Introduction

Somatic cell population like human dermal fibroblasts (HDFs) undergoes a limited number of cell divisions before the dormant stage in which cell division stops. This phenomenon is called ‘Hayflick limit’, showing that a normal human fetal cell population can divide up to 60 times in certain cell culture condition (Hayflick and Moorhead 1961). However, it is generally accepted that the somatic cells are mortal except for stem cells. In contrast to dividing cells, senescent or late senescent (just before dormancy) cells display characteristic phenotypic alterations, such as G0/G1 cell cycle arrest (Demidenko and Blagosklonny 2008), higher senescence-associated β-galactosidase activity (Kurz et al. 2000), and hypo-responsiveness to growth factors (Kim et al. 2010). These changes contribute to the decline in cellular function accompanying the aging process.

Cellular senescence has been suggested to reduce the risk of neoplastic transformation by preventing cell proliferation. However, accumulating evidence presents that the senescent cells can have deleterious effects on tissue microenvironment, thereby contributing to late-life cancers. Increased expression of vascular endothelial growth factor (VEGF) is a frequent characteristic of senescent human and mouse fibroblasts in culture (Coppe et al. 2006). Increased secretion of VEGF, much more in senescent fibroblasts than pre-senescent cells, stimulates tumor vascularization in aged mice facilitating age-associated cancer development. Other factors secreted from senescent fibroblasts are also known to alter senescent fibroblasts into pro-inflammatory cells (Coppe et al. 2010). Therefore, knowing what kinds of proteins are secreted by senescent somatic cells is very important for understanding and retarding the aging process.

Cells influence their own or neighboring microenvironment through the synthesis and secretion of a variety of extracellular matrix (ECM) components. Several ECM components are known to undergo changes during aging process (Lener et al. 2006) and in aging-related diseases, such as type II diabetes and Werner’s syndrome (Labat-Robert and Robert 1988). The insoluble ECM factors secreted by senescent cells also contribute to cancer progression in aged organisms, suggesting the active role of ECM in determining cellular fate. As the human lifespan increases, interests and necessities in the wound healing of the elderly’s skin are growing. In particular, the secreted ECM proteins provide a fundamental clue for skin regeneration. This study intends to attempt proteomic analysis of ECM of late senescent HDF cells. Although there were general reports related to ECM proteomics in HDFs (Yang et al. 2011; Won et al. 2012), a systematic ECM proteome analysis for late senescent skin cells has not been performed yet. Therefore, we performed comparative proteomic analysis of ECM proteins from young and late senescent HDFs to identify secreted proteins and to suggest their functional roles in aging process.

Materials and methods

Materials

DMEM, FBS, antibiotics (penicillin and streptomycin) were purchased from Gibco/BRL Life Technologies, Inc. (Carlsbad, CA). Monoclonal antibodies against fibronectin 1 (FN1) and vitronectin (VTN) were purchased from Chemicon (Bedford, MA), and monoclonal antibody against high-temperature requirement factor 1 (HTRA1) was from Abcam (Cambridge, MA). In addition, monoclonal antibodies against caveolin 1 (CAV1) and apolipoprotein B (APOB) were obtained from BD Biosciences Pharmingen (Mountain View, CA) and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), respectively. HRP-conjugated anti-rabbit and anti-mouse secondary antibodies were obtained from Vector Laboratories (Burlingame, CA). The protein assay kit was purchased from Bio-Rad Laboratories (Hercules, CA), and the ECL system was purchased from GE Healthcare (Buckinghamshire, UK).

Cell culture and preparation of ECM

Neonatal HDFs were isolated from neonates as described previously (Tomela et al. 2021). Briefly, cells were obtained from Seoul National University (SNU) and cultured at 37 °C, 5% CO2 in DMEM supplemented with 10% FBS, 100 units/ml penicillin, and 100 μg/ml streptomycin. Cells with passage #15, #30 and #40 were considered young, early senescent, and late senescent (or near quiescent) state, respectively. ECM proteins were prepared from cultured population doublings (PD) 15 and PD 40 HDFs as described previously (Yeo et al. 2000). The matrices after trypsinized HDFs were scraped out and resuspended in 10 mM Tris–HCl buffer (pH 7.2) containing 1% Triton X-100, 10% glycerol, 5% sodium dodecyl sulfate (SDS), 100 mM DTT, and 1% protease inhibitor cocktail. Protein concentrations were estimated using the BCA™ protein assay kit. The filtrates were passed through the 1 kDa cut-off Centricon filter and vacuum dried samples were stored at -70 °C prior to LC–MS analysis.

