SUMMARY
The senescence-associated secretory phenotype (SASP) has recently emerged as both a driver of, and promising therapeutic target for, multiple age-related conditions, ranging from neurodegeneration to cancer. The complexity of the SASP, typically monitored by a few dozen secreted proteins, has been greatly underappreciated, and a small set of factors cannot explain the diverse phenotypes it produces in vivo. Here, we present ‘SASP Atlas’, a comprehensive proteomic database of soluble and exosome SASP factors originating from multiple senescence inducers and cell types. Each profile consists of hundreds of largely distinct proteins, but also includes a subset of proteins elevated in all SASPs. Based on our analyses, we propose several candidate biomarkers of cellular senescence, including GDF15, STC1 and SERPINs. This resource will facilitate identification of proteins that drive specific senescence-associated phenotypes and catalog potential senescence biomarkers to assess the burden, originating stimulus and tissue of senescent cells in vivo.
INTRODUCTION
Cellular senescence is a complex stress response that causes an essentially irreversible arrest of cell proliferation and development of a multi-component senescence-associated secretory phenotype (SASP) (Acosta et al., 2008; Coppé et al., 2008, 2010a; Kuilman et al., 2008). The SASP consists of myriad cytokines, chemokines, growth factors, and proteases that initiate inflammation, wound healing, and growth responses in nearby cells and tissues (Neves et al., 2015; Tchkonia et al., 2013). In young and healthy tissues, the SASP is typically transient and tends to contribute to the preservation or restoration of tissue homeostasis (Neves et al., 2015). However, the increase in senescent cells with age and a chronic SASP are now known to be key drivers of many pathological hallmarks of aging, including chronic inflammation, tumorigenesis, impaired stem cell renewal, and others (Neves et al., 2015; Tominaga, 2015). Using either or both of two transgenic mouse models that allow the selective elimination of senescent cells (Baker et al., 2011; Demaria et al., 2014), or compounds that mimic the effect of these transgenes, data from several laboratories strongly support the idea that the presence of senescent cells drives multiple age-related phenotypes and pathologies, including age-related atherosclerosis (Childs et al., 2016), osteoarthritis (Jeon et al., 2018), cancer metastasis and cardiac dysfunction (Baar et al., 2017; Demaria et al., 2017), myeloid skewing in the bone marrow (Abdul-Aziz et al., 2018; Chang et al., 2016), kidney dysfunction (Valentijn et al., 2018), and overall decrements in healthspan (Baker et al., 2016).
Several types of stress elicit a senescence and secretory response, which in turn can drive multiple phenotypes and pathologies associated with aging in mammals. Some of these stressors have shared effects. For example, telomere attrition resulting from repeated cell division (replicative senescence), elevated reactive oxygen species, chromatin disruption, and even the activation of certain oncogenes all can cause genotoxic stress, as can a number of therapeutic drug treatments, such as anti-cancer chemotherapies (Demaria et al., 2017) and certain highly active antiretroviral therapies for HIV treatment or prevention (Hernandez-Vallejo et al., 2013). However, whether these stressors produce similar or distinct SASPs is at present poorly characterized. Therefore, a comprehensive characterization of SASP components is critical to understanding how senescent response can drive such diverse pathological phenotypes in vivo. It is also a critical step in clarifying how various stimuli, all acting through senescence, differentially affect health.
The SASP was originally characterized by antibody arrays, which are necessarily biased, to measure the secretion of a small set of pro-inflammatory cytokines, proteases and protease inhibitors, and growth factors (Acosta et al., 2008; Coppé et al., 2008; Kuilman et al., 2008; Rodier et al., 2009). Subsequently, numerous large-scale unbiased gene expression studies were performed on different tissues and donors of varying ages, all of which indicated the existence of a complex and robust secretory program in senescent cells (Coppé et al., 2010b). A recent meta-analysis of senescent cell transcriptomes confirmed the expression of a few dozen of the originally characterized SASP factors in multiple senescent cell types (Hernandez-Segura et al., 2017). While unbiased comprehensive genomic analyses are valuable in describing the overall phenotype of senescent cells, the transcriptome only modestly correlates with the proteome, and cannot accurately and quantitatively assess the secreted protein profiles. Thus, transcriptomic profiles are unlikely to identify novel SASP factors without protein verification. Recently, the SASP induced by several genotoxic stressors was qualitatively assessed using mass spectrometry-based proteomics (Özcan et al., 2016). However, an in-depth, quantitative assessment of the SASP originating from multiple stimuli and cell types is missing. In addition, the secretion of exosomes and their contents by senescent cells has also been recently reported (Takasugi, 2018). Others have investigated the miRNA components of senescent exosomes but, aside from pro-tumorigenic effects (Takasugi, 2018), the proteomic content and function of exosomes and small extracellular vesicles (EVs) secreted by senescent cells remains largely unknown.
