Endoplasmic reticulum stress-induced senescence in human lung fibroblasts

Abstract

Loss of proteostasis and cellular senescence have been previously established as characteristics of aging; however, their interaction in the context of lung aging and potential contributions to aging-associated lung remodeling remains understudied. In this study, we aimed to characterize endoplasmic reticulum (ER) stress response, cellular senescence, and their interaction in relation to extracellular matrix (ECM) production in lung fibroblasts from young (25–45 yr) and old (>60 yr) humans. Fibroblasts from young and old patients without significant preexisting lung disease were exposed to vehicle, MG132, etoposide, or salubrinal. Afterward, cells and cell lysates or supernatants were analyzed for ER stress, cellular senescence, and ECM changes using protein analysis, proliferation assay, and senescence-associated beta-galactosidase (SA-β-Gal) staining. At baseline, fibroblasts from aging individuals showed increased levels of ER stress (ATF6 and PERK), senescence (p21 and McL-1), and ECM marker (COL1A1) compared to those from young individuals. Upon ER stress induction and etoposide exposure, fibroblasts showed an increase in senescence (SA-β-Gal, p21, and Cav-1), ER stress (PERK), and ECM markers (COL1A1 and LUM) compared to vehicle. Additionally, IL-6 and IL-8 levels were increased in the supernatants of MG132- and etoposide-treated fibroblasts, respectively. Finally, the ER stress inhibitor salubrinal decreased the expression of p21 compared to vehicle and MG132 treatments; however, salubrinal inhibited COL1A1 but not p21 expression in MG132-treated fibroblasts. Our study suggests that ER stress response plays an important role in establishment and maintenance of a senescence phenotype in lung fibroblasts and therefore contributes to altered remodeling in the aging lung.

NEW & NOTEWORTHY The current study establishes functional links between endoplasmic reticulum (ER) stress and cellular senescence per se in the specific context of aging human lung fibroblasts. Recognizing that the process of aging per se is complex, modulated by the myriad of lifelong and environmental exposures, it is striking to note that chronic ER stress may play a crucial role in the establishment and maintenance of cellular senescence in lung fibroblasts.

INTRODUCTION

Senescence is a well-defined cellular state that is a feature of tissue aging (1). Lung tissues from patients with age-related chronic pulmonary diseases such as chronic obstructive pulmonary diseases (COPD), pulmonary fibrosis (PF), and asthma in the elderly show a greater level of cellular senescence compared to their age-matched controls (2–4). The normal aging lung, as well as those with age-related diseases, have increased airway and parenchymal remodeling with altered cell proliferation (5–7). Accordingly, understanding the relationship between aging, senescence, and remodeling becomes relevant for identifying biomarkers of health versus disease in aging, and toward the development of novel therapeutic approaches.

A myriad of stimuli including telomere attrition, genotoxic drugs (e.g., etoposide), irradiation, oxidative stress, proteostasis loss, mitochondrial dysfunction, genomic instability, and epigenetic alterations are known as triggers for cellular senescence (8, 9). Telomere shortening leads to breaks in the double-stranded DNA, followed by recruitment of DNA damage response factors including phosphorylated histone H2AX (γ-H2AX) which upregulates expression of p21 that causes cell cycle arrest in the G₁ phase (10).

Increased levels of p21 and/or p16, DNA damage, and increased levels of antiapoptotic proteins characterize senescence (11–13). Senescent cells remain metabolically active and secrete different molecules commonly named senescence-associated secretory profile (SASP), which include proinflammatory cytokines and chemokines, and extracellular matrix (ECM) proteins (11, 12). SASP is considered a critical aspect of how senescent cells can change their milieu to influence their microenvironment. ECM remodeling is an important part of lung aging, and previous studies have demonstrated a decreased expression of decorin in senescent lung fibroblasts (3). Furthermore, higher levels of collagen type I alpha 1 (COL1A1) with higher age were demonstrated in different lung compartments including parenchymal regions (14). Accordingly, understanding the impact of aging per se, or of extrinsic factors, on senescence, and furthermore, the impact of senescence on tissue health and disease is critical to identifying novel approaches to attenuate detrimental effects of aging.

Proteostasis comprises mechanisms such as protein synthesis, folding, stability, trafficking, and degradation which contribute to proper cellular protein processing and prevention of abnormal protein accumulation (15). Maintenance of proteostasis requires optimal functioning of the cellular translational machinery; chaperones, autophagy, and the ubiquitin-proteasome system (16). An imbalance between these processes results in loss of proteostasis, which is characterized by increased endoplasmic reticulum (ER) stress caused by accumulation of misfolded or unfolded proteins. The ER stress response can be regulated through three different pathways including the inositol-requiring element 1 alpha (IRE1α), activating transcription factor 6 (ATF6), and protein kinase RNA-like ER kinase (PERK) pathway. The IRE1α and PERK pathways have been shown to be impaired with aging including in the lung (17, 18), which are strong indicators of an ER stress environment. Chronic ER stress does not necessarily lead to cell death as it can also lead to cellular senescence. For example, normal human dermal fibroblasts show increased levels of senescence markers upon exposure to ER stress inducers such as thapsigargin, tunicamycin, and dithiothreitol (19). Accordingly, recognizing that ER stress can promote cellular senescence, and that senescence is related to ECM changes in lung fibroblasts derived from COPD patients (3), exploring interactions between ER stress and cellular senescence in the context of the aging lung becomes significant toward understanding aging-associated lung remodeling. Lung fibroblasts are structural cells that are considered one of the main sources of ECM proteins. Investigating the role of fibroblasts in aging-associated lung remodeling will be critical. Therefore, in this study we aimed to characterize the ER stress response, cellular senescence, and their interaction in relation to ECM production in lung fibroblasts from humans with increasing age.

