Cellular senescence alters the morphology, cell proliferation and cytokines secreted by ADSCs
To assess cellular senescence in ADSCs, cell cultures from five 1-month-mice and five 20-month-mice in vitro cultivation. After 7 days, most cells adhered to the tissue culture plastic, assuming a fibroblast-like phenotype as flat and spindle shaped cells. Observed with phase-contract microscope, there were no morphological differences between the two donors (Fig. 1A). With increasing passage number, ADSCs had some senescence-associated changes: the increase of average cell size, flattening of cells, and accumulation of granular inclusions in cytoplasm.
ADSCs were fixed to further analyze its inner structure and cell composition. They presented large and round cell bodies with pseudopodia protruded. Homogeneous round lipid droplets and a lot of ribosomes were found in the cytoplasm. ADSCs have a relatively lobular or polygonal large nucleus and a high karyoplasmic ratio. In addition, there are abundant projections and depressions on the nuclear membrane and 1–3 nucleoli in the nucleus. In a further magnification view, the mitochondrion in the cytoplasm had a bilayer membrane structure with an inner membrane folded into cristae. Compared with ADSCs from 1-month mice (1M ADSCs), ADSCs from 20-month mice (20M ADSCs) showed swollen rough endoplasmic reticulum, a large number of medullary corpuscles in cytoplasm (Fig. 1B, B6) and lipid droplets increased (Fig. 1B, B7). Mitochondrion was swollen with altered crest structure (Fig. 1B, B8).
To further investigate whether ADSCs from aged donors are functional, the differentiation capacity of 1M and 20M ADSCs into the adipogenic, osteogenic lineages was analysed. The results confirmed that the 20M ADSCs still have the two lineages differention ability as 1M ADSCs (Fig. 1C, D and E). In addition, the functionality of cytokines secreting was examined by employing a cytokine array. As shown in Fig. 1G, the levels of chemokines in cell culture supernatant were obviously increased with age, including granulocyte-macrophage colony-stimulating factor (GM-CSF), interleukin-4 (IL-4), IL-13, IL-17, macrophage inflammatory protein-1α (MIP-1α) or C-C motif chemokine ligand 3 (CCL3) and thymus-expressed chemokine (TECK) or CCL25.
When cultured to P3, cells were seeded in 24-well plate at a density of 1\(\times\)103 per well, and measured every day to draw the growth curve of ADSCs from 1M and 20M mice. As shown in Fig. 1F, the young ADSCs entered exponential growth period in a shorter time and grew in a fast speed compared to the aged.
A ADSCs morphology under optical microscope. Scale bars of 10× groups, 100µm. Scale bars of 20× groups, 75µm. B Ultrastructural changes of ADSCs under transmission electron microscope(TEM). B1-B4 represent ADSCs from 1-month (1M) mice while B5-B8 represent 20-month(20M). C Osteogenic differentiation differences of ADSCs from 1M and 20M mice. Scale bars, 100µm D, E Adipogenicity differences and quantitative analysis of ADSCs from 1M and 20M mice. Scale bars, 100µm F Growth curve of ADSCs. G Cytokine assay of ADSCs from 1M and 20M donors. ∗, p < 0.05.
An elevated number of apoptotic cells accompanied by an increase of cells in the quiescent G0 phase from 20M mice
All the cells matched the ADSCs markers——positive in CD29, CD105 and negative in CD45, CD11b. As the Fig. 2A showed, the average percentage of CD29, CD105, CD45, CD11b in 1M ADSCs is 96.9, 28.5, 25.0, 11.8, while the average percentage in 20M ADSCs is 97.2, 42.0, 19.1, 4.7. 1M and 20M cells have no significant difference in these surface markers, which means the adipose tissue obtained from young mice has the same yield of the old under a same condition.
SA–β-Gal staining indicated few senescence cells at early passages among the two types of ADSCs (Fig. 2B). After two serial cell passages every 3–4 days, the number of SA–β-Gal-positive cells was determined, and the two groups have no significant difference (Fig. 2B and E). The senescent cells number was increased as the passage number increased, indicating that passage time has a greater effect on senescence than donor age.
The influence of donor age on the cell cycle and apoptosis was measured by flow cytometry (Fig. 2C, D, G, and H). An increase in apoptotic cells was detected in 20M ADSCs compared with 1M ADSCs (Fig. 2C and G). 20M Cells in G0 phase increase and cells in G1 and S/G2/M phases decrease were noted compared with 1M ADSCs (Fig. 2D and H).
Finally, analysis of aging-related genes expression, using qRT-PCR, respectively, demonstrated a drastic increase expression of p19 in 20M ADSCs, compared to 1M cells. p16 and p21 are two markers frequently associated with cellular senescence33, both of which were highly expressed in cells isolated from the old donors (Fig. 2F).