Studies of the SASP have made it clear that the SASP is not a single phenotype, but a series of highly complex and dynamic phenotypes that vary depending on the inducer of senescence and cell type. Given the growing size and complexity of the SASP, a central resource is needed for researchers to reference and more precisely interrogate context-dependent senescent phenotypes. Here we present “SASP Atlas” (www.SASPAtlas.com), a comprehensive, curated and expanding online database of the soluble secretomes of senescent cells (sSASPs) induced by various stimuli and in several cell types. We also present the first comprehensive proteomic analysis of the exosomal SASP (eSASP), which is largely distinct from the sSASP. Our proteomic analysis leverages a modern data-independent acquisition (DIA or SWATH) mass spectrometry workflow, which comprehensively acquires label-free, quantitative peptide (MS1) and fragment-level (MS2) data for all peptides in each sample (Bruderer et al., 2015; Egertson et al., 2015; Gillet et al., 2012; Rardin et al., 2015; Schilling et al., 2017). DIA workflows are not limited by stochastic peptide MS/MS sampling biases introduced by more traditional data-dependent acquisition (DDA) mass spectrometry. In addition to our interactive, online database (SASP Atlas), we assembled and deposited proteomic panels of SASP factors on Panorama Web, a freely-available web-repository for targeted mass spectrometry assays (Sharma et al., 2014, 2018) as a resource for others to detect and quantify the SASP in mass spectrometric studies.
The physiological effects of senescent cells derive largely from their ability to alter the behavior and function of neighboring cells through the secretion of proteins and other bioactive molecules, both locally within the tissue of residence and systemically throughout the organism. Comprehensive characterization of the SASPs originating from multiple stresses and cell types will help identify biomarkers associated with a number of diseases driven by the presence of senescent cells. Inhibiting the SASP or selectively removing senescent cells is a promising strategy to combat many age-related diseases and improve healthspan (Baker et al., 2016), and has driven efforts to develop drugs, such as senolytics and SASP inhibitors, for use in humans.
RESULTS AND DISCUSSION
Cellular senescence drives robust and variable changes in the secreted proteome
The secretomes of senescent cells and quiescent control cells were obtained from cultured primary human lung fibroblasts (IMR90) and renal cortical epithelial cells, as shown in Figure 1. Briefly, senescence was induced in cultured cells by one of three senescence-inducing stimuli: X-ray irradiation (IR), inducible RAS overexpression (RAS), or treatment with the protease inhibitor atazanavir (ATV). Secreted soluble proteins and exosomes/EVs were obtained from the conditioned media of senescent and control cells. Protein identification and quantification was then performed using a label-free, comprehensive and unbiased data-independent acquisition (DIA) approach, which enables more sensitive and accurate quantification of proteins by integration of MS2 fragment ion chromatograms. For quantification of relative abundances, proteins secreted by senescent cells were compared with their respective controls, and significantly changed proteins (q-value <0.05) were identified. Proteins secreted at higher levels by senescent, relative to quiescent, cells were defined as SASP components. Most SASP components were secreted at very high levels by senescent cells versus non-senescent cells (Figure 2). Each treatment group and respective controls contained between 4 and 10 biological replicates (see STAR Methods for replicate details). Relative protein quantification and statistical details for all secretomes are presented in Table S1. Induction of senescence was independently validated by measuring senescence-associated β-galactosidase (SA-β-Gal) staining and p16INK4a and IL-6 mRNA levels (Figure S1), as described (Coppé et al., 2008). There was no detectable cell death, as measured by a Sytox Green viability dye assay (Figure S2). X-irradiation and RAS overexpression induced senescence in >90% of cells and ATV induced senescence in about 65% of cells, as measured by SA-β-Gal staining (Figure S1A-B).
This unbiased proteomic profiling identified up to ~1700 secreted proteins, a large fraction of which were up-or down-regulated following induction of senescence by IR, RAS, or ATV (Figure 2), ranging between 376 to 747 significant changes per inducer. As expected, most of the significantly changed secreted proteins were markedly upregulated in senescent, compared to quiescent, cells, but interestingly, a minority were downregulated (Figure 2A). Notably, the protein cargo of exosomes/EVs released by senescent cells was substantially distinct compared to non-senescent cells (Figure 2A), indicating the existence of an exosome/EV SASP (eSASP) in addition to the sSASP.
The majority of changes in the sSASP, independent of inducer, exhibited increased secretion by senescent cells, with only 2-6% of proteins secreted at lower levels. In contrast, one-half to two-thirds of all significant protein changes in exosomes/EVs from senescent fibroblasts were decreased relative to quiescent cells (Figure 2A). For senescent renal epithelial cells, the sSASP comprised a more even mix of proteins with significantly lower or higher relative secretion. While the renal epithelial cell sSASP contained more overall changes than the fibroblast sSASP (830 versus 592 significant protein changes following IR), the magnitude of the fold-changes in the sSASP were higher in fibroblasts regardless of inducer (Figure 2B). For example, 95% of significant changes in the fibroblast sSASP were >2-fold, compared to only 17% of changes in the renal epithelial cell sSASP.
For each cell type and fraction, we checked for previously identified markers of senescence and the SASP (Figure 2C). These markers included chemokines (CXCLs), high mobility group box 1 protein (HMGB1), IGF binding proteins (IGFBPs), matrix metalloproteinases (MMPs), lamin b1 (LMNB1), and tissue inhibitors of metallopeptidase (TIMPs). In addition, we measured mRNA levels of p16INK4a (CDKN2A) and interleukin-6 (IL-6), and SA-β-Gal staining (Figure S1C). In fibroblasts, all previously identified senescence and SASP markers were elevated, regardless of the senescence inducer. However, while expression of p16INK4a, IL-6 and SA-β-Gal were also elevated in renal epithelial cells, the proteins secreted by fibroblasts were either decreased or unchanged, except for IGFBP2. This finding suggests that fibroblast SASP markers are not sufficient to detect senescence and that profiling the SASP in different cell types and biological contexts is required to detect senescent cells in vivo with confidence. Within exosomes/EVs secreted by senescent fibroblasts, nearly all previously identified SASP factors were either absent, unchanged or decreased, and none were consistently elevated in response to more than one inducer (Figure 2C).