MATERIALS AND METHODS

Ethics Statement and Acquisition of Primary Lung Fibroblasts

Primary human lung fibroblasts were isolated from leftover lung tissues of patients without chronic lung diseases including COPD, idiopathic PF, and asthma undergoing resection for focal lung carcinoma at St. Mary’s Hospital (Mayo Clinic Rochester, MN). Patients were classified as young (25 - 45 yr) or old (≥ 60 yr) and both sexes were included. The study was approved by Mayo Clinic Institutional Review Board No. 16-009655. Patient consent (written or video/verbal) was obtained during clinic visits before surgical decisions, and relevant clinical data were collected upon acquisition of lung tissues. For anonymization/deidentification and storage, samples were assigned unique numbers unrelated to any patient identifier (patient number, age, sex, date of birth, date of surgery, etc.). Primary lung fibroblasts were isolated following a previously described protocol (20).

Patient Characteristics

Lung fibroblasts were isolated from young versus old patients without significant chronic lung disease, with no history of smoking. Fibroblasts were isolated from macroscopically normal lung tissues taken distant from the tumor. The young group had a median age of 40 yr, while the old group had a median age of 74 yr. Detailed patient characteristics are depicted in Table 1.

Table 1.

Patient characteristics

	Age (median)	Number of Patients	Sex (male/female)	Smoking Status	FEV1, mean ± SEM
Young patients	40 (33–45)	5	2/3	Nonsmokers	82.75 ± 3.97
Old patients	74 (62–84)	11	8/3	Nonsmokers	75.98 ± 7.01

Values are median for age, male/female for sex, and means ± SE for forced expiratory volume in 1 s (FEV₁). Pulmonary function tests indicate forced expiratory volume in 1 s above 80% of the predicted value for 2 old patients. For 1 patient (old patient), the pulmonary function tests are abnormal. No pulmonary function data were available for 1 young patient and 2 old patients. The FEV₁ of these 6 patients are not included in the calculated FEV₁ mean.

Cell Culture and Treatments

Primary lung fibroblasts were cultured in Dulbecco’s minimum essential medium (DMEM) with low glucose (1 g/L; Gibco, lot no. 2323513) supplemented with 10% fetal bovine serum (R&D Systems, lot no. G22143) and 1% antibiotics and antimycotic (AA; Gibco, lot no. 2441424) (growth media) at 37°C and 5% CO₂. Cells at less than passage five of culture were used. For experiments, 12-well plates were seeded with 50,000 cells per well and incubated for 24 h at 37°C and 5% CO₂. Then, fibroblasts were made quiescent by incubation in DMEM low glucose supplemented with 0.1% bovine serum albumin and 1% antibiotics and antimycotic for 24 h. Subsequently, fibroblasts were cultured in growth media containing 0.01% DMSO (vehicle), the ER stress inducer MG132 (300 nM; treatments every 24 h for 3 days), and etoposide (10 µM; treatment for 2 days) as a positive control for cellular senescence. Finally, media were changed to normal growth medium, and cells were incubated for an additional 5 days at 37°C and 5% CO₂. Samples were harvested at two time points, early (56 h) and late (8 days), and analyzed for protein expression levels of ER stress, ECM, and senescence markers (Fig. 1). To investigate the effect of ER stress inhibitor salubrinal on the expression of senescence markers and ECM protein, fibroblasts were treated in parallel at different time points with MG132 (300 nM; treatments every 24 h for 3 days) and salubrinal (12.5 µM; treatment every 24 h for 3 days) (see Fig. 10A). In the second experimental design (see Fig. 10B), we treated fibroblasts sequentially with MG132 and salubrinal.

Figure 1.

In vitro ER stress-senescence model in primary lung fibroblasts. This schematic depicts the exposure protocol; the different time points for sample harvesting for measurement of ER stress, extracellular matrix, and senescence markers; and the time points for the determination of cell viability. [Image created with a licensed version of BioRender.com.]

Figure 10.

Salubrinal inhibits the expression of p21 and COL1A1. (A) Lysates from fibroblasts treated with vehicle (0.01% DMSO), ER stress inducer MG132 (300 nM), or ER stress inhibitor salubrinal (12.5 μM) harvested at day 8 (experimental design 1) were analyzed for p21, Cav-1, and COL1A1 expression. Salubrinal treatment showed a significant reduction in p21 and COL1A1 levels compared to vehicle and MG132 treatments. For statistical analysis, one-way ANOVA, Friedman test was applied. Data from 7 (4 old and 3 young) patients shown. (B) Lysates from fibroblasts treated with vehicle (0.01% DMSO), ER stress inducer MG132 (300 nM), and MG132 treatment followed by salubrinal (12.5 μM) (experimental design 2) were analyzed for the expression levels of p21, Cav-1, and COL1A1. No significant change was observed in p21 expression after salubrinal treatment. Additionally, salubrinal reduced COL1A1 and increased Cav-1 levels in MG132-treated fibroblasts. For statistical analysis, one-way ANOVA, Friedman test was applied. Data from 9 (8 old and 1 young) patients shown. [Image created with a licensed version of BioRender.com.]