Senescent-associated β-galactosidase assay

Senescence indices using senescence-associated β-galactosidase (SA-β-Gal) in passage #15, #30, and #40 HDFs were detected by Senescence Cells Histochemical Staining kit (Sigma-Aldrich, Seoul, Korea). SA-β-Gal analysis was performed as described previously (Pflieger et al. 2006).

Proteomic analysis of ECM protein extracts

Proteins (50 µg) were loaded on 3‑8% Tris–acetate gel. After running on SDS-PAGE, gels were fixed in 30% ethanol/7.5% acetic acid for 2 h followed by staining with Coomassie Brilliant Blue R250. Each gel lane was cut into 12 pieces and each piece was transferred to a new Eppendorf tube. The sliced gel piece was washed with 10 mM ammonium bicarbonate (ABC) and 50% acetonitrile and incubated in tryptic digestion buffer containing 50 mM ABC, 5 mM CaCl2, and 1 mg/ml trypsin at 37 °C for 16 h. Peptides were recovered by two cycles of extraction with 50 mM ABC and 100% acetonitrile. The lyophilized peptide extracts were dissolved in 0.1% formic acid and analyzed by LC–MS/MS as described previously (Kim et al. 2005). Eluted peptides were applied to a 7-Tesla Fourier Transform-Ion Cyclotron Resonance (FT-ICR) MS (Thermo Fisher, San Jose, CA). For LC separation, a microcapillary column (75 μm × 150 mm) packed with C18 silica resin was used. The mobile phases, A and B, were composed of 0 and 100% acetonitrile, respectively, each containing formic acid (0.1%). MS-derived peptides were searched against the human database of International Protein Index (http://www.ebi.ac.uk/IPI) using the Mascot program version 2.2 (http://www.matrixscience.com, Matrixscience, London, UK). Peptide MS/MS assignments were filtered as the criteria: peptide tolerance of parental ion was set at 50 ppm and MS/MS tolerance at 0.8 Da. Carbamidomethylation of cysteine (+ 57 Da) and oxidation of methionine (+ 16 Da) were chosen as variable modifications. In addition, one missed cleavage of trypsin was allowed for peptide identification. MS and MS/MS spectra with Mascot scores higher than the probability (P < 0.01 and P < 0.05, respectively) were considered valid. As an indication of identification certainty, the false discovery rate for peptide and protein matches above identity threshold was calculated by Peptide Validator at 1.0%. The peptide score is − 10 × Log (P), where P is the probability that anobserved match is a random event. Individual peptide scores were considered as identity or extensive homology (P < 0.01).

For statistical evaluation, a two-sided t-test was used. The p value was corrected using false discovery rate (FDR) based multiple hypothesis testing. Both t-test and FDR based multiple hypothesis testing were carried out with the default settings of the Perseus statistics software. All data were normalized using a linear regression analysis.

Protein–protein interaction analysis

To infer the biological functions of plasma membrane and ECM proteins through their interaction network, protein network analysis was performed using Ingenuity Pathways Analysis (IPA, http://www.ingenuity.com). IPA provides the identified protein interaction network on the basis of a regularly updated “Ingenuity Pathways Knowledge-base.” Ingenuity Pathways Knowledge-base is an updatable database containing millions of individual protein–protein relationships retrieved from literature (Rather et al. 2022). Network generation is optimized for the inclusion of many proteins as possible from the input expression profile and aimed at the production of highly connected networks.