A The proportion of ADSCs expressing positive and negative surface marker. B, E SA-β-gal staining and quantitative analysis of 1M and 20M ADSCs. Scale bars, 100µm. C, D, G, and H Flow cytometry and quantitative analysis of Annexin V and 7-AAD between 1M and 20M ADSCs. ∗, p < 0.05. ∗∗, p < 0.01. F Age-related gene expression of 1M and 20M ADSCs. ∗∗, p < 0.01. ∗∗∗∗, p < 0.0001.
Lef1 and Ddx3y were the specific elements for the down-regulated and up-regulated genes, respectively, in mADSCs with age.
To explore the underlying mechanism, we performed RNA-seq on ADSCs from 1M and 20M mice. Sequencing yielded on average 39.2 million raw reads per 1M ADSCs and 33.5 million raw reads per 20M ADSCs. As shown in Supporting Information Table S2, the Q20 of 1M ADSCs and 20M ADSCs were more than 95%, and the Q30 were more than 90%. Alignment analysis resulted in an average total read mapping rate of 95.51% and a proper pair alignment rate of 88.35% in 1M ADSCs, while 20M’s average total read mapping rate was 95.39% and a proper pair alignment rate was 88.26%. The results were listed as Supporting Information Table S3. The RPKM value of two groups were listed in Supporting Information Table S4. It is commonly supposed to be a marker of significant difference in gene expression when it’s more than 1.
1M and 20M ADSCs have 11,041 genes in common (Fig. 3A). The overall distribution of differentially expressed genes (DEGs) between the two groups were selected according to the value of fold change and the significance level, which could be inferred completely in volcano plot. DEGs from 20M mice compared to that of the 1M mice were colored with green and red, regarding 431 significantly down-regulated and 1,481 significantly up-regulated genes, respectively (Fig. 3C). Top 10 of the most down-regulated genes were Lef1, Lingo2, Gpbar1, Mcpt1, Hecw1, Dthd1, Gpr149, Ajap1, Zfp82, Fibcd1 and top 10 of the most up-regulated genes were Ddx3y, S100a9, S100a8, Eif2s3y, Ngp, Chil3, Ltf, Saa3, Kdm5d and Uty according to the absolute value of log2Foldchange with age (Fig. 3D). The DEGs were verified by qRT-PCR, and the gene expression pattern was consistent with the transcriptome results(Fig. 3B).
A Venn chart showing 11041 differentially expressed genes (DEGs) in common between 1M and 20M ADSCs. Corrected P-value of 0.05 and absolute fold change of 2 were set as the threshold for significantly differential expression. B Verification of DEGs through qRT-PCR. ∗, p < 0.05. ∗∗, p < 0.01∗, p < 0.0001. C Volcano plot showing DEGs significantly down-regulated (Green) and up-regulated (Red) compared to 1M ADSCs. D Top 10 of DEGs profiles. Green bar indicates significantly down-regulated DEGs, while red indicates significantly up-regulated DEGs.
Chemokine signaling pathway and Hippo signaling pathway were the top potential signaling pathways, indicating that these two pathways related to age-related alterations in mADSCs.
To explore the changing profile of gene expression level, up-regulated and down-regulated genes were analyzed separately. The DEGs were grouped in three GO categories: Biological Process(BP), Cellular Component(CC) and Molecular Function(MF). The up-regulated DEGs were mostly associated with innate immune response, inflammatory response, leukocyte migration, leukocyte chemotaxis, cell chemotaxis in BP. As for CC, the genes are mostly expressed differentially in contractile fiber, myofiril. Cytokine activity, chemokine activity are the mostly related terms of differentially expressed genes in MF (Fig. 4A). KEGG is a database resource for understanding high-level functions and utilities of the biological system, such as the cell, the organism and the ecosystem, from molecular-level information, especially large-scale molecular datasets generated by genome sequencing and other high-throughput experimental technologies34. KEGG analysis further identified the cytokine-cytokine receptor interaction and chemokine signaling pathway as the top scoring pathway among the age-related up-regulated genes (Fig. 4C).
GO and KEGG pathway enrichment of age-related down-regulated DEGs were showed in Fig. 4B and D. The results of GO showed that DEGs were enriched in axon devilment, mesenchymal cell development in BP, and growth factor activity in MF (Fig. 4B). Seven KEGG pathway were significantly enriched, namely gastric cancer, Hippo signaling pathway, vascular smooth muscle contraction, mircoRNAs in cancer, axon guidance, TGF-beta signaling pathway, and dilated cardiomyopathy (DCM). Gastric cancer, Hippo signaling pathway and axon guidance were the highest ratio of the number of DEGs annotated to corresponding pathway to the number of annotated DEGs i.e., 11 out of 124 genes in response to aging (Fig. 4D).