Cell Viability Assay

Fibroblasts exposed to vehicle, MG132 (300 nM; 3 days) or etoposide (10 µM; 2 days) were mixed 1:1 ratio with Trypan Blue and analyzed for cell viability on days 5 and 8, using Countess II from Invitrogen (Carlsbad, CA).

Proliferation Assay

To examine the effect of MG132 or etoposide on fibroblast proliferation, the CyQUANT NF Cell Proliferation Assay kit (no. C35006) from Invitrogen (Carlsbad, CA) was used. Fibroblasts (1,000 cells/well) were seeded in a flat bottom clear, black polystyrene 96-well plate and cultured as described above (Fig. 1). Before the assay was performed, the growth media were replaced with 50 µL/well of 1× dye binding solution which contains 1× Hanks’ balanced salt solution, 1× dye delivery reagent, and 1× CyQUANT NF dye reagent. The 1× dye binding solution was prepared according to the manufacturer’s protocol. Fibroblasts were incubated for 1 h at 37°C to allow the equilibration of dye-DNA binding, resulting in a stable fluorescence end point. Then, the plate was read with excitation at 485 nm and emission at 530 nm using a FlexStation 3 from Molecular Devices (San Jose, CA).

Protein Analysis

Protein expression was quantified using a capillary-based electrophoresis system (ProteinSimple WES, Santa Clara, CA). Samples were prepared following the manufacturer’s protocol. ER stress, senescence markers, and ECM proteins were analyzed using the primary antibodies listed in Table 2. All of the following secondary antibodies were purchased from ProteinSimple: Anti-Mouse Secondary HRP Antibody (no. 042-205), Anti-Rabbit Secondary HRP Antibody (no. 042-206), and Anti-Goat Secondary HRP Antibody (no. 043-522-2). We used the Simple Western Total Protein Assay (no. DM-TP01) to normalize protein expression. The images of the capillary showing the specificity of the used primary antibodies are available in Supplemental File S3.

Table 2.

Antibodies used for the detection of ER stress, senescence, and ECM markers

Markers/Primary Antibody	Catalog Number	Vendor	Dilution
ER stress markers
ATF6 (21)	65880	Cell Signaling Technology	1:25
BiP (21)	3177	Cell Signaling Technology	1:100
IRE1α (21)	3294	Cell Signaling Technology	1:25
PERK (21)	5683	Cell Signaling Technology	1:50
Ubiquitin (22)	43124	Cell Signaling Technology	1:50
XBP1s (21)	40435S	Cell Signaling Technology	1:50
Senescence markers
p16 (23)	ab108349	abcam	1:50
p21 (24)	ab109520	abcam	1:50
LMNB1 (25)	ab133741	abcam	1:50
Cav-1 (26)	ab211503	abcam	1:50
Mcl-1 (27)	5453	Cell Signaling Technology	1:50
Bcl-xL (28)	2764	Cell Signaling Technology	1:50
ECM markers
COL1A1 (21)	NBP1-30054	NOVUS Biologicals	1:50
COL6A2 (29)	NBP1-90951	NOVUS Biologicals	1:50
FN1 (21)	ab2413	abcam	1:100
LUM (30)	ab168348	abcam	1:50
MMP2 (31)	4022	Cell Signaling Technology	1:50

Numbers in parentheses are references. ATF-6, activating transcription factor 6; BiP, binding immunoglobulin protein; COL1A1, collagen type I alpha 1; COL6A2, collagen type VI alpha 2; ECM, extracellular matrix; ER, endoplasmic reticulum; FN1, fibronectin 1; IRE1α, inositol-requiring element 1 alpha; LUM, lumican; MMP2, matrix metalloproteinase 2; PERK, protein kinase RNA-like ER kinase; XBP1s, spliced X-box binding protein 1.

ELISA

Duoset ELISA kits (R&D Systems, no. DY206 anno. ab108345d no. DY208) were used to measure interleukin-6 (IL-6) and interleukin-8 (IL-8), respectively, following the manufacturer’s protocol. Absorbance was read at 450 and 530 nm using the FlexStation 3 from Molecular Devices (San Jose, CA). A standard curve was generated to calculate the cytokine concentrations in samples, and the concentration was normalized to cell number.

Beta-Galactosidase Staining

The Senescence Beta-Galactosidase Staining Kit from Cell Signaling (no. 9860) was used. Lung fibroblasts cultured in six-well plates and exposed to 0.01% DMSO (vehicle), MG132, and etoposide were measured at time point day 8 by rinsing once with 1× phosphate-buffered saline (PBS), then 1× fixative solution was added to each well, and the samples were incubated for 15 min at room temperature. The plate was then washed twice with 1× PBS; 1 mL of the beta-galactosidase (β-Gal) staining solution, pH 6.0, was added to each well; and the plate was sealed with parafilm and incubated overnight for 18 h at 37°C in a dry incubator without CO₂. The staining solution was discarded and the plate was washed once with 1× PBS. Cells were stained with DAPI (4′,6-diamidino-2-phenylindole dihydrochloride) (ThermoFisher; no. D1306) for 15 min and washed once with 1× PBS and 1 mL of 70% glycerol solution was added to each well. Total and β-Gal-positive cell numbers were counted manually and the percentage of β-Gal-positive cells was determined.

Statistical Analyses

Data were analyzed using GraphPad Prism 9.4.1 (GraphPad, San Diego, CA). For statistical analysis, Mann-Whitney U tests were used to test differences between fibroblasts from young control patients compared with old control patients. The effects of MG132, etoposide, and salubrinal treatments were analyzed using one-way ANOVA. The statistical methods used to analyze each set of data are highlighted in the figure legends. P < 0.05 was considered statistically significant.

