aging hall- these caveats, our research provides initial evidence that human
mark.45,109–113 We observed transcriptome-proteome decou- aging can be analyzed, quantified, and potentially modulated
pling, indicating reduced central dogma-based information through a protein-centric approach.
flow with age. We identified widespread amyloid protein aggre-
gation as a common aging trait across human organs and RESOURCE AVAILABILITY
confirmed SAP’s role in promoting vascular endothelial damage.
Lead contact
Amyloid-induced chronic inflammation is closely linked to the
Further information and requests for resources and reagents should be
onset of cardiovascular and cerebrovascular diseases.114–122 directed to Dr. Guang-Hui Liu ([email protected]).
Additionally, we have observed an increase in complement pro-
teins and immunoglobulins, which serve as a universal aging Materials availability
signature across tissues.26,42,123–125 Interestingly, SAP and This study did not generate new unique reagents.
immunoglobulin G (IgG) seem to utilize similar Fcγ receptor acti-
vation to mediate pro-inflammatory signaling.42,126,127,128 These Data and code availability
findings constitute an amyloid-centric inflammaging framework • The proteomics data of human tissues have been deposited at the iProX
underlying human organ aging and degeneration. database consortium. Bulk RNA sequencing (RNA-seq) data of human
tissues have been deposited at the Genome Sequence Archive (GSA) in
In translational medicine, our study offers new strategies for
the National Genomics Data Center, Beijing Institute of Genomics
protein-targeted geroprotective interventions.101,129 Identified (China National Center for Bioinformation) of the Chinese Academy of
aging-associated cell surface proteins may serve as seno-anti- Sciences. Accession numbers are listed in the key resources table.
gens for developing senolytic vaccines or CAR-T therapies to • This paper does not report original code.
clear senescent cells.99,129 Moreover, proteolysis targeting • Any additional information required to reanalyze the data reported in this
chimera (PROTAC)-mediated degradation or neutralizing anti- paper is available from the lead contact upon request.
bodies against soluble senoproteins could help block aging
drivers at the protein level. ACKNOWLEDGMENTS
In summary, the human aging protein navigator established in
this study not only reveals aging biomarkers but also identifies We thank the Aging Biomarker Consortium for their invaluable feedback and
organ-associated aging drivers, thereby enhancing our under- guidance throughout this study. We thank Tong Chen, Zhao Wang, Li Yi,
standing of the complex and heterogeneous biology of aging. and Zhenru Wu from West China Hospital for their assistance with human tis-
sue procurement and processing; Jingyi Jia, Jiayi Liu, and Feifei Liu for their
These insights may facilitate the development of targeted inter-
support in animal management; Ansheng Tan, Tingting Yan, Peng Yang,
ventions for aging and age-related diseases, paving the way to
Xiaoyan Sun, Beibei Liu, Ying Jing, and Kejie Zheng for their assistance with
improve the health of older adults. mouse tissue harvesting; Lei Bai, Jing Chen, Jing Lu, Qun Chu, Ying Yang,
Luyang Tian, Shangyi Qiao, Xiuping Li, and Xuewei Chen for their administra-
tive support; and Meng Wan and Ying Zhou from the Institute of Biophysics,
Limitations of the study Chinese Academy of Sciences, for their assistance with slide image acquisi-
Our study provides a comprehensive proteomic perspective on tion. We thank the Human Organ Physiopathology Emulation System
human aging, yet there is room for improvement. Though we (HOPE) for assisting in the development of human cell senescence models.
strived for representative sampling across age groups and We thank the Proteomic Navigator of the Human Body (π-HuB) Project for their
sexes, limitations in human tissue availability restricted sample support and guidance in proteomic analysis. The graphical elements in this
size and data completeness. This study represents a proof of study’s figures are sourced partially from BioRender.com, with some designed
by Yizhu Wang. This work was supported by the National Natural Science
concept, and further analysis of organ aging dynamics and aging
Foundation of China (82488301), the National Key R&D Program of China
clocks in larger cohorts is necessary to enhance its robustness. (2022YFA1103700), the National Natural Science Foundation of China
Notably, data from critical organs such as the brain, kidneys, and (82125011, 82361148131, 92168201, 82330044, 32341001, 82322025,
reproductive system were absent, highlighting important areas 32121001, 82192863, 82361148130, 8231101626, 82422031, 82271600,

Figure 7. Plasma senoproteins induce vascular and systemic aging

(A) Left: IHC staining of p21Cip1 in human aorta from young to old (n = 21). Right: Spearman’s correlation was applied. Scale bars, 50 and 10 μm.
(B) Schematic of vascular endothelial cell treatment with recombinant senoproteins.
(C) SA-β-Gal staining in hAECs treated with vehicle (n = 6) or indicated proteins (n = 6). Scale bars, 50 μm.
(D) Western blot in hAECs treated with vehicle (n = 3) or indicated proteins (n = 3).
(E and F) IF staining of H3K9me3 and LAP2 in hAECs treated with vehicle (n = 300 nuclei) or indicated proteins (n = 300 nuclei). Scale bars, 20 μm.
(G–J) RT-qPCR of senescence-related genes, migration assay, tube formation, and monocyte adhesion assay in hAECs treated with vehicle (n = 6) or indicated
proteins (n = 6). Scale bars, 100 μm.
(K) Grip strength, beam-crossing time, and foot slips in mice treated with vehicle (n = 7) or GPNMB (n = 10).
(L–R) IHC or IF staining in aorta from vehicle (n = 6–7)- or GPNMB (n = 9–10)-treated mice, including 8-OHdG (L), MMTV (M), p21Cip1 (N), vWF (O), S100A8 (P),
BODIPY (Q), and oil red O (R). Scale bars, 20 μm.
Data are presented as mean ± SEM. Statistical significance between multiple groups was assessed using a one-way ANOVA test (C and E–J). Between two
groups, unpaired two-tailed t test or Mann-Whitney test was applied (D and K–R).
See also Figure S8.

Cell 188, 1–22, October 2, 2025 15

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article
82471586, 92468303, 82401822, and 32088101), the National Key R&D Pro- ○ Oil red O staining
gram of China (2020YFA0804000 and 2022YFA1103800), the NCD-Major Proj- ○ BODIPY staining
ect (2024ZD0530400), the STI2030-Major Projects (2021ZD0202400), the Bei- ○ RNA-seq library construction and data processing

jing Natural Science Foundation (JQ24044, Z240018, F251011, Z230011, ○ Quality control of proteomics data

5242024, and 5244046), the CAS Basic Research Program (YSBR-076 and ○ Identification of high and low abundance proteins

YSBR-012), the Strategic Priority Research Program of CAS (XDA0460403 ○ Classification of proteins based on chromosome location of protein-

and XDC0200000), XPLORER PRIZE (2021-1045), the Shenzhen Medical coding genes
Research Fund (C2406001 and D2401003), the Beijing Anzhen Hospital ○ Classification of proteins based on subcellular location, cellular

Research Fund (2024AZB3002), the CAS Informatization Plan (CAS- component, and molecular function
WX2022SDC-XK14), Beijing Medical Research Institute Pilot Project ○ Classification of proteins based on differential expression between

(JYY2023-13), the CAS Youth Interdisciplinary Team, the Key Lab of Alz- tissues
heimer’s Disease of Zhejiang Province (ZJAD-2024001), the Institute of ○ Identification of sex-biased proteins

Zoology Initiative Program (2023IOZ0202, 2024IOZ0103, and 2023IOZ0102), ○ Bulk RNA-seq data processing

the Youth Innovation Promotion Association of CAS (2022083), the Talent Pro- ○ Decoupling analysis of protein abundance and mRNA expression

grams of Capital Medical University (12500825 and 12300927), the Beijing Uni- ○ Translational noise analysis

versity Faculty Training Program (BPHR202203105), the CAST Young Elite ○ Gene set variation analysis (GSVA)

Scientists Program (2021QNRC001), the Hainan Provincial Natural Science ○ Identification of age-dependent differentially expressed proteins

Foundation (825QN316), Space Medical Experiment Project of CMSP (DEPs)

(HYZHXMH01012), Chinese Academy of Sciences President’s International ○ Module analysis of DEPs

Fellowship Initiative (2024PG0022), and the CAS International Partnership Pro- ○ Functional enrichment analysis

gram (073GJHZ2023019FN). ○ Protein complex annotations

○ Identification of age-related protein complexes

AUTHOR CONTRIBUTIONS ○ Disease-associated pathway analysis

○ Proteomic aging clock analysis

G.-H.L., W.Z., J.Q., and J.Y. designed the project and supervised all experi- ○ Age-group-specific DEPs analysis

ments. J.Y., Y.D., G.X., X.X., Q.W., M.J., M.X., and Z.H. performed human ○ Sliding window analysis of the senescence proteome

sample collection and clinical coordination. Y.D., Y.Z., B.Z., Y.F., Y.J., and ○ Analysis of secretory proteins

S.M. conducted proteomic profiling and data acquisition. Z.C. provided intel- ○ Module analysis of secretory proteins

lectual and technical support in proteomic mass spectrometry analysis. Y.Z., ○ Senescence-associated secretory phenotype (SASP) analysis

B.Z., and Y.D. carried out bioinformatic and statistical analysis. Y.Z., B.Z., Y. ○ Tissue-tissue interaction analysis

D., and J.L. developed multi-organ aging clock modeling. Y.D., Y.F., S.F., A. ○ Tissue-tissue interactions mediated by circulating proteins

T., and Z.H. completed tissue immunostaining and histopathology. Y.D., Y. ○ Disease enrichment analysis of tissue-tissue interacted protein pairs

F., S.F., and A.T. executed plasma protein functional validation, including ○ Plasma protein proxy aging clock analysis

in vitro and in vivo. Y.D., S.F., and D.G. performed plasma ELISA experiments. • QUANTIFICATION AND STATISTICAL ANALYSIS
S.M., S.S., S.W., and F.H. supervised the project. G.-H.L., W.Z., J.Q., Y.D., Y.
Z., B.Z., Y.F., S.M., S.F., and A.T. contributed to manuscript writing, review,
and editing. SUPPLEMENTAL INFORMATION

DECLARATION OF INTERESTS Supplemental information can be found online at https://doi.org/10.1016/j.cell.

2025.06.047.
Z.C. is an employee of Jingjie PTM BioLab (Hangzhou) Co. Inc.
Received: November 22, 2024
STAR★METHODS Revised: April 24, 2025
Accepted: June 30, 2025
Detailed methods are provided in the online version of this paper and include
the following: REFERENCES
• KEY RESOURCES TABLE
• EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS 1. Partridge, L., Deelen, J., and Slagboom, P.E. (2018). Facing up to the
○ Ethical statements global challenges of ageing. Nature 561, 45–56. https://doi.org/10.
○ Human samples 1038/s41586-018-0457-8.
○ Experimental mouse models
2. Tyshkovskiy, A., Ma, S., Shindyapina, A.V., Tikhonov, S., Lee, S.G., Bo-
○ Cell culture
zaykut, P., Castro, J.P., Seluanov, A., Schork, N.J., Gorbunova, V., et al.
• METHOD DETAILS (2023). Distinct longevity mechanisms across and within species and
○ Human samples collection
their association with aging. Cell 186, 2929–2949.e20. https://doi.org/
○ Cell treatment with recombinant protein
10.1016/j.cell.2023.05.002.
○ Tissue protein extraction and mass spectrometry analysis
○ ELISA assay
3. Campisi, J., Kapahi, P., Lithgow, G.J., Melov, S., Newman, J.C., and
○ Immunofluorescence and immunohistochemical staining
Verdin, E. (2019). From discoveries in ageing research to therapeutics
○ Western blotting
for healthy ageing. Nature 571, 183–192. https://doi.org/10.1038/
○ SA-β-Gal staining
s41586-019-1365-2.
○ Endothelial cell tube formation assay 4. Cai, Y., Song, W., Li, J., Jing, Y., Liang, C., Zhang, L., Zhang, X., Zhang,
○ Vascular endothelial cell migration assay W., Liu, B., An, Y., et al. (2022). The landscape of aging. Sci. China Life
○ Monocyte adhesion assay Sci. 65, 2354–2454. https://doi.org/10.1007/s11427-022-2161-3.
○ RNA isolation and real-time quantitative PCR (RT-qPCR) 5. Su, S.Y., He, C., Ren, J., and Song, M. (2024). Global insights into aging:
○ Recombinant protein treatment of mice a multidisciplinary approach to understanding and addressing age-
○ Mouse behavioral assays related challenges. Life Med. 3, lnae029. https://doi.org/10.1093/life-
○ Hematoxylin and eosin (H&E) staining medi/lnae029.