Senescence-inducing stresses drive largely distinct secretory phenotypes
To determine the effect of senescence-inducing stimuli on the SASPs of senescent cells, we compared the sSASP from human primary fibroblasts induced to senesce by IR, RAS or ATV. Strikingly, the sSASP was largely distinct among inducers, with an overlap of 172 proteins among 1107 total increased proteins and no overlap among decreased proteins (Table S2). Thus, most protein changes were not shared among inducers (Figure 3A). Among the overlapping increased proteins, levels varied among inducers (Figure 3B). In particular, the ATV-induced sSASP had fewer and generally smaller changes than either the RAS- or IR-induced sSASP, perhaps corresponding to less efficient induction of senescence, as determined by SA-β-Gal staining (Figure S1).
To determine whether there is a core of pathways or functions associated with the SASPs, we performed pathway and network analysis on proteins that overlapped in the sSASPs of each inducer (Figure 3C). The largest pathway networks associated with all inducers related to tissue and cell structure and organization, including extracellular matrix, cytoskeleton, integrins, and peptidase activity. Interestingly, neurodegeneration and three related pathways with high agreement (kappa score >40%) – apoptosis, ROS signaling and TP53-regulated metabolism - were also enriched among overlapping sSASP proteins (Figure 3C-D). Among the neurodegeneration proteins were amyloid precursor protein (APP) and cystatin 3 (CST3), related to Alzheimer’s pathogenesis and risk (Kaur and Levy, 2012; O’Brien and Wong, 2011), as well as Parkinsonism-associated deglycase (PARK7 or DJ1) (Bonifati et al., 2003) (Figure 3D). This enrichment of neurodegeneration-associated proteins and highly connected pathways suggests that senescence may contribute to pathogenesis of neurodegenerative diseases, for which these sSASP factors might serve as biomarkers, regardless of the senescence-inducing stimuli. Further research will be needed to assess the value of these proteins as biomarkers.
To distill the overlapping ‘core’ sSASP proteins into primary components in an unbiased fashion, we performed an unsupervised machine learning analysis on the levels of secreted proteins (Figure 3E). K-means clustering analysis uncovered three primary clusters among core sSASP components. Strikingly, one cluster consisting of just three proteins – chemokine (C-X-C motif) ligand 1 (CXCL1), matrix metallopeptidase 1 (MMP1), and stanniocalcin 1 (STC1) – were very highly present in the sSASP of all inducers, suggesting that these proteins are best qualified to serve as surrogate markers of the sSASP. Of note, among the top expressed sSASP proteins was STC1, a previously unidentified SASP factor and a secreted hormone with many disease associations (Chang et al., 2015; Du et al., 2016; Ohkouchi et al., 2015; Pan et al., 2015; Shahim et al., 2017; Su et al., 2015), demonstrating the utility of robust unbiased approaches to uncovering novel molecules, biological functions and pathways. Our approach also validates MMP1 and CXCL1 as SASP markers.
The sSASP is largely distinct in composition and regulation in fibroblasts and epithelial cells
We also compared the secretomes of fibroblasts and renal epithelial cells to understand the cell-type specificity of the sSASP. Indeed, the sSASP of fibroblasts and epithelial cells were largely distinct (Figure 4A-B). Among the proteins increased in the sSASP of each cell type, 20-30% overlapped, and the magnitude of the change by renal epithelial cells was, in most cases, lower than in fibroblasts, although it is possible that senescent fibroblasts inherently secrete more proteins than epithelial cells. Senescent epithelial cells also secreted significantly lower levels of many proteins, compared to senescent fibroblasts, regardless of senescence inducer. Interestingly, 20-30% of the proteins significantly decreased in the sSASP of renal epithelial cells overlapped with proteins significantly increased in the fibroblast sSASP (Figure 4B), suggesting a negative correlation of a portion of the sSASP between cell types. Among the epithelial sSASP factors that changed oppositely to fibroblast sSASP factors were IGFBPs, TIMPs 1 and 2, CXCL1, and most SERPINs (Figure 2C). In all, 54 sSASP factors were shared between all senescence inducers and cell types that we examined (Table S3).
Pathway and network analysis of proteins increased in the sSASPs of fibroblasts and epithelial cells (Figure 4C) showed that most pathways belonged to one of three general categories: protein turnover and secretion, primary metabolism, and cellular detoxification. While not as apparent on a molecule-by-molecule basis, on the pathway level many functions were commonly enriched in both the epithelial and fibroblast sSASPs (Figure 4C and Figure 3C), including vesicle-mediated transport and exosomes, glycolytic metabolism, and cellular detoxification. The notable exceptions enriched uniquely by epithelial cells are protein translation and degradation (lysosome and phagosome).