RESULTS

Aging Lung Fibroblasts Show Higher Levels of ER Stress and Senescence Markers

Cell lysates from lung fibroblasts of young versus old patients that were cultured for 8 days in normal growth medium were examined for expression of ER stress, senescence markers, anti-apoptotic proteins (e.g., Mcl-1 and Bcl-xL), and ECM proteins. Fibroblasts from old patients had increased levels of ATF6, PERK, p21, and Mcl-1 compared to fibroblasts from young patients (Fig. 2, A–C). Additionally, a trend toward increased COL1A1 was noted in fibroblasts from old patients compared to those from young. No significant differences were observed in IRE1α, spliced X-box binding protein 1 (XBP1s), binding immunoglobulin protein (BiP), p16, Bcl-xL, caveolin-1 (Cav-1), collagen type VI alpha 2 (COL6A2), fibronectin 1 (FN1), lumican (LUM), and matrix metalloproteinase 2 (MMP2) levels between the two age groups.

Figure 2.

Lung fibroblasts from old patients express higher levels of ER stress and senescence markers and extracellular matrix (ECM) proteins. Lung fibroblasts from young (25–45 yr) and old (>60 yr) patients were cultured for 8 days and cell lysates were analyzed for protein expression using a capillary-based electrophoresis system (Wes) (A) ER stress markers (ATF6, PERK, IRE1α, BiP and XBP1s) (B) senescence markers (p16, p21, Bcl-xL, Cav-1, and Mcl-1); and (C) ECM proteins (COL1A1, COL6A2, FN1, LUM and MMP2). For statistical analysis, Mann-Whitney test was applied; n = 5–7 patients per group.

MG132 Induces ER Stress Response and Senescence-Like Phenotype in Lung Fibroblasts

To investigate the role of ER stress in the development of cellular senescence, fibroblasts were exposed to the proteasome inhibitor MG132 (300 nM). As a positive control for senescence, the well-known senescence inducer etoposide (10 μM) was used. Protein expression of ER stress and senescence markers was measured in cell lysates (Fig. 3). The expression levels of ER stress markers IRE1α, PERK, ATF6, XBP1s, BiP, and ubiquitinated proteins (Fig. 3A) and senescence markers p21 and LMNB1 (Fig. 3B) were examined.

Figure 3.

MG132 induces ER stress response and senescence-like phenotype in lung fibroblasts. Cells from young and old patients were treated with vehicle (0.01% DMSO), ER stress inducer MG132 (300 nM), or etoposide (10 μM). Cell lysates collected at 56 h were analyzed for (A) ER stress markers (ubiquitinated proteins, IRE1α, ATF6, PERK, XBP1s and BiP) and (B) senescence markers (p21 and LMNB1). Protein expression was normalized to total protein. For statistical analysis, one-way ANOVA, Friedman test was applied. Data are presented for 12 (7 old and 5 young) patients.

MG132-treated fibroblasts showed an increased level of ubiquitinated proteins compared to vehicle, while no difference was observed between etoposide-treated fibroblasts and vehicle. MG132- and etoposide-treated fibroblasts showed increased levels of IRE1α. However, the downstream effector of IRE1α, XBP1s, was significantly decreased only in MG132-treated fibroblasts compared to vehicle control. The ER stress markers ATF6, PERK, and BiP showed no changes upon MG132 and etoposide treatment compared to vehicle.

Quantification of the senescence markers p21 and LMNB1 at an early time point 56 h (injury response) showed an increased level of p21 in MG132 and etoposide-treated fibroblasts compared to vehicle. LMNB1 level was decreased in MG132- and etoposide-treated fibroblasts compared to vehicle. The response after exposure to MG132 or etoposide was similar in young and old fibroblasts (Supplemental File S2; Fig. 1).

MG132 Inhibits Proliferation of Lung Fibroblasts

Cellular senescence is characterized by cell cycle exit which manifests as cell growth arrest. In addition to the cellular senescence markers that have been previously analyzed, we investigated the effects of MG132 and etoposide on cell proliferation. We found MG132 and etoposide significantly inhibited proliferation of lung fibroblasts compared to vehicle (Fig. 4A). MG132- and etoposide-treated fibroblasts showed comparable growth rates. The percentage of living cells was significantly lower at day 5 or 8 following etoposide treatment compared to vehicle (Fig. 4B). To be sure that the lower number of cells after etoposide and MG132 were not related to a higher rate of cell death, we performed a cell viability assay. The percentage of dead cells was significantly higher at day 5 or 8 following etoposide treatment compared to vehicle. No significant differences were observed in the percentage of living and dead cells at day 5 or 8 following MG132 treatment compared to the vehicle.

Figure 4.

MG132 inhibits proliferation of lung fibroblasts. Lung fibroblasts from young and old patients were treated with vehicle (0.01% DMSO), ER stress inducer MG132 (300 nM), or etoposide (10 μM). Fibroblasts were treated on days 1, 2, and 3. (A) CyQUANT proliferation assay was performed at day 0, 3, 5, and 8 which showed substantial and progressive reduction in proliferation by both MG132 and etoposide. (B) Cell viability was assessed by Trypan Blue staining which showed increased cell death in these groups. For statistical analysis, two-way ANOVA multiple comparisons was applied. Data are presented for 5 (3 old and 2 young) patients.

MG132 Induces Cellular Senescence in Lung Fibroblasts

Fibroblasts treated with MG132 or etoposide and harvested at day 8 showed a significant increase in SA-β-Gal activity, compared to vehicle (Fig. 5A). Similar to the early time point, increased levels of p21 were found at day 8 in MG132- and etoposide-treated cells together with increased levels of caveolin-1 (Cav-1). The response after exposure to MG132 or etoposide was similar in young and old fibroblasts (Supplemental File S2; Fig. 2). As LMNB1 protein expression was not detectable at day 8, we measured the mRNA level of LMNB1 and found it decreased in MG132- and etoposide-treated fibroblasts compared to vehicle (Supplemental File S1; Fig. 1). No differences were observed for p16, Bcl-xL, and Mcl-1 levels in these groups in comparison to vehicle (Fig. 5B). The quantitative RT-PCR analysis showed decreased mRNA for CDKN2A (p16) in both MG132- and etoposide-treated fibroblast compared to vehicle, whereas a trend toward an increase was detected for CDKN1A (p21) mRNA expression in etoposide-treated fibroblasts (Supplemental File S1; Fig. 1). Young fibroblasts exposed to MG132 and etoposide did not show a different response compared to old fibroblasts.

Figure 5.

MG132 induces cellular senescence in lung fibroblasts. Cells from young and old patients were treated with vehicle (0.01% DMSO), ER stress inducer MG132 (300 nM), or etoposide (10 μM). (A) MG132- and etoposide-treated fibroblasts showed an increased percentage of SA-β-Gal positive cells. SA-β-Gal staining was performed at day 8 and normalized to total cell number. For statistical analysis, repeated measures one-way ANOVA was applied. Data are presented for 6 (3 young and 3 old) patients. (B) Levels of p16, p21, Cav-1, Mcl-1, and Bcl-xL were analyzed in cell lysates at day 8. p21 and Cav-1 levels were significantly increased by MG132 and etoposide. For statistical analysis, one-way ANOVA, Friedman test was applied. Data are presented for 12 (7 old and 5 young) patients. Scale bar = 10 µm.

The ER Stress Response Is Modulated in Lung Fibroblasts following Senescence Induction

Given our findings that MG132 and etoposide induce senescence in fibroblasts, we investigated if there is a subsequent, sustained ER stress response in the cells after cessation of MG132 or etoposide, which would suggest a downstream effect of senescence. We measured IRE1α, PERK, ATF6, XBP1s, and BiP in cell lysates from lung fibroblasts at day 8 (i.e., 5 days following the last exposure to MG132 or etoposide; see Fig. 1). At day 8, PERK was significantly increased in MG132- and etoposide (10 out of 12 cell lines showed an increased level of PERK) treated lung fibroblasts; however, the level of BiP, an important regulator of the activity of ER stress markers, was significantly decreased in these groups in comparison to vehicle-treated fibroblasts (Fig. 6). Additionally, etoposide-induced senescent fibroblasts showed a decreased level of IRE1α compared to vehicle-treated fibroblasts.

Figure 6.

The ER stress response is modulated in lung fibroblasts following senescence induction. Lung fibroblasts from young and old patients were initially treated with the ER stress inducer MG132 and etoposide for 3 days and allowed to grow out further until day 8. Fibroblasts at day 8 were analyzed for expression of ER stress markers (IRE1α, ATF6 and PERK) and ER stress response (XBP1s and BiP) markers. For statistical analysis, one-way ANOVA, Friedman test was applied. Data from 12 (7 old and 5 young) patients shown.

No significant difference was observed in the response of young versus old fibroblasts upon exposure to MG1312 or etoposide (Supplemental File S2; Fig. 3). Finally, the mRNA level of two additional chaperones important for the folding of proteins in the ER including calnexin (CANX) and heat shock protein 47 (HSP47) was assessed. There were no significant changes in the mRNA levels of CANX and HSP47 in MG132 and etoposide-treated fibroblasts compared to vehicle (Supplemental File S1; Fig. 2).

Susceptibility to ER Stress-Induced Senescence Is Not Age Dependent

To determine if fibroblasts from old patients are more susceptible to senescence-induction upon MG132 treatment compared to young, we compared the proteins p21, Cav-1, and BiP that showed a significant change upon exposure to MG132. A significant decrease of BiP was observed in old fibroblasts exposed to MG132 compared to vehicles. However, young fibroblasts exposed to MG132 showed no change compared to matched vehicle control (Fig. 7). No significant difference was observed for p21, Cav-1, or BiP levels while comparing the young to old fibroblasts exposed to MG132.

Figure 7.

Susceptibility to senescence induced by MG132 does not differ by age. Separate analyses were performed for the markers p21, Cav-1, and BiP which showed a significant change in fibroblasts [young and old patients combined (see Figure 6)] following exposure to ER stress inducer MG132 (300 nM). While changes in BiP differed by age, changes in p21 or Cav-1 did not. For statistical analysis, one-way ANOVA, Friedman test was applied for the comparison between vehicle and MG132 (*P < 0.05). Mann-Whitney test was applied for the comparison between young and old fibroblasts exposed to MG132. Data from 7 old and 5 young patients shown. The black dotted line represents the vehicle.

