Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Apr 23;27(4):997-1007.e5.
doi: 10.1016/j.celrep.2019.03.104.

Notch Signaling Mediates Secondary Senescence

Affiliations
Free PMC article

Notch Signaling Mediates Secondary Senescence

Yee Voan Teo et al. Cell Rep. .
Free PMC article

Abstract

Oncogene-induced senescence (OIS) is a tumor suppressive response to oncogene activation that can be transmitted to neighboring cells through secreted factors of the senescence-associated secretory phenotype (SASP). Currently, primary and secondary senescent cells are not considered functionally distinct endpoints. Using single-cell analysis, we observed two distinct transcriptional endpoints, a primary endpoint marked by Ras and a secondary endpoint marked by Notch activation. We find that secondary oncogene-induced senescence in vitro and in vivo requires Notch, rather than SASP alone, as previously thought. Moreover, Notch signaling weakens, but does not abolish, SASP in secondary senescence. Global transcriptomic differences, a blunted SASP response, and the induction of fibrillar collagens in secondary senescence point toward a functional diversification between secondary and primary senescence.

Keywords: CEBPB; Notch; TGFB; bystander senescence; oncogene induced senescence; paracrine senescence; secondary senescence; senescence; senescence associated secretory phenotype; single-cell RNA sequencing.

Figures

None
Figure 1
Figure 1
Secondary Senescent Cells Only Partially Resemble Paracrine-Induced Senescence (A) Schematic representation of the time course experiment. (B) Number of senescent cells with reads mapping to the G > T mutation site of RAS gene. (C) Monocle2 plot for time course experiment. The presence of the mutated RAS gene is indicated. Pie charts for the percentage of Ras+/Ras− cells in the top and bottom clusters. (D) Boxplots for the expression of senescence genes in the time course experiment. The top and bottom bounds of the boxplot correspond to the 75th and 25th percentile, respectively. p values were obtained using differential analysis in SCDE. (E) Unsupervised clustering using SC3 for senescent cells. Cells were annotated as either OIS (top senescence branch, purple), secondary senescence (bottom branch, green), or NA (neither, pink). (F) Schematic representation of the co-culture experiment. (G) t-Distributed Stochastic Neighbor Embedding (tSNE) visualization of co-culture scRNA-seq. (H) tSNE visualization of single cells grouped into 3 clusters. (I) Boxplots for the expression of senescence genes in the co-culture experiment. The top and bottom bounds of the boxplot correspond to the 75th and 25th percentile, respectively. p values were obtained using differential analysis in SCDE. (J) Integration analysis of the two senescence clusters from time course and co-culture experiments. (K) Overlap of differentially expressed (DE) genes between paracrine/OIS, time course, and co-culture experiments. Related to Figure S1 and Table S1. Alignment Rates and Quality Control of RNA Sequencing Data, Related to Figures 1 and 4, Table S2. Differential Expression of RNA Sequencing Data, Related to Figures 1, 2, 3, and 4, Table S3. Presence of Construct and qPCR Primer, Related to Figures 1, 2, 3, and 4, Table S4. Genes for Venn Diagrams, Related to Figures 1 and 2.
Figure 2
Figure 2
Secondary Senescence Comprises NIS Signature in the Majority of Cells (A) Boxplots for the expression of genes COL1A1, COL3A1, and COL5A2 in the time course and co-culture experiments (p < 0.05). The top and bottom bounds of the boxplots correspond to the 75th and 25th percentile, respectively. p values were obtained using differential analysis in SCDE. (B) Model suggesting NIS and RIS are regulated by Notch1 through TGFB and CEBPB, respectively. (C) IPA analysis of the two senescence clusters from the time course and co-culture experiments relative to growing. (D) Boxplots for the expression of TGFB1I1, CTGF, and CEBPB genes in the time course (top) and co-culture experiments (middle). The top and bottom bounds of the boxplot correspond to the 75th and 25th percentile, respectively. p values were obtained using differential analysis in SCDE. Bar graphs denoting expression of TGFB1 (n = 6), TGFBL (n = 6), and CEBPB (n = 3) mRNA as measured by qPCR in OIS and GFP cells (bottom) (TGFB1: t = −3.2317, df = 5.5117, p = 0.02; TGFBI: t = −2.2567, df = 9.8141, p = 0.05; CEBPB: t = 0.068192, df = 3.2294, p = 0.95, unpaired Student’s t test. Error bars represent SEM). (E) Representative image of GFP (secondary senescence) and CEBPB (red) immunofluorescence in the co-culture experiment. Mean intensity for primary (ER:Ras) and secondary senescent cells (GFP) was measured (p = 0.016, unpaired Student’s t test). Error bars are displayed as SEM. (F) GSEA plots for the enrichment of secondary and primary OIS DE genes (time course and co-culture experiments) in Hoare et al. (2016) NIS and RIS log2FC preranked genes. Normalized enrichment score (NES) and false discovery rate (FDR) are shown. (G) Venn diagrams overlapping expression signatures from time course (top) and co-culture (bottom) with NIS signature genes. (Secondary senescence: Secondary senescence/OIS upregulated genes; NIS: Hoare et al. (2016) NIS/RIS upregulated genes; RIS: Hoare et al. (2016) RIS/NIS upregulated genes.) Related to Figure S2 and Table S4.
Figure 3
Figure 3
NIS Mediates Secondary Senescence In Vitro (A) Schematic representation of co-cultures with perturbed Notch signaling. (B) Bar plot for EdU incorporation in growing (black) or senescent (gray) EV or dnMAML1 cells co-cultured with ER:Ras as proportion of all cells scored. Error bars are displayed as SEM; F[7,16] = 20.63, p < 0.001, one-way ANOVA with Tukey’s test. (n = 3 per condition). Representative images are shown. (C) Scmap cluster projection of the dnMAML1 and EV 10× scRNA-seq dataset to the GFP co-culture 10× scRNA-seq dataset (see Figure 1H). (D) tSNE plot of single cells colored by the projection toward the GFP co-culture 10× dataset (see Figure 1H). Pie charts show percentage of cells. (E) GSEA pre-ranked test for enrichment of Notch signaling in mVenus:EV identified as secondary senescence by scmap. (F) Heatmap of single-cell data comparing mVenus:EV and mVenus:dnMAML1 for collagens and SASP genes. Red, upregulated and blue, downregulated. (G) GSEA pre-ranked test for enrichment of SASP genes in mVenus:dnMAML1 identified as secondary senescence by scmap. (H) GSEA pre-ranked test for enrichment of E2F targets in mVenus:dnMAML1 identified as secondary senescence by scmap. (I) Schematic representation of transwell co-culture assay of OIS and GFP cells. (J) Heatmap of significantly differentially expressed genes (p < 0.05) between GFP contact and GFP no contact cells. (K) GSEA pre-ranked analysis for enrichment of Notch signaling in GFP contact cells compared to GFP no contact cells. (L) Pathway analysis for DE genes between GFP contact/GFP no contact (p < 0.05). (M) GSEA pre-ranked analysis for enrichment of E2F targets in GFP no contact compared to GFP contact cells. Leading edge genes are indicated. Related to Figure S3 and Table S2.
Figure 4
Figure 4
Notch Signaling Mediates Secondary Senescence In Vivo (A) Schematic representation of in vivo single-cell experiment. (B) Representative immunofluorescence images of liver section from induced AhCre+Mdm2fl/fl and control AhCreWTMdm2fl/fl mice stained for p53 and CDKN1A. Intrinsically induced senescence (arrowhead) and secondary senescence (arrow) are indicated. Boxplot for CDKN1A intensity in primary versus secondary senescent cells. (senescence: F[1,50291] = 2766, p < 0.0001; biological replicates: F[2,50291] = 283.2, p < 0.0001; senescence × biological replicates: F[2,50291] = 280.5, p < 0.0001, two-way ANOVA). Scale bar, 22 μm. (C) Pathway analysis for Mdm2+ (secondary) genes. (D) GSEA for Mdm2+/Mdm2− cells (NES = 1.07). Leading edge genes are indicated. (E) Heatmap for Notch pathway, hepatocyte markers, and Cdkn1a genes in Mdm2+ and Mdm2− cells. Constitutive genes and Cdkn1a were colored by their expression relative levels (binary: red expressed, white not expressed). (F) SCDE for Maml1, Rfng, and Smad3 in Mdm2+ cells (orange lines) and Mdm2− cells (blue lines). Joint posterior is marked by black line. Fold change of the genes in Mdm2+/Mdm2− is indicated in red, and dotted lines mark the 95% confidence interval. MLE, maximum likelihood estimation; CI, confidence interval; Z, Z score. (G) Representative immunofluorescence images of liver section from induced and control mice. Primary senescent cells (arrowheads) and secondary senescent cells (arrows) are indicated (CDKN1A: F[1,60145] = 353.3, p < 0.0001; biological replicates: F[2,60145] = 1044, p < 0.0001; CDKN1A × biological replicates: F[2,60145] = 8.96, p < 0.0001, two-way ANOVA). Scale bar, 22 μm. Related to Figure S4 and Tables S1 and S2.

