Stimuli-specific senescence of primary human lung fibroblasts modulates alveolar stem cell function

Abstract Aging is the main risk factor for chronic lung diseases (CLDs) including idiopathic pulmonary fibrosis (IPF) and chronic obstructive pulmonary disease (COPD). Accordingly, hallmarks of aging such as cellular senescence are present in different lung cell types such as fibroblasts in these patients. However, whether the senescent phenotype of fibroblasts derived from IPF or COPD patients differs is still unknown. Therefore, we characterized senescence at baseline and after exposure to disease-relevant insults (H 2 O 2 , bleomycin, and TGF-β1) in cultured primary human lung fibroblasts (phLