Conserved transcriptomic signatures and protein markers in cellular senescence models

doi:10.21203/rs.3.rs-3110821/v1

Download PDF

Article

Conserved transcriptomic signatures and protein markers in cellular senescence models

https://doi.org/10.21203/rs.3.rs-3110821/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Cellular senescence is described as an irreversible cell cycle arrest induced in response to various stresses. Senescent cells are characterised by heterogeneous signalling alterations, complex secretory phenotype, known as senescence-associated secretory phenotype (SASP), and diverse transcriptomic profile. With the aim to investigate senescence heterogeneity and identify conserved transctiptomic alterations and senescence markers, we performed RNA-seq and multiplex proteomic analysis in proteasome inhibition-induced and stress-induced premature senescence models of HFL1 and BJ human fibroblasts. Our data revealed diverse transcriptomic signatures, but also, 231 common differentially expressed genes related to cell division and ECM remodelling, and enriched pathways that remained conserved among the different models with senescence onset. Moreover, we identified a subset of conserved protein senescence markers and validated them in replicative senescent models. These proteins are involved in cell cycle arrest and promote a pro-inflammatory environment in premature and replicative senescence models. We suggest that the simultaneous analysis of p21, p-c-JUN, BCL-xL and survivin in cellular lysates, and IL-8, GM-CSF, GDF-15 and GROa in culture supernatants can provide a powerful tool for the identification and monitoring of senescent cells and can support the assessment of the efficacy of potential senotherapeutic approaches.

Biological sciences/Systems biology

Health sciences/Biomarkers

Biological sciences/Cell biology/Senescence

Biological sciences/Biotechnology/Sequencing/Rna sequencing

Aging is an inevitable process characterised by the functional deterioration of cells, tissues, and organisms. In humans and mammals, it is directly linked to the appearance of age-related conditions, such as chronic inflammation, cancer, neurodegenerative and cardiovascular diseases. One of the hallmarks of aging is cellular senescence, described as a state of irreversible cell cycle arrest¹. Hayflick and Moorhead first observed that cells ceased proliferation following successive replication cycles², a phenomenon termed replicative senescence which was attributed to telomere shortening³. Senescence is a physiological cell-autonomous mechanism that normally prevents cells from malignant transformation, but is also implicated in embryonic development, tissue repair, wound healing and immune surveillance⁴. Nevertheless, the abnormal accumulation of senescent cells in tissues can have detrimental effects on their regenerative capacity and creates an inflammatory environment that promotes the development of several age-related diseases, such osteoarthritis, atherosclerosis, neurodegenerative diseases, liver fibrosis, chronic inflammation and cancer^5,6.

Apart from telomere erosion, senescence is induced in response to various stresses, including oxidative stress, DNA damage, telomere dysfunction, chromatin disorganisation and oncogene activation⁷. Senescent cells exhibit various non-exclusive characteristics independently of the senescence type and inducer, such as enlarged flattened morphology, increased lysosomal content presented with increased Senescence-associated beta-galactosidase (SA-β-Gal) activity at pH 6, activation of the p53–p21^CIP1 and p16^INK4A–retinoblastoma (Rb) pathways and reduced expression of the nuclear lamina protein Lamin B1 (LMNB1)^8–10. They are also characterised by a complex, diverse, and intensive secretory phenotype, known as senescence-associated secretory phenotype (SASP). SASP consists of several pro-inflammatory cytokines, chemokines and growth factors, including Il-6, IL-8, MCP-1, IL-10, IL-1a, IL-1b, MMP-1, MMP-3, GROa, CXCL2, GM-CSF, M-CSF, GDF-15 and STC1^4,11,12. Some of these extracellular moieties are released in response to DNA damage via NF-kB activation, in an IL-1-dependent manner, while others are secreted as soluble molecules or in exosomes¹².

Senescent cells evade apoptosis through the activation of prosurvival factors of BCL2 family, particularly BCL-xL and BCL-W, p21, PI3K^13–16, which promote apoptosis resistance and further cause senescent cell accumulation in several tissues^17,18. Their accumulation promotes a prolonged SASP activity and stimulates immune response. Moreover, SASP factors enable senescent cells to act both in an autocrine and paracrine manner, and therefore, their prolonged activation affects surrounding cells and causes alterations in the microenvironment that benefits age-related pathologies^5,6,19,20.

Although cellular senescence has gained research attention, senescence programme heterogeneity, along with SASP diversity and complexity have impeded the identification of universal markers for senescent cell detection and therapeutic approaches that target senescent cells and delay age-related conditions^21–23. The different senescence inducers may activate similar or distinct signaling cascades and cause secretion of extracellular moieties, that may depend on the inducer itself, the extent of damage and the cell type and origin. Whole transcriptome analysis of different senescence types has depicted this heterogeneity, aiming to identify common SASP factors^21,23,24.

In this study, we developed two cellular senescence models in human HFL1 and BJ fibroblasts from lung and skin, respectively, and investigated senescence heterogeneity and SASP complexity at a transcriptomic and proteomic level, with the aim to identify conserved transcriptomic signatures and potential universal senescence markers. Firstly, a proteasome inhibition-induced premature senescence (PIIPS) model was developed by exposing cells to the proteasome inhibitor bortezomib, which leads to loss of proteostasis and accumulation of misfolded proteins observed during aging^25–28. A stress induced-premature senescence (SIPS) model was also developed where cells were exposed to oxidative stress using H₂O₂ that has been previously shown to induce premature senescence in various cell types^29–32. Transcriptomic screening results revealed distinct transcriptomic profiles, but also a conserved senescence signature consisting of 217 common downregulated and 14 common upregulated genes across PIIPS and SIPS in both cell lines. Senescence heterogeneity was also observed at the protein level following multiplex analysis of signalling molecules and extracellularly secreted SASP factors and a panel of eight proteins with consistent alterations across senescence models and cell lines was identified and further validated in Replicative Senescence (RS) models of the same cell lines.

Induction of cellular senescence with proteasome inhibition and oxidative stress in human lung and dermal fibroblasts and phenotype assessment

Firstly, we established the cellular senescence models by exposing the two fibroblast strains to the proteasome inhibitor bortezomib and the oxidant H₂O_2, under certain conditions, until they reached a senescence state. We used an ATP assay to assess cell viability and determine IC₅₀ doses for both treatments and cell lines (Fig. 1a). Several treatment durations and concentrations lower or equal to the IC₅₀ doses were used for each treatment scheme. Optimal conditions for senescence induction were selected based on IC₅₀ doses and SA-β-gal results. HFL1 and BJ cells ceased proliferation following treatment with 9.7nM and 10.5nM bortezomib for 9 and 10 days, respectively. In the SIPS model, HFL1 and BJ cells presented a senescence phenotype upon treatment with 75 µM and 125 µM of H₂O₂ for 2 hours, respectively, followed by recovery for 72 hours in complete culture media. These treatment conditions were used for all subsequent experiments conducted in PIIPS and SIPS models.

Next, we measured SA-β-Gal activity, a typical senescence marker¹³, under the optimized treatment conditions to confirm that cells’ exit from cell cycle was linked to senescence. We observed enhanced SA-β-Gal activity in both models and cell lines, with approximately 70–80% of cells stained positive, whereas less than 10% of cells were positive in the control non-senescent groups (Fig. 1b).