Changes in ECM Expression and SASP Expression by MG132

Cell lysates and supernatants from cells treated with vehicle, MG132, or etoposide were collected on day 8 and analyzed for ECM proteins COL1A1, COL6A2, FN1, and LUM, the regulator of ECM MMP2, and secretion of interleukin IL-6 and chemokine CXCL8. MG132- and etoposide-induced senescent fibroblasts showed a significant increase of COL1A1 level, and the LUM level was significantly increased in MG132-exposed fibroblasts compared to vehicle (Fig. 8A). No changes were observed for FN1, COL6A2, and MMP2 in MG132- or etoposide-treated groups compared to vehicle. Furthermore, the response after exposure to MG132 or etoposide was similar in young and old fibroblasts (Supplemental File S2; Fig. 4). MG132 exposure resulted in increased IL-8 levels at day 8 but not of IL-6, whereas etoposide treatment led to elevated IL-6 levels at day 8 (Fig. 8B). Further analysis of COL1A1, COL6A2 and FN1 gene expression showed a significant increase in COL1A1 gene expression in MG132-treated fibroblasts (Supplemental File S1; Fig. 3). This result correlates with the protein level of COL1A1 in MG132-treated fibroblasts compared to vehicle. Gene expression of COL6A2 was significantly decreased in etoposide-treated fibroblasts compared to vehicle, while no change was identified for FN1 (Supplemental File S1; Fig. 3).

Figure 8.

ECM and SASP are altered by MG132 and etoposide treatments. Lung fibroblasts from young and old patients were treated with vehicle (0.01% DMSO), ER stress inducer MG132 (300 nM), or etoposide (10 μM) for 3 days. (A) Cell lysates at day 8 were analyzed for expression of ECM proteins COL1A1, COL6A2, FN1, LUM and MMP2. COL1A1 and LUM were significantly changed at day 8 by earlier ER stress/senescence induction. Data from 12 (5 young and 7 old) patients shown. (B) Supernatants were analyzed for secretion of IL-6 and IL-8. Early exposure to MG-132 led to elevated IL-8 at day 8, while early etoposide exposure increased IL-6 at that time point. For statistical analysis, one-way ANOVA, Friedman test was applied. Data from 9 (5 old and 4 young) patients shown.

ER Stress Contributes to Altered Expression of p21 in Lung Fibroblasts

An ER stress response was observed in MG132-treated fibroblasts (Figs. 3 and 6). To determine whether ER stress leads to establishment of a senescent profile, we examined the association of the ER stress markers (IRE1α, ATF6, and PERK) with the senescence marker p21 at day 8 in MG132 and etoposide samples using a Spearman correlation analysis. ATF6 and PERK expression levels but not those of IRE1α were significantly positively correlated with p21 in MG132 and etoposide-treated fibroblasts (Fig. 9). Toward establishing causation per se, we used salubrinal, an ER stress inhibitor of the PERK pathway, and found decreased levels of p21 compared to vehicle and MG132 treatments (Fig. 10A). Additionally, we investigated the effect of salubrinal on Cav-1 and COL1A1 expression, and noticed that salubrinal significantly decreased level of COL1A1 compared to MG132 treatment; in contrast, salubrinal treatment showed an increased trend of Cav-1. Overall, no differences were observed in young versus old fibroblasts exposed to MG132 or salubrinal compared to vehicle. Following this experiment, we investigated the effect of salubrinal on MG132-induced ER stress and assessed the levels of p21, Cav-1, and COL1A1 (Fig. 10B). For this purpose, a different experimental design was established (Fig. 10B). Here we observed that salubrinal did not significantly decrease the MG132-induced p21 and Cav-1 levels, whereas it did significantly reduce the MG132-induced increase in COL1A1, supporting the contribution of ER stress to regulation of COL1A1 expression in primary lung fibroblasts.

Figure 9.

ER stress contributes to altered p21 expression. The association between the senescence marker p21 and ER stress markers was examined using Spearman’s correlation test which showed a positive correlation with p21 for ATF6 and PERK, but not IRE1α. Data from 12 (7 old and 5 young) patients shown.

DISCUSSION

Recognizing that senescence occurs in aging tissues and that aging-associated pulmonary diseases such as COPD, pulmonary fibrosis, and even asthma in the elderly that involve lung remodeling (including fibrosis) show greater cellular senescence, we sought to understand the links between ER stress and senescence and how this impacts ECM production in lung fibroblasts from young versus old patients. We investigated changes in expression of ER stress and senescence markers and of ECM production. We demonstrated that aging does promote an increase in ER stress and senescence in lung fibroblasts. We demonstrated a role for ER stress induction in the establishment of a senescence phenotype in fibroblasts, although the susceptibility for this link does not appear to be age-dependent. Some ER stress markers are positively associated with expression of senescence markers, but importantly modulation of ER stress modulates senescence suggesting a common link. Furthermore, senescence per se may maintain ER stress or altered proteostasis. Overall, our results suggest that ER stress response plays an important role in establishment and maintenance of a senescence phenotype in lung fibroblasts. In turn, the senescent fibroblast may contribute to the altered remodeling of the aging lung, which is relevant to aging-associated lung diseases.