Immunoblot analysis

ECM protein expression levels were confirmed by immunoblot analysis as described previously (Kwon et al. 2010). ECM protein lysates were prepared in a lysis buffer containing 50 mM Tris–HCl (pH 7.5), 150 mM NaCl, 2 mM EDTA, 1 mM EGTA, 1 mM Na3VO4, 10 mM NaF, 1 mM DTT, 1 mM PMSF, 25 μg/ml leupeptin, 25 μg/ml aprotinin, 5 mM benzamidine, and 1% Ingepal CA630. Protein concentrations were determined using the Bio-Rad protein assay kit according to the manufacturer’s protocol. Equal amounts of ECM proteins (45 μg) isolated from passage #15 and #40 HDFs were resolved by SDS-PAGE and transferred onto Immobilon PVDF membranes. Blots were blocked with a solution containing 5% nonfat dried milk or 5% BSA and 0.1% Tween 20 and incubated with primary antibody in blocking solution overnight. And the blots were washed and further probed with HRP-conjugated anti-rabbit IgG (1: 5,000). The immune complexes were visualized using the ECL detection system according to the manufacturer’s protocol. Antibodies against FN1, VTN, HTRA1, CAV1, and APOB were used for immunoblot analysis.

Statistical analysis

Significant differences between the mean values of expressed protein levels from passage #15 and #40 HDFs were determined by Student t-test with Graph-Pad Prism (GraphPad, San Diego, CA). Data were presented as means ± standard deviation from triplicate performances. Statistical P values less than 0.05 were considered as significant.

Results and discussion

In the present study, we examined the senescence states of passage #15, #30, and #40 HDFs using SA-β-Gal staining. According to previous reports, SA-β-Gal positive cell% in young HDFs showed less than 5% in young cells to ~ 80% in senescent cells (Jang et al. 2019; Yang et al. 2020). Likewise, SA-β-Gal positive cell% of passage #15 was 14%, while those cell% of passage #30 and #40 HDFs were 46% (P < 0.01) and 83% (P < 0.001) stained, respectively (Fig. 1). The passage number that reaches senescence depends on the origin of the HDF cells. ATCC-derived HDFs have a passage number of 17 to reach senescence, which is shorter than the passage number 35 to reach senescence in SNU-derived cells (Yang et al. 2020; Jang et al. 2020). In our study, we defined passage #30 and #40 cells as early and late senescent state on the basis of SA-β-Gal staining.

Fig. 1
figure 1

SA-β-Gal staining of HDFs. A Light microscopic images of young HDFs (passage #15) and late senescent HDFs (passage #40) are presented. B Quantification of SA-β-Gal-positive cells% of HDFs. Data represent the mean ± standard deviation of triplicate experiments. **Statistical significance P < 0.001

Full size image

Next, we characterized the ECM proteins of young and late senescent HDFs to identify late senescence-related proteins (Additional file 1: Table S1, Additional file 2: Table S2, Additional file 3: Table S3 and Additional file 4: Table S4). Protein separation and analysis were accomplished by SDS-PAGE and in-gel tryptic digestion, followed by high-resolution 7-Tesla FT-ICR. The use of these techniques is known as shotgun proteomics as described previously (Cargile et al. 2004). ECM proteins from young and late senescent HDFs were resolved by SDS-PAGE and each lane was cut into 12 gel pieces for subsequent in-gel tryptic digestion (data not shown). The trypsin-digested peptides were separated by reverse-phase LC and identified by 7-Tesla FT-ICR MS analysis. Shotgun proteomic analysis revealed that MS/MS spectra from young (111,817 spectra) and late senescent HDFs (99,255 spectra) resulted in 3458 and 3287 peptides, respectively. Using Mascot search against human IPI dataset, a total of 189 proteins were finally identified, which the proteins were categorized on the basis of intracellular localization of IPA annotation: cytoplasm (40%), ECM (35%), nucleus (2%), plasma membrane (10%), and unknown (14%). Semi-quantitative analysis of ECM proteins was performed to calculate the mol% of individual proteins derived from the exponentially modified protein abundance index (Ishihama et al. 2005). Fold changes in mol% were also calculated to compare the relative abundance of ECM proteins between young and late senescent HDFs, as listed in Table 1. As shown in Table 1, 50 ECM proteins were differentially expressed: 15 ECM proteins were up-regulated and 50 ECM proteins were down-regulated in late senescent HDFs compared to young HDFs.