We then assessed the levels of cyclin dependent kinase inhibitor p21 as an additional marker for cell cycle arrest and senescence, using a custom-developed ELISA xMAP assay. Bortezomib caused a significant 11-fold and 37-fold increase in p21 levels in HFL1 and BJ cells, respectively, whilst H₂O₂ treatment caused a 5-fold increase in p21 levels in both cell lines (Fig. 1c). Senescence phenotype was also monitored microscopically, with senescent cells presenting a flattened cell morphology and proliferation arrest in all senescence models. Altogether, these results indicate that proteasome inhibition with bortezomib and oxidative stress with H₂O₂ under the selected conditions induce a premature senescent phenotype in both fibroblast strains.

PIIPS and SIPS cause a pre-redox state with elevated ROS levels

Exposure to oxidants such as H₂O₂ induces changes in intracellular ROS levels which have been reported to promote a pre-redox state leading to a premature senescence phenotype^29–32. We therefore assessed the accumulation of ROS as an additional marker of premature senescence in the established PIIPS and SIPS models using a fluorescence-based cellular assay. Significant accumulation of ROS was observed in both cell lines and senescence models, compared to control untreated cells, (Fig. 2). In SIPS model, cells showed significantly elevated ROS levels straight after exposure to H₂O₂, as well as after cell recovery for 72 hours. These results indicate that oxidative stress acts as a mechanism accelerating senescence in both PIIPS and SIPS models, and that cellular senescence is not presented as an instant response to oxidative stress but may be observed at a later stage following exposure to the oxidant.

Transcriptomic signature of PIIPS and SIPS models in HFL1 and BJ cells

In the current study, we performed RNA sequencing to investigate transcriptomic alterations and identify commonly affected genes across the different cell lines and senescence models. Results from differential gene expression and enrichment analysis are collectively presented in Fig. 3 and Supplementary Fig. S1-S5. Samples of the same group (treated/senescent or control/non-senescent) showed a good correlation and clustered together with the only exception of one replicate from SIPS BJ cells (sample ID 1151) (Supplementary Fig. S1). Senescence onset caused mainly downregulation of gene expression, in all comparison groups, while bortezomib, as senescence stimulant, caused more profound alterations and a stronger phenotype compared to H₂O₂, in both fibroblast lines (Fig. 3a & Supplementary Fig. S2). Multidimensional scaling (MDS), a dimensionality reduction method, indicated that samples were clearly separated in the 2D space based on the fibroblast strain and bortezomib as senescence inducer (Fig. 3b), further demonstrating that transcriptomic signatures in senescence are mainly dependent upon the cell line and, in part, senescence type.

A total of 231 altered transcripts were common across the senescence models and cell lines tested, of which 217 were downregulated and 14 were upregulated (Supplementary Fig. S3 & Tables S3-S4). Commonly downregulated genes included BIRC5, HMGB1, HMGB2, LMNB1, ELN, MMP11, MKI67, COL12A1, COL14A1, COL15A1, COL16A1, COL1A1, COL3A1 and other cell cycle, DNA replication and extracellular matrix-related genes, while upregulated transcripts included MMP1, MMP3 and GDF-15. When applying less stringent thresholds (FC > = 1.5 or < = 0.67), we could identify 35 shared upregulated DEGs (Supplementary Tables S4-5), among which was CDKN1A which encodes for p21. Principal Component Analysis (PCA) using only the 217 downregulated and 35 upregulated DEGs that were conserved across the four senescence models revealed that senescent and non-senescent samples remained separated and were slightly influenced by the senescence inducer or the cell line. (Supplementary Fig. S3). This subset of genes is implicated in processes related to the cell cycle, cell cycle checkpoints, p21 and p53 signalling, cytoskeletal changes occurring during cell mitosis, ECM degradation and cellular senescence, as indicated in Supplementary Fig. S5.

Enriched pathways from DEGs identified in PIIPS and SIPS models

GSEA was carried out to identify the pathways that were enriched in SIPS and PIIPS and conserved across the cell lines and senescence inducers. A heatmap of the 26 most significant up- and down-regulated KEGG pathways of all comparison groups based on their median rank values is presented in Fig. 3c. All significantly enriched pathways from the 40 most enriched KEGG pathways for each group, categorized and separated as downregulated or upregulated, are demonstrated in Supplementary Fig. S4, along with their adjusted p values. Pathways and processes related to senescence such as “Cell cycle”, “DNA replication”, “oocyte meiosis”, “protein digestion and absorption” were downregulated in all groups (Fig. 3c). “Lysosome”, “synaptic vesicle cycle” and “Rheumatoid arthritis” represented some of the upregulated pathways in all senescence models, while ‘’Proteasome’’ was mainly upregulated in PIIPS and may represent a compensatory mechanism in response to proteasome inhibition.

Phospho-proteomic profile of PIIPS and SIPS reveal the signalling programmes employed with senescence onset

Signalling profiles of the different senescence groups were compared following analysis of the phosphorylated levels of 19 phosphoproteins by multiplex ELISA assays (Fig. 4a & Supplementary Table S1). Significantly increased levels of p-PRAS40, p-c-JUN and p-GSK3 and decreased levels of p-STAT3 were observed across all senescence models and cell lines. Phosphorylation of MEK1 and p53 was specifically activated by PIIPS and SIPS, respectively. No change in phosphorylation levels was found to be dependent upon cell type only. Other phosphoproteins with altered phosphorylation profile across one or more models or cell lines included ERK1/2, p38, CREB1, Smad3, HSPB1 and AKT1.

The levels of an additional set of intracellular proteins implicated in prosurvival pathways were also analysed, including survivin, Bcl-2, BcL-xL and caspase 3. Survivin was found consistently downregulated in both cell types and senescence inducers. On the contrary, the anti-apoptotic factor BcL-xL was significantly upregulated in all senescence models, while caspase 3 and Bcl-2 levels remained unchanged in most cases. Analysis of p21 levels was used as positive control throughout this experiment. As expected, p21 levels showed significant increase across all conditions tested.

Screening of extracellular protein markers reveal SASP profile and common markers shared among different senescence types

We then investigated whether the different senescence inducers caused similar alterations in the secretory profile of the cell lines tested. Based on previously characterised SASP-related proteins and literature evidence, the levels of 75 secreted proteins in the condition media from the different models were analysed by multiplex ELISA assays (Supplementary Table S2). These included cytokines, chemokines, MMPs, interleukins and growth factors.

Proteins with significantly altered levels in each senescence group compared to untreated control are presented in Fig. 4b. PIIPS cells demonstrated an intense secretory phenotype compared to SIPS cells, in both cell lines, based on the number and fold change of altered proteins. Several senescence related markers were secreted in the condition media following senescence onset, regardless of cell type or inducer, including VEGF, GM-CSF, MIF, GDF-15, IL-8 and GROa. Elevated levels of IL-6, IL-11, FGF-b, MMP7 and CCL20, as well as decreased levels of FST were observed in PIIPS models only. Finally, changes in the levels of ST2, MMP3 and VCAM1 were exclusive to HFL1 cells, and decrease in the PROK1 levels exclusive to BJ cells irrespective of the senescence type. All other changes observed were specific to a senescence type and cell line.