Similar articles

See all similar articles

Cited by 4 articles

References

    1. Acosta J.C., O’Loghlen A., Banito A., Guijarro M.V., Augert A., Raguz S., Fumagalli M., Da Costa M., Brown C., Popov N. Chemokine signaling via the CXCR2 receptor reinforces senescence. Cell. 2008;133:1006–1018. - PubMed
    1. Acosta J.C., Banito A., Wuestefeld T., Georgilis A., Janich P., Morton J.P., Athineos D., Kang T.W., Lasitschka F., Andrulis M. A complex secretory program orchestrated by the inflammasome controls paracrine senescence. Nat. Cell Biol. 2013;15:978–990. - PMC - PubMed
    1. Adams P.D. Remodeling of chromatin structure in senescent cells and its potential impact on tumor suppression and aging. Gene. 2007;397:84–93. - PMC - PubMed
    1. Adams P.D. Healing and hurting: molecular mechanisms, functions, and pathologies of cellular senescence. Mol. Cell. 2009;36:2–14. - PubMed
    1. Aksoy O., Chicas A., Zeng T., Zhao Z., McCurrach M., Wang X., Lowe S.W. The atypical E2F family member E2F7 couples the p53 and RB pathways during cellular senescence. Genes Dev. 2012;26:1546–1557. - PMC - PubMed

Publication types

LinkOut - more resources

Feedback