A, B GO enrichment of age-related DEGs. The top ten terms enriched for each category were listed. A Up-regulated genes enrichment; B Down-regulated genes enrichment. C, D KEGG pathway enrichment of age-related DEGs. The top twenty pathways enriched were listed. Count, differential gene count in indicated pathway; padj, adjusted P value of indicated pathway. C Up-regulated genes enrichment; D Down-regulated genes enrichment.
Chemokine signaling pathway was the crucial signaling pathway in mADSCs with age.
We further performed GO and KEGG pathway enrichment analysis on all genes we enriched. The functional categorization revealed that most of the DEGs with age function in inflammatory response, contractile fiber, and calcium ion binding(Table 1). Overall DEGs were found to be associated with the following signaling pathways: Chemokine, Calcium, Rap1, PI3K-Akt, MAPK and cGMP-PKG (Fig. 5A).
To further investigate the involved signaling pathway from an overall perspective, a gene set analysis was conducted in significance level at FDR < 0.05. Based on the parameters, 49 positive related categories and 1 negative related categories are identified as enriched categories, in which 5 most significant categories and representatives in the reduced sets are shown in Fig. 5B. Clustering analysis showed that chemokine signaling pathway clustered with the highest normalized enrichment score (NES,1.8068) while the NES of negative related one DNA replication was − 1.7139. Enriched genes were shown in red in the map of chemokine signaling pathway (Fig. 5C, Supporting Information Table S5). Meanwhile, screened differential genes were imported into the PANTHER tool for pathway enrichment analysis. Inflammation mediated by chemokine and cytokine signaling pathway was the most significant positive category (Supporting Information Fig.S1).
Table 1
GO functional annotation of overall DEGs
Category
|
GO ID
|
Description
|
GeneRatioa
|
Padjb
|
Upc
|
Downd
|
BP
|
GO:0006954
|
inflammatory response
|
144/1636
|
1.08E-26
|
130
|
14
|
|
GO:0045087
|
innate immune response
|
135/1636
|
1.62E-25
|
126
|
9
|
|
GO:0003012
|
muscle system process
|
103/1636
|
2.47E-23
|
86
|
17
|
|
GO:0006935
|
chemotaxis
|
123/1636
|
2.47E-23
|
100
|
23
|
|
GO:0050900
|
leukocyte migration
|
89/1636
|
2.47E-23
|
83
|
6
|
CC
|
GO:0043292
|
contractile fiber
|
82/1676
|
5.36E-26
|
66
|
16
|
|
GO:0030016
|
myofibril
|
75/1676
|
2.13E-23
|
62
|
13
|
|
GO:0044449
|
contractile fiber part
|
73/1676
|
2.78E-23
|
58
|
15
|
|
GO:0030017
|
sarcomere
|
68/1676
|
4.63E-22
|
55
|
13
|
|
GO:0009897
|
external side of plasma membrane
|
77/1676
|
1.09E-18
|
68
|
9
|
MF
|
GO:0005509
|
calcium ion binding
|
112/1622
|
1.43E-12
|
89
|
23
|
|
GO:0008009
|
chemokine activity
|
18/1622
|
5.38E-09
|
16
|
2
|
|
GO:0005125
|
cytokine activity
|
41/1622
|
1.92E-08
|
36
|
5
|
|
GO:0030414
|
peptidase inhibitor activity
|
39/1622
|
4.59E-08
|
33
|
6
|
|
GO:0005126
|
cytokine receptor binding
|
54/1622
|
5.04E-08
|
46
|
8
|
Abbreviations: GeneRatio, differential gene count in indicated pathway versus total differential gene count; Padj, adjusted P value of indicated pathway; Up, up-regulated differential gene count in indicated pathway; Down, down-regulated differential gene count in indicated pathway.
A KEGG pathway enrichment of age-related DEGs. B Volcano and enrichment plot of GSEA in significance level at FDR < 0.05. NES, normalized enrichment score. FDR, false discovery rate. C Map of chemokine signaling pathway in KEGG database. Mapped DEGs were shown in red.
The CCL7-CCL2-CCR2 axis regulates age-related changes in mADSCs.
During the evaluation of the clusters enriched in chemokine signaling pathway for Biogrid data, it was found that the Biogrid PPI network had one node CCL2 of very high degree corresponding to the chemokine. Enrichment analysis showed total number of 18 genes in the expanded sub-network, and top 5 neighbors based on the probability of random walk method was listed in Supporting Information Table S5. All seeds and top ranking neighbors in the expanded sub-network can enrich to 5 GO BP categories (Fig. 6A and B). Results showed total number of 3 genes, CCL7, CCL2, CCR2 in the retrieved sub-network, which enrich to 5 GO BP categories (Fig. 6C and D).
A Expanded PPI network of DEGs mapped chemokine signaling pathway. B Top 5 of GO BP categories that all seeds and top ranking neighbors in A could be enriched. C Retrieved PPI network of DEGs mapped chemokine signaling pathway. D Top 5 of GO BP categories that Ccl7, Ccl2, Ccr2 could be enriched.