Table 1 List of differentially abundant ECM proteins between passage #15 and #40 HDFs
Full size table

In order to explore hub proteins in the extracellular region of late senescent HDFs, a total of 83 proteins expressed in the plasma membrane and ECM were queried to IPA, resulting in a distinct network inter-connected with 59 proteins (Fig. 2). IPA is a web-based software that facilitates the network of protein–protein interactions in silico based on the collated interactions and associations reported in the literature (Deighton et al. 2010). As shown in IPA analysis, FN1 was at the center of the protein network. FN1, a major ECM glycoprotein, binds to integrin and other ECM proteins either directly or indirectly and is involved in various biological functions, such as adhesion, growth, migration, and differentiation (Ksiazek et al. 2009). FN1-associated proteins included VTN, CAV1, APOB, and HTRA1 (Fig. 2). FN1 was exclusively observed in young HDFs, whereas APOB and HTRA1 were mainly present in late senescent HDFs (Table 1). The relative abundance of VTN and CAV1 was not significantly different between young and late senescent HDFs. However, VTN and CAV1 are directly linked to FN1 hub protein. Thus, biochemical validations of FN1, APOB, HTRA1 including VTN and CAV1 were conducted. Additionally, the possible roles of target ECM proteins directed connected to FN1 were described hereafter.

Fig. 2
figure 2

Interaction network of differentially abundant ECMs and secretory proteins between passage #15 and #40 HDFs. Protein network analysis was performed by Ingenuity Pathway Analysis. Up- and down-regulated proteins are colored as red and green in late senescent HDFs, respectively. Gray-colored symbols indicate proteins with net fold changes less than 1.5. Solid and dotted lines indicate direct and indirect relationship, respectively

Full size image

The protein levels of VTN and CAV1 were significantly altered in late senescent HDFs. The immunoblot analysis revealed that VTN was slightly reduced but CAV1 dramatically increased in late senescent ECM (Fig. 3). The differences between passage #15 and #40 HDFs were detected significantly at P < 0.001. Interestingly, FN1 is known to be an age-related marker protein due to its increased protein and mRNA levels in senescent cells (Matos et al. 2012). Contrast to the previous report, the expression levels of FN1 in late senescent HDFs were observed down-regulated. We suggest that the cellular FN1 is involved in a fibrillary network forming detergent-insoluble fibrils and the fibrillary fibronectin has been reduced or absent on the surface of late-passage cells as cited previously (Wierzbicka-Patynowski et al. 2004; Wang et al. 2022). Furthermore, cellular aging is associated with the proteolytic degradation of ECM proteins (Labat-Robert 2003). Increased elastase-type activity, gelatinase, Ca2+-dependent fibronectinase, matrix metalloproteinases (MMPs), and cathepsin G have been implicated in FN fragmentation in the ECM (Hibbert et al. 2019). Therefore, proteolysis of FN may be responsible for the reduction of FN in ECM of senescent fibroblasts.

Fig. 3
figure 3

Immunoblot analysis of ECM proteins. A Images of expressed ECM proteins in passage #15 and #40 HDFs. Equal amount of ECM proteins (45 μg) from sub-confluent young and late senescent fibroblasts were used for immunoblot analysis with antibodies against FN1, VTN, HTRA1, CAV1, and APOB. B Quantification of protein expression levels. Relative expression levels of ECM proteins were quantified with β-actin as a house-keeping protein. Quantification was performed using Image J program. ***Statistical significance P < 0.001