16 Cell 188, 1–22, October 2, 2025

Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article
6. Kennedy, B.K., Berger, S.L., Brunet, A., Campisi, J., Cuervo, A.M., Epel, 21. Fu, X., Wang, S., Wu, Y., Sun, Y., Liu, W., Xi, X., Li, G.L., Liu, K., Yuan, W.,
E.S., Franceschi, C., Lithgow, G.J., Morimoto, R.I., Pessin, J.E., et al. et al.; Aging; Biomarker Consortium (2025). A biomarker framework for
(2014). Geroscience: linking aging to chronic disease. Cell 159, auditory system aging: the Aging Biomarker Consortium consensus
709–713. https://doi.org/10.1016/j.cell.2014.10.039. statement. Life Med. 4, lnaf011. https://doi.org/10.1093/lifemedi/
7. Prince, M.J., Wu, F., Guo, Y., Gutierrez Robledo, L.M., O’Donnell, M., lnaf011.
Sullivan, R., and Yusuf, S. (2015). The burden of disease in older people 22. Schaum, N., Lehallier, B., Hahn, O., Pálovics, R., Hosseinzadeh, S., Lee,
and implications for health policy and practice. Lancet 385, 549–562. S.E., Sit, R., Lee, D.P., Losada, P.M., Zardeneta, M.E., et al. (2020).
https://doi.org/10.1016/S0140-6736(14)61347-7. Ageing hallmarks exhibit organ-specific temporal signatures. Nature
8. Zhang, W., Che, Y., Tang, X., Chen, S., Song, M., Wang, L., Sun, A.J., 583, 596–602. https://doi.org/10.1038/s41586-020-2499-y.
Chen, H.Z., Xu, M., et al.; Aging; Biomarker Consortium (2023). A 23. Rando, T.A., and Wyss-Coray, T. (2021). Asynchronous, contagious and
biomarker framework for cardiac aging: the Aging Biomarker Consortium digital aging. Nat Aging 1, 29–35. https://doi.org/10.1038/s43587-020-
consensus statement. Life Med. 2, lnad035. https://doi.org/10.1093/life- 00015-1.
medi/lnad035.
24. Li, J., Xiong, M., Fu, X.H., Fan, Y., Dong, C., Sun, X., Zheng, F., Wang, S.
9. Kroemer, G., Maier, A.B., Cuervo, A.M., Gladyshev, V.N., Ferrucci, L., W., Liu, L., Xu, M., et al. (2023). Determining a multimodal aging clock in a
Gorbunova, V., Kennedy, B.K., Rando, T.A., Seluanov, A., Sierra, F., cohort of Chinese women. Med. 4, 825–848.e13. https://doi.org/10.
et al. (2025). From geroscience to precision geromedicine: Understand- 1016/j.medj.2023.06.010.
ing and managing aging. Cell 188, 2043–2062. https://doi.org/10.1016/j.
25. Zhao, R., Lu, H., Yuan, H., Chen, S., Xu, K., Zhang, T., Liu, Z., Jiang, Y.,
cell.2025.03.011.
Suo, C., and Chen, X. (2025). Plasma proteomics-based organ-specific
10. Guo, J., Huang, X., Dou, L., Yan, M., Shen, T., Tang, W., and Li, J. (2022). aging for all-cause mortality and cause-specific mortality: a prospective
Aging and aging-related diseases: from molecular mechanisms to inter- cohort study. GeroScience 47, 1411–1423. https://doi.org/10.1007/
ventions and treatments. Signal Transduct. Target. Ther. 7, 391. https:// s11357-024-01411-w.
doi.org/10.1038/s41392-022-01251-0.
26. Bao, H., Cao, J., Chen, M., Chen, M., Chen, W., Chen, X., Chen, Y., Chen,
11. Tian, Y.E., Cropley, V., Maier, A.B., Lautenschlager, N.T., Breakspear, Y., Chen, Y., et al.; Aging; Biomarker Consortium (2023). Biomarkers of
M., and Zalesky, A. (2023). Heterogeneous aging across multiple organ aging. Sci. China Life Sci. 66, 893–1066. https://doi.org/10.1007/
systems and prediction of chronic disease and mortality. Nat. Med. 29, s11427-023-2305-0.
1221–1231. https://doi.org/10.1038/s41591-023-02296-6.
27. Zheng, Z., Li, J., Liu, T., Fan, Y., Zhai, Q.C., Xiong, M., Wang, Q.R., Sun,
12. Suryadevara, V., Hudgins, A.D., Rajesh, A., Pappalardo, A., Karpova, A., X., Zheng, Q.W., Che, S., et al. (2024). DNA methylation clocks for esti-
Dey, A.K., Hertzel, A., Agudelo, A., Rocha, A., Soygur, B., et al. (2024). mating biological age in Chinese cohorts. Protein Cell 15, 575–593.
SenNet recommendations for detecting senescent cells in different tis- https://doi.org/10.1093/procel/pwae011.
sues. Nat. Rev. Mol. Cell Biol. 25, 1001–1023. https://doi.org/10.1038/
28. Sun, S., Jiang, M., Ma, S., Ren, J., and Liu, G.H. (2025). Exploring the het-
s41580-024-00738-8.
erogeneous targets of metabolic aging at single-cell resolution. Trends
13. Zheng, F., Lei, J., He, Z., Ning, T., Sun, S., Cai, Y., Zhao, Q., Ma, S., Endocrinol. Metab. 36, 133–146. https://doi.org/10.1016/j.tem.2024.
Zhang, W., Qu, J., et al. (2025). FOXO3-engineered human mesenchymal 07.009.
stem cells efficiently enhance post-ischemic stroke functional rehabilita-
tion. Protein Cell 16, 365–373. https://doi.org/10.1093/procel/pwaf004. 29. Ma, S., Chi, X., Cai, Y., Ji, Z., Wang, S., Ren, J., and Liu, G.H. (2023). De-
coding Aging Hallmarks at the Single-Cell Level. Annu. Rev. Biomed.
14. Yang, Y., Lu, X., Liu, N., Ma, S., Zhang, H., Zhang, Z., Yang, K., Jiang, M.,
Data Sci. 6, 129–152. https://doi.org/10.1146/annurev-biodatasci-
Zheng, Z., Qiao, Y., et al. (2024). Metformin decelerates aging clock in
020722-120642.
male monkeys. Cell 187, 6358–6378.e29. https://doi.org/10.1016/j.cell.
2024.08.021. 30. Liu, B., Qu, J., Zhang, W., Izpisua Belmonte, J.C., and Liu, G.H. (2022). A
stem cell aging framework, from mechanisms to interventions. Cell Rep.
15. Yan, H., Ren, J., and Liu, G.-H. (2023). Fecal microbiota transplantation:
41, 111451. https://doi.org/10.1016/j.celrep.2022.111451.
A new strategy to delay aging. hLife 1, 8–11. https://doi.org/10.1016/j.
hlife.2023.06.002. 31. Li, J., Zheng, Y., Yan, P., Song, M., Wang, S., Sun, L., Liu, Z., Ma, S., Iz-
pisua Belmonte, J.C., Chan, P., et al. (2021). A single-cell transcriptomic
16. Zhang, Z., Tian, X., Lu, J.Y., Boit, K., Ablaeva, J., Zakusilo, F.T., Emmrich,
atlas of primate pancreatic islet aging. Natl. Sci. Rev. 8, nwaa127. https://
S., Firsanov, D., Rydkina, E., Biashad, S.A., et al. (2023). Increased hya-
doi.org/10.1093/nsr/nwaa127.
luronan by naked mole-rat Has2 improves healthspan in mice. Nature
621, 196–205. https://doi.org/10.1038/s41586-023-06463-0. 32. Lei, J., Jiang, X., Li, W., Ren, J., Wang, D., Ji, Z., Wu, Z., Cheng, F., Cai,
Y., Yu, Z.R., et al. (2022). Exosomes from antler stem cells alleviate
17. Lei, J., Xin, Z., Liu, N., Ning, T., Jing, Y., Qiao, Y., He, Z., Jiang, M., Yang,
mesenchymal stem cell senescence and osteoarthritis. Protein Cell 13,
Y., Zhang, Z., et al. (2025). Senescence-resistant human mesenchymal
220–226. https://doi.org/10.1007/s13238-021-00860-9.
progenitor cells counter aging in primates. Cell. https://doi.org/10.
1016/j.cell.2025.05.021. 33. May, M. (2023). Biomarkers of aging remain elusive as researchers try to
slow the biological clock. Nat. Med. 29, 2673–2676. https://doi.org/10.
18. Suo, J., Gan, Y., Xie, Y., Xu, S., Wang, J., Chen, D., Chen, L., Deng, L.,
1038/s41591-023-02560-9.
Feng, S., et al.; Aging; Biomarker Consortium (2023). A framework of bio-
markers for skeletal aging: a consensus statement by the Aging 34. Moqri, M., Herzog, C., Poganik, J.R., Justice, J., Belsky, D.W., Higgins-
Biomarker Consortium. Life Med. 2, lnad045. https://doi.org/10.1093/ Chen, A., Moskalev, A., Fuellen, G., Cohen, A.A., et al.; Biomarkers; Ag-
lifemedi/lnad045. ing Consortium (2023). Biomarkers of aging for the identification and
evaluation of longevity interventions. Cell 186, 3758–3775. https://doi.
19. Zhang, L., Wang, Z., Zhang, Y., Ji, R., Li, Z., Zou, J., and Gao, B. (2024).
org/10.1016/j.cell.2023.08.003.
Regulatory cellular and molecular networks in the bone microenviron-
ment during aging. Life Med. 3, lnae019. https://doi.org/10.1093/life- 35. Ren, J., Song, M., Zhang, W., Cai, J.P., Cao, F., Cao, Z., Chan, P., Chen,
medi/lnae019. C., Chen, G., Chen, H.Z., et al. (2023). The Aging Biomarker Consortium
20. Lu, H., Jing, Y., Zhang, C., Ma, S., Zhang, W., Huang, D., Zhang, B., Zuo, represents a new era for aging research in China. Nat. Med. 29, 2162–
Y., Qin, Y., Liu, G.H., et al. (2024). Aging hallmarks of the primate ovary 2165. https://doi.org/10.1038/s41591-023-02444-y.
revealed by spatiotemporal transcriptomics. Protein Cell 15, 364–384. 36. Moqri, M., Herzog, C., Poganik, J.R., Ying, K., Justice, J.N., Belsky, D.W.,
https://doi.org/10.1093/procel/pwad063. Higgins-Chen, A.T., Chen, B.H., Cohen, A.A., Fuellen, G., et al. (2024).

Cell 188, 1–22, October 2, 2025 17

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article
Validation of biomarkers of aging. Nat. Med. 30, 360–372. https://doi.org/ coverage with narrow-window data-independent acquisition. Nat. Bio-
10.1038/s41591-023-02784-9. technol. 42, 1855–1866. https://doi.org/10.1038/s41587-023-02099-7.
37. Herzog, C.M.S., Goeminne, L.J.E., Poganik, J.R., Barzilai, N., Belsky, D. 53. Stewart, H.I., Grinfeld, D., Giannakopulos, A., Petzoldt, J., Shanley, T.,
W., Betts-LaCroix, J., Chen, B.H., Chen, M., Cohen, A.A., et al.; Bio- Garland, M., Denisov, E., Peterson, A.C., Damoc, E., Zeller, M., et al.
markers; Aging Consortium (2024). Challenges and recommendations (2023). Parallelized Acquisition of Orbitrap and Astral Analyzers Enables
for the translation of biomarkers of aging. Nat Aging 4, 1372–1383. High-Throughput Quantitative Analysis. Anal. Chem. 95, 15656–15664.
https://doi.org/10.1038/s43587-024-00683-3. https://doi.org/10.1021/acs.analchem.3c02856.
38. Shen, X., Wang, C., Zhou, X., Zhou, W., Hornburg, D., Wu, S., and 54. Eldjarn, G.H., Ferkingstad, E., Lund, S.H., Helgason, H., Magnusson, O.
Snyder, M.P. (2024). Nonlinear dynamics of multi-omics profiles during T., Gunnarsdottir, K., Olafsdottir, T.A., Halldorsson, B.V., Olason, P.I.,
human aging. Nat Aging 4, 1619–1634. https://doi.org/10.1038/ Zink, F., et al. (2023). Large-scale plasma proteomics comparisons
s43587-024-00692-2. through genetics and disease associations. Nature 622, 348–358.
39. Wolf, J., Rasmussen, D.K., Sun, Y.J., Vu, J.T., Wang, E., Espinosa, C., Bi- https://doi.org/10.1038/s41586-023-06563-x.
gini, F., Chang, R.T., Montague, A.A., Tang, P.H., et al. (2023). Liquid-bi- 55. Oh, H.S.H., Rutledge, J., Nachun, D., Pálovics, R., Abiose, O., Moran-
opsy proteomics combined with AI identifies cellular drivers of eye aging Losada, P., Channappa, D., Urey, D.Y., Kim, K., Sung, Y.J., et al.
and disease in vivo. Cell 186, 4868–4884.e12. https://doi.org/10.1016/j. (2023). Organ aging signatures in the plasma proteome track health
cell.2023.09.012. and disease. Nature 624, 164–172. https://doi.org/10.1038/s41586-
40. Shen, Y., Timsina, J., Heo, G., Beric, A., Ali, M., Wang, C., Yang, C., 023-06802-1.
Wang, Y., Western, D., Liu, M., et al. (2024). CSF proteomics identifies 56. Ogrodnik, M., Carlos Acosta, J., Adams, P.D., d’Adda di Fagagna, F.,
early changes in autosomal dominant Alzheimer’s disease. Cell 187, Baker, D.J., Bishop, C.L., Chandra, T., Collado, M., Gil, J., Gorgoulis,
6309–6326.e15. https://doi.org/10.1016/j.cell.2024.08.049. V., et al. (2024). Guidelines for minimal information on cellular senes-
41. Li, L., Xu, Y., Yan, L., Li, X., Li, F., Liu, Z., Zhang, C., Lou, Y., Gao, D., cence experimentation in vivo. Cell 187, 4150–4175. https://doi.org/10.
Cheng, X., et al. (2023). Dynamic network biomarker factors orchestrate 1016/j.cell.2024.05.059.
cell-fate determination at tipping points during hESC differentiation. 57. Wilhelm, M., Schlegl, J., Hahne, H., Gholami, A.M., Lieberenz, M., Savit-
Innovation (Camb.) 4, 100364. https://doi.org/10.1016/j.xinn.2022. ski, M.M., Ziegler, E., Butzmann, L., Gessulat, S., Marx, H., et al. (2014).
100364. Mass-spectrometry-based draft of the human proteome. Nature 509,
42. Ma, S., Ji, Z., Zhang, B., Geng, L., Cai, Y., Nie, C., Li, J., Zuo, Y., Sun, Y., 582–587. https://doi.org/10.1038/nature13319.
Xu, G., et al. (2024). Spatial transcriptomic landscape unveils immuno- 58. Niu, L., Thiele, M., Geyer, P.E., Rasmussen, D.N., Webel, H.E., Santos,
globin-associated senescence as a hallmark of aging. Cell 187, 7025– A., Gupta, R., Meier, F., Strauss, M., Kjaergaard, M., et al. (2022). Nonin-
7044.e34. https://doi.org/10.1016/j.cell.2024.10.019. vasive proteomic biomarkers for alcohol-related liver disease. Nat. Med.
43. Kaushik, S., and Cuervo, A.M. (2015). Proteostasis and aging. Nat. Med. 28, 1277–1287. https://doi.org/10.1038/s41591-022-01850-y.
21, 1406–1415. https://doi.org/10.1038/nm.4001. 59. Choughule, K.V., Locuson, C.W., and Coughtrie, M.W.H. (2015). Charac-
44. Hipp, M.S., Kasturi, P., and Hartl, F.U. (2019). The proteostasis network terization of bovine phenol sulfotransferases: evidence of a major role for
and its decline in ageing. Nat. Rev. Mol. Cell Biol. 20, 421–435. https:// SULT1B1 in the liver. Xenobiotica 45, 495–502. https://doi.org/10.3109/
doi.org/10.1038/s41580-019-0101-y. 00498254.2014.997325.
45. López-Otı́n, C., Blasco, M.A., Partridge, L., Serrano, M., and Kroemer, G. 60. Zhu, G., Fang, Q., Zhu, F., Huang, D., and Yang, C. (2021). Structure and
(2023). Hallmarks of aging: An expanding universe. Cell 186, 243–278. Function of Pancreatic Lipase-Related Protein 2 and Its Relationship
https://doi.org/10.1016/j.cell.2022.11.001. With Pathological States. Front. Genet. 12, 693538. https://doi.org/10.
3389/fgene.2021.693538.
46. Aebersold, R., and Mann, M. (2016). Mass-spectrometric exploration of
proteome structure and function. Nature 537, 347–355. https://doi.org/ 61. Frigeri, A., Nicchia, G.P., Balena, R., Nico, B., and Svelto, M. (2004).
10.1038/nature19949. Aquaporins in skeletal muscle: reassessment of the functional role of
47. Keele, G.R., Zhang, J.G., Szpyt, J., Korstanje, R., Gygi, S.P., Churchill, G. aquaporin-4. FASEB J. 18, 905–907. https://doi.org/10.1096/fj.03-
A., and Schweppe, D.K. (2023). Global and tissue-specific aging effects 0987fje.
on murine proteomes. Cell Rep. 42, 112715. https://doi.org/10.1016/j. 62. Zhang, B., Wu, Y., Yang, X., Xiang, Y., and Yang, B. (2023). Molecular
celrep.2023.112715. insight into arrhythmogenic cardiomyopathy caused by DSG2 mutations.
48. Kim, M.S., Pinto, S.M., Getnet, D., Nirujogi, R.S., Manda, S.S., Chaer- Biomed. Pharmacother. 167, 115448. https://doi.org/10.1016/j.biopha.
kady, R., Madugundu, A.K., Kelkar, D.S., Isserlin, R., Jain, S., et al. 2023.115448.
(2014). A draft map of the human proteome. Nature 509, 575–581. 63. Wei, Y.-N., Hu, H.-Y., Xie, G.-C., Fu, N., Ning, Z.-B., Zeng, R., and Khai-
https://doi.org/10.1038/nature13302. tovich, P. (2015). Transcript and protein expression decoupling reveals
49. He, F., Aebersold, R., Baker, M.S., Bian, X., Bo, X., Chan, D.W., Chang, RNA binding proteins and miRNAs as potential modulators of human ag-
C., Chen, L., Chen, X., Chen, Y.J., et al. (2024). π-HuB: the proteomic ing. Genome Biol. 16, 41. https://doi.org/10.1186/s13059-015-0608-2.
navigator of the human body. Nature 636, 322–331. https://doi.org/10. 64. Cai, N., Li, M., Qu, J., Liu, G.H., and Izpisua Belmonte, J.C. (2012). Post-
1038/s41586-024-08280-5. translational modulation of pluripotency. J. Mol. Cell Biol. 4, 262–265.
50. Yan, P., Ren, J., Zhang, W., Qu, J., and Liu, G.H. (2020). Protein quality https://doi.org/10.1093/jmcb/mjs031.
control of cell stemness. Cell Regen. 9, 22. https://doi.org/10.1186/ 65. Cai, N., Li, M., Qu, J., Liu, G.H., and Izpisua Belmonte, J.C. (2012). Post-
s13619-020-00064-2. translational modulation of pluripotency. J. Mol. Cell Biol. 4, 262–265.
51. Sun, B.B., Chiou, J., Traylor, M., Benner, C., Hsu, Y.H., Richardson, T.G., https://doi.org/10.1093/jmcb/mjs031.
Surendran, P., Mahajan, A., Robins, C., Vasquez-Grinnell, S.G., et al. 66. Rahman, M.M., and Lendel, C. (2021). Extracellular protein components
(2023). Plasma proteomic associations with genetics and health in the of amyloid plaques and their roles in Alzheimer’s disease pathology. Mol.
UK Biobank. Nature 622, 329–338. https://doi.org/10.1038/s41586- Neurodegener. 16, 59. https://doi.org/10.1186/s13024-021-00465-0.
023-06592-6. 67. Schartz, N.D., and Tenner, A.J. (2020). The good, the bad, and the oppor-
52. Guzman, U.H., Martinez-Val, A., Ye, Z., Damoc, E., Arrey, T.N., Pash- tunities of the complement system in neurodegenerative disease.
kova, A., Renuse, S., Denisov, E., Petzoldt, J., Peterson, A.C., et al. J. Neuroinflammation 17, 354. https://doi.org/10.1186/s12974-020-
(2024). Ultra-fast label-free quantification and comprehensive proteome 02024-8.