There are currently limited data on ER stress and unfolded protein response (UPR) in the aging lung. An age-related decline in UPR has been reported in the lungs of old rats where UPR markers including BiP, activating transcription factor 4 (ATF4), and total and phosphorylated eukaryotic translation initiation factor 2A (p-eIF2a) are decreased, while PERK is increased (18). Alveolar type 2 cells (AEC2) from old mice express low levels of BiP and increased levels of glucose-regulated protein 94 (GRP94) and C/EBP homologous protein (CHOP) to UPR (32). From a replicative perspective, fibroblasts at a greater number of passages express less calnexin compared to those of earlier passages (33). Levels of heat shock protein 70 (HSP70) are also decreased with replicative cycles in lung fibroblasts (34). These data indicate an aging-related decline in chaperone function in lung cells that leads to a downstream impact on the ER and/or its functionality. Conversely, as the key protein quality control network, the proteasomal system is crucial for degradation of misfolded or unfolded and damaged proteins (35). This mechanism also becomes impaired with age as shown in other organ systems (36, 37) and in dermal fibroblasts and T cells (38, 39). In the rat lung, proteasomal activity is decreased in 2-yr-old animals compared to very young rats (37). However, in some studies, there was maintenance of lung proteasomal activity with aging (40). Our findings in old lung fibroblasts are in line with a previous study in rats that showed increased levels of the ER stress marker PERK in lung homogenates from aged animals (>18 mo old) compared to young rats (18). Additionally, we found that ATF6 expression is higher in old lung fibroblasts compared to young. However, our study shows no difference in BiP between age groups, which differs from a previous study in rat lung homogenates (18). These may reflect species differences in the response to aging and cell type-specific context to the ER stress response or its changes with age.

Cellular senescence is characterized by permanent cell-cycle arrest and can be a result of stimuli such as DNA damage, telomere shortening, genomic and epigenomic changes, mitochondrial dysfunction, and impaired proteostasis. Resistant to apoptosis, senescent cells can remain metabolically active, and their SASP includes proinflammatory cytokines, growth factors, chemokines, and ECM proteins. While there has been much focus on the detrimental effects of senescent cells, they are known to have beneficial effects in terms of tissue repair and tumor suppression. Normally, the immune system clears excessive senescent cells, but with aging, reduced or impaired immune responses can result in the accumulation of senescent cells, leading to exacerbated effects and contributions to diseases of aging. Compared to other organ systems where the contributions of senescence/SASP to aging or aging-associated diseases have been explored to a substantial extent, there is relatively less data on aging lungs or diseases such as COPD, pulmonary fibrosis, and asthma. Woldhuis and colleagues (4) showed that senescent fibroblasts derived from COPD patients showed increased levels of SASP compared to their matched controls. Additionally, Aghali and colleagues (2) also showed that airway smooth muscle cells from old asthmatic patients showed higher levels of senescence and secrete higher levels of SASP compared to their age-matched controls. There is now evidence that senescent cells accumulate in aging lungs (2, 31) and exacerbate lung diseases, but the mechanisms by which enhanced senescence occurs in these conditions and how senescent cells or SASP induce downstream signaling to promote disease including contributing to remodeling are still under investigation. The present study provides insights regarding the potential upstream role of ER stress specific to lung fibroblasts in the context of aging and further suggests a downstream interaction between senescence and ER stress that might help maintain disease-promoting pathways.

A recent study in primary lung fibroblasts from young (3–4 mo) and old (18–24 mo) C57BL/6 mice showed a higher level of senescence markers including p16 and p21 in fibroblasts from old mice (41). Our results regarding p21 in human lung fibroblasts are in accordance with these previous findings; however, we did not observe an age-related change in p16 levels. Additionally, we did observe increases in the expression of the antiapoptotic protein Mcl-1 in old fibroblasts. These results do show that aging involves fibroblast senescence. Our studies using MG132 then illustrate the involvement of ER stress.

The role of ER stress in cell cycle regulation is still under investigation, especially in the field of lung aging. A previous study has shown that induction of chronic ER stress using thapsigargin, tunicamycin, or dithiothreitol results in senescence of normal human dermal fibroblasts (19). Using the ER stress inducer MG132, we demonstrate ER stress in lung fibroblasts 3 days after treatment, with an increased level of IRE1α but a decreased level of XBP1s. The role of XBP1s has been demonstrated in keratinocyte proliferation (42). Importantly, the XBP1s level has been shown to be decreased in keratinocytes undergoing cellular senescence (42). Our data suggest that MG132 inhibits proliferation of lung fibroblasts through the IRE1α-XBP1s axis. However, the level of XBP1s was not found to remain low after the removal of MG132, e.g., at day 8. Surprisingly, the level of BiP that remained unchanged at an earlier time point (day 3), was significantly decreased not only in MG132-treated lung fibroblasts but also in etoposide-treated cells. Ei et al. (43) showed that BiP level was downregulated in cisplatin-induced premature senescence in nonsmall lung cell lung cancer cell lines A549 and H460 and that overexpression of BiP facilitates their reentry in the cell cycle. Similarly, Borok et al. (32) depicted the role of BiP in the establishment of cellular senescence profile. Their BiP-knockout mouse model showed increased levels of senescence markers, and the use of ER stress inhibitor tauroursodeoxycholic acid (TUDCA) decreased the expression levels of p16, p21, and p53 (32). Thus, there may be a dynamic regulation of the cell cycle by XBP1s at earlier time points, but by BiP at a later time point in lung fibroblasts. These findings highlight the role of BiP in the maintenance of lung fibroblast senescence as opposed to its initiation.