Surprisingly, the majority of renal epithelial sSASP proteins with significantly lower secretion were enriched in pathways related to tissue and cell structure, adhesion and motility (Figure 4D). This is in contrast to previous reports and our own findings in fibroblasts (Figure 3C), in which these functions were most highly increased, regardless of inducer. The irradiated epithelial sSASP also had significantly lower levels of proteins involved in RNA processing, in contrast to increased RNA metabolism in the irradiated fibroblast sSASP. Additionally, the epithelial sSASP was significantly depleted in proteins related to proteasome degradation, antigen processing, and the complement system.
Damage-associated molecular patterns (DAMPs, also known as alarmins or danger signals) are released from cells in response to internal and external stressors, and have been identified as components of the SASP (Davalos et al., 2013). Pattern recognition receptor-bearing cells, including cells of the innate immune system, recognize extracellular DAMPs as signals to promote inflammatory and fibrotic responses in tissues. An increase in circulating DAMPs is hypothesized to play a role in aging (Feldman et al., 2015; Huang et al., 2015), particularly in the age-related inflammation termed ‘inflammaging’ (Franceschi and Campisi, 2014). DAMPs may also be biomarkers of a number of diseases, including traumatic, cardiovascular, metabolic, neurodegenerative, malignant, and infectious diseases (Feldman et al., 2015; Fucikova et al., 2015; Garg et al., 2015). HMGB1, a founding member of the DAMPs, is a prominent SASP marker (Davalos et al., 2013). Our proteomic analysis identified increased secretion of multiple DAMPs, including the prototypical molecules HMGB1 and calreticulin (CALR), by senescent fibroblasts under all senescence-inducing stresses (Table 1). However, in senescent epithelial cells the secretion of DAMPs was either unchanged or significantly reduced, demonstrating that some defining SASP components may vary depending on cell type.
Exosome/EV characteristics and proteomic contents are altered by cellular senescence
Because exosome and vesicle-mediated transport proteins are enriched in the soluble SASP, we hypothesized that senescent cells would show significant changes in exosome biology and proteomic content, or the exosome/EV SASP (eSASP). We used an ultracentrifugation-based protocol to enrich conditioned media for exosomes and small EVs released by quiescent and senescent fibroblasts induced by X-irradiation and RAS overexpression (Figure 1, see methods). We confirmed the presence of exosome/EV markers, such as CD63 and CD9 (Keerthikumar et al., 2016), in the purified fractions and the absence of exosome/EV markers in EV-depleted soluble fractions (Figure S3A). Exosomes/EVs from senescent fibroblasts had a robustly altered proteomic composition compared with exosomes from non-senescent cells (Figure 2A).
To determine whether the secretion and characteristics of exosomes/EVs are altered in senescent cells, we performed particle counting and size distribution analysis of exosomes/EVs secreted into the culture medium of senescent and non-senescent cells over a 24-hour period. On average, senescent cells released a smaller number of vesicles -- about 275 vesicles per cell compared to 374 vesicles per cell in controls (Figure 5A). However, the mean diameter, size distribution and protein content of senescent exosomes/EVs was significantly greater than controls (Figure 5B-C, Figure S3B-C). Of note, a recent study examining EVs in humans reported a decline in the concentration and increase in size of EVs circulating in the plasma of older, compared to younger, individuals (Eitan et al., 2017). These findings are consistent with an expected increase in senescent cell burden with age. Further work with senolytics may validate whether exosome/EV concentration and size are indicators of senescent cell burdens in humans.
The protein content of exosomes/EVs released by IR-vs RAS-induced senescent fibroblasts were largely distinct, sharing only 9 significantly altered proteins (Figure 5D). Exosomes/EVs were reported to contain protein signatures of their originating cells (Belov et al., 2016; Takasugi, 2018), presenting a unique opportunity to develop senescence biomarkers with a degree of specificity. As suggested by our proteomic analysis, exosome/EV cargo proteins might distinguish senescent cells of different origins or originating from different stressors. The membrane protein topology of exosomes is also representative of their originating cells (Belov et al., 2016; Takasugi, 2018). About 30% of all the exosome/EV proteins that increased upon senescence are plasma membrane proteins (Figure 5E). Therefore, the identification of plasma membrane proteins associated with exosomes/EVs might also serve to identify cell-surface proteins expressed on senescent cells. There are currently no reliable senescence-specific cell-surface markers, and so the identification of these markers would be very valuable for isolating or therapeutically targeting senescent cells.
The SASP Contains Potential Aging and Disease Biomarkers
As a driver of many aging and disease phenotypes, we hypothesized that the SASP would include known biomarkers of aging and age-related diseases. A recent proteomic biomarker study in a human cohort reported 217 proteins that are significantly associated with age in human plasma (Tanaka et al., 2018). Of these markers, the authors noted that 20 proteins (9.2%) are present in the originally-defined SASP (Coppé et al., 2008). Strikingly, multiple newly identified SASP factors from our study were also identified by Tanaka et al. as age-associated. Considering our newly identified “core SASP” -- SASP components resulting from all senescence inducers -- the number of age-associated plasma proteins that are also SASP proteins, doubles to 40 proteins (18.4%, Table S4). If all novel SASP proteins we identified are considered, over a third (36.9%) of age-related changes in plasma proteins are also highly secreted by senescent cells, (Table S4). Thus, plasma biomarkers of aging are highly enriched with SASP factors. Notably, the protein having the strongest association with aging reported by Tanaka et al., GDF15 (r=0.82), was among the most highly secreted proteins in the sSASP induced by IR, RAS and ATV in fibroblasts, and epithelial cells induced by IR (Figure 6A). In addition to aging, various large cohort studies identified GDF15 as a biomarker of cardiovascular disease, cardiovascular and cancer mortality and morbidity, renal disease, and all-cause mortality independent of cardiovascular mortality (Bidadkosh et al., 2017; Daniels et al., 2011; Ho et al., 2012, 2013; Rohatgi et al., 2012; Wallentin et al., 2013; Wollert et al., 2017).
Complement and coagulation cascade proteins, particularly those with protease inhibitor activity such as SERPINs, were also noted as prominent plasma biomarkers of aging (Tanaka et al., 2018). Notably, these proteins and their pathway networks were robustly altered in the SASPs of cells induced to senesce by multiple stressors (Figure 3C, Figure 6B). Additionally, two of the top “core sSASP” proteins identified by an unbiased k-means clustering algorithm – STC1 and MMP1 (Figure 6C-D) – were reported as significant aging biomarkers (Tanaka et al., 2018). In addition to aging, MMP1 has been identified as a biomarker for several cancers, pulmonary fibrosis, and potentially Alzheimer’s disease (Bhat et al., 2012; Chen et al., 2016; Rosas et al., 2008; Roy et al., 2009), whereas STC1 has been identified as a diagnostic and prognostic biomarker for cancers, pulmonary fibrosis, renal ischemia/reperfusion injury, and Alzheimer’s disease (Chang et al., 2015; Du et al., 2016; Ohkouchi et al., 2015; Pan et al., 2015; Shahim et al., 2017; Su et al., 2015). The enrichment of aging and disease biomarkers in the secretomes of senescent cells supports their links to a wide spectrum of age-related diseases and highlights the need for comprehensive secretome profiles of senescent cells in multiple biological contexts.
SASP Atlas and Proteomics Assays for Biomarker Discovery and Identification of Secretory Phenotypes
Here we present SASP Atlas (www.SASPAtlas.com), the first proteome-based database of SASPs. This comprehensive and unbiased database contains exosome/EV and soluble secretomes, in addition to SASPs originating from multiple senescence-inducing stresses and cell types. The SASP Atlas will be continuously updated with SASP profiles from new cell types and inducers of senescence, including paracrine (or “bystander”) senescence (Acosta et al., 2013; Nelson et al., 2012), as well as temporal dynamics of the SASP – all generated by our laboratories. In addition to a database, we created proteomic ‘secretome panels’ on Panorama Web (Sharma et al., 2018), a freely-available repository server for targeted mass spectrometry assays as a resource for others to detect and quantify the SASP in mass spectrometry studies. These resources can be used both as a reference and a guide to pinpoint and quantify SASP factors that may be associated with diseases of interest, and to develop aging and disease-related biomarkers (Figure 7).
SASP profiles are needed to develop senescence biomarkers in human plasma or other biofluids and are critical for identifying individuals to treat with, and measuring the efficacy of, senescence-targeted therapies such as senolytics. Translating senescence- and SASP-targeting interventions to humans will require a comprehensive profile of SASPs, both to identify the deleterious components of the SASP and to develop human biomarkers to assess senescent cell burden. The SASP, as originally identified, comprised ~50 cytokines, chemokines, growth factors, and proteases that were detected by biased methods (e.g., antibody arrays) and/or transcriptional analyses (Acosta et al., 2008; Coppé et al., 2008, 2010a; Kuilman et al., 2008). Transcriptomic profiling has also been performed on senescent cells (Hernandez-Segura et al., 2017). While these unbiased comprehensive analyses are valuable in describing the overall phenotype of senescent cells, they cannot accurately portray the secreted protein profile with quantitative accuracy. For example, a recent meta-analysis of senescent cell transcriptomes identified 21 core SASP mRNAs expressed by senescent fibroblasts, whereas we found 172 core SASP proteins, with only a small fraction of all protein-level changes being identified by transcriptomics (Figure S4). Of course, the number and nature of these core proteins will change as we and others interrogate additional cell types and senescence inducers. Additionally, the secretion of SASP factors such as HMGB1 and other DAMPs is not generally transcriptionally driven.
Our quantitative and unbiased proteomic analysis of senescent fibroblasts and epithelial cells clearly reveals a much larger and diverse SASP than initially reported. These SASP profiles contribute a number of new potential senescence, aging, and disease biomarkers. In addition to general senescence biomarkers, many protein markers will likely be specific to cell-type and originating stimulus. Thus, biomarkers present in human patients in vivo will likely vary depending on the affected tissue, originating cell types, and senescence stimuli. Therefore, comprehensive quantitative profiles of the SASP under a variety of physiological conditions will provide biomarker candidates with a higher degree of selectivity to specific in vivo pathologies in humans.
METHODS
KEY RESOURCES TABLE
CONTACT FOR REAGENT AND RESOURCE SHARING
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Birgit Schilling (bschilling{at}buckinstitute.org).