18 Cell 188, 1–22, October 2, 2025

Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article
68. Reichwald, J., Danner, S., Wiederhold, K.H., and Staufenbiel, M. (2009). framework for liver aging: the Aging Biomarker Consortium consensus
Expression of complement system components during aging and amy- statement. Life Med. 3, lnae004. https://doi.org/10.1093/lifemedi/
loid deposition in APP transgenic mice. J. Neuroinflammation 6, 35. lnae004.
https://doi.org/10.1186/1742-2094-6-35. 85. Argentieri, M.A., Xiao, S., Bennett, D., Winchester, L., Nevado-Holgado,
69. Doni, A., Parente, R., Laface, I., Magrini, E., Cunha, C., Colombo, F.S., A.J., Ghose, U., Albukhari, A., Yao, P., Mazidi, M., Lv, J., et al. (2024). Pro-
Lacerda, J.F., Campos, A., Jr., Mapelli, S.N., Petroni, F., et al. (2021). teomic aging clock predicts mortality and risk of common age-related
Serum amyloid P component is an essential element of resistance diseases in diverse populations. Nat. Med. 30, 2450–2460. https://doi.
against Aspergillus fumigatus. Nat. Commun. 12, 3739. https://doi.org/ org/10.1038/s41591-024-03164-7.
10.1038/s41467-021-24021-y. 86. Kuo, C.L., Chen, Z., Liu, P., Pilling, L.C., Atkins, J.L., Fortinsky, R.H., Ku-
70. Liu, X., Liu, Z., Wu, Z., Ren, J., Fan, Y., Sun, L., Cao, G., Niu, Y., Zhang, B., chel, G.A., and Diniz, B.S. (2024). Proteomic aging clock (PAC) predicts
Ji, Q., et al. (2023). Resurrection of endogenous retroviruses during aging age-related outcomes in middle-aged and older adults. Aging Cell 23,
reinforces senescence. Cell 186, 287–304.e26. https://doi.org/10.1016/j. e14195. https://doi.org/10.1111/acel.14195.
cell.2022.12.017. 87. Lee, W.T., Wu, P.Y., Cheng, Y.M., and Huang, Y.F. (2024). Tissue Inhib-
71. Li, M.H., Jiang, X., Jing, Y., Yan, K., Bi, S.J., Wang, S., Ma, S., Liu, G.H., itor of Metalloproteinase 3: Unravelling Its Biological Function and Signif-
Zhang, W., Sun, S., et al. (2025). CRISPR-based screening pinpoints icance in Oncology. Int. J. Mol. Sci. 25, 3191. https://doi.org/10.3390/
H2AZ1 as a driver of senescence in human mesenchymal stem cells. ijms25063191.
Protein Cell 16, 296–305. https://doi.org/10.1093/procel/pwae035. 88. Qi, J.H., and Anand-Apte, B. (2015). Tissue inhibitor of
72. Zhang, W., Qu, J., Liu, G.H., and Belmonte, J.C.I. (2020). The ageing epi- metalloproteinase-3 (TIMP3) promotes endothelial apoptosis via a cas-
genome and its rejuvenation. Nat. Rev. Mol. Cell Biol. 21, 137–150. pase-independent mechanism. Apoptosis 20, 523–534. https://doi.org/
https://doi.org/10.1038/s41580-019-0204-5. 10.1007/s10495-014-1076-y.
73. Wu, Z., Zhang, W., Qu, J., and Liu, G.H. (2024). Emerging epigenetic in- 89. Wang, Q., Wang, X., Liu, B., Ma, S., Zhang, F., Sun, S., Jing, Y., Fan, Y.,
sights into aging mechanisms and interventions. Trends Pharmacol. Sci. Ding, Y., Xiong, M., et al. (2024). Aging induces region-specific dysregu-
45, 157–172. https://doi.org/10.1016/j.tips.2023.12.002. lation of hormone synthesis in the primate adrenal gland. Nat Aging 4,
74. Ren, R., Deng, L., Xue, Y., Suzuki, K., Zhang, W., Yu, Y., Wu, J., Sun, L., 396–413. https://doi.org/10.1038/s43587-024-00588-1.
Gong, X., Luan, H., et al. (2017). Visualization of aging-associated chro- 90. Jiao, H., Li, X., Li, Y., Guo, Y., Hu, X., Sho, T., Luo, Y., Wang, J., Cao, H.,
matin alterations with an engineered TALE system. Cell Res. 27, Du, W., et al. (2024). Localized, highly efficient secretion of signaling pro-
483–504. https://doi.org/10.1038/cr.2017.18. teins by migrasomes. Cell Res. 34, 572–585. https://doi.org/10.1038/
75. Liang, C., Ke, Q., Liu, Z., Ren, J., Zhang, W., Hu, J., Wang, Z., Chen, H., s41422-024-00992-7.
Xia, K., Lai, X., et al. (2022). BMAL1 moonlighting as a gatekeeper for 91. Chen, J., Fang, M., Li, Y., Ding, H., Zhang, X., Jiang, X., Zhang, J., Zhang,
LINE1 repression and cellular senescence in primates. Nucleic Acids C., Lu, Z., and Luo, M. (2024). Cell surface protein–protein interaction
Res. 50, 3323–3347. https://doi.org/10.1093/nar/gkac146. profiling for biological network analysis and novel target discovery. Life
76. Wu, Z., Qu, J., and Liu, G.-H. (2024). Roles of chromatin and genome Med. 3, lnae031. https://doi.org/10.1093/lifemedi/lnae031.
instability in cellular senescence and their relevance to ageing and 92. Feizi, A., Gatto, F., Uhlen, M., and Nielsen, J. (2017). Human protein
related diseases. Nat. Rev. Mol. Cell Biol. 25, 979–1000. https://doi. secretory pathway genes are expressed in a tissue-specific pattern to
org/10.1038/s41580-024-00775-3. match processing demands of the secretome. npj Syst. Biol. Appl. 3,
77. Fan, Y., Zheng, Y., Zhang, Y., Xu, G., Liu, C., Hu, J., Ji, Q., Zhang, S., 22. https://doi.org/10.1038/s41540-017-0021-4.
Fang, S., Lei, J., et al. (2025). ARID5A orchestrates cardiac aging and 93. De Cecco, M., Ito, T., Petrashen, A.P., Elias, A.E., Skvir, N.J., Criscione,
inflammation through MAVS mRNA stabilization. Nat CardioVasc Res S.W., Caligiana, A., Brocculi, G., Adney, E.M., Boeke, J.D., et al. (2019).
4, 602–623. https://doi.org/10.1038/s44161-025-00635-z. L1 drives IFN in senescent cells and promotes age-associated inflamma-
78. Wang, C., Jiang, X., Li, H.Y., Hu, J., Ji, Q., Wang, Q., Liu, X., Huang, D., tion. Nature 566, 73–78. https://doi.org/10.1038/s41586-018-0784-9.
Yan, K., Zhao, L., et al. (2025). RIG-I-driven CDKN1A stabilization rein- 94. Biswas, L., Chen, J., De Angelis, J., Singh, A., Owen-Woods, C., Ding, Z.,
forces cellular senescence. Sci. China Life Sci. 68, 1646–1661. https:// Pujol, J.M., Kumar, N., Zeng, F., Ramasamy, S.K., et al. (2023).
doi.org/10.1007/s11427-024-2844-8. Lymphatic vessels in bone support regeneration after injury. Cell 186,
79. Wu, Z., Qu, J., Zhang, W., and Liu, G.H. (2024). Stress, epigenetics, and 382–397.e24. https://doi.org/10.1016/j.cell.2022.12.031.
aging: Unraveling the intricate crosstalk. Mol. Cell 84, 34–54. https://doi. 95. Garcia-Alonso, L., Lorenzi, V., Mazzeo, C.I., Alves-Lopes, J.P., Roberts,
org/10.1016/j.molcel.2023.10.006. K., Sancho-Serra, C., Engelbert, J., Marec �ková, M., Gruhn, W.H., Bot-
80. Zhao, H., Ji, Q., Wu, Z., Wang, S., Ren, J., Yan, K., Wang, Z., Hu, J., Chu, ting, R.A., et al. (2022). Single-cell roadmap of human gonadal develop-
Q., Hu, H., et al. (2022). Destabilizing heterochromatin by APOE mediates ment. Nature 607, 540–547. https://doi.org/10.1038/s41586-022-
senescence. Nat. Aging 2, 303–316. https://doi.org/10.1038/s43587- 04918-4.
022-00186-z. 96. Wu, G., Ma, Z., Cheng, Y., Hu, W., Deng, C., Jiang, S., Li, T., Chen, F., and
81. Piñero, J., Ramı́rez-Anguita, J.M., Saüch-Pitarch, J., Ronzano, F., Cen- Yang, Y. (2018). Targeting Gas6/TAM in cancer cells and tumor microen-
teno, E., Sanz, F., and Furlong, L.I. (2020). The DisGeNET knowledge vironment. Mol. Cancer 17, 20. https://doi.org/10.1186/s12943-018-
platform for disease genomics: 2019 update. Nucleic Acids Res. 48, 0769-1.
D845–D855. https://doi.org/10.1093/nar/gkz1021. 97. Happonen, K.E., and Dahlbäck, B. (2016). Gas6 fueling tumor-mediated
82. Ye, Y., Yang, K., Liu, H., Yu, Y., Song, M., Huang, D., Lei, J., Zhang, Y., thrombosis. Blood 127, 672–673. https://doi.org/10.1182/blood-2015-
Liu, Z., Chu, Q., et al. (2023). SIRT2 counteracts primate cardiac aging 12-683474.
via deacetylation of STAT3 that silences CDKN2B. Nat Aging 3, 1269– 98. Borgel, D. (2008). Gas6 inflames cell interactions. Blood 111, 3915.
1287. https://doi.org/10.1038/s43587-023-00486-y. https://doi.org/10.1182/blood-2008-01-133314.
83. Chen, C., Liu, J., and Yu, W. (2024). Unlocking heart anti-aging potential: 99. Suda, M., Shimizu, I., Katsuumi, G., Yoshida, Y., Hayashi, Y., Ikegami, R.,
the SIRT2-STAT3-CDKN2B pathway as a bridge between fiction and re- Matsumoto, N., Yoshida, Y., Mikawa, R., Katayama, A., et al. (2021). Se-
ality. Life Med. 3, lnae020. https://doi.org/10.1093/lifemedi/lnae020. nolytic vaccination improves normal and pathological age-related phe-
84. Jiang, M., Zheng, Z., Wang, X., Chen, Y., Qu, J., Ding, Q., Zhang, W., Liu, notypes and increases lifespan in progeroid mice. Nat Aging 1, 1117–
Y.S., Yang, J., et al.; Aging; Biomarker Consortium (2024). A biomarker 1126. https://doi.org/10.1038/s43587-021-00151-2.

Cell 188, 1–22, October 2, 2025 19

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article
100. Liu, Y.Y., Pang, J., Zhang, C., Zeng, L.T., Wang, Y., Wang, S.B., Fan, G. 117. Gillmore, J.D., and Hawkins, P.N. (2013). Pathophysiology and treatment
Q., Zhang, L.Q., Shen, T., Li, X.F., et al. (2024). Biofluid GPNMB/osteoac- of systemic amyloidosis. Nat. Rev. Nephrol. 9, 574–586. https://doi.org/
tivin as a potential biomarker of ageing: A cross-sectional study. Heliyon 10.1038/nrneph.2013.171.
10, e36574. https://doi.org/10.1016/j.heliyon.2024.e36574. 118. Skovronsky, D.M., Lee, V.M.Y., and Trojanowski, J.Q. (2006). Neurode-
101. Sun, S., Li, J., Wang, S., Li, J., Ren, J., Bao, Z., Sun, L., Ma, X., Zheng, F., generative diseases: new concepts of pathogenesis and their therapeu-
Ma, S., et al. (2023). CHIT1-positive microglia drive motor neuron ageing tic implications. Annu. Rev. Pathol. 1, 151–170. https://doi.org/10.1146/
in the primate spinal cord. Nature 624, 611–620. https://doi.org/10.1038/ annurev.pathol.1.110304.100113.
s41586-023-06783-1. 119. Iadanza, M.G., Jackson, M.P., Hewitt, E.W., Ranson, N.A., and Radford,
102. Adams, R.A., Fernandes-Cerqueira, C., Notarnicola, A., Mertsching, E., S.E. (2018). A new era for understanding amyloid structures and disease.
Xu, Z., Lo, W.S., Ogilvie, K., Chiang, K.P., Ampudia, J., Rosengren, S., Nat. Rev. Mol. Cell Biol. 19, 755–773. https://doi.org/10.1038/s41580-
et al. (2021). Serum-circulating His-tRNA synthetase inhibits organ-tar- 018-0060-8.
geted immune responses. Cell. Mol. Immunol. 18, 1463–1475. https:// 120. Gao, Y., Hong, Y., Huang, L., Zheng, S., Zhang, H., Wang, S., Yao, Y.,
doi.org/10.1038/s41423-019-0331-0. Zhao, Y., Zhu, L., Xu, Q., et al. (2023). β2-microglobulin functions as an
endogenous NMDAR antagonist to impair synaptic function. Cell 186,
103. Galatolo, D., Kuo, M.E., Mullen, P., Meyer-Schuman, R., Doccini, S., Bat-
1026–1038.e20. https://doi.org/10.1016/j.cell.2023.01.021.
tini, R., Lieto, M., Tessa, A., Filla, A., Francklyn, C., et al. (2020). Bi-allelic
mutations in HARS1 severely impair histidyl-tRNA synthetase expression 121. Langworth-Green, C., Patel, S., Jaunmuktane, Z., Jabbari, E., Morris, H.,
and enzymatic activity causing a novel multisystem ataxic syndrome. Thom, M., Lees, A., Hardy, J., Zandi, M., and Duff, K. (2023). Chronic ef-
Hum. Mutat. 41, 1232–1237. https://doi.org/10.1002/humu.24024. fects of inflammation on tauopathies. Lancet Neurol. 22, 430–442.
https://doi.org/10.1016/S1474-4422(23)00038-8.
104. Ozcelik, A., Rufo, J., Guo, F., Gu, Y., Li, P., Lata, J., and Huang, T.J.
(2018). Acoustic tweezers for the life sciences. Nat. Methods 15, 1021– 122. Richards, D.B., Cookson, L.M., Berges, A.C., Barton, S.V., Lane, T., Rit-
1028. https://doi.org/10.1038/s41592-018-0222-9. ter, J.M., Fontana, M., Moon, J.C., Pinzani, M., Gillmore, J.D., et al.
(2015). Therapeutic Clearance of Amyloid by Antibodies to Serum Amy-
105. Rutledge, J., Oh, H., and Wyss-Coray, T. (2022). Measuring biological
loid P Component. N. Engl. J. Med. 373, 1106–1114. https://doi.org/10.
age using omics data. Nat. Rev. Genet. 23, 715–727. https://doi.org/
1056/NEJMoa1504942.
10.1038/s41576-022-00511-7.
123. Ma, S., Sun, S., Geng, L., Song, M., Wang, W., Ye, Y., Ji, Q., Zou, Z.,
106. Moqri, M., and Snyder, M. (2023). Organ-specific aging and the risk of Wang, S., He, X., et al. (2020). Caloric Restriction Reprograms the
chronic diseases. Nat. Med. 29, 1068–1069. https://doi.org/10.1038/ Single-Cell Transcriptional Landscape of Rattus Norvegicus Aging. Cell
s41591-023-02338-z. 180, 984–1001.e22. https://doi.org/10.1016/j.cell.2020.02.008.
107. Furman, D., Campisi, J., Verdin, E., Carrera-Bastos, P., Targ, S., France- 124. Yu, L., Wan, Q., Liu, Q., Fan, Y., Zhou, Q., Skowronski, A.A., Wang, S.,
schi, C., Ferrucci, L., Gilroy, D.W., Fasano, A., Miller, G.W., et al. (2019). Shao, Z., Liao, C.Y., Ding, L., et al. (2024). IgG is an aging factor that
Chronic inflammation in the etiology of disease across the life span. Nat. drives adipose tissue fibrosis and metabolic decline. Cell Metab. 36,
Med. 25, 1822–1832. https://doi.org/10.1038/s41591-019-0675-0. 793–807.e5. https://doi.org/10.1016/j.cmet.2024.01.015.
108. Li, Y., Wu, A., Jin, X., Shen, H., Zhao, C., Yi, X., Nie, H., Wang, M., Yin, S., 125. Wang, S., Zheng, Y., Li, J., Yu, Y., Zhang, W., Song, M., Liu, Z., Min, Z.,
Zuo, H., et al. (2024). DDO1002, an NRF2-KEAP1 inhibitor, improves he- Hu, H., Jing, Y., et al. (2020). Single-Cell Transcriptomic Atlas of Primate
matopoietic stem cell aging and stress response. Life Med. 3, lnae043. Ovarian Aging. Cell 180, 585–600.e19. https://doi.org/10.1016/j.cell.
https://doi.org/10.1093/lifemedi/lnae043. 2020.01.009.
109. Sonninen, T.M., Goldsteins, G., Laham-Karam, N., Koistinaho, J., and 126. Du Clos, T.W., and Mold, C. (2011). Pentraxins (CRP, SAP) in the process
� (2020). Proteostasis Disturbances and Inflammation in
Lehtonen, S. of complement activation and clearance of apoptotic bodies through Fcγ
Neurodegenerative Diseases. Cells 9, 2183. https://doi.org/10.3390/ receptors. Curr. Opin. Organ Transplant. 16, 15–20. https://doi.org/10.
cells9102183. 1097/MOT.0b013e32834253c7.