Validation of shared markers from PIIPS and SIPS in Replicative senescence models for the development of multiplex assay panels

We next investigated if the senescence markers shared among the four cellular senescence models are also conserved in replicative senescence (RS) models. For RS model establishment, early passage HFL1 and BJ cells were grown and sub-cultured at optimal conditions. Cells initially divided every 2–3 days until they reached RS, at passage 45 or more, when they ceased proliferation and exhibited an enlarged flattened morphology when observed microscopically. The senescence phenotype was evaluated in RS cells, by assessing SA-β-Gal activity and p21 levels (Fig. 5a-b). Late passage (P > 45) HFL1 and BJ cells showed a significantly increased SA-β-Gal activity compared to early passage cells (P < 18), with more than 70% of cells stained positive (Fig. 5a). p21 levels showed an 11-fold and 3-fold level increase in RS HFL1 and BJ cells, respectively (Fig. 5b). We also sought to examine if the pre-redox state acquired in PIIPS and SIPS was also observed in RS models. HFL1 at P49 and BJ cells at P47 demonstrated elevated intracellular ROS levels (Fig. 5c). Such alterations were of a similar level to those observed in SIPS models, or even more profound for RS HFL1 cells, further depicting the relationship between oxidative damage and replicative senescence.

Out of the 6 common signalling proteins and intracellular pro-survival markers identified in PIIPS and SIPS models, we identified four proteins, namely, p-cJun, Bcl-xL, Survivin and p-STAT3 which, together with p21, exhibited significant alterations in RS HFL1 and BJ cells compared to control early passage cells (Fig. 6a). Significantly increased p-GSK3 levels were observed in RS HFL1 cells but not BJ cells whilst no significant changes were observed for p-PRAS40 (Supplementary Fig. S6). Similarly, analysis of the secreted proteins in culture supernatants from RS HFL1 and BJ cells using multiplex ELISA identified increased levels of IL-8, GROa, GM-CSF, MIF and GDF-15 consistent with PIIPS and SIPS profiles (Fig. 6b), while VEGF secretion did not show significant alterations (Supplementary Fig. S6)

Senescence phenotype is highly dynamic and its heterogeneity has been attributed to several factors including cell type, senescence stimulant, and recently, time-dependent changes following senescence induction^21,33. In the current study, we performed RNA-seq analysis to investigate senescence heterogeneity and describe the transcriptomic signature of PIIPS and SIPS in two cell lines. Multiplex protein analysis was also performed to delineate the signalling footprint and SASP profile of PIIPS and SIPS models, with the aim to identify commonly altered proteins across all models. We identified shared genes and pathways associated with senescence onset that were conserved across all senescence models, together with common signalling mediators and extracellular markers that were further validated in RS models. Finally, we propose the development of a panel consisting of four intracellular and four extracellular proteins, that will enable their simultaneous assessment and monitoring of the senescence phenotype.

Transcriptome profiling of different senescence types in various models has become a powerful tool for identifying senescence-related genes and pathways or interrogate candidate senolytics^{21,23, 34–38}. Although several types of senescence have been analysed at a transcriptome level, the transcriptomic signature of PIIPS has not yet been delineated. Thus, we performed transcriptomic analysis in PIIPS along with SIPS models of two fibroblast strains. PIIPS models exhibited a more profound effect on the transcriptome of the two fibroblast cell lines tested compared to SIIPS, suggesting that different or additional mechanisms may be activated to promote and maintain senescence. Cell type also had a strong influence on transcriptomic profiles which is in accordance with previous studies, comparing senescent fibroblasts of different origin or fibroblasts and endothelial cells^21,23. While these data reflect the inherent complexity of senescence mechanisms, enrichment analysis showed that the affected processes are similar across the senescence models tested and are related to loss of protein homeostasis and genomic instability which represent two of the hallmarks of aging¹. These data are in accordance with previous studies from several strains of RS fibroblasts^23,24.

A total of 217 downregulated and 14 upregulated transcripts were identified to be common among PIIPS and SIPS groups of HFL1 and BJ cells. This subset of genes is presumably related to senescence onset independently of senescence inducer or fibroblast strain. The most prevalent commonly upregulated genes included MMP1 and MMP3, well described MMPs that act in different senescence types and degrade collagen and ECM^39–42. MMPs are established SASP factors secreted in senescence state with a potential tumorigenic function ^43,44. ECM environment has been reported to play a crucial role in tissue homeostasis, but also to control age-related diseases associated with dysregulated matrix remodelling. ECM remodelling is displayed either as aberrant ECM deposition observed in fibrotic diseases and cancer, or as aberrant ECM degradation observed in osteoarthritis and COPD^45–49. The consistent upregulation of MMP3 transcripts was also translated to elevated protein levels in culture supernatants of HFL1 but not BJ cells. Inconsistencies in transcriptomic and protein data between HFL1 and BJ cells may reflect tissue-dependent mechanisms as seen in previously reported data⁵⁰. Several other members of MMP family transcripts were differentially expressed, either increased or reduced, further indicating the disruption of homeostatic mechanisms occurring in senescence and SASP^41,51.

Apart from MMPs dysregulation, senescent cells burden ECM structure by downregulating genes encoding for major ECM components, such as Collagens⁵². In our dataset, COL12A1, COL14A1, COL15A1, COL16A1, COL1A1, COL3A1 significantly decreased with senescence onset in all four groups, while genes COL5A3, COL6A3, COL5A1, COL8A2, COL4A1, COL1A2, COL5A2, COL11A1, COL4A2, COL27A1 were reduced in PIIPS HFL1 and BJ cells (Supplementary Table S3). Decrease of collagen types I and III has been associated with chronological aging and the disruption of structural integrity observed in older skin^43,44. Our data are in line with previous studies that reported a decline in ECM components in RS and SIPS⁵³, and revealed similar profile in PIIPS. Consistent results were also obtained for LMNB1 and ELN transcripts, which were diminished in all senescence models, in accordance with previous data highlighting the role of Lamin B1 loss as a senescence biomarker in vitro and in vivo^8,54,55, and underlining the implication of ELN senescence-associated downregulation in tissue dysfunction^56,57.

Common senescence features and pathways were also confirmed at the protein level when a targeted list of 100 protein markers representing signalling pathways, pro-survival markers and secreted inflammatory molecules was interrogated in PIIPS and SIPS models. Common MAPK stressor and proliferation signaling cascades were activated as a result of ROS formation observed across all senescence models. Some heterogeneity was evident on the mediators of these pathways, for example, p-p38 MAPK levels were increased in all but the SIPS BJ cells, pERK1/2 was elevated in all but the PIIPS HFL1 cells and p-MEK1 and p-p53 levels were increased in a PIIPS-specific manner. Our results are in line with previous reports suggesting ERK1 and p38 MAPK pathways play a crucial role in cell cycle arrest and SASP regulation in senescent cells^58,59. These stressor pathways also regulate the aberrant activation of CDK inhibitors including the classical senescence marker p21 which was overexpressed in all groups at both transcript and protein levels^24,31,60. Among the affected mediators, c-JUN and STAT3 were also confirmed to be significantly altered in RS models suggesting a more conserved role in senescence regulation. c-JUN is a component of the AP-1 transcription factor, a downstream effector of the SAPK/JNK MAPK stressor pathway and has been reported to enhance SASP by promoting the expression of MMPs^61,62. STAT3 is another transcription factor acting downstream of SAPK/JNK, ERK and JAK kinases with a yet inconclusive role in senescence. In line with our results, inhibition of Stat3-Y705 phosphorylation was reported to accelerate senescence in triple negative breast cancer and melanoma cells⁶³. Other studies, however, have reported enhanced STAT3 activity in IL-6- and H₂O₂-induced senescent lung fibroblasts and attenuation of senescence features upon STAT3 inhibition with a small molecule inhibitor^64,65 suggesting different signalling cascades may be activated depending on the inducer or the cell type. Additional work would be needed before STAT3 activity could be regarded as a conserved senescence feature.

In addition to c-JUN and STAT3, Bcl-XL and survivin, two proteins of the anti-apoptotic pathway, had significantly altered levels in PIIPS, SIPS and RS models. The increased Bcl-XL levels identified in our study support previous data that highlight the activation of the anti-apoptotic machinery during senescence^{34, 66–68}. The decrease in the levels of survivin were consistent with transcriptomic results where BIRC, the gene encoding for survivin, was downregulated in PIIPS and SIPS. Survivin, a member of the inhibitor of apoptosis protein (IAP) family, is required for cell division but not for cell survival^69,70, while its upregulation seems to be responsible for senescence escape, reentry in the cell cycle and subsequent cell viability⁷¹. Several studies have associated chemoresistance with survivin overexpression and its depletion or pharmacological inhibition has been shown to reduce apoptosis, cause cell growth arrest and increase cellular senescence^72,73.

Analysis of extracellular proteins revealed a diverse SASP profile with several cytokines and growth factors being affected by one or more inducers in the two cell lines tested. Our analysis identified a core set of secreted markers that were conserved between PIIPS and SIPS in both cell lines and included GDF-15, GROa, IL-8, GM-CSF, MIF and VEGF. Interestingly, these proteins, with the exception of VEGF, were also shown to be secreted in RS models further confirming the presence of commonly activated pathways and common features of senescence development and maintenance.

GDF-15 is a cytokine with a known dual role in the aging process acting as a stress responsive cytokine in age-related pathologies, but also as an anti-inflammatory factor that inhibits T cell activation and tissue fibrosis^74,80. Previous studies have also shown correlation between GDF-15 levels and biological age and senescence^33,75. The role of GROa as SASP component has also been reported in replicative and oncogene-induced senescence, where it acts in an autocrine and paracrine manner through continuous activation of NF-Kb^19,76. GROa overexpression has been reported in fibroblasts following induction of senescence by IR, RAS, or ATV³³. Similar to GROa, IL-8 promotes the creation of inflammatory network^77,78 and maintains senescence and SASP via its CXCR2-binding activity⁷⁹. MIF is considered a proinflammatory cytokine that plays a key role in inflammation related to joint contracture, cellular senescence and ageing⁸⁰. However, its rejuvenating role in aged mesenchymal stem cells is contradictory with its role in senescence and ageing^81,82. Finally, GM-CSF is thought to play a role in the immune system's response to senescent cells although its exact role in the SASP is not fully understood. GM-CSF can activate and recruit immune cells, such as macrophages, to sites of inflammation and tissue damage⁸³. It has been previously implicated in RS fibroblasts²⁰ and is at least partially regulated by the stressor pathway p38MAPK⁵⁹. Collectively, these results demonstrate the potential utility of GDF-15, GROa, IL-8 and GM-CSF as markers of senescence.

To conclude, our study describes for the first time the transcriptomic and targeted proteomic profile of PIIPS models in comparison to SIPS and RS in two cell lines and highlights the common senescence signatures acquired in response to the different types of stresses. Overall, our data suggests that simultaneous analysis of the levels of p21, p-c-JUN, survivin and BCL-xL in cellular lysates and SASP factors GDF-15, IL-8, GM-CSF and GROa in culture supernatants can aid the identification of a senescent phenotype across a range of senescence models. More importantly, multiplex analysis of these protein markers can provide a quick screening tool to monitor the efficacy of therapeutic approaches that target senescent cells, senescence pathways and/or SASP factors to delay senescence onset.

Cell cultures

Normal human lung HFL1 and newborn skin BJ fibroblasts were acquired from ATCC (CCL-153 and CRL-2522, respectively). Both cell lines were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) with sodium pyruvate (PanBiotech, #P04-03590), supplemented with 10% Fetal Bovine Serum (FBS) (Gibco, #10270106), 1% penicillin -streptomycin (Life Technologies, #15140-122). Cells were cultured and maintained following manufacturer’s instructions at 37°C in a humidified chamber with 5% CO₂. They were sub-cultured when they reached approximately 80% confluency. For subculturing, culture medium was aspirated, and cells were washed with Ca++/Mg + + free Phosphate-buffered Saline (PBS, pH 7.4) (Gibco, #10010015). Cells were incubated with 0.05% trypsin-0.02% EDTA (Life Technologies, #15400054) in PBS to detach, resuspended in fresh media and seeded in 75cm² flasks. Both cell lines were used within Passage (P) 18 for PIIPS and SIPS models.

Cell viability assay

HFL1 and BJ cells were sub-cultured in flat bottom 96-well plates at different cell densities depending on the treatment. Cells were treated with various bortezomib concentrations and cell viability was assessed using CellTiter-Glo® (Promega, #G7570) Luminescent Cell Viability Assay as per manufacturer’s instructions at various time points; 12h, 24h, 48, 5 days, 10 days, 13 days for HFL1 cells, 5, 7 and 10 days for BJ cells. Similarly, cell viability was assessed following exposure to H₂O₂ for up to 2 hours and recovery in complete culture media for 24, 48 and 72 hours. Luminescence signal was measured in VarioskanTM LUX multi-mode microplate reader (Thermo Fisher Scientific). The experimental data were then extrapolated onto a 4-parameter logistic regression model in GraphPad Prism 9.0 and IC₅₀ doses were estimated. Treatments doses up to the IC₅₀ values were used for subsequent experiments.

Proteasome inhibition-induced premature senescence (PIIPS)

Early passage HFL1 and BJ cells (P13-P18) were seeded in 60 x 15mm petri dishes and treated with bortezomib (Calbiochem, #PS-341) concentrations up to IC₅₀ doses. Several treatment durations were tested for optimisation purposes. Treatment media was renewed every day to ensure bortezomib’s activity was the same throughout treatment duration. Cells were sub-cultured in new petri dishes when they reached approximately 80% confluency to prevent cell growth arrest due to cell contact.

Stress-induced premature Senescence (SIPS)

HFL1 and BJ cells were seeded in 6-well plates and incubated with H₂O₂ (Honeywell Fluka, #31642) dissolved in culture media at concentrations up to IC₅₀ doses. Several treatment and recovery durations were tested for optimisation purposes. Treatment media was removed, and cells were washed and left to recover in complete culture media at 37°C and 5% CO₂. Cells presented a senescence phenotype following recovery, at which point all subsequent experiments were carried out.

Replicative Senescence (RS)

HFL1 and BJ cells were cultured and allowed to proliferate until they reached the Hayflick limit. 2 x 10⁶ cells were seeded in 75cm² flask after every trypsinization. Early passage cells had a doubling time of 2–3 days, while late passage cells (P > 45) did not divide for approximately 1 month. Samples from RS cells were collected for validation analysis.

Senescence-associated-beta-galactosidase staining

The senescence-associated-β-galactosidase (SA-β-Gal) staining was detected with Cellular Senescence assay kit (Merck Millipore, #KAA002RF) according to the manufacturer’s instructions. Briefly, cells were washed with PBS and fixed at room temperature for 15 minutes. Following removal of fixing solution, two washing steps with PBS were performed. SA-β-Gal detection solution was then added, and cells were incubated at 37°C overnight, in the absence of CO₂ and protected from light. Next day, SA-β-Gal detection solution was aspirated, and blue stained SA-β-Gal positive cells were counted using an Olympus CK30/CK40 Phase Contrast Microscope. Positive cells were expressed as a percentage of the total number of cells measured in each microscope field. At least 5 random fields were observed with at least five biological replicates for each condition.

Sample collection: Cell lysis and supernatant collection

Cells were cultured in 6-well plates and 35mm petri dishes for PIIPS and SIPS, respectively. For secretome analysis, conditioned media was collected and stored at -80°C. For phoshoproteomic and intracellular protein analysis, plates and petri dishes were placed on ice, washed with ice cold PBS (Gibco, #10010015) and incubated with Lysis buffer (Protavio, #PR-LYSB) supplemented with 100x protease inhibitors (Protavio, #PR-PIC) and 50x PMSF (Sigma, #93482) for 1 minute. Cells were then detached from the plate using a cell scraper, collected in Eppendorf tubes and cell debris was removed by centrifugation at 12,000xg for 10 minutes at 4°C. Supernatants were collected, snap-frozen in liquid nitrogen and stored at -80°C for protein quantification and subsequent multiplex analysis.

Total protein content estimation

Total protein quantification was carried out with Pierce^™ BCA Protein Assay Kit (Pierce, Thermo Scientific, #23225) following manufacturer’s instructions and absorbance was measured at 562 nm using the absorbance plate reader Tecan INFINITE F50 with Magellan Software for INFINITE F50.

Measurement of Intracellular ROS Levels

Intracellular Reactive oxygen species (ROS) levels were assessed using CM-H2DCFDA fluorescence dye (Invitrogen, #C6827), in PIIPS, SIPS and RS samples. Treatment and culture media were removed, cells were washed with PBS and incubated with CM-H2DCFDA reagent diluted in anhydrous DMSO at a target concentration of 10µmol/L. Incubation took place at 37°C for 45 minutes, protected from light. Fluorescence signal was measured on a Varioskan LUX Multimode Microplate Reader (Thermo Scientific), with peak excitation wavelength at 490nm and peak emission wavelength at 530 nm. After removing the fluorescent dye, cells were washed with PBS, lysed following cell lysis protocol described earlier, and total protein content was estimated. Fluorescence signal was normalised to total protein concentration for each sample and results were expressed as relative fluorescence units per mg protein (RFU/mg).

Multiplex protein analysis

For intracellular protein analysis, the concentration of cell lysates was adjusted to 500µg/mL prior to sample analysis. Custom xMAP Luminex assays (Protavio, Athens, Greece) developed as previously described⁸⁴ were used to analyse phosphorylated or total levels of 25 intracellular protein targets (listed in Supplementary Table S1). Data are presented as Median Fluorescence intensities (MFI).

For secretome analysis, conditioned media were collected following senescence induction (PIIPS, SIPS and RS) and stored at -80°C until further analysis. A custom-developed xMAP Luminex assay panel (Protavio Ltd, Athens, Greece) was used to assess the levels of 75 extracellular protein targets (listed in Supplementary Table S2). Data are presented as MFIs.

RNA isolation and RNA sequencing analysis

Cells were cultured in petri dishes for total RNA extraction (3 biological replicates per condition per cell line) and processed accordingly until they acquired PIIPS and SIPS. Cells were lysed using Trizol reagent (Invitrogen, #15596026) according to manufacturer’s instruction. Briefly, cells were incubated with Trizol at room temperature for 30 minutes and transferred to Eppendorf tubes. Samples were incubated with chloroform (Sigma, #C2432) for 15 minutes. Phase separation was performed with sample centrifugation at 12,000 xg for 15 minutes at 4°C. The aqueous phase containing the RNA was collected, incubated with isopropanol (Honeywell Fluka, # 33539) and centrifuged at 12,000 ×g for 10 minutes at 4°C. Supernatant was discarded and RNA precipitate formed a white gel-like pellet. The RNA pellet was washed with 75% Ethanol, air-dried, resuspended in RNase-free water, and stored at -80°C until further use. To determine RNA yield and sample purity, absorbance was measured at 260nm and 280nm in GeneQuant II RNA/DNA calculator (Pharmacia Biotech) and a ratio greater than 2, was required to qualify samples for further use.

Libraries were prepared using the QuantSeq 3’ mRNA-Seq Library prep Kit (Lexogen, Vienna, Austria) according to the manufacturer’s instructions. Briefly, up to 500 ng total RNA was subjected to first-strand synthesis using oligo dT primers followed by RNA removal and second strand synthesis using random primers containing Illumina-compatible linker sequences. The resulting double-stranded cDNA was purified with magnetic beads and amplified using the QuantSeq 3’mRNA Library prep kit FWD for Illumina (Lexogen, Vienna, Austria). During the amplification process, Lexogen i7 6-nucleotide indices and adapter sequences required for cluster generation were incorporated into the cDNA fragments. The amplified library was purified with magnetic beads, quantified using a Qubit-4 fluorometer (Invitrogen) and analyzed on an Agilent 2100 bioanalyzer, according to manufacturers' instructions. Finished libraries were diluted to the appropriate concentration, multiplexed, denatured, and sequenced on an Illumina NextSeq 550, according to manufacturer's instructions, to produce 75 bp single-end reads.

The quality of the FASTQ files was evaluated following the recommendations of FastQC (available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc). To trimming and map the reads to the GRCh38 human reference genome, we used Bbduk (Available online at: https://jgi.doe.gov/data-and-tools/software-tools/bbtools/bb-tools-user-guide/bbduk-guide/) and STAR⁸⁵, respectively, as part of the recommended by Lexogen and BlueBee pipeline. Genome tracks were constructed and manually inspected. The raw BAM files were summarised to create a 3'UTR read counts table, using the Bioconductor package Genomic Ranges, through the metaseqR2 wrapper⁸⁶ in R.

In the final table, each row represented a gene, each column represented a sample, and each cell contained the corresponding read counts associated with that gene and sample. The gene counts table was normalized to account for any inherent systematic or experimental biases, such as sequencing depth, gene length, or GC content bias, using the Bioconductor package EDASeq⁸⁷. The output of the normalization algorithm was a table with normalized counts, which was further used for differential expression analysis with statistical algorithms.

Differential gene expression analysis

The resulting gene counts table was subjected to differential expression analysis to compare senescent versus non-senescent groups for each cell line (HFL1 and BJ cells) and senescence stimulant (bortezomib and H₂O₂). This analysis was conducted using the Bioconductor packages DESeq⁸⁸, edgeR⁸⁹, NOISeq⁹⁰, limma⁹¹ and NBPSeq⁹². To combine the statistical significance from multiple algorithms and perform meta-analysis, we applied the PANDORA weighted P-value across results method using the metaseqR2 R package. Genes with a log₂ fold change ≥ 1 were considered significantly upregulated, while genes with log₂ fold change ≤-1 were considered significantly downregulated, with a meta p value < 0.05. Additionally, we created a Multidimensional scaling (MDS) plot and heatmaps of Differentially Expressed Genes (DEGs) for all samples using the same package to demonstrate the correlation of samples from different conditions and how they cluster together according to their gene expression profile. These heatmaps also depict potential clusters of co-expressed genes and reveal any experimental quality issue within the replicates of the same group, since hierarchical clustering was applied.

Gene Set Enrichment Analysis (GSEA)

The Bioconductor package Ensemble of Gene Set Enrichment Analyses (EGSEA) was used to perform GSEA and identify significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. EGSEA was used for each of the contrast groups and as a combination to show direct comparisons between them. This package performs Ensemble Gene Set Enrichment Analysis and uses the analysis results of twelve GSE algorithms in the literature to calculate collective significance scores for each gene set. These methods include: ora, globaltest, plage, safe, zscore, gage, ssgsea, roast, fry, PADOG, camera and GSVA. Genes that had at least 4 counts in each group in at least 60% of the samples and at least 10 counts in total in both control and treatment groups were included in this analysis. All 4 contrast groups were included in the analysis. Pathways were sorted by median rank and only those with a p value < 0.05 using Wilkinson test were considered significantly enriched.

Data presentation and statistical analysis

Data are presented as the means of at least three independent experiments ± SD. Statistical significance was assessed by unpaired, two-tailed Student’s t-test. Statistical analysis and data representation were performed using Microsoft excel, GraphPad Prism 9 (GraphPad Software) and R (R Core Team).

Acknowledgements

This work was supported by HealthAge which has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 812830. We also acknowledge support for part of this work by the project “The Greek Research Infrastructure for Personalised Medicine (pMedGR)” (MIS 5002802) which is implemented under the Action “Reinforcement of the Research and Innovation Infrastructure”, funded by the Operational Programme Competitiveness, Entrepreneurship and Innovation (NSRF 2014-2020) and co-financed by Greece and the European Union (European Regional Development Fund). N.T., CF and VP have been co-financed by the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation under the call Research-Create-Innovate B (Project code: T2EDK-01291) for the research and development of the multiplex assay panels.

Additional Information
The author(s) declare no competing interests.

Data availability

The RNA sequencing data have been deposited to the Gene Expression Omnibus (GEO) prior to manuscript submission with Accession Number GSE235768. To allow review of record GSE235768 while it remains private, the following secure token should be used: exankwiwtvophuh. The rest of the data generated during this study are graphically demonstrated in this article and its Supplementary Information files.

Author Contributions

The authors confirm contribution to the paper as follows: study conception: L.G.A.; study design: experiments: L.G.A., D.S.; xMAP assay development: D.S., V.P., N.T.; analysis and visualisation of results: D.S., C.F., K.M.V.; interpretation of results: D.S, L.G.A., N.T.; draft manuscript preparation: D.S.; manuscript review: L.G.A., N.T. All authors reviewed the results and approved the final version of the manuscript.

López-Otín, C., Blasco, M. A., Partridge, L., Serrano, M. & Kroemer, G. The Hallmarks of Aging. Cell 153, 1194–1217 (2013).
Hayflick, L. & Moorhead, P. S. The serial cultivation of human diploid cell strains. Experimental Cell Research 25, 585–621 (1961).
Nelson, G. et al. A senescent cell bystander effect: senescence‐induced senescence. Aging Cell 11, 345–349 (2012).
Lopes-Paciencia, S. et al. The senescence-associated secretory phenotype and its regulation. Cytokine 117, 15–22 (2019).
Neves, J., Demaria, M., Campisi, J. & Jasper, H. Of Flies, Mice, and Men: Evolutionarily Conserved Tissue Damage Responses and Aging. Developmental Cell 32, 9–18 (2015).
Tchkonia, T., Zhu, Y., van Deursen, J., Campisi, J. & Kirkland, J. L. Cellular senescence and the senescent secretory phenotype: therapeutic opportunities. J. Clin. Invest. 123, 966–972 (2013).
Muñoz-Espín, D. & Serrano, M. Cellular senescence: from physiology to pathology. Nat Rev Mol Cell Biol 15, 482–496 (2014).
En, A., Takauji, Y., Miki, K., Ayusawa, D. & Fujii, M. Lamin B receptor plays a key role in cellular senescence induced by inhibition of the proteasome. FEBS Open Bio 10, 237–250 (2020).
Loaiza, N. & Demaria, M. Cellular senescence and tumor promotion: Is aging the key? Biochimica et Biophysica Acta (BBA) - Reviews on Cancer 1865, 155–167 (2016).
Childs, B. G., Baker, D. J., Kirkland, J. L., Campisi, J. & Deursen, J. M. Senescence and apoptosis: dueling or complementary cell fates? EMBO Rep 15, 1139–1153 (2014).
Maciel-Barón, L. A. et al. Senescence associated secretory phenotype profile from primary lung mice fibroblasts depends on the senescence induction stimuli. AGE 38, 26 (2016).
Di Micco, R., Krizhanovsky, V., Baker, D. & d’Adda di Fagagna, F. Cellular senescence in ageing: from mechanisms to therapeutic opportunities. Nat Rev Mol Cell Biol 22, 75–95 (2021).
Dimri, G. P. et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc. Natl. Acad. Sci. U.S.A. 92, 9363–9367 (1995).
Krishnamurthy, J. et al. Ink4a/Arf expression is a biomarker of aging. J. Clin. Invest. 114, 1299–1307 (2004).
Liu, Y. et al. Expression of p16 ^INK4a in peripheral blood T-cells is a biomarker of human aging. Aging Cell 8, 439–448 (2009).
Melk, A. et al. Expression of p16INK4a and other cell cycle regulator and senescence associated genes in aging human kidney. Kidney International 65, 510–520 (2004).
Prata, L. G. P. L., Ovsyannikova, I. G., Tchkonia, T. & Kirkland, J. L. Senescent cell clearance by the immune system: Emerging therapeutic opportunities. Seminars in Immunology 40, 101275 (2018).
Lee, K.-A., Flores, R. R., Jang, I. H., Saathoff, A. & Robbins, P. D. Immune Senescence, Immunosenescence and Aging. Front. Aging 3, 900028 (2022).
Acosta, J. C. et al. Chemokine Signaling via the CXCR2 Receptor Reinforces Senescence. Cell 133, 1006–1018 (2008).
Coppé, J.-P. et al. Senescence-Associated Secretory Phenotypes Reveal Cell-Nonautonomous Functions of Oncogenic RAS and the p53 Tumor Suppressor. PLoS Biol 6, e301 (2008).
Hernandez-Segura, A. et al. Unmasking Transcriptional Heterogeneity in Senescent Cells. Curr Biol 27, 2652-2660.e4 (2017).
van Deursen, J. M. The role of senescent cells in ageing. Nature 509, 439–446 (2014).
Casella, G. et al. Transcriptome signature of cellular senescence. Nucleic Acids Research 47, 7294–7305 (2019).
Marthandan, S. et al. Conserved Senescence Associated Genes and Pathways in Primary Human Fibroblasts Detected by RNA-Seq. PLoS ONE 11, e0154531 (2016).
Chondrogianni, N. et al. Central Role of the Proteasome in Senescence and Survival of Human Fibroblasts. Journal of Biological Chemistry 278, 28026–28037 (2003).
Chondrogianni, N. & Gonos, E. S. Proteasome inhibition induces a senescence-like phenotype in primary human fibroblasts cultures. Biogerontology 5, 55–61 (2004).
Ukekawa, R., Maegawa, N., Mizutani, E., Fujii, M. & Ayusawa, D. Proteasome Inhibitors Induce Changes in Chromatin Structure Characteristic of Senescent Human Fibroblasts. Bioscience, Biotechnology, and Biochemistry 68, 2395–2397 (2004).
Torres, C., Lewis, L. & Cristofalo, V. J. Proteasome inhibitors shorten replicative life span and induce a senescent-like phenotype of human fibroblasts. J. Cell. Physiol. 207, 845–853 (2006).
Frippiat, C. et al. Subcytotoxic H2O2 Stress Triggers a Release of Transforming Growth Factor-β1, Which Induces Biomarkers of Cellular Senescence of Human Diploid Fibroblasts. Journal of Biological Chemistry 276, 2531–2537 (2001).
Frippiat, C., Dewelle, J., Remacle, J. & Toussaint, O. Signal transduction in H2O2-induced senescence-like phenotype in human diploid fibroblasts. Free Radical Biology and Medicine 33, 1334–1346 (2002).
Furukawa, A., Tada-Oikawa, S., Kawanishi, S. & Oikawa, S. H₂O₂ Accelerates Cellular Senescence by Accumulation of Acetylated p53 via Decrease in the Function of SIRT1 by NAD⁺ Depletion. Cell Physiol Biochem 20, 045–054 (2007).
Liu, D.-H. et al. Oncogenic ras Provokes Premature Cell Senescence. 59, 9 (2012).
Basisty, N. et al. A proteomic atlas of senescence-associated secretomes for aging biomarker development. PLoS Biol 18, e3000599 (2020).
Zhu, Y. et al. The Achilles’ heel of senescent cells: from transcriptome to senolytic drugs. Aging Cell 14, 644–658 (2015).
Crowe, E. P. et al. Changes in the Transcriptome of Human Astrocytes Accompanying Oxidative Stress-Induced Senescence. Front. Aging Neurosci. 8, (2016).
Herranz, N. et al. mTOR regulates MAPKAPK2 translation to control the senescence-associated secretory phenotype. Nat Cell Biol 17, 1205–1217 (2015).
Ogrodnik, M. et al. Whole‐body senescent cell clearance alleviates age‐related brain inflammation and cognitive impairment in mice. Aging Cell 20, (2021).
Chen, Y. et al. Transcriptome and lipidome profile of human mesenchymal stem cells with reduced senescence and increased trilineage differentiation ability upon drug treatment. Aging 13, 9991–10014 (2021).
Millis, A. J. T., McCue, H. M., Kumar, S. & Baglioni, C. Metalloproteinase and TIMP-1 gene expression during replicative senescence. Experimental Gerontology 27, 425–428 (1992).
Krizhanovsky, V. et al. Senescence of Activated Stellate Cells Limits Liver Fibrosis. Cell 134, 657–667 (2008).
Coppé, J.-P. et al. A Human-Like Senescence-Associated Secretory Phenotype Is Conserved in Mouse Cells Dependent on Physiological Oxygen. PLoS ONE 5, e9188 (2010).
Pitiyage, G. N. et al. Senescent mesenchymal cells accumulate in human fibrosis by a telomere-independent mechanism and ameliorate fibrosis through matrix metalloproteinases: Senescent fibroblast accumulation may ameliorate human fibrosis. J. Pathol. 223, 604–617 (2011).
Varani, J. et al. Decreased Collagen Production in Chronologically Aged Skin. The American Journal of Pathology 168, 1861–1868 (2006).
Farage, M. A., Miller, K. W. & Maibach, H. I. Degenerative Changes in Aging Skin. in Textbook of Aging Skin (eds. Farage, M. A., Miller, K. W. & Maibach, H. I.) 15–30 (Springer Berlin Heidelberg, 2017). doi:10.1007/978-3-662-47398-6_4.
Estany, S. et al. Lung fibrotic tenascin-C upregulation is associated with other extracellular matrix proteins and induced by TGFβ1. BMC Pulm Med 14, 120 (2014).
Shi, Y. et al. A small molecule promotes cartilage extracellular matrix generation and inhibits osteoarthritis development. Nat Commun 10, 1914 (2019).
Zhen, G. & Cao, X. Targeting TGFβ signaling in subchondral bone and articular cartilage homeostasis. Trends in Pharmacological Sciences 35, 227–236 (2014).
Hunter, D. J. Pharmacologic therapy for osteoarthritis—the era of disease modification. Nat Rev Rheumatol 7, 13–22 (2011).
Sand, J. M. B. et al. Accelerated extracellular matrix turnover during exacerbations of COPD. Respir Res 16, 69 (2015).
Hudgins, A. D. et al. Age- and Tissue-Specific Expression of Senescence Biomarkers in Mice. Front. Genet. 9, 59 (2018).
Kletsas, D. Human lung fibroblasts prematurely senescent after exposure to ionizing radiation enhance the growth of malignant lung epithelial cells in vitro and in vivo. Int J Oncol (2011) doi:10.3892/ijo.2011.1132.
Gerasymchuk, M., Robinson, G. I., Kovalchuk, O. & Kovalchuk, I. Modeling of the Senescence-Associated Phenotype in Human Skin Fibroblasts. IJMS 23, 7124 (2022).
Schnabl, B. Replicative senescence of activated human hepatic stellate cells is accompanied by a pronounced inflammatory but less fibrogenic phenotype. Hepatology 37, 653–664 (2003).
Freund, A., Laberge, R.-M., Demaria, M. & Campisi, J. Lamin B1 loss is a senescence-associated biomarker. MBoC 23, 2066–2075 (2012).
Matias, I. et al. Loss of lamin‐B1 and defective nuclear morphology are hallmarks of astrocyte senescence in vitro and in the aging human hippocampus. Aging Cell 21, (2022).
Levi, N., Papismadov, N., Solomonov, I., Sagi, I. & Krizhanovsky, V. The ECM path of senescence in aging: components and modifiers. FEBS J 287, 2636–2646 (2020).
Chilosi, M., Carloni, A., Rossi, A. & Poletti, V. Premature lung aging and cellular senescence in the pathogenesis of idiopathic pulmonary fibrosis and COPD/emphysema. Translational Research 162, 156–173 (2013).
Anerillas, C., Abdelmohsen, K. & Gorospe, M. Regulation of senescence traits by MAPKs. GeroScience 42, 397–408 (2020).
Freund, A., Patil, C. K. & Campisi, J. p38MAPK is a novel DNA damage response-independent regulator of the senescence-associated secretory phenotype: p38 regulates the senescence secretory phenotype. The EMBO Journal 30, 1536–1548 (2011).
Chen, Q. M., Prowse, K. R., Tu, V. C., Purdom, S. & Linskens, M. H. K. Uncoupling the Senescent Phenotype from Telomere Shortening in Hydrogen Peroxide-Treated Fibroblasts. Experimental Cell Research 265, 294–303 (2001).
Fitsiou, E., Pulido, T., Campisi, J., Alimirah, F. & Demaria, M. Cellular Senescence and the Senescence-Associated Secretory Phenotype as Drivers of Skin Photoaging. Journal of Investigative Dermatology 141, 1119–1126 (2021).
Dasgupta, J. et al. Reactive oxygen species control senescence-associated matrix metalloproteinase-1 through c-Jun-N-terminal kinase. J. Cell. Physiol. 225, 52–62 (2010).
Kojima, H., Inoue, T., Kunimoto, H. & Nakajima, K. IL-6-STAT3 signaling and premature senescence. JAK-STAT 2, e25763 (2013).
Kojima, H., Kunimoto, H., Inoue, T. & Nakajima, K. The STAT3-IGFBP5 axis is critical for IL-6/gp130-induced premature senescence in human fibroblasts. Cell Cycle 11, 730–739 (2012).
Waters, D. W. et al. STAT3 Regulates the Onset of Oxidant-induced Senescence in Lung Fibroblasts. Am J Respir Cell Mol Biol 61, 61–73 (2019).
Chang, J. et al. Clearance of senescent cells by ABT263 rejuvenates aged hematopoietic stem cells in mice. Nat Med 22, 78–83 (2016).
Zhu, Y. et al. Identification of a novel senolytic agent, navitoclax, targeting the Bcl‐2 family of anti‐apoptotic factors. Aging Cell 15, 428–435 (2016).
Rochette, P. J. & Brash, D. E. The Achilles’ heel of senescent cells: from transcriptome to senolytic drugs. Mechanisms of Ageing and Development 129, 207–214 (2008).
Lens, S. M. A. Survivin is required for a sustained spindle checkpoint arrest in response to lack of tension. The EMBO Journal 22, 2934–2947 (2003).
Yang, D., Welm, A. & Bishop, J. M. Cell division and cell survival in the absence of survivin. Proc. Natl. Acad. Sci. U.S.A. 101, 15100–15105 (2004).
Wang, Q. et al. Survivin and escaping in therapy-induced cellular senescence. Int. J. Cancer 128, 1546–1558 (2011).
Han, T.-L. et al. Depletion of Survivin suppresses docetaxel-induced apoptosis in HeLa cells by facilitating mitotic slippage. Sci Rep 11, 2283 (2021).
Shi, X. et al. GDP366, a novel small molecule dual inhibitor of survivin and Op18, induces cell growth inhibition, cellular senescence and mitotic catastrophe in human cancer cells. Cancer Biology & Therapy 9, 640–650 (2010).
Chung, H. K. et al. GDF15 deficiency exacerbates chronic alcohol- and carbon tetrachloride-induced liver injury. Sci Rep 7, 17238 (2017).
Park, H., Kim, C.-H., Jeong, J.-H., Park, M. & Kim, K. S. GDF15 contributes to radiation-induced senescence through the ROS-mediated p16 pathway in human endothelial cells. Oncotarget 7, 9634–9644 (2016).
Acosta, J. C. et al. A complex secretory program orchestrated by the inflammasome controls paracrine senescence. Nat Cell Biol 15, 978–990 (2013).
Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an Interleukin-Dependent Inflammatory Network. Cell 133, 1019–1031 (2008).
You, K. et al. Moderate hyperoxia induces senescence in developing human lung fibroblasts. American Journal of Physiology-Lung Cellular and Molecular Physiology 317, L525–L536 (2019).
Kwon, J. H. et al. Senescence-Associated Secretory Phenotype Suppression Mediated by Small-Sized Mesenchymal Stem Cells Delays Cellular Senescence through TLR2 and TLR5 Signaling. Cells 10, 63 (2021).
Zhang, Y. et al. Macrophage migration inhibitory factor activates the inflammatory response in joint capsule fibroblasts following post-traumatic joint contracture. Aging 13, 5804–5823 (2021).
Zhang, Y. et al. Macrophage migration inhibitory factor rejuvenates aged human mesenchymal stem cells and improves myocardial repair. Aging 11, 12641–12660 (2019).
Xia, W., Zhang, F., Xie, C., Jiang, M. & Hou, M. Macrophage migration inhibitory factor confers resistance to senescence through CD74-dependent AMPK-FOXO3a signaling in mesenchymal stem cells. Stem Cell Res Ther 6, 82 (2015).
Hamilton, J. A. GM-CSF in inflammation and autoimmunity. Trends in Immunology 23, 403–408 (2002).
Poussin, C. et al. The species translation challenge—A systems biology perspective on human and rat bronchial epithelial cells. Sci Data 1, 140009 (2014).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Fanidis, D. & Moulos, P. Integrative, normalization-insusceptible statistical analysis of RNA-Seq data, with improved differential expression and unbiased downstream functional analysis. Briefings in Bioinformatics 22, bbaa156 (2021).
Risso, D., Schwartz, K., Sherlock, G. & Dudoit, S. GC-Content Normalization for RNA-Seq Data. BMC Bioinformatics 12, 480 (2011).
Anders, S. & Huber, W. Differential expression analysis for sequence count data. (2010).
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR : a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Tarazona, S., García-Alcalde, F., Dopazo, J., Ferrer, A. & Conesa, A. Differential expression in RNA-seq: A matter of depth. Genome Res. 21, 2213–2223 (2011).
Bioinformatics and computational biology solutions using R and Bioconductor. (Springer Science+Business Media, 2005).
Di, Y., Schafer, D. W., Cumbie, J. S. & Chang, J. H. The NBP Negative Binomial Model for Assessing Differential Gene Expression from RNA-Seq. Statistical Applications in Genetics and Molecular Biology 10, (2011).

No competing interests reported.

SupplementaryFiguresandTables.pdf

Download PDF

Version 1

posted

You are reading this latest preprint version

Conserved transcriptomic signatures and protein markers in cellular senescence models

Status:

Version 1

Abstract

Figures

Introduction

Results

PIIPS and SIPS cause a pre-redox state with elevated ROS levels

Transcriptomic signature of PIIPS and SIPS models in HFL1 and BJ cells

Enriched pathways from DEGs identified in PIIPS and SIPS models

Phospho-proteomic profile of PIIPS and SIPS reveal the signalling programmes employed with senescence onset

Discussion

Materials and methods

Cell cultures

Cell viability assay

Proteasome inhibition-induced premature senescence (PIIPS)

Stress-induced premature Senescence (SIPS)

Replicative Senescence (RS)

Senescence-associated-beta-galactosidase staining

Sample collection: Cell lysis and supernatant collection

Total protein content estimation

Measurement of Intracellular ROS Levels

Multiplex protein analysis

RNA isolation and RNA sequencing analysis

Differential gene expression analysis

Gene Set Enrichment Analysis (GSEA)

Data presentation and statistical analysis

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1