Full size image

In the present study, VTN was slightly down-regulated in the matrix of late senescent HDFs (Table 1 and Fig. 3). Similar to FN1, cellular aging-associated proteolysis of VTN may be responsible for decreased VTN in late senescent ECM. VTN plays a role in several processes, such as hemostasis, vascular remodeling, phagocytosis, tumor growth and metastasis, wound healing, and viral infection (Budatha et al. 2021). Indeed, the binding of urokinase plasminogen activator (uPA) to uPA receptor (uPAR) induces the activation of plasminogen to plasmin, which favors proteolysis of ECM proteins and converts VTN into defined fragments (Kost et al. 1996). VTN is considered a promising rejuvenation marker due to its anti-aging effect. Since CAV1 is overexpressed in dermal and subdermal cells during chronological and UV-induced aging, CAV1 has been suggested to play a key role in cellular senescence and skin aging (Jang et al. 2019). CAV1 directly interacts with receptors and signaling molecules, such as transforming growth factor beta (TGF-β), toll-like and glucocorticoid receptors, MMPs, and heat shock proteins (Kruglikov et al. 2019). Since CAV1 is highly expressed in different hyper-proliferative and inflammatory skin diseases, the regulation of CAV1 is considered a treatment strategy in dermatology (Kruglikov and Scherer 2019). Although CAV1 has been known to be age-related protein marker, we did not detect CAV1 by shotgun proteomics in the present study. Presumably, it is due to the limitation of protein sample preparation. However, the immunoblot analysis of passage #15 and #40 HDFs revealed to be significantly increased in late senescent HDFs (Fig. 3). These results are consistent with those of previous reports (Kim et al. 2010; Jang et al. 2019).

APOB, the main protein found in low-density lipoprotein (LDL), selectively interacts with the constituents of ECM, such as collagen, proteoglycans, and glycosaminoglycans (Hevonoja et al. 2000). The present study showed that five apolipoprotein isoforms, APOA1, APOA2, APOB, APOC3, and APOE, were directly or indirectly linked to FN1 subnet hub protein (Fig. 2). Among them, APOA1, APOC3, and APOE were mainly present in young HDFs, whereas APOB was exclusively present in late senescent HDFs (Table 1). The protein levels of APOB were significantly higher in late senescent HDFs than in young cells (Fig. 3). These results are consistent with the previous report demonstrating that the accumulation of APOB was associated with aging in retinal pigment epithelium (Curcio et al. 2010). HTRA1 is a secretory protein that is indirectly linked to FN1 and functions as a heat shock serine protease activated by external stress (Yang et al. 2006). HTRA1 has been validated as an aging marker of age-related macular degeneration (Gu et al. 2009). Herein, the expression levels of HTRA1 were approximately 4.6-fold higher in the ECM of late senescent fibroblasts than in young cells (Table 1). As expected, the protein levels of HTRA1 were significantly higher in late senescent HDFs than in young cells (Fig. 3). Interestingly, HTRA1 is known to inhibit the signaling of TGF-β family proteins (Oka et al. 2004). Recently, HTRA1 has been associated with frailty status in older adults (Lorenzi et al. 2016). Since chronic inflammation may be involved in the pathogenesis of frailty, HTRA1, as an inhibitor of anti-inflammatory cytokine TGF-β, seems to be a causative factor for frailty in older adults. HTRA1 has also been implicated in several age-associated diseases, such as Alzheimer's disease (Grau et al. 2005).

Besides the proteins mentioned above, it is notable to describe several proteins expressed exclusively in young or late senescent state of HDFs. The expression levels of GDF15 belonged to proteins expressed exclusively in late senescent state are increased in UV-irradiated senescent fibroblasts (Kim et al. 2020). The cytoskeletal protein, desmosome (JUP), belonged to exclusive expressions in late senescent HDFs, were reduced in aged skin tissue (Oender et al. 2008). Next, the proteins expressed exclusively in early stage contain PON1, Eln, IGFBP-3, and COL6A1. Paraoxonase 1 (PON1)-targeted human dermal microvascular endothelial cells are turned into cellular senescence (Lee et al. 2012). Likewise, the elastin (Eln) is associated with a loss in aging process (Lasio and Kozel 2018) and the function of IGFBP-3 is considered important at early stage of aging (Song et al. 2018). Mutation of collagen VI containing COL6A1 may lead to skin abnormality (Lettmann et al. 2014).

Conclusions

Shotgun proteomics combined with IPA-based protein network analysis presented that FN1-centered ECM and secretory proteins were differentially expressed in young and late senescent HDFs. The finally identified proteins including FN1, VTN, APOB, and HTRA1 were significantly altered in late senescent HDFs compared to young cells. Comparative proteomic approach using late senescent HDFs clearly revealed potential aging-related markers. The applications to these ECM proteins might give some clues to cure the aged human skin.