110. Korovila, I., Hugo, M., Castro, J.P., Weber, D., Höhn, A., Grune, T., and 127. Geng, L., Ping, J., Wu, R., Yan, H., Zhang, H., Zhuang, Y., Ning, T., Wang,
Jung, T. (2017). Proteostasis, oxidative stress and aging. Redox Biol. J., Liang, C., Zhang, J., et al. (2025). Systematic profiling reveals betaine
13, 550–567. https://doi.org/10.1016/j.redox.2017.07.008. as an exercise mimetic for geroprotection. Cell. https://doi.org/10.1016/j.
cell.2025.06.001.
111. Jiménez-Loygorri, J.I., and Boya, P. (2024). Aging STINGs: mitophagy at
128. Sun, G., Li, K., Ping, J., Zhao, L., Cui, C., Wu, J., Xie, L., Yao, X., Xu, G.,
the crossroads of neuroinflammation. Autophagy 20, 1684–1686. https://
Ma, S., et al. (2025). A single-cell transcriptomic atlas reveals senescence
doi.org/10.1080/15548627.2024.2322421.
and inflammation in the post-tuberculosis human lung. Nat. Microbiol.
112. Franceschi, C., Garagnani, P., Parini, P., Giuliani, C., and Santoro, A. https://doi.org/10.1038/s41564-025-02050-3.
(2018). Inflammaging: a new immune-metabolic viewpoint for age-
129. Wu, R., Sun, F., Zhang, W., Ren, J., and Liu, G.H. (2024). Targeting aging
related diseases. Nat. Rev. Endocrinol. 14, 576–590. https://doi.org/10.
and age-related diseases with vaccines. Nat Aging 4, 464–482. https://
1038/s41574-018-0059-4.
doi.org/10.1038/s43587-024-00597-0.
113. Yan, H., Jiang, M., Qu, J., and Liu, G.-H. (2025). IgG: An Emerging Hall-
130. Schneider, C.A., Rasband, W.S., and Eliceiri, K.W. (2012). NIH Image to
mark of Aging. Oral Sci. Homeost. Med. 1. https://doi.org/10.26599/
ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675. https://
OSHM.2025.9610001.
doi.org/10.1038/nmeth.2089.
114. Wechalekar, A.D., Gillmore, J.D., and Hawkins, P.N. (2016). Systemic 131. Demichev, V., Messner, C.B., Vernardis, S.I., Lilley, K.S., and Ralser, M.
amyloidosis. Lancet 387, 2641–2654. https://doi.org/10.1016/S0140- (2020). DIA-NN: neural networks and interference correction enable deep
6736(15)01274-X. proteome coverage in high throughput. Nat. Methods 17, 41–44. https://
115. Westermark, G.T., Fändrich, M., and Westermark, P. (2015). AA amyloid- doi.org/10.1038/s41592-019-0638-x.
osis: pathogenesis and targeted therapy. Annu. Rev. Pathol. 10, 132. Feng, Z., Fang, P., Zheng, H., and Zhang, X. (2023). DEP2: an upgraded
321–344. https://doi.org/10.1146/annurev-pathol-020712-163913. comprehensive analysis toolkit for quantitative proteomics data. Bioin-
116. Gertz, M.A., Dispenzieri, A., and Sher, T. (2015). Pathophysiology and formatics 39, btad526. https://doi.org/10.1093/bioinformatics/btad526.
treatment of cardiac amyloidosis. Nat. Rev. Cardiol. 12, 91–102. 133. Hao, Z., Lv, D., Ge, Y., Shi, J., Weijers, D., Yu, G., and Chen, J. (2020).
https://doi.org/10.1038/nrcardio.2014.165. RIdeogram: drawing SVG graphics to visualize and map genome-wide

20 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article
data on the idiograms. PeerJ Comput. Sci. 6, e251. https://doi.org/10. human synovium through single-nucleus transcriptomic profiling. Protein
7717/peerj-cs.251. Cell 15, 441–459. https://doi.org/10.1093/procel/pwad060.
134. Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., and Smyth, 152. Yang, S., Liu, C., Jiang, M., Liu, X., Geng, L., Zhang, Y., Sun, S., Wang, K.,
G.K. (2015). limma powers differential expression analyses for RNA- Yin, J., Ma, S., et al. (2024). A single-nucleus transcriptomic atlas of pri-
sequencing and microarray studies. Nucleic Acids Res. 43, e47. mate liver aging uncovers the pro-senescence role of SREBP2 in hepa-
https://doi.org/10.1093/nar/gkv007. tocytes. Protein Cell 15, 98–120. https://doi.org/10.1093/procel/
135. Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis (New pwad039.
York: Springer-Verlag). https://doi.org/10.1007/978-3-319-24277-4. 153. Hu, Q., Zhang, B., Jing, Y., Ma, S., Hu, L., Li, J., Zheng, Y., Xin, Z., Peng,
136. Gu, Z. (2022). Complex heatmap visualization. Imeta 1, e43. https://doi. J., Wang, S., et al. (2024). Single-nucleus transcriptomics uncovers a ger-
org/10.1002/imt2.43. oprotective role of YAP in primate gingival aging. Protein Cell 15,
137. Shannon, P., Markiel, A., Ozier, O., Baliga, N.S., Wang, J.T., Ramage, D., 612–632. https://doi.org/10.1093/procel/pwae017.
Amin, N., Schwikowski, B., and Ideker, T. (2003). Cytoscape: a software 154. Huang, D., Zuo, Y., Zhang, C., Sun, G., Jing, Y., Lei, J., Ma, S., Sun, S.,
environment for integrated models of biomolecular interaction networks. Lu, H., Cai, Y., et al. (2023). A single-nucleus transcriptomic atlas of pri-
Genome Res. 13, 2498–2504. https://doi.org/10.1101/gr.1239303. mate testicular aging reveals exhaustion of the spermatogonial stem cell
138. Lehallier, B., Gate, D., Schaum, N., Nanasi, T., Lee, S.E., Yousef, H., reservoir and loss of Sertoli cell homeostasis. Protein Cell 14, 888–907.
Moran Losada, P., Berdnik, D., Keller, A., Verghese, J., et al. (2019). Un- https://doi.org/10.1093/procel/pwac057.
dulating changes in human plasma proteome profiles across the lifespan. 155. Lei, J., Jiang, X., Huang, D., Jing, Y., Yang, S., Geng, L., Yan, Y., Zheng,
Nat. Med. 25, 1843–1850. https://doi.org/10.1038/s41591-019-0673-2. F., Cheng, F., Zhang, W., et al. (2024). Human ESC-derived vascular cells
139. Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., promote vascular regeneration in a HIF-1α dependent manner. Protein
Batut, P., Chaisson, M., and Gingeras, T.R. (2013). STAR: ultrafast uni- Cell 15, 36–51. https://doi.org/10.1093/procel/pwad027.
versal RNA-seq aligner. Bioinformatics 29, 15–21. https://doi.org/10. 156. Uras, F., Küçük, B., Bingöl Özakpınar, Ö., and Demir, A.M. (2015). Growth
1093/bioinformatics/bts635. Arrest-Specific 6 (Gas6) and TAM Receptors in Mouse Platelets. Turk. J.
140. Liao, Y., Smyth, G.K., and Shi, W. (2014). featureCounts: an efficient gen- Haematol. 32, 58–63. https://doi.org/10.4274/tjh.2013.0097.
eral purpose program for assigning sequence reads to genomic features. 157. Rejas-González, R., Montero-Calle, A., Valverde, A., Salvador, N.P., Car-
Bioinformatics 30, 923–930. https://doi.org/10.1093/bioinformatics/ ballés, M.J.C., Ausı́n-González, E., Sánchez-Naves, J., Campuzano, S.,
btt656. Barderas, R., and Guzman-Aranguez, A. (2024). Proteomics Analyses
141. Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of of Small Extracellular Vesicles of Aqueous Humor: Identification and Vali-
fold change and dispersion for RNA-seq data with DESeq2. Genome dation of GAS6 and SPP1 as Glaucoma Markers. Int. J. Mol. Sci. 25,
Biol. 15, 550. https://doi.org/10.1186/s13059-014-0550-8. 6995. https://doi.org/10.3390/ijms25136995.
142. Wu, T., Hu, E., Xu, S., Chen, M., Guo, P., Dai, Z., Feng, T., Zhou, L., Tang, 158. Belgiovine, C., Digifico, E., Erreni, M., Putignano, A.R., Mannarino, L.,
W., Zhan, L., et al. (2021). clusterProfiler 4.0: A universal enrichment tool Valentino, S., Grizzi, F., Pasqualini, F., Recordati, C., Bertola, L., et al.
for interpreting omics data. Innovation (Camb.) 2, 100141. https://doi. (2025). Malignant mesothelioma-associated inflammatory microenviron-
org/10.1016/j.xinn.2021.100141. ment promotes tumor progression via GPNMB. J. Transl. Med. 23, 454.
143. Hänzelmann, S., Castelo, R., and Guinney, J. (2013). GSVA: gene set https://doi.org/10.1186/s12967-025-06407-4.
variation analysis for microarray and RNA-seq data. BMC Bioinformatics 159. Fukaura, M., Ishitsuka, Y., Shirakawa, S., Ushihama, N., Yamada, Y.,
14, 7. https://doi.org/10.1186/1471-2105-14-7. Kondo, Y., Takeo, T., Nakagata, N., Motoyama, K., Higashi, T., et al.
144. Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, (2021). Intracerebroventricular Treatment with 2-Hydroxypropyl-β--
O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., et al. Scikit- Cyclodextrin Decreased Cerebellar and Hepatic Glycoprotein Nonmeta-
learn: Machine Learning in Python. J. Mach. Learn. Res. 12, 2825–2830. static Melanoma Protein B (GPNMB) Expression in Niemann-Pick Dis-
145. Peng, Y., Ding, L., Xiao, Z., Song, M., Lv, J., and Liu, G.H. (2024). Ethical ease Type C Model Mice. Int. J. Mol. Sci. 22, 452. https://doi.org/10.
concerns in aging research: perspectives of global frontline researchers. 3390/ijms22010452.
Sci. China Life Sci. 67, 2149–2156. https://doi.org/10.1007/s11427-024- 160. Eskes, E.C.B., van der Lienden, M.J.C., Sjouke, B., van Vliet, L., Brands,
2650-y. M., Hollak, C.E.M., and Aerts, J. (2023). Glycoprotein non-metastatic
146. Peng, Y., Ding, L., Song, M., Xiao, Z., Lv, J., and Liu, G.H. (2023). Acting protein B (GPNMB) plasma values in patients with chronic visceral acid
on ethics and governance of aging research. Trends Mol. Med. 29, sphingomyelinase deficiency. Mol. Genet. Metab. 139, 107631. https://
419–421. https://doi.org/10.1016/j.molmed.2023.03.004. doi.org/10.1016/j.ymgme.2023.107631.
147. Zhang, W., Zhang, S., Yan, P., Ren, J., Song, M., Li, J., Lei, J., Pan, H., 161. Sainaghi, P.P., Alciato, F., Carnieletto, S., Castello, L., Bergamasco, L.,
Wang, S., Ma, X., et al. (2020). A single-cell transcriptomic landscape Sola, D., Bongo, A.S., Inglese, E., Polosa, R., and Avanzi, G.C. (2009).
of primate arterial aging. Nat. Commun. 11, 2202. https://doi.org/10. Gas6 evaluation in patients with acute dyspnea due to suspected pulmo-
1038/s41467-020-15997-0. nary embolism. Respir. Med. 103, 589–594. https://doi.org/10.1016/j.
148. Patel, J.J., Srivastava, S., and Siow, R.C.M. (2016). Isolation, Culture, rmed.2008.10.018.
and Characterization of Vascular Smooth Muscle Cells. Methods Mol. 162. Sun, S., Ma, S., Cai, Y., Wang, S., Ren, J., Yang, Y., Ping, J., Wang, X.,
Biol. 1430, 91–105. https://doi.org/10.1007/978-1-4939-3628-1_6. Zhang, Y., Yan, H., et al. (2023). A single-cell transcriptomic atlas of ex-
149. Niu, L., Stinson, S.E., Holm, L.A., Lund, M.A.V., Fonvig, C.E., Cobuccio, ercise-induced anti-inflammatory and geroprotective effects across the
L., Meisner, J., Juel, H.B., Fadista, J., Thiele, M., et al. (2025). Plasma pro- body. Innovation (Camb.) 4, 100380. https://doi.org/10.1016/j.xinn.
teome variation and its genetic determinants in children and adolescents. 2023.100380.
Nat. Genet. 57, 635–646. https://doi.org/10.1038/s41588-025-02089-2. 163. Diamanti, K., Cavalli, M., Pereira, M.J., Pan, G., Castillejo-López, C., Ku-
150. Koh, B., Liu, M., Almonte, R., Ariad, D., Bundalian, G., Chan, J., Choi, J., mar, C., Mundt, F., Komorowski, J., Deshmukh, A.S., Mann, M., et al.
Chou, W.-F., Cuaresma, R., et al. (2024). Multi-omics profiling with untar- (2022). Organ-specific metabolic pathways distinguish prediabetes,
geted proteomics for blood-based early detection of lung cancer. Pre- type 2 diabetes, and normal tissues. Cell Rep. Med. 3, 100763. https://
print at medRxiv. https://doi.org/10.1101/2024.01.03.24300798. doi.org/10.1016/j.xcrm.2022.100763.
151. Liu, F., Lu, Y., Wang, X., Sun, S., Pan, H., Wang, M., Wang, Z., Zhang, W., 164. Wei, R., Wang, J., Su, M., Jia, E., Chen, S., Chen, T., and Ni, Y. (2018).
Ma, S., Sun, G., et al. (2024). Identification of FOXO1 as a geroprotector in Missing Value Imputation Approach for Mass Spectrometry-based

Cell 188, 1–22, October 2, 2025 21

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article
Metabolomics Data. Sci. Rep. 8, 663. https://doi.org/10.1038/s41598- 167. Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of
017-19120-0. fold change and dispersion for RNA-seq data with DESeq2. Genome
Biol. 15, 550. https://doi.org/10.1186/s13059-014-0550-8.
165. Frankish, A., Diekhans, M., Jungreis, I., Lagarde, J., Loveland, J.E., 168. Marakulina, D., Vorontsov, I.E., Kulakovskiy, I.V., Lennartsson, A.,
Mudge, J.M., Sisu, C., Wright, J.C., Armstrong, J., Barnes, I., et al. Drabløs, F., and Medvedeva, Y.A. (2023). EpiFactors 2022: expansion
(2021). GENCODE 2021. Nucleic Acids Res. 49, D916–D923. https:// and enhancement of a curated database of human epigenetic factors
doi.org/10.1093/nar/gkaa1087. and complexes. Nucleic Acids Res. 51, D564–D570. https://doi.org/10.
1093/nar/gkac989.
166. Uhlén, M., Fagerberg, L., Hallström, B.M., Lindskog, C., Oksvold, P., 169. Steinkamp, R., Tsitsiridis, G., Brauner, B., Montrone, C., Fobo, G., Frish-
Mardinoglu, A., Sivertsson, Å., Kampf, C., Sjöstedt, E., Asplund, A., man, G., Avram, S., Oprea, T.I., and Ruepp, A. (2025). CORUM in 2024:
et al. (2015). Proteomics. Tissue-based map of the human proteome. Sci- protein complexes as drug targets. Nucleic Acids Res. 53, D651–D657.
ence 347, 1260419. https://doi.org/10.1126/science.1260419. https://doi.org/10.1093/nar/gkae1033.