Our study also showed increased expression of p21, Cav-1, and SA-β-Gal activity, with inhibition of cell proliferation and decreased expression of LMNB1 in MG132- and etoposide-treated lung fibroblasts. To ensure that the lower cell number observed after treatment with MG132 and etoposide was not linked to apoptosis, we quantified the number of dead cells, and no difference was observed between MG132 or etoposide treatment and vehicle-treated lung fibroblasts. An unconventional marker, Cav-1, was measured to characterize senescent fibroblasts. Cav-1 is involved in vesicular trafficking and interacts with multiple proteins including those involved in the MAP kinase signaling pathway, which is relevant to both senescence (12) and to fibroblast proliferation and ECM production (12, 44, 45). We have previously shown the relevance of Cav-1 in airway mesenchymal cells in the context of cell proliferation and ECM production(46). The expression of Cav-1 is lower in fast-proliferating tumors (47). However, the overexpression of Cav-1 in mouse embryonic fibroblasts results in establishment of a senescence phenotype (48). Conversely, the senescence phenotype can be reversed by downregulating expression of Cav-1 in human dermal fibroblasts (49). Our findings showed a significantly increased level of Cav-1 in MG132- or etoposide-treated lung fibroblasts, again highlighting the presence of senescence and its promotion by ER stress induction. Interestingly, we did not observe changes in the expression of antiapoptotic proteins Bcl-xL or Mcl-1 upon exposure to MG132 or etoposide; however, a significant change was observed in Mcl-1 level in old lung fibroblasts compared to young. In contrast, DNA damage-, oncogene-, and replication-induced senescence in primary lung fibroblasts showed increased levels of antiapoptotic proteins Bcl-2-like protein 2 (Bcl-W) and Bcl-xL (13). This highlights a possible stimulus-dependent expression of antiapoptotic proteins in lung fibroblasts.

MG132 and etoposide-induced senescent lung fibroblasts showed different SASPs, with MG132 exposure increasing IL-8, and etoposide increasing IL-6. These results suggest a stimulus-dependent SASP. Additionally, in MG132-induced senescence, changes were observed in COL1A1 and LUM expression. ECM proteins have been previously identified as being part of SASP; however, it is not possible to discriminate whether COL1A1 and LUM are part of SASP as the measurement was made in lysates. Regardless, an increase in SASP relevant to inflammation and remodeling (ECM) by ER stress suggests a causative link between ER stress and senescence. This was further supported by the inhibitory effect of salubrinal that specifically blocks the PERK pathway, where it decreased (or showed a decreased trend) p21 and COL1A1 expression. These findings suggest that the PERK pathway is involved in p21 expression; however, the combined treatment of MG132 followed by salubrinal showed no significant effect on p21 in contrast to the COL1A1 level that was significantly reduced. This indicates the potential of salubrinal to reverse the senescence-associated phenotype in MG132-induced senescent fibroblasts, in particular for COL1A1. In contrast to our study, a previous study has shown the inhibitory effects of salubrinal on p16 and p21 in hydrogen peroxide-induced cellular senescence in the mouse (50). Perhaps interestingly, while we expected an age-related difference in the susceptibility to ER stress induction in terms of senescence, we found that fibroblasts from young and old lungs were comparable in their sensitivities to MG132. However, we did note that BiP was significantly decreased in MG132-induced senescent old fibroblasts (i.e., at day 8 after the initial exposure to MG132) but not in young fibroblasts. Considering the role of BiP in reentry into the cell cycle, our results suggest that older senescent fibroblasts can remain in the cell cycle exit mode, while younger senescent fibroblasts could reenter the cell cycle (i.e., what may appear as senescent may not be permanent as it is seen in older individuals).

The current study establishes functional links between ER stress and cellular senescence per se in the specific context of aging human lung fibroblasts. Recognizing that the process of aging per se is complex, modulated by the myriad of lifelong and environmental exposures, it is striking to note baseline differences in ER stress and senescence markers in older fibroblasts that presumably reflect a stable marker despite individual variations in external exposures. The physiological implications of these age-related differences and the impact of senescent cells at different ages in the context of ER stress or remodeling remain to be well established. Since aging lung shows increased fibrosis/remodeling, which has been shown to be exacerbated in age-related lung diseases such as asthma, COPD, or pulmonary fibrosis, having a stable marker such as ER stress or senescence may help identify novel therapeutic targets. For example, targeting senescent cells could reduce ER stress and blunt remodeling or modulate ER stress with a downstream impact on senescence and again remodeling. Hereby, the use of ER stress inhibitors, senolytics, or senomorphics could be beneficial.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL MATERIAL

Supplemental Files S1, S2, and S3: https://doi.org/10.6084/m9.figshare.23648634.

GRANTS

This work was supported by the Abel Tasman Talent Program Fellowship in association with the Healthy Aging Alliance, provided by the University Medical Center Groningen and the Mayo Clinic (to M. L. Koloko Ngassie), by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO; Aspasia 015.013.010) (to J. K. Burgess), by the NWO Subsidienummer Aspasia: 015.015.044 (to C.-A. Brandsma), and by the National Institutes of Health Grants R01 HL158532 (to Y. S. Prakash), R01 HL088029 (to Y. S. Prakash), and R01 HL142061 (to C. M. Pabelick).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.L.K.N., C.M.P., R.G., C.-A.B., J.K.B., and Y.S.P. conceived and designed research; M.L.K.N., L.Y.D., B.B.R., and A.K.-K. performed experiments; M.L.K.N. and L.Y.D. analyzed data; M.L.K.N.., L.Y.D., C.M.P., R.G., C.-A.B., J.K.B., and Y.S.P. interpreted results of experiments; M.L.K.N. prepared figures; M.L.K.N. drafted manuscript; M.L.K.N., L.Y.D., B.B.R., A.K.-K., C.M.P., R.G., C.-A.B., J.K.B., and Y.S.P. edited and revised manuscript; M.L.K.N., L.Y.D., B.B.R., A.K.-K., C.M.P., R.G., C.-A.B., J.K.B., and Y.S.P. approved final version of manuscript.