EXPERIMENTAL MODEL AND SUBJECT DETAILS
Human Cell Culture and Primary Cell Lines
IMR90 primary human lung fibroblasts (Female, ATCC #CCL186) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco #12430-054) supplemented with Penicillin Streptomycin (5000 U/mL penicillin and 5000 μg/mL streptomycin, Gibco #15070063) and 10% Fetal Bovine Serum (Gibco #2614079) and maintained at 37° C, 10% CO2, and 3% O2. Primary human renal epithelial cells (ATCC PCS400011) were cultured in Renal Epithelial Cell Basal Medium (Female, ATCC #PCS-400-030) and maintained at 37° C, 10% CO2, and 3% O2.
Induction of Senescence
X-ray irradiation
Induction of senescence by ionizing radiation (IR) was performed with 10 Gy X-ray. Quiescent control cells were mock irradiated. Senescent cells were cultured for 10 days to allow development of the senescent phenotype, and quiescent cells were cultured in 0.2% serum for 3 days. Subsequently, cells were washed with PBS (Gibco #10010-023) and placed in serum-free and phenol-red-free DMEM (Gibco #21063-029) and conditioned media was collected after 24 hours to obtain the secreted protein fractions.
RAS overexpression
RASv12 was cloned in pLVX vector (Lenti-X™ Tet-On from Clontech #632162) to make lentivirus. The Lenti-X Tet-On Advanced Inducible Expression System is a lentiviral-based gene transfer technology to deliver inducible gene expression system into dividing and nondividing target cells, primary cells, and stem cells. Lentivirus were then generated and collected from 293FT cells and used for transduction to IMR-90 cells (PD-30). Transduced cells were selected with puromycin (1 μg/ml) for 24 hours. Stable IMR-90 cells carrying inducible RASv12 were generated and used for further experiments. For induction of RASv12, cells were treated with 1 μg/ml Doxycycline dissolved in DMSO (Sigma # D9891) for 4 days (early time point) or 7 days, which is optimal for senescence induction. Doxycycline was replaced after every 48 hours in the induction medium Subsequently, cells were washed with PBS (Gibco #10010-023) and placed in serum-free and phenol-red-free DMEM (Gibco #21063-029) and conditioned media was collected after 24 hours to obtain the secreted protein fractions.
Atazanivir treatment
Induction of senescence was performed with a clinically relevant dose (20 µM) of Atazanavir. Cells were grown in appropriate media containing Atazanavir or vehicle (DMSO) for 9 days (early timepoint) or 14 days, which is optimal for senescence induction. Subsequently, cells were washed with PBS (Gibco #10010-023) and placed in serum-free and phenol-red-free DMEM (Gibco #21063-029) and conditioned media was collected after 24 hours to obtain the secreted protein fractions.
Isolation of Secreted Soluble Proteins and Exosomes/EVs
Proteins secreted into serum-free medium over a 24-hr period were collected for each experimental condition. An ultracentrifugation-based protocol was used to separate the exosome and small extracellular vesicle fraction from the soluble protein fraction (Théry et al., 2006). Briefly, conditioned medium containing secreted proteins was centrifuged at 10,000 x g at 4° C for 30 minutes to remove cellular debris. The supernatant was then centrifuged at 20,000 x g at 4° C for 70 minutes to remove microvesicles. The supernatant was then ultracentrifuged at 100,000 x g at 4° C for 70 minutes to pellet exosomes. The exosome-depleted supernatant was saved for proteomic sample preparation. The exosome pellet was then washed and ultracentrifuged at 100,000 x g at 4° C for 70 minutes two times with PBS before resuspending the exosome pellet in PBS and saved for proteomic sample preparation.
Proteomic Sample Preparation
Chemicals
Acetonitrile (#AH015) and water (#AH365) were obtained from Burdick & Jackson (Muskegon, MI). Reagents for protein chemistry including iodoacetamide (IAA, #I1149), dithiothreitol (DTT, #D9779), formic acid (FA, #94318-50ML-F), and Triethylammonium bicarbonate buffer 1.0 M, pH 8.5±0.1 (#T7408) were purchased from Sigma Aldrich (St. Louis, MO). Urea (#29700) was purchased Thermo Scientific (Waltham, MA). Sequencing grade trypsin (#V5113) was purchased from Promega (Madison, WI). HLB Oasis SPE cartridges (#186003908) were purchased from Waters (Milford, MA).
Protein concentration and quantification
Samples were concentrated using Amicon Ultra-15 Centrifugal Filter Units with 3 kDa molecular weight cutoff (MilliporeSigma #UFC900324) per manufacturer instructions and transferred into a solution of 8M urea/50 mM Triethylammonium bicarbonate buffer at pH 8. Protein quantitation was then performed using a BCA Protein Assay Kit (Pierce #23225, Waltham, MA).
Digestion
An aliquot of 25-100 μg from each sample was then brought to equal volume with 50 mM Triethylammonium bicarbonate buffer at pH 8. The protein mixtures were reduced with 20 mM DTT (37°C for 1 hour), and subsequently alkylated with 40 mM iodoacetamide (30 minutes at RT in the dark). Samples were diluted 10-fold with 50 mM Triethylammonium bicarbonate buffer at pH 8 and incubated overnight at 37°C with sequencing grade trypsin (Promega) added at a 1:50 enzyme:substrate ratio (wt/wt).
Desalting
The peptide supernatants were then collected and desalted with Oasis HLB 30 mg Sorbent Cartridges (Waters #186003908, Milford, MA), concentrated, and re-suspended in a solution containing mass spectrometric ‘Hyper Reaction Monitoring’ peptide standards (HRM, Biognosys #Kit-3003, Switzerland) and 0.2% formic acid in water.
Mass Spectrometry Analysis
Samples were analyzed by reverse-phase HPLC-ESI-MS/MS using the Eksigent Ultra Plus nano-LC 2D HPLC system (Dublin, CA) combined with a cHiPLC System, which was directly connected to a quadrupole time-of-flight SCIEX TripleTOF 6600 or a TripleTOF 5600 mass spectrometer (SCIEX, Redwood City, CA). Typically, mass resolution in precursor scans was ~ 45,000 (TripleTOF 6600), while fragment ion resolution was ~15,000 in ‘high sensitivity’ product ion scan mode. After injection, peptide mixtures were transferred onto a C18 pre-column chip (200 μm x 6 mm ChromXP C18-CL chip, 3 μm, 300 Å, SCIEX) and washed at 2 μl/min for 10 min with the loading solvent (H2O/0.1% formic acid) for desalting. Subsequently, peptides were transferred to the 75 μm x 15 cm ChromXP C18-CL chip, 3 μm, 300 Å, (SCIEX), and eluted at a flow rate of 300 nL/min with a 3 h gradient using aqueous and acetonitrile solvent buffers.
For label-free relative quantification all study samples were analyzed by data-independent acquisitions (DIA), or specifically variable window SWATH acquisitions. In these SWATH acquisitions, instead of the Q1 quadrupole transmitting a narrow mass range through to the collision cell, windows of variable width (5 to 90 m/z) are passed in incremental steps over the full mass range (m/z 400-1250). The cycle time of 3.2 sec includes a 250 msec precursor ion scan followed by 45 msec accumulation time for each of the 64 SWATH segments. The variable windows were determined according to the complexity of the typical MS1 ion current observed within a certain m/z range using a SCIEX ‘variable window calculator’ algorithm (i.e. more narrow windows were chosen in ‘busy’ m/z ranges, wide windows in m/z ranges with few eluting precursor ions) (Schilling et al., 2017). SWATH MS2 produces complex MS/MS spectra which are a composite of all the analytes within each selected Q1 m/z window. This method requires high scan speeds and high-resolution capabilities, as well as a ‘reference spectral library’ that can be either generated in-house or taken from existing comprehensive and large-scale studies. All collected mass spectrometry data was processed in Spectronaut using pan-human library that provides SWATH assays for ~10,000 human proteins (Rosenberger et al., 2014).
Cell Viability Assays
Cell viability was assessed with SYTOX Green Nucleic Acid Stain (Invitrogen #S7020). Senescent and control cells over a 24-hour period in serum-free medium containing SYTOX Green with continuous imaging. Cell death was quantified by counting total SYTOX Green positive nuclei appearing in a 24-hour time-lapse video.
Senescence-Associated β-Galactosidase Staining
Senescence-associated beta-galactosidase (SA-β-gal) activity was determined using the BioVision Senescence Detection Kit (Cat# K320-250) protocol. For each experiment, approximately 100–150 cells were counted.
RNA Extraction and quantitative Real-Time PCR
Total RNA was prepared using the PureLink Micro-to-Midi total RNA Purification System (Invitrogen # 12183018A), according to the manufacturer’s protocol. Samples were treated with DNase I Amp Grade (Invitrogen #18068015) before retrotranscription to eliminate genomic DNA contamination. RNA was reverse transcribed into cDNA using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems #4368813), according to the manufacturer’s protocol. Quantitative RT-PCR (qRT-PCR) reactions were performed as described using the Universal Probe Library system (Roche). The Actin and Tubulin predeveloped TaqMan assays (Applied Biosystems) were used to control for cDNA quantity. Quantitative RT-PCR assays were performed on the LightCycler 480 Sytem (Roche, Basel, Switzerland). The primers and probes were as follows:
Human actin F 5’-CCAACCGCGAGAAGATGA; R 5’-TCCATCACGATGCCAGTG, UPL probe #64
Human tubulin F 5’-CTTCGTCTCCGCCATCAG; R 5’-TTGCCAATCTGGACACCA, UPL Probe #58
Human IL-6 F 5’-GCCCAGCTATGAACTCCTTCT; R 5’-GAAGGCAGCAGGCAACAC, UPL Probe #45
Human p16INK4a F 5’-GAGCAGCATGGAGCCTTC; R 5’-CGTAACTATTCGGTGCGTTG, UPL Probe #34
Exosome Size Distribution Analysis
Particle diameter and concentration were assessed using an IZON qNano Nanoparticle Characterization instrument (Zen-bio, Inc., Research Triangle Park, NC).