22 Cell 188, 1–22, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

STAR★METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies
Rabbit Polyclonal anti-SULT1B1 Proteintech Cat# 16050-1-AP; RRID: AB_2302767
Rabbit Polyclonal anti-PNLIPRP2 Proteintech Cat# 26218-1-AP; RRID: AB_2880430
Rabbit Polyclonal anti-AQP4 Proteintech Cat# 16473-1-AP; RRID: AB_2827426
Rabbit Polyclonal anti-DSG2 Proteintech Cat# 21880-1-AP; RRID: AB_10836933
Mouse Monoclonal anti-SGF29 (CCDC101) Santa Cruz Cat# sc-515286; RRID: N/A
Rabbit Polyclonal anti-MMP3 Abcam Cat# ab53015; RRID: AB_881242
Rabbit Monoclonal anti-SAA1/2 Abcam Cat# ab190802; RRID: N/A
Mouse Monoclonal anti-Cardiac Troponin T Abcam Cat# ab8295; RRID: N/A
Rabbit Polyclonal anti-Lambda Light Chain Bioss Cat# bs-3801R; AB_10855843
Rabbit Polyclonal anti-IgA Immunoway Cat# YT2281; RRID: N/A
Rabbit Monoclonal anti-C4A Abclonal Cat# A3545; RRID: AB_3096350
Mouse Monoclonal anti-Serum amyloid Proteintech Cat# 66084-1-Ig; RRID: AB_11182271
P component (SAP)
Rabbit Monoclonal anti-p21Cip1 Cell Signaling Technology Cat# 2947; RRID: AB_823586
Mouse Monoclonal anti-IL-6 Abcam Cat# ab9324; RRID: AB_307175
Rabbit Polyclonal anti-Collagen IV Abcam Cat# ab6586
Rabbit Monoclonal anti-Phospho-Histone H2AX Cell signaling technology Cat# 9718; RRID: AB_2118009
Rabbit Polyclonal anti-H3K9me3 Abcam Cat# 8898; RRID: AB_306848
Rabbit Polyclonal anti-TIMP3 Proteintech Cat# 10858-1-AP; RRID: AB_2204973
Rabbit Polyclonal anti-CXCL12 Proteintech Cat# 17402-1-AP; RRID: AB_2878404
Rabbit Polyclonal anti-COMP Abclonal Cat# A13963; RRID: AB_2760817
Rabbit Polyclonal anti-NOTCH3 Abclonal Cat# A3115; RRID: AB_2764913
Rabbit Monoclonal anti-GPNMB Cell Signaling Technology Cat# 90205s
Rabbit Polyclonal anti-LTBP2 Bioss Cat# bs-18440R; RRID: N/A
Rabbit Polyclonal anti-HTRA1 Proteintech Cat# 55011-1-AP; RRID: AB_10859830
Mouse Monoclonal anti-LAP2 BD Cat# 611000; RRID: AB_10571596
Mouse Monoclonal anti-GAPDH Santa Cruz Biotechnology Cat# sc-365062; RRID: AB_10847862
Rabbit Polyclonal anti-GAS6 Proteintech Cat# 13795-1-AP
Rabbit Polyclonal anti-CCT5 Proteintech Cat# 11603-1-AP
Rabbit Polyclonal anti-CCT7 Proteintech Cat# 15994-1-AP
Rabbit Recombinant anti-HARS Proteintech Cat# 83497-1-RR
Rabbit Polyclonal anti-PSMB1 Proteintech Cat# 11749-1-AP
Rabbit Polyclonal anti-YTHDC1 Proteintech Cat# 14392-1-AP
Rabbit Polyclonal anti-MMTV Novus Biologicals Cat# NBP2-44179
Mouse Monoclonal anti-8-OHdG Santa Cruz Cat# sc-66036
Rabbit Polyclonal anti-S100A8 Abcam Cat# ab180735
Sheep Polyclonal anti-CD31 R&D system Cat# AF806
Rabbit Polyclonal anti-Von Willebrand Factor (vWF) Abcam Cat# ab6994
BODIPY™ 493/503 Invitrogen Cat# D3922; RRID: N/A
Alexa Fluor 488 Donkey Anti-Rabbit IgG (H+L) Invitrogen Cat# A-21206; RRID: AB_2535792
Alexa Fluor 568 Donkey Anti-Rabbit IgG (H+L) Invitrogen Cat# A-10042; RRID: AB_2534017
Alexa Fluor 568 Donkey Anti-Sheep IgG (H+L) Invitrogen Cat# A-21099; RRID: AB_10055702
(Continued on next page)

Cell 188, 1–22.e1–e12, October 2, 2025 e1

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Alexa Fluor 647 Donkey Anti-Rabbit IgG (H+L) Invitrogen Cat# A-31573; RRID: AB_2536183
Alexa Fluor 488 Donkey Anti-Mouse IgG (H+L) Invitrogen Cat# A-21202; RRID: AB_141607
Alexa Fluor 568 Donkey Anti-Mouse IgG (H+L) Invitrogen Cat# A-10037; RRID: AB_2534013
Alexa Fluor(R) 488 anti-mouse MAdCAM-1 BioLegend Cat# A-120708; RRID: AB_493398
Alexa Fluor 488 Donkey Anti-Rat IgG (H+L) Invitrogen Cat# A-21208; RRID: AB_2535794
Biological samples
Human heart, aorta, lung, muscle, spleen, West China Hospital of N/A
lymph, adrenal, skin, liver, intestine, Sichuan University
pancreas, adipose and peripheral
blood for proteomics analysis
Human peripheral blood for ELISA analysis Quzhou People’s Hospital, N/A
Quzhou, China
Experimental models: Organisms/strains
C57BL/6J male mice Specific Pathogen Free (SPF) N/A
Biotechnology Co., Ltd.,
Beijing, China
Experimental models: Cell lines
Human aortic endothelial cells Lonza Cat# CC-2535
Human vascular smooth muscle cells This study, see STAR Methods N/A
Human acute monocytic leukemia cells (THP-1) Pricella Cat# CL-0233
Chemicals, peptides, and recombinant proteins
Hoechst 33342 Thermo Fisher Scientific Cat# H3570; RRID: N/A
VECTASHIELD Antifade Mounting Medium Neobioscience Cat# H-1000
Calcein AM Invitrogen Cat# C3099
EBM®-2 Basal Medium Lonza Cat# CC-3156
EGM®-2 Endothelial SingleQuots® Kit Lonza Cat# CC-4176
Fetal bovine serum GEMINI Cat# 900-108
DMEM/High Glucose HyClone Cat# SH30243.01
Accumax Millipore Cat# SCR006
Accutase Sigma Cat# A6964
Penicillin/streptomycin (P/S) GIBCO Cat# 15140-163
TrypLE Express Enzyme (1×), no phenol red GIBCO Cat# 12563029
HiScript III RT SuperMix for qPCR (+gDNA wiper) Vazyme Cat# R323-01
NovoStart®SYBR qPCR SuperMix Plus Novoprotein Cat# E096-01B
4% paraformaldehyde, PFA Dingguo, China Cat# AR-0211
Triton X-100 Sigma-Aldrich Cat# T9284
Trypsin Promega Cat# V5113
Protease Inhibitor Cocktail Merck Millipore Cat# P8340
Tetraethylammonium bromide Sigma Cat# TX0255
Trichloroacetic acid Sigma Cat# 822342
DL-Dithiothreitol Sigma Cat# LEYH9ACF09F3
Iodoacetamide Sigma Cat# I1149
Protease inhibitor cocktail Roche Cat# 4693159001
TRIzol™ Reagent Thermo Fisher Scientific Cat# 15596018
NP-40 Dingguo, China Cat# DH218
pH 6.0 sodium citrate buffer ZSGB-BIO Cat# ZLI-9064
Normal Donkey Serum Jackson ImmunoResearch Cat# 017-000-121; RRID: AB_2337258
X-Gal Amresco Cat# 0428-25G
(Continued on next page)

e2 Cell 188, 1–22.e1–e12, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
PVDF membranes Millipore Cat# IPVH00010
GAS6 recombinant protein MCE Cat# HY-P70780
SLPI recombinant protein MCE Cat# HY-P73414
HTRA1 recombinant protein MCE Cat# HY-P701005
COMP recombinant protein MCE Cat# HY-P72945
IGFBP7 recombinant protein MCE Cat# HY-P70726
NOTCH3 recombinant protein MCE Cat# HY-P78009
GPNMB recombinant protein MCE Cat# HY-P75951
NEGR1 recombinant protein MCE Cat# HY-P76508
SAP recombinant protein MCE Cat# HY-P71067
Recombinant mouse GPNMB protein MCE Cat# HY-P77116
Recombinant mouse GAS6 protein MCE Cat# HY-P75174A
Critical commercial assays
Top 14 Abundant Protein Thermo Fisher Scientific Cat# A36371
Depletion Spin Columns Kit
BCA Protein Assay Kit Dingguo, China Cat# BCA02
DAB Staining Kit ZSGB-BIO Cat# ZLI-9018; PV-9001; PV-9002
Human HTRA1 ELISA Kit Elabscience Cat# E-EL-H6300
Human CHI3L1 ELISA Kit CUSABIO Cat# CSB-E13608h
Human NOTCH3 ELISA Kit CUSABIO Cat# CSB-EL015952HU
Human EFEMP1 ELISA Kit CUSABIO Cat# CSB-EL007450HU
Human LTBP2 ELISA Kit CUSABIO Cat# CSB-E11268h
Human ITLN1 ELISA Kit Elabscience Cat# E-EL-H2028
Human COMP ELISA Kit Elabscience Cat# E-EL-H0654
Human IL-6 ELISA Kit BioLegend Cat# 430515
Human GAS6 ELISA Kit Elabscience Cat# E-EL-H6014
Human HARS ELISA Kit CUSABIO Cat# CSB-EL010137HU
Deposited data
The raw proteomic data This study IPX0010296000
The raw bulk RNA-seq data This study GSA: HRA009355, HRA011245
Oligonucleotides
Primers for qRT-PCR, see Table S5 This paper N/A
Software and algorithms
GraphPad Prism (version 8.0.2) GraphPad Software Inc. https://www.graphpad.com/
Image J Schneider et al.130 https://imagej.nih.gov/ij/
DIA-NN (version 1.8) Demichev et al.131 https://github.com/vdemichev/DiaNN
DEP2 (version 0.5.4.02) Feng et al.132 https://github.com/mildpiggy/DEP2
RIdeogram (version 0.2.2) Hao et al.133 https://cran.r-project.org/web/
packages/RIdeogram/vignettes/
RIdeogram.html
limma (version 3.58.1) Ritchie et al.134 https://bioconductor.org/packages/
release/bioc/html/limma.html
ggplot2 (version 3.4.2) Wickham et al.135 https://ggplot2.tidyverse.org/
ComplexHeatmap (version 2.18.0) Gu136 https://github.com/jokergoo/ComplexHeatmap
Cytoscape (version 3.10.2) Shannon et al.137 https://cytoscape.org/
DEswan Benoit et al.138 https://github.com/lehallib/DEswan
Trim Galore (version 0.6.10) N/A https://github.com/FelixKrueger/TrimGalore
STAR (version 2.7.11a) Dobin et al.139 https://github.com/alexdobin/STAR
(Continued on next page)

Cell 188, 1–22.e1–e12, October 2, 2025 e3

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
140
featureCounts (version 2.0.6) Liao et al. https://subread.sourceforge.net/
featureCounts.html
DESeq2 (version 1.44.0) Love et al.141 https://github.com/mikelove/DESeq2
pheatmap (version 1.0.12) N/A https://cran.r-project.org/web/packages/
pheatmap/index.html
clusterProfiler (version 4.10.0) Wu et al.142 https://bioconductor.org/packages/
release/bioc/html/clusterProfiler.html
GSVA (version 1.50.0) Hänzelmann et al.143 https://bioconductor.org/packages/
devel/bioc/vignettes/GSVA/inst/
doc/GSVA.html
glmnet (version 4.1_8) N/A https://doi.org/10.18637/jss.v033.i01
mlr3verse (version 0.2.8) N/A https://github.com/mlr-org/mlr3verse
iml (version 0.11.1) N/A https://github.com/giuseppec/iml
CellPhoneDB (version 5.0.1) Garcia-Alonso et al.95 https://github.com/ventolab/CellphoneDB
Tidygraph (version 1.2.3) N/A https://github.com/thomasp85/tidygraph
Scikit-learn (version 1.3.0) Pedregosa et al.144 https://github.com/scikit-learn/scikit-learn

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Ethical statements
This study adheres to the guiding principles of the Declaration of Helsinki and is conducted in accordance with ethical princi-
ples.145,146 Prior approval was obtained for the use of human genetic resources from the Ethics Committee of West China Hospital,
Sichuan University (NO. 2023-801), Quzhou People’s Hospital (2020-12-001), Beijing Institute of Genomics (China National Center for
Bioinformation) of the Chinese Academy of Sciences (NO. 2023H045) and Beijing Anzhen hospital (NO. 2021-28). All mouse exper-
imental procedures in this study were conducted in accordance with the guidelines set by the Animal Care and Use Institutional Com-
mittee of the Institute of Zoology, Chinese Academy of Sciences (IOZ-IACUC-2025-215).

Human samples
A total of 76 organ transplant donors (aged 14-68 years, 57 males and 19 females) who suffered accidental traumatic brain injury were
included. All samples in this study were obtained from individuals of Chinese ethnicity. Human tissue donors were not subjected to
any experimental procedures prior to donation. All samples were collected postmortem following the standard organ procurement
procedures. Prior to sample collection, detailed explanations regarding the purpose and processing procedures of the samples were
provided to the donors’ families, and written informed consent was obtained. The plasma samples for the independent validation of
ELISA experiments were sourced from a separate cohort in Quzhou.24
Sample size was determined based on data availability and constraints of human tissue acquisition. Given the exploratory nature of
multi-organ human proteomic profiling, formal statistical power calculations were not feasible. Nonetheless, we included biological
replicates for each tissue and age group to ensure representation across decades for trajectory analyses. All human tissue and
plasma samples were grouped based on chronological age according to donor metadata. Detailed sample information was summa-
rized in Table S1.

Experimental mouse models

Male wild-type C57BL/6J mice in this study were procured from Biotechnology Co., Ltd., Beijing, China. All mice were housed under
SPF barrier conditions. Environmental parameters were strictly maintained at a temperature of 20-25◦ C, relative humidity of 30-70%,
and a 12-hour light/dark cycle. The mice were randomly divided into the control group and the experimental group for subsequent
analysis.

Cell culture
Human aortic endothelial cells (hAECs) were purchased from Lonza (Lonza, CC-2535), and cultured in accordance with recommen-
ded standard endothelial medium, containing EBM®-2 Basal Medium (Lonza, CC-3156) and EGM®-2 MV Microvascular Endothelial
Cell Growth Medium SingleQuots® supplements (FBS, Hydrocortisone, hFGF-B, VEGF, R3-IGF-1, Ascorbic Acid, hEGF, GA-1000
and Heparin) (Lonza, CC-4176). Endothelial cells are persistently cultured in Advanced BioMatrix-precoated collagen dishes, with the
fresh culture medium replaced every 48 hours to maintain an optimal environment for cell growth. Digestion of cells is performed us-
ing Accumax (Millipore), and passaging is conducted when the cell confluence reaches 80%.147

e4 Cell 188, 1–22.e1–e12, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Human vascular smooth muscle cells (hVSMCs) were isolated as follows.148 Tunica intima was scratched from the aorta, and
tunica media was reserved for hVSMCs isolation. The tunica media was diced into 0.5 mm-sized fragments, supplemented with
0.25% trypsin in a 37◦ C water-bath for 3.5 hours. This enzymatic digestion process was ceased by adding an equal volume of
DMEM medium. The mixture was passed through a 120 μm sterilized nylon mesh filter (Millipore) to remove undigested debris.
The strained solution was centrifuged at 500 rpm for five minutes, after which the pellet was reconstituted and cultured in recommen-
ded culturing medium DMEM medium (DMEM, HyClone) supplemented with 10% fetal bovine serum (FBS, GEMINI), 2 mM
GlutaMAXTM (Gibco), 0.1 mM MEM non-essential amino acids (NEAA, Gibco) and 1% penicillin/streptomycin (Gibco). Fresh culture
medium was replaced every 48 hours to maintain an optimal environment for cell growth. Cell passaging was conducted when the
cell confluence reached 80%.
The THP-1 cells were purchased from Pricella (Pricella) and cultured in DMEM/High Glucose (HyClone) medium supplemented
with 10% FBS (GEMINI), 1% penicillin/streptomycin (Gibco), 0.1 mM NEAA (Gibco) and 2 mM GlutMaxTM (Gibco). The cells were
centrifuged every 48 hours to remove the spent medium and replace it with fresh medium to maintain the optimal growth
environment.

METHOD DETAILS

Human samples collection

All tissue samples were collected by certified medical professionals immediately following surgical resection. Samples were either
flash-frozen in liquid nitrogen for proteomic analysis or fixed in 4% paraformaldehyde for subsequent dehydration and embedding
procedures. Blood samples were processed within 30 minutes of collection, and plasma was aliquoted and stored at -80◦ C to mini-
mize freeze-thaw cycles.

Cell treatment with recombinant protein

Experimental analyses of hAECs were conducted at the fifth generation (considered early passage). hAECs were treated with COMP
(200 ng/ml), NOTCH3 (30 ng/ml), HTRA1 (100 ng/ml), IGFBP7 (100 ng/ml), GPNMB (100 ng/ml), NEGR1 (100 ng/ml), SLPI (100 ng/ml),
GAS6 (30 ng/ml), SAP (100 ng/ml) or vehicle (10% FBS in PBS, refer to instruction manual).
Experimental analyses of hVSMCs were conducted at the sixth generation (considered early passage). hVSMCs were treated with
GAS6 (30 ng/ml), HTRA1 (100 ng/ml), IGFBP7 (100 ng/ml), GPNMB (100 ng/ml) or vehicle (10% FBS in PBS, refer to instruction
manual).

Tissue protein extraction and mass spectrometry analysis

Tissue protein extraction and mass spectrometry were performed with reference to published articles, and were simply performed as
follows.58,149,150 Technical support for the above work was provided by PTM BioLab.
Protein extraction
Protein extraction protocols were optimized for various tissue types. For plasma, samples were centrifuged at 12,000 g for 10 minutes
at 4◦ C to remove cellular debris. High-abundance proteins were depleted using the Top 14 Abundant Protein Depletion Spin Columns
Kit. For other tissues, samples were ground into a fine powder under liquid nitrogen. For adrenal glands, lung, liver, muscle, heart and
aorta, the powdered tissue was resuspended in four volumes of lysis buffer (1% Triton X-100 containing 1% protease inhibitor cock-
tail), followed by sonication for cell disruption. Centrifugation at 12,000 × g for 10 minutes at 4◦ C removed debris, and the supernatant
was collected. For skin and adipose tissues, the powdered tissue was resuspended in lysis buffer (1% Triton X-100 containing 1%
protease inhibitor cocktail). After sonication, an equal volume of Tris-buffered phenol was added, and the mixture was centrifuged at
5,500 × g for 10 minutes at 4◦ C. The supernatant was mixed with five volumes of 0.1 M ammonium acetate in methanol, followed by
protein precipitation overnight and the pellet was washed with methanol and acetone, then solubilized in 8 M urea. For lymph, intes-
tine, pancreas, and spleen, the powdered tissue was resuspended in five volumes of 10% trichloroacetic acid (TCA) in acetone and
incubated at -20◦ C for 4 hours. Then the samples were centrifuged at 4,500 × g for 5 minutes at 4◦ C. The supernatant was discarded,
and the pellet was washed three times with ice-cold acetone and subsequently dried under vacuum. The dried pellet was solubilized
in 8 M urea lysis buffer (for pancreas and spleen) or 1% SDS (for lymph) containing 1% protease inhibitor cocktail. After heating at
95◦ C for 20 minutes (only for pancreas), the lysate was centrifuged at 12,000 × g for 10 minutes, and the supernatant was collected.
All protein concentration was determined using a BCA assay kit.
Trypsin digestion
Equal protein aliquots from each sample were enzymatically digested as follows: Sample volumes were adjusted to uniformity using
lysis buffer. For heart, adrenal, liver, aorta, lung, lymph, muscle and intestine, TCA was added dropwise to a final concentration of
20% with vortex mixing, followed by precipitation at 4◦ C for 2 hours. For pancreas, skin and adipose, one volume of pre-chilled
acetone was added and mixed by vortexing, followed by addition of four volumes of pre-chilled acetone. The mixture was precip-
itated at -20◦ C for 2 hours. Then samples were centrifuged at 4,500 × g for 5 minutes. The supernatant was discarded, and the pellet
was washed 2-3 times with ice-cold acetone. After air-drying, the pellet was resuspended in 200 mM triethylammonium bicarbonate
(TEAB) and dispersed by ultrasonication. Trypsin was added at a 1:50 (w/w, enzyme-to-protein) ratio for overnight digestion at 37◦ C.
Dithiothreitol (DTT) was then added to a final concentration of 5 mM and incubated at 56◦ C for 30 minutes for disulfide bond

Cell 188, 1–22.e1–e12, October 2, 2025 e5

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

reduction. Subsequently, iodoacetamide (IAA) was added to 11 mM final concentration and incubated at room temperature in the
dark for 15 minutes for alkylation.
LC-MS/MS Analysis
Peptides were dissolved in mobile phase A and separated using the Vanquish Neo ultra-high performance liquid chromatography
(UHPLC) system. Mobile phase A consisted of an aqueous solution containing 0.1% formic acid, while mobile phase B was an
aqueous solution containing 0.1% formic acid and 80% acetonitrile. The liquid chromatography gradient for plasma was set as fol-
lows: 0-1.6 minutes, 4%-22.5% B; 1.6-2.0 minutes, 22.5%-35% B; 2.0-2.6 minutes, 35%-55% B; 2.6-2.7 minutes, 55%-99% B;
2.7-6.8 minutes, 99% B; 6.8-7.6 minutes, 99% B, with a flow rate maintained at 300 nL/minutes. For tissues other than plasma,
the gradient was as follows: 0-0.5 minutes, 4% B; 0.5-0.6 minutes, 4%-8% B; 0.6-13.6 minutes, 8%-22.5% B; 13.6-20.5 minutes,
22.5%-35% B; 20.5-20.9 minutes, 35%-55% B; 20.9-21.4 minutes, 55%-99% B; 21.4-22.6 minutes, 99% B, with a flow rate main-
tained at 400 nL/minutes. After separation by the UHPLC system, the peptides were introduced into the NSI ion source for ionization,
followed by analysis using the Orbitrap Astral mass spectrometer. The ion source voltage was set at 1,900 V, with parent peptide ions
detected and analyzed by the Orbitrap detector, and the secondary fragment ions detected by the Astral detector. The primary mass
spectrometry scan range was set from 380 to 980 m/z, with a scanning resolution of 240,000; the secondary mass spectrometry scan
range was fixed at a starting point of 150 m/z, with a scanning resolution of 80,000. Data acquisition was conducted in data-inde-
pendent acquisition (DIA) mode, where multiple continuous m/z windows followed the primary scan, and peptide ions were intro-
duced into the higher-energy collisional dissociation (HCD) collision pool with 25% fragmentation energy for collision-induced disso-
ciation, followed by secondary mass spectrometry analysis. To improve the mass spectrometry efficiency, the automatic gain control
(AGC) was set to 500%, and the maximum injection time was set to 3 milliseconds.
Database search
The raw data from the mass spectrometer were imported into the search software, and the DIA-NN (version 1.8) search engine was
used with default parameters. The database used was Homo_sapiens_9606_SP_20230103.fasta (20,389 sequences), with ‘‘trypsin/
P’’ as the cleavage enzyme and a maximum missed cleavage set to 1. Fixed modifications were set to: ‘‘N-term M excision’’ and ‘‘C
carbamidomethylation’’. A theoretical spectral library was constructed using deep learning algorithms, and a decoy database was
included to calculate the false discovery rate (FDR) caused by random matches. The FDR for precursor identification was set at
1%. To obtain high-quality results, the search results were further filtered with the criteria set as follows: precursor and protein
FDR were both set to 1%; proteins must be identified with at least one peptide.

ELISA assay
The ELISA experiment was performed following the standard protocol of the reference kit. To put it succinctly, plasma samples with
EDTA as an anticoagulant were gently thawed on ice and centrifuged. The resulting supernatant was then diluted according to the
recommended ratio and, along with the standards, was added to an ELISA plate that has been pre-coated with antibodies or antigens
for incubation. A sequence of reagents including biotinylated antibodies, horseradish peroxidase-conjugated streptavidin, TMB sub-
strate solution, and stopping solution was added, with the reaction proceeding for an appropriate time. Absorbance was measured at
450 nm, and the standard curve was constructed using the recommended professional software Curve Expert to determine the sam-
ple concentration. The statistical analysis was conducted with GraphPad (version 8.0.2) software. The plasma samples used for
ELISA are detailed in Table S1.

Immunofluorescence and immunohistochemical staining

Immunofluorescence staining of human multi-tissue arrays was conducted as previously described methods.77,151,152 In general, the
paraffin-embedded human multi-tissue arrays were deparaffinized in xylene and rehydrated through concentration gradient ethanol.
Then, tissue arrays were boiled in 10 mM sodium citrate antigen retrieval buffer (pH 6.0) for 15 minutes. Upon cooling down to RT,
tissue arrays were permeabilized with 0.4% Triton X-100 in PBS for 40 minutes followed by blocking with 10% donkey serum for 1
hour at room temperature. Tissue arrays were then incubated with primary antibodies overnight at 4◦ C and fluorescence-labeled sec-
ondary antibodies at room temperature for 1 hour. Images were captured using the PerkinElmer Vectra Polaris automated quantita-
tive pathology imaging system. Detailed information of the antibodies and reagents is listed in key resources table.
For immunohistochemical staining, paraffin-embedded tissue arrays were deparaffinized and rehydrated, followed by the process
of antigen retrieval and permeabilization described above. Then, 3% H2O2 in methanol solution was used to remove endogenous
peroxidase of human tissues. After blocking with 10% donkey serum for 1 hour, sections were incubated with the appropriate con-
centration of primary antibodies overnight at 4◦ C and corresponding HRP-labeled secondary antibodies for 1 hour at RT. The
following step was DAB-HRP color development and hematoxylin staining. Images were captured using the PerkinElmer Vectra
Polaris automated quantitative pathology imaging system. Detailed information about the antibodies and reagents is listed in key
resources table.

Western blotting
Western blotting assay was performed with reference to published procedures.78,153 For the processing of tissue samples, retrieved
the tissue samples from liquid nitrogen, wrapped them in aluminum foil to prevent splashing, and ground them into a fine powder
using a mortar and pestle. Added RIPA lysis buffer and lysed at 4◦ C for 1 hour. After centrifugation, the supernatant was collected.

e6 Cell 188, 1–22.e1–e12, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

For cell samples, washed the adherent cultured cells with PBS, then added 1×SDS protein lysis buffer and lysed on ice for 30 minutes,
collecting the lysate. Boiled the aforementioned lysate at 95◦ C for 5 minutes to denature the proteins. Quantified the proteins using a
BCA assay kit, and separated proteins of different molecular weights by SDS-PAGE gel electrophoresis. Subsequently, the samples
were incubated with specific primary antibodies and horseradish peroxidase (HRP)-labeled secondary antibodies, and collected the
protein luminescent signals using a Bio-Rad system. The protein grayscale was analyzed by using Image J software, and statistical
analysis and presentation were carried out with GraphPad (version 8.0.2) software.

SA-β-Gal staining
SA-β-Gal staining assay was performed with reference to published procedures.77,154 Briefly, after rinsing with PBS, adherent cells
were treated with fixative solution (2% formaldehyde and 0.2% glutaraldehyde solution) and left to stand for 5 minutes at room tem-
perature to fix the cells. Subsequently, X-Gal staining solution was added under a pH 6.0 buffer system and incubated in the dark at
37◦ C for an appropriate time. Representative image capture was performed using the microscope system (Nikon, Y-TV55), and the
proportion of SA-β-Gal positive cells was calculated using Image J software, with statistical analysis conducted using GraphPad
(version 8.0.2) software.

Endothelial cell tube formation assay

The tube formation assay was performed according to the published protocol.147,155 In brief, human aortic endothelial cells treated
with indicated purified proteins for 4 days were digested with Accutase (Sigma), after which cell counting was performed. The cells
were then seeded at a density of 1×105 cells per well into a 12-well plate precoated with Matrigel. After 3 hours culture, the cells were
stained with Calcein AM (Invitrogen) for visualization. Images of the cells were captured using a Zeiss microscope (ZEISS, Axio
Observer 3), The lengths and node numbers of the tubes were calculated using Image J, and the statistical results were presented
and analyzed using GraphPad (version 8.0.2) software.

Vascular endothelial cell migration assay

The endothelial cell migration assay was conducted following the methods we previously published.77 Human aortic endothelial cells
in both the control and experimental groups were seeded at a density of 1×105 cells per well in 12-well plates precoated with collagen
and cultured at 37◦ C. Once the cells were completely confluent, a clear scratch was made using the tip of a 10 μL pipette. The cells
were then gently washed twice with PBS to remove non-adherent cells, preventing them from reattaching to the scratched area. After
washing, the cells were further cultured at 37◦ C using endothelial cell medium containing 0.5% serum. The migration status of the
cells was observed and documented at 0, 4, 8, and 12 hours. Images of the designated areas were captured using an inverted mi-
croscope (Olympus, CKX53) and analyzed using Image J software.

Monocyte adhesion assay

Following a four-day incubation period with designated recombinant proteins, human aortic endothelial cells are subjected to diges-
tion with Accutase (Sigma), followed by enumeration and reseeding into 24-well plates precoated with collagen. The adhesion assay
was conducted once the cellular confluence reached 90%. THP-1 cells were harvested and pre-stained with calcium-AM fluorescent
dye. Resuspended in DMEM medium to a concentration of 1×106 cells per milliliter, 200 microliters of this cell suspension were then
introduced into the 24-well plates that contained the endothelial cells, which were subsequently incubated at 37◦ C for a duration of
30 minutes. To eliminate non-adherent THP-1 cells, cultures were gently washed with PBS. Cell images were captured using a Zeiss
microscope (Axio Observer 3). THP-1 cell numbers were quantified using Image J, and statistical results were analyzed with
GraphPad Prism (version 8.0.2).

RNA isolation and real-time quantitative PCR (RT-qPCR)

Total RNA was extracted from cultured hAECs using TRIzol reagent. After determining the total RNA concentration with a NanoDrop
one C (Thermo Fisher), genomic DNA was removed, and 1 μg of RNA was reverse-transcribed into cDNA. RT-qPCR was conducted
using iTaq Universal SYBR Green SuperMix (Bio-Rad) on a CFX384 Real-Time PCR system (Bio-Rad), with GAPDH included as an
internal control in each reaction. The relative RNA expression was calculated using the ΔΔCq method. The primer sequences used
for RT-qPCR are shown in Table S5.

Recombinant protein treatment of mice

Given that endogenous GAS6 and GPNMB are chronically elevated in aged human tissues and fluids, while exogenously adminis-
tered proteins rapidly decline to low levels in vivo, we administered a dosage approximately three times the reported physiological
concentration, with injections every five days, to better mimic an aging-associated state.156–161 In brief, 8-month-old male C57BL/6
mice, used for protein tail-vein injection, were randomly assigned to experimental groups. These mice received either PBS (as vehicle
control), GAS6 (8 μg/kg) or GPNMB (40 μg/kg) every 5 days. The injections were administered four times, and tissues were collected
for subsequent analyses.

Cell 188, 1–22.e1–e12, October 2, 2025 e7

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Mouse behavioral assays

Behavioral tests for mice were performed as in previous studies.162
Grip strength test
The mouse grip strength was detected using a grip strength meter (Panlab Grid Strength meter, LE902). Briefly, the mouse was
placed on the grid and pulled at a constant speed until it released the grid. This procedure was repeated six times, the monitor values
were recorded, and the average of the recorded values was calculated.
Inverted screen test
Position the mice at the epicenter of the gridded lattice screen, initiate the chronometer, and swiftly invert the grid within a two-sec-
ond interval, ensuring the rodent’s orientation is inverted. Subsequently, the lattice was securely positioned 20–30 centimeters above
the substrate, suspending the mice in the meshed grid as they navigated, with the descent duration meticulously documented.
Balanced beam test
A square wooden beam was suspended 40 cm above the cage floor, with an enclosed dark chamber positioned at one terminus. A
distance of 90 cm was marked along the beam, and the mice were tested by timing the traversal from the start point to the finish. A
stopwatch was used to measure the time, meanwhile the foot slips of mice were recorded. Repeat the test 3 times, record each time
used and foot slips and take the average of the values.

Hematoxylin and eosin (H&E) staining

For H&E staining, sections were deparaffinized and rehydrated using xylene and a gradient concentration of ethanol. The sections
were then immersed in hematoxylin for 5 minutes, excess staining solution was removed by rinsing in running water, and differen-
tiated for 3 seconds in 1% hydrochloric acid (in 70% alcohol). Sections were washed with running water for 10 minutes, stained
with eosin solution for 2 minutes, and rinsed in running tap water. Finally, sections were dehydrated and sealed with neutral gum.
Images were captured using a Leica CS2 slice scanner, and Image J software was used to quantify the inflammation areas of mouse
livers and spleens.

Oil red O staining

Frozen tissue sections (30 μm thick) were air-dried for 15 minutes and washed three times with PBS. Subsequently, the sections were
fixed with 4% PFA for 10 minutes and washed three more times with PBS. After exposure to 60% isopropanol for 5 seconds, the
sections were stained with 60% Oil Red O for 15 minutes. The sections were then washed three times with PBS. Hematoxylin staining
was performed for 5 minutes, followed by sequential immersion in 1% hydrochloric acid (in 70% alcohol) and 1% ammonia water.
Finally, the sections were sealed with 50% glycerol.

BODIPY staining
Frozen tissue sections (10 μm thick) were air-dried for 15 minutes, washed three times with PBS, fixed with 4% PFA for 10 minutes,
permeabilized with 0.4% Triton X-100 for 30 minutes, and incubated with primary antibodies overnight. During secondary antibody
incubation, 10 μg/mL BODIPY was added and incubated at room temperature for 1 hour. The tissues were stained with Hoechst
33342 and mounted using anti-fade mounting medium. Images were acquired using the Zeiss LSM 980 confocal microscope system.

RNA-seq library construction and data processing

Total RNA was extracted from cells or approximately 0.2 g of tissue using TRIzol, following the manufacturer’s protocol. The RNA
concentration was measured using a fragment analyzer (Advanced Analytical). Around 1-2 μg of total RNA was used to construct
a strand-specific transcriptome library, which was sequenced on the DNBSEQ-T7 platform.

Quality control of proteomics data

The raw protein abundance matrix, containing data from 13 tissues and 516 samples, was processed using the DEP2 package
(version 0.5.4.02).132 Proteins were considered detected only if they were identified in at least 50% of the samples from either females
or males within each tissue type. After data filtering, the abundance of each protein was Log2-transformed to better approximate a
normal distribution. Normalization was performed using the ‘‘normalize_vsn’’ function, which transformed the data to stabilize the
variance across the dynamic range of measurements, enhancing the comparability of samples across batches. Missing values for
proteins were imputed using Quantile Regression Imputation of Left-Censored data (QRILC) with the ‘‘impute’’ function and the
parameter fun = ‘‘QRILC’’, which is specifically suited for left-censored data distributions.163,164 Detailed information about the num-
ber of proteins identified in each tissue is listed in Table S1.

Identification of high and low abundance proteins

The identification of high and low abundance proteins was similar to that described in previous work.57 Briefly, the high abundance
proteins were defined as the top 10% high abundance levels across all tissues or in each tissue. And the low abundance proteins
were defined as the top 10% low abundance levels across all tissues or in each tissue.

e8 Cell 188, 1–22.e1–e12, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Classification of proteins based on chromosome location of protein-coding genes

The gene annotation file (GRCh38) containing chromosome location information was archived from GENCODE (https://www.
gencodegenes.org/).165 The distribution of chromosome locations was calculated and visualized using the RIdeogram (version
0.2.2) package.133

Classification of proteins based on subcellular location, cellular component, and molecular function
The subcellular location information of each protein was derived from the predicted protein class obtained from the Human Protein
Atlas (https://www.proteinatlas.org/).166 Gene sets of different cellular components and molecular functions were obtained from the
Molecular Signatures Database (MSigDB) (https://www.gsea-msigdb.org/gsea/msigdb). The number and proportion of proteins
belonging to different categories across tissues were calculated and visualized using the ggplot2 package (version 3.4.2).

Classification of proteins based on differential expression between tissues

We modified the definition provided by the Human Protein Atlas (https://www.proteinatlas.org/) to classify proteins based on differ-
ential expression between tissues.166 Specifically, proteins were classified as ‘‘tissue-enriched’’ if their expression level in a single
tissue was at least 10-fold higher than the mean expression level across all tissues. ‘‘Tissue-enhanced’’ proteins were defined as
those expressed at least 10-fold higher in several (more than two) tissues compared to the average expression across all tissues.
‘‘Housekeeping’’ proteins were defined as those detected in all tissues that did not fall into either the ‘‘tissue-enriched’’ or ‘‘tis-
sue-enhanced’’ categories. Proteins that did not fall into any of the above categories were classified as ‘‘others’’. Detailed information
about the tissue specificity categories of different proteins is listed in Table S1.

Identification of sex-biased proteins

Only tissues with at least two samples from both female and male groups were selected for sex-biased analysis. Sex-biased proteins
of each tissue were calculated with a cutoff of p values < 0.05 and |log2FC| > 2 between female and male groups using the limma
package (version 3.58.1).134 Age effect was corrected via the ‘‘lmFit’’ function. Sex-biased proteins were also identified in young
(samples aged 14–30) group to estimate the changes of GSVA scores between young (samples aged 14–30) and old (samples
aged 45–68) groups.

Bulk RNA-seq data processing

For the processing of bulk RNA-seq data, we used Trim Galore (version 0.6.10) (https://github.com/FelixKrueger/TrimGalore) to re-
move sequencing adapters and filter out low-quality reads. The cleaned reads were then aligned to the human reference genome
(GRCh38) using STAR (version 2.7.11a).139 Gene expression levels were quantified using featureCounts (version 2.0.6).140 The
read counts were converted into Transcripts Per Million (TPM) using a custom script. Differentially expressed genes (DEGs) were
identified using DESeq2 (version 1.44.0),167 applying a cutoff of Benjamini-Hochberg adjusted p values < 0.05 and |log2FC| > 1.

Decoupling analysis of protein abundance and mRNA expression

First, we selected samples aged 14–30 as young samples and samples aged 45-68 as old samples. We aligned the normalized pro-
tein abundance with mRNA expression levels (TPM) using gene symbol. Only tissues with at least 5 samples in each group were
selected for subsequent analysis.
Next, we calculated the Spearman correlation between protein and mRNA for each protein in different groups. Proteins with p
values < 0.05 were selected for further analysis. Proteins with R > 0 were defined as positive correlation proteins, and those with
a correlation R < 0 were defined as negative correlation proteins.
Finally, we defined proteins that were positive correlation in young samples but showed a decrease in correlation during aging as
decoupling proteins. The change in correlation was calculated by Rold − Ryoung . GO (Gene Ontology) enrichment analysis for decou-
pling proteins was performed using the clusterProfiler package (version 4.10.0).142 Detailed information about the decoupling pro-
teins is listed in Table S2.

Translational noise analysis

We divided samples aged 14-30 into young group and samples aged 45-68 into old group. We independently classified the ‘‘tissue-
enriched’’ proteins as described in the previous section, in the young group for each tissue. We then calculated the standard devi-
ation and coefficient of variation in samples belonging to the same age group using the ‘‘tissue-enriched’’ proteins from the young
group as a measurement of translational noise.

Gene set variation analysis (GSVA)

The GSVA score was calculated using the GSVA package (version 1.50.0) based on normalized protein abundance or gene
counts.143 Gene sets were obtained from the MSigDB (https://www.gsea-msigdb.org/gsea/msigdb), the Kyoto Encyclopedia of
Genes and Genomes (KEGG) (https://www.kegg.jp/kegg/), the EpiFactors database (https://epifactors.autosome.org/) and the
CORUM database (https://mips.helmholtz-muenchen.de/corum/).168,169

Cell 188, 1–22.e1–e12, October 2, 2025 e9

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Identification of age-dependent differentially expressed proteins (DEPs)

Correlation analysis was performed using chronological age and Log2-transformed protein abundance via Spearman method in each
tissue. Upregulated DEPs were identified with a cutoff of p values < 0.05 and Spearman’s rank correlation coefficient > 0. Downre-
gulated DEPs were identified with a cutoff of p values < 0.05 and Spearman’s rank correlation coefficient < 0. DEPs exhibiting uniform
upregulation across more than six tissues without conflicting trends in any organ were designated as ubiquitous upregulated
proteins with aging (UUPA), while DEPs exhibiting uniform downregulation across more than six tissues without conflicting trends
in any organ were designated as ubiquitous downregulated proteins with aging (UDPA). Detailed information about the DEPs is listed
in Table S2.

Module analysis of DEPs

We clustered the Spearman’s rank correlation coefficients of upregulated or downregulated DEPs across 13 tissues into 10 modules
using the ‘‘pheatmap’’ function from the pheatmap package (version 1.0.12) with the parameter kmeans_k = 10.

Functional enrichment analysis

GO enrichment analysis was performed using the ‘‘compareCluster’’ function of clusterProfiler package (version 4.10.0) with a cutoff
of p values < 0.05. Gene set enrichment analysis (GSEA) was performed using the ‘‘gseGO’’ and ‘‘gseKEGG’’ functions of cluster-
Profiler package (version 4.10.0).142 The results were further visualized with the ggplot2 (version 3.4.2) and enrichplot (version
1.24.0) packages.

Protein complex annotations

To define the set of protein complexes for downstream analysis, we used annotations from the CORUM database (https://mips.
helmholtz-muenchen.de/corum/),169 a comprehensive resource of mammalian protein complexes. We excluded protein complexes
in which fewer than half of the subunits were detected in the corresponding tissues. For protein complexes consisting of only two
subunits, both subunits were required to be observed in corresponding tissues. We retained protein complexes that belong to
one of the functional complex groups, which resulted in 816 protein complexes across the 13 tissues.

Identification of age-related protein complexes

We calculated the average abundance of all subunits in a protein complex as the abundance of that protein complex. We performed
correlation analysis using chronological age and Log2-transformed complex abundance via Spearman method in each tissue. Up-
regulated complexes were identified with a cutoff of p values < 0.05 and Spearman’s rank correlation coefficient > 0. Downregulated
complexes were identified with a cutoff of p values < 0.05 and Spearman’s rank correlation coefficient < 0.

Disease-associated pathway analysis

GSEA was performed using disease-related gene sets from the DisGeNET database,81 implemented via the GSEA function from the
clusterProfiler package (version 4.10.0).142 The input was a ranked protein expression matrix based on Spearman correlation with
age. Parameters were set as minGSSize=3, maxGSSize=2000, eps=1e-100, and pvalueCutoff=0.05. Proteins associated with
age-related diseases were further visualized as a protein-disease network in Cytoscape (version 3.10.2).137

Proteomic aging clock analysis

The methodology for constructing the proteomic clock is adapted from previously published studies.24,27 To construct the aging
clock, we utilized normalized protein abundance data from various tissues. Initially, we selected DEPs (see Methods Identification
of DEPs) specific to each tissue as input data and used the chronological age of each sample as the training label. A multivariate linear
regression model was employed using the glmnet package (version 4.1_8) (https://doi.org/10.18637/jss.v033.i01) in R (version 4.3.3),
with elastic net algorithm for model training. Hyperparameter selection and training were conducted using random search via the
mlr3verse package (version 0.2.8) (https://github.com/mlr-org/mlr3verse) in R (version 4.3.3). Leave-one-out cross-validation was
applied by removing one sample at a time from the dataset to predict its biological age. The final model was chosen based on
the lowest mean absolute error (MAE). Important proteins within the model were identified using the ‘‘FeatureImp’’ function from
the iml package (version 0.11.1) (https://github.com/giuseppec/iml) in R (version 4.3.3), which is based on permutation importance.
Detailed information about the clock importance proteins is listed in Table S3.

Age-group-specific DEPs analysis

We identified age-group-specific differentially expressed proteins (age-group-specific DEPs) for three age stages: early (14-35
years), middle (14-55 years), and late (14-68 years). This identification employed the same methodological approach detailed in
the ‘‘Identification of age-dependent differentially expressed proteins (DEPs)’’ section. To distinguish them from the age-dependent
DEPs, we designated the latter as lifespan-wide DEPs. To quantify the cumulative occurrence of lifespan-wide DEPs within each age
stage, we calculated the overlap between the age-group-specific DEPs and the lifespan-wide DEPs for each respective stage.
To assess the commonality and uniqueness of lifespan-wide DEPs appearing across different age groups, we performed pairwise
comparisons between the overlap results from each stage. Finally, we performed enrichment analysis on the overlap DEPs between

e10 Cell 188, 1–22.e1–e12, October 2, 2025

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

age-group-specific DEPs and lifespan-wide DEPs. This analysis utilized the clusterProfiler R package (version 4.10.0) within R soft-
ware (version 4.3.3).142

Sliding window analysis of the senescence proteome

We performed sliding window analysis using the DEswan package (https://www.biorxiv.org/content/10.1101/751115v1) in R (version
4.3.3). Aging-related DEPs were used as input, with a window size of 20 years and step size of 5 years. Proteins exhibiting p
values < 0.05 at a given time point were defined as time-resolved DEPs at that specific time point.24 To analyze tissue-specific dy-
namic during aging, we normalized the time-resolved DEPs by the number of detected proteins in each tissue. This generated tissue-
specific temporal change profiles. To identify tissues undergoing the most pronounced changes during aging, we calculated the cu-
mulative count of unique time-resolved DEPs per tissue across all ages (counting each protein only once regardless of recurrence).
To determine the peak of proteomic during aging for each tissue, we normalized the tissue-specific time-resolved DEP counts by
their maximum value across all time points. Finally, we performed functional enrichment analysis on significantly changed proteins
(time-resolved DEPs) using the clusterProfiler R package (version 4.10.0) within R (version 4.3.3),142 aiming to identify pathway acti-
vation dynamics. It should be noted that due to the relatively small sample size in this study, future validation in larger, more demo-
graphically balanced cohorts is warranted to confirm the robustness of these findings.
Considering the limited sample size in this study, further validation in larger and more representative cohorts is necessary to ensure
the reliability and broader applicability of these results.

Analysis of secretory proteins

Based on the subcellular location information, we extracted the ‘‘Secreted & Membrane & Intracellular’’, ‘‘Secreted & Membrane’’,
‘‘Secreted & Intracellular’’, and ‘‘Secreted only’’ for further analysis of secreted proteins. The categories of secreted proteins were
archived from the Human Protein Atlas (https://www.proteinatlas.org/),166 which can be categorized into 4 major groups and 21 sub-
groups. Detailed information about the secreted proteins is listed in Table S4.

Module analysis of secretory proteins

Correlation analysis was performed for all secreted proteins using chronological age and Log2-transformed protein abundance via
Spearman method in each tissue. We clustered and classified the Spearman’s rank correlation coefficients of all secreted proteins
into 5 modules using the ‘‘Heatmap’’ function of ComplexHeatmap package (version 2.18.0) with parameter ‘‘row_km= 5’’.136

Senescence-associated secretory phenotype (SASP) analysis

We defined all upregulated secreted proteins as our in-house generated SASP factors. We also applied joint analysis using an inte-
grated gene list of previous study and Reactome database (R-HSA-2559582: Senescence-Associated Secretory Phenotype).42,93
The joint analysis of SASP network was visualized in Cytoscape (version 3.10.2).137

Tissue-tissue interaction analysis

First, we selected samples aged 14–30 years as young group and those aged 45–68 years as old group. Then, treating each sample
as a cell, we converted the proteome expression matrix into the input format required by CellPhoneDB (version 5.0.1) to calculate
interaction profiles for each group.95 Ligand-receptor pairs with p values < 0.001 were identified as significantly interacting pairs. Us-
ing the tidygraph package (version 1.2.3) (https://github.com/thomasp85/tidygraph) in R (version 4.3.3), we visualized the number of
ligand-receptor interactions between tissues within each age group. To identify differential interactions, we specifically screened for
the interactions with difference mean more than 0.5. Detailed information about the tissue-tissue interaction pairs is listed in Table S4.

Tissue-tissue interactions mediated by circulating proteins

Ligands secreted from tissues into plasma
To identify ligands secreted from tissues into the plasma, we filtered DEPs using the list of secreted proteins provided by
CellPhoneDB (version 5.0.1),95 retaining only those DEPs that showed concordant changes in both plasma and tissues. We hypoth-
esized that the changes in the expression levels of these proteins in the plasma with age are due to secretion from tissues.
Interactions between ligands in plasma and receptors in tissues
Following the methodology for tissue-tissue interaction analysis, we grouped the samples and calculated the interaction profiles for
each group. We identified ligand-receptor pairs with p values < 0.05 as significantly interacting pairs, focusing exclusively on inter-
actions where ligands in the plasma interacted with receptors in other tissues for downstream analysis. To understand the changes in
interactions during aging, we subtracted the expression levels of ligand-receptor pairs in the young group from those in the old group,
obtaining the expression differences of ligand-receptor pairs during aging. Expression levels involving non-significant interactions
were set to zero.
Tissue-plasma-tissue interactions
To confirm the changes in tissue interactions mediated by plasma during aging, we overlapped the ligands obtained from the first
step with the ligand-receptor pairs from the second step, retaining only those ligand-receptor pairs that showed consistent changes
in direction across the two steps. Detailed information about the plasma-mediated tissue-tissue interaction pairs is listed in Table S4.

Cell 188, 1–22.e1–e12, October 2, 2025 e11

 Please cite this article in press as: Ding et al., Comprehensive human proteome profiles across a 50-year lifespan reveal aging trajectories and
signatures, Cell (2025), https://doi.org/10.1016/j.cell.2025.06.047

ll
Article

Disease enrichment analysis of tissue-tissue interacted protein pairs

GSEA was performed using disease-related gene sets from the DisGeNET database, implemented with the GSEA function from the
clusterProfiler package. For each tissue pair, all the interacted protein pair were combined and enriched. To ensure biological rele-
vance, the diseases with p values < 0.05 were selected, and only protein pairs where both ligand and receptor proteins showed co-
enrichment within the same disease pathway were retained for downstream analysis.

Plasma protein proxy aging clock analysis

Data processing
First, individuals with both tissue and plasma data are selected, and proteins common to both tissue and plasma are chosen. After
normalized using min-max scaling separately, the processed tissue and plasma data are merged to form the training dataset.
Model training
The merged training dataset is used to train an elastic net regression model, with tissue age (calculated using the tissue aging clock
developed in this study) as the regression target. The model is implemented using the ElasticNet module from the Python Scikit-learn
library (version 1.3.0).144 Leave-one-out cross-validation is employed to reduce overfitting. Recursive feature elimination is applied,
where in each iteration, the least important feature (based on absolute importance) is removed from the clock. As the number of fea-
tures decreases, the model performance (measured by MAE) initially increases and then decreases. The optimal number of features,
corresponding to the best model performance, is selected. After hyperparameter optimization, the final model is trained, and the fea-
tures within the model are considered the optimal combination for simulating tissue aging in plasma.
Model performance evaluation
Evaluation. After model training, plasma proteomics data are input into the specific tissue’s plasma protein proxy clock. The pre-
dicted age obtained through leave-one-out cross-validation is compared with the tissue age for that specific tissue. The mean ab-
solute errors, Spearman’s correlation coefficients, and p values are then calculated to assess model performance.
Given the relatively small sample size, further validation in larger and more diverse datasets will be required to confirm the robust-
ness and generalizability of these findings.

QUANTIFICATION AND STATISTICAL ANALYSIS

All statistical analyses and data handling were performed using GraphPad Prism software (version 8.0.2). Data are presented
as mean ± standard error of the mean (SEM). Statistical differences between two groups were assessed using two-tailed t-tests
(passed normality test) or Mann-Whitney test (not passed normality test), whereas those among more than two groups were as-
sessed using one-way analysis of variance (ANOVA). In histological staining analysis, fold-change data were calculated relative to
the mean of all samples, and Spearman’s rank correlation test was used to calculate the correlation coefficient and p values of
age-related data. In plasma ELISA experiments, Pearson correlation analysis was used to calculate the correlation coefficient and
p values of age-related data. Significant differences between groups are indicated either by p values directly in the figure or by as-
terisks: * p value < 0.05, ** p value < 0.01, *** p value < 0.001.

e12 Cell 188, 1–22.e1–e12, October 2, 2025

 ll
Article

Supplemental figures

(legend on next page)

 ll
Article

Figure S1. Gender, tissue bias, and functional attributes of identified proteins, related to Figure 1
(A) Histograms of protein abundance distribution in each tissue. Red: before quantile regression imputation of left-censored data (QRILC) imputation. Blue: after
QRILC imputation.
(B) UMAP plot showing the distribution of samples across different tissues.
(C) Heatmap of Spearman’s correlation coefficients between protein expression levels in each tissue. Color key: white to red (low to high correlation).
(D) Left: pie charts of protein proportions in different categories across tissues. Right: bar plot of ‘‘tissue-enriched’’ proteins.
(E) Left: heatmap of tissue-enriched protein expression across tissues. Right: representative GO terms enriched in tissue-enriched proteins. Color key: white to
red (− log10(p value) low to high).
(F) Western blot of tissue-specific proteins, with Ponceau S as a loading control. Left: protein structures predicted using AlphaFold.
(G) Left: heatmap of tissue-enhanced protein expression across tissues, categorized into modules. Right: representative GO terms enriched in the top 10
modules. Color key: blue to red (scaled expression level low to high).
(H) Left: heatmap of ‘‘housekeeping’’ and ‘‘others’’ protein expression across tissues. Right: representative GO terms enriched in these categories. Color key:
blue to red (protein abundance low to high).
(I) Radar chart of sex-biased protein counts across tissues.
(J) Top: pie charts of sex-biased proteins (p < 0.05, |log2FC| > 2) across tissues. Bottom: scatterplot of sex-biased protein differences between females and males,
with top 2 labeled.
(K) Western blot of proteins with sex biases: SGF29 (female biases in heart) and MMP3 (male biases in adipose). Ponceau S as loading control.
 ll
Article

(legend on next page)

 ll
Article

Figure S2. Aging-related mRNA-protein decoupling and proteostasis loss, related to Figure 1
(A) Schematic of protein homeostasis maintenance. Protein quality control occurs at transcription, translation, folding, modification, transport, localization, and
function as complexes. Proteostasis ensures normal tissue function and specificity and contributes to sex-based differences.
(B) Density plot of Spearman’s correlation loss between protein abundance and mRNA expression during aging. Blue: young group; red: old group.
(C) Sankey plot of decoupling proteins showing loss of Spearman’s correlation between protein abundance and mRNA expression in the aged group. Red:
positive correlation; blue: negative correlation; light gray: other proteins.
(D) Bar plot of GO terms for decoupling genes in each tissue. Color: white to red (− log10(p value) low to high).
(E) x axis: age-related correlation; y axis: decoupling difference. Red: age-positively correlated decoupling proteins. Blue: age-negatively correlated decoupling
proteins. Gray: other decoupling proteins.
(F) Dot plot of GSEA for protein homeostasis pathways for DEPs. Dot size: − log10(p value); color: blue to red (NES low to high).
(G) IHC staining of indicated proteins in human aorta tissues from young to old (n = 25). Spearman’s rank correlation was applied. Each point: one individual. Scale
bars: 50 μm.
(H) Boxplots of translational noise increase in tissue-enriched proteins during aging. Metrics: log2(1 + SD) and log2(1 + CV). SD, CV. Wilcoxon rank-sum test
applied.
(I) Left: density plots of GSVA scores using male-biased proteins in young groups across tissues (colors indicate sex groups). Right: density plots of GSVA scores
using female-biased proteins in young groups across tissues (colors indicate sex groups).
 ll
Article

(legend on next page)

 ll
Article

Figure S3. Age-related protein profile changes across tissues, related to Figure 2
(A) Top: pie charts of upregulated DEP subcellular location proportions. Top 2 Spearman’s correlation proteins are noted on the right. Bottom: abundance and
linear relationship with age for indicated proteins. Colors: tissues. Each point: one individual (Table S1).
(B) Top: pie charts of downregulated DEP subcellular location proportions. Bottom 2 Spearman’s correlation proteins are noted on the right. Bottom: abundance
and linear relationship with age for indicated proteins. Colors: tissues from (A). Each point: one individual (Table S1).
(C) Left: boxplots of Spearman’s correlation across tissues and aging upregulated DEP modules. Right: LOESS fitting trajectories of representative proteins in
each module. Significant tissues highlighted with colors from (A).
(D) Left: boxplots of Spearman’s correlation across tissues and aging downregulated DEP modules. Right: LOESS fitting trajectories of representative proteins in
each module. Significant tissues highlighted with colors from (A).
(E) Network diagram of GO terms enriched in upregulated DEP modules. Node pie charts: protein proportions. Pie sizes: enriched protein count; pie colors:
modules.
(F) Network diagram of GO terms enriched in downregulated DEP modules. Node pie charts: protein proportions. Pie sizes: enriched protein count; pie colors:
modules.
(G) Schematic of treating young hAECs with SAP protein.
(H) SA-β-Gal staining in hAECs treated with vehicle (n = 6) or SAP (n = 6). Scale bars, 100 μm.
(I) Tube formation analysis was performed in hAECs treated with vehicle (n = 6) or SAP (n = 6). Scale bars, 200 μm.
(J) Adhesion of monocytes assay was performed in hAECs treated with vehicle (n = 6) or SAP (n = 6). Scale bars, 100 μm.
(K) ELISA analysis of IL-6 was performed in hAECs treated with vehicle (n = 6) or SAP (n = 6).
Data are presented as mean ± SEM. Statistical significance between two groups was assessed using an unpaired two-tailed t test (H–K).
 ll
Article

(legend on next page)

 ll
Article

Figure S4. Protein hallmarks linked to human aging and aging-related diseases, related to Figure 3
(A) Immunostaining of p21Cip1, H3K9me3, IL-6, and collagen IV in various tissues (n = 22–62). Spearman’s correlation for coefficients and p values. Scale bars:
50 μm. Each point: one individual.
(B) Heatmap of downregulated epigenetic factor profile across tissues. Top 30 shared factors shown on right. Color key: blue to red (correlation coefficients low to
high).
(C) IHC staining of YTHDC1 in human liver (n = 50) and adrenal (n = 62) tissues from young to old. Spearman’s correlation for coefficients and p values. Each point:
one individual. Scale bars: 50 μm.
(D) Left: bar plot of complexes identified in each tissue. Right: pie charts of complex proportions in different functional groups across tissues.
(E) Left: dot plot of GSEA for aging-related chronic diseases. Dot sizes: − log10(p value); color: NES (light red to dark red). Right: network diagram of proteins
associated with aging-related chronic diseases. Node sizes: number of diseases; color: mean Spearman’s correlation coefficient (white to red).
 ll
Article

(legend on next page)

 ll
Article

Figure S5. Dynamic changes in proteomic clocks across tissues, related to Figure 4
(A) Bar plot of permutation importance proteins (importance > 1) (left) and top 5 importance proteins for each tissue aging clock (right).
(B) Heatmap of important proteins present in ≥4 tissue aging clocks. Color: permutation importance score (gray to red: low to high).
(C) Forest plot of TIMP3 importance across tissue clocks. Filled dots: importance >1; hollow points: importance ≤1; error bars: 90% confidence intervals.
(D) IHC staining of TIMP3 in spleen (n = 34), lymph (n = 25), adrenal (n = 62), pancreas (n = 34), and heart (n = 27). Each point: one individual. Spearman’s correlation
for coefficients and p values. Scale bars, 50 μm.
(E) We identified age-group-specific DEPs for early, middle, and late stages of aging. The percentage of these DEPs among detected proteins in each tissue is
shown (see STAR Methods).
(F) Bar graph of overlap between age-group-specific DEPs and lifespan-wide DEPs. Red: upregulated overlap; blue: downregulated overlap.
(G) Venn diagrams of overlap between age-group-specific and lifespan-wide DEPs. Left: upregulated overlap. Right: downregulated overlap. Circles and outer
box: different age stages.
(H) Enrichment analysis of DEPs overlapping between age-group-specific and lifespan-wide DEPs. Left: upregulated overlap. Right: downregulated overlap. Pie
chart sizes: protein count. Pie chart colors: proportion in different tissues.
(I) Relative number of time-resolved DEPs across tissues, normalized by the maximum in each tissue (see STAR Methods).
(J) Snapshots of time-resolved DEPs (normalized to detected proteins per tissue) at ages 30, 45, and 55 years. Colors: different tissues.
(K) Functional enrichment analysis of time-resolved DEPs at ages 30, 45, and 55 years. Pie charts: GO terms; pie size: enriched protein count; color: tissue
contribution.
 ll
Article

Figure S6. Aging-related secretome dynamics and inter-organ communication, related to Figure 4
(A) Schematic of secretome categories. Secreted proteins are divided into 4 major groups and 21 subgroups based on biological functions.
(B) Pie chart of secreted protein proportions across categories. Number of proteins in each category is noted beside the name.
(C) Left: expression patterns of secreted proteins across tissues during aging (5 modules). Colors: mean Spearman’s correlation coefficient (blue to red). Middle:
boxplots of Spearman’s correlation across tissues and modules. Category proportions (colors from B) annotated on right. Right: representative GO terms en-
riched in each module. Dot sizes: protein counts; dot colors: − log10(p value) (white to black).
(D) Heatmap of SASP factors across tissues. Number of SASP factors per tissue on top; factors shared in ≥5 tissues on right. Colors: mean Spearman’s cor-
relation coefficient (white to red).
(E) Network diagram of tissue interactions during aging. Line thickness and color (white to red): interaction strength (low to high).

(legend continued on next page)

 ll
Article

(F) Heatmap of ligand-receptor pairs across tissues during aging. Top: upregulated pairs (white to red: mean expression low to high). Bottom: downregulated
pairs (blue to white: mean expression low to high).
(G) Dot plot of interaction protein pairs and associated diseases. Left: enriched diseases from tissue pairs during aging (color and dot size: − log10(p value) low to
high). Right: interacting proteins of tissue pairs (color: interaction difference during aging.
(H) Ridge plot of GSEA for aging-related diseases. x axis: age correlation (Spearman’s correlation coefficient) of genes in pathways. Color: NES.
 ll
Article

Figure S7. Plasma protein proxy aging clocks for different organs, related to Figure 6
(A) Bar plot of proteins with similar age-related trends in tissue and plasma (plasma-tissue paired DEPs). y axis: number of paired DEPs (below zero: commonly
downregulated; above zero: upregulated). x axis: different tissues. Gray bars: proteins with the same trend as plasma in a specific tissue; yellow bars: proteins
with similar trends in ≥2 tissues.
(B) Dot plot of plasma-tissue paired DEPs commonly downregulated with aging (Spearman’s rank correlation coefficient < 0, p < 0.05). Proteins differentially
expressed in ≥5 tissues are shown. x axis: proteins; y axis: tissues. Dot color: correlation of age-related change. Line plot above shows average plasma
abundance.
(C) ELISA analysis of HARS1 expression in plasma with age. Pearson’s correlation for coefficients and p values (n = 55, randomly selected individuals).
(D) Schematic of predicting tissue age using plasma proteins. Plasma and tissue proteomics data were integrated, and common proteins were selected. Elastic
net regression model built with tissue age as target. Plasma surrogate tissue clock developed after optimization and feature selection (see STAR Methods).
(E) Scatter plot of chronological age vs. predicted age (tissue age and plasma protein proxy tissue age). Correlation coefficient, p value, and MAE calculated using
Spearman’s correlation. Right bar chart: the top 10 proteins ranked by descending feature importance (see STAR Methods) in the plasma protein proxy clock.
(F) Dot plot of the Spearman’s correlation between plasma protein proxy aging clock predicted age and tissue age. x axis: predicted age by plasma protein proxy
aging clock using plasma protein abundance. y axis: tissue age predicted by tissue proteomic aging clocks using tissue protein abundance. Colors: white to red
(correlation coefficients low to high). Size: − log10(p value).
(G) Bar chart of feature counts for each plasma protein proxy aging clock.
(H) Dot plot of correlation between protein abundance (used in plasma protein proxy aging clock) and age in tissue, plasma, and their importance in the clock.
Spearman’s correlation for coefficients and p values. Dot color: correlation coefficient; dot size: − log10(p value).
 ll
Article

Figure S8. Plasma senoproteins promote vascular endothelial cell senescence, related to Figure 7
(A–D) SA-β-Gal staining (A), western blot (B), migration assay (C), and tube formation (D) in hAECs treated with vehicle (n = 6 for A and C–E; n = 3 for B) or NOTCH3
(n = 6 for A and C–E; n = 3 for B). Scale bars, 100 μm.
(E) Dot plots of ELISA for IL-6 in hAECs treated with vehicle or recombinant proteins. Dot color: IL-6 concentration (fold induction, normalized by vehicle). Dot size:
p value (one-way ANOVA), n = 6.
(F) Top: bar plot of DEGs between recombinant protein groups (COMP, NOTCH3, HTRA1, SLPI, IGFBP7, GPNMB, and NEGR1) and the vehicle group. Upre-
gulated DEGs: red; downregulated DEGs: blue. Bottom: heatmap of relative GSVA scores for senescence-associated pathways. Colors: white to red (GSVA
scores low to high).
(G–I) SA-β-Gal (G), H3K9me3 (H), and LAP2 (I) staining in hVSMCs treated with vehicle or GPNMB. n = 6 (G), n = 300 nuclei (H and I). Scale bars: 200 μm (G), 20 μm
(H and I).
(J–L) SA-β-Gal (J), H3K9me3 (K), and LAP2 (L) staining in hVSMCs treated with vehicle or HTRA1. n = 6 (J), n = 300 nuclei (K and L). Scale bars: 200 μm (J), 20 μm (K
and L).
(M–O) SA-β-Gal (M), H3K9me3 (N), and LAP2 (O) staining in hVSMCs treated with vehicle or IGFBP7. n = 6 (M), n = 300 nuclei (N and O). Scale bars: 200 μm (M),
20 μm (N and O).
(P) Dot plots of ELISA for IL-6 in hVSMCs treated with GPNMB, HTRA1, or IGFBP7. Dot color: IL-6 concentration (fold induction, normalized by vehicle). Dot size:
p value (one-way ANOVA), n = 6.
Data are presented as mean ± SEM. Statistical significance between two groups was assessed using an unpaired two-tailed t test or Mann-Whitney test (A–E and
G–O).