QUANTIFICATION AND STATISTICAL ANALYSIS
Processing, Quantification, and Statistical Analysis of MS Data
SWATH acquisitions were quantitatively processed using the proprietary Spectronaut v12 (12.020491.3.1543) software (Bruderer et al., 2015) from Biognosys. Quantitative SWATH MS2 data analysis was based on extracted ion chromatograms (XICs) of 6-10 of the most abundant fragment ions in the identified spectra. Relative quantification was performed comparing different conditions (senescent versus control) assessing fold changes for proteins from the investigated cells and conditions. The number of replicates from proteomic each experiment are as follows: X-ray irradiated fibroblasts, 4 senescent and 4 control replicates; X-ray irradiated renal epithelial cells, 5 senescent and 5 control replicates; 4 day RAS-induction in fibroblasts, 10 senescent and 10 control replicates; 7 day RAS-induced fibroblasts, 6 senescent and 6 control replicates; atazanavir-treated fibroblasts, 3 senescent (9 days treatment), 3 senescent (14 days treatment), and 4 control replicates; X-ray irradiated fibroblast exosomes, 5 senescent and 5 control replicates; 7 day RAS-induced fibroblast exosomes, 6 senescent and 6 control replicates. Significance was assessed using FDR corrected q-values<0.05. See Data Accession (below) for quantitative results from mass spectrometric SWATH analysis.
Pathway and Network Analysis
Gene ontology, pathway, and network analysis was performed using the GlueGO package version 2.5.3 in Cytoscape version 3.7.1 (Bindea et al., 2009; Shannon et al., 2003). Curated pathways for enrichment analysis were referenced from the following databases: GO Biological Function, GO Cellular Compartment, Kegg pathways, WikiPathways, and Reactome Pathways. For gene ontology data, testing was restricted to pathways with experimental evidence (EXP, IDA, IPI, IMP, IGI, IEP). The statistical cutoff for enriched pathways was Bonferroni-adjusted p-values < 0.01 by right-sided hypergeometric testing. Pathway-connecting edges were drawn for kappa scores > 40%. Kappa scores are a measure of inter-pathway agreement among observed proteins that indicate whether pathway agreement is greater than expected by chance based on shared proteins. Pathways with the same color indicate >= 50% similarity in terms.
K-Means Clustering
Unsupervised clustering was performed in Python with Scikit-learn, a module integrating a wide range of machine learning algorithms (Pedregosa et al., 2011). Datasets were pre-processed with the StandardScaler function and clustered with the KMeans algorithm.
Data Visualization
Heatmaps were visualized in R using the heatmap.2 function in the ‘gplots’ package (Warnes et al., 2019). Venn diagrams were visualized with the “VennDiagram” package in R (Chen, 2018). Color palettes in R were generated with the “RColorBrewer” package (Neuwirth, 2014). Pathway and network visualizations were generated and modified using the GlueGO package in Cytoscape (Bindea et al., 2009; Shannon et al., 2003).
DATA AND SOFTWARE AVAILABILITY
Data availability
All raw files are uploaded to the Center for Computational Mass Spectrometry, MassIVE, and can be downloaded using the following ftp link ftp://massive.ucsd.edu/MSV000083468 (MassIVE ID number: MSV00008346). Data uploads include the protein identification and quantification details, spectral library, and FASTA file used for mass spectrometric analysis. Proteomic assays of SASP profiles are available on Panorama (https://panoramaweb.org/project/Schilling/SASP_Atlas_Buck/begin.view?), a repository for targeted mass spectrometry assays generated in Skyline software (Sharma et al., 2014, 2018). All quantitative data is available for viewing and downloading on SASP Atlas (www.saspatlas.com).
MassIVE
MSV000083468 (ftp://massive.ucsd.edu/MSV000083468) SASP Atlas: www.SASPAtlas.com
SUPPLEMENTAL INFORMATION
Supplemental Information includes 4 figures, 4 tables:
Supplemental Table 1: Mass spectrometry quantification outputs for each dataset, as separate worksheets in a single excel workbook.
Supplemental Table 2: Proteins with increased secretion in all senescence-inducing stresses.
Supplemental Table 3: Proteins with increased secretion in all cell types and senescence-inducing stresses.
Supplemental Table 4: Age-associated plasma proteins in the SASP
Supplemental Figure 1: Senescence-associated β-galactosidase staining and expression of senescence markers.
Supplemental Figure 2: Cell Viability Assays.
Supplemental Figure 3: Exosome/EV proteomic markers and size distribution analysis.
Supplemental Figure 4: Comparison of our proteomic data with transcriptional changes in senescent cells.
AKNOWLEDGMENTS
This work was supported by the National Institute on Aging (U01 AG060906-01, PI: Schilling; and P01AG017242 and R01AG051729, PI: Campisi) and National Institute of Health Shared Instrumentation Grant (1S10 OD016281, Buck Institute). N.B. and O.J. were supported by postdoctoral fellowships from the Glenn Foundation for Medical Research. A.K. was supported by the SENS Foundation. We thank John C.W. Carroll for help creating figures.
AUTHOR CONTRIBUTIONS
Conceptualization, N.B., B.S., J.C. L.F.; Cell Culture, N.B., A.K., O.J., C.K., T.P.; RNA expression and Activity Assays, N.B., A.K., O.J., C.K. T.P.; Proteomic Sample Preparation, N.B., T.P., A.H., S.S.; Data Analysis, N.B., B.S., A.H., S.S.; Pathway and Network Analysis, N.B., C.R.; Visualization, N.B., C.R.; Web Database, C.R., N.B, B.S; Writing and Editing, N.B., B.S., J.C., L.F.; Funding Acquisition, B.S., J.C.
DECLARATION OF INTERESTS
The authors declare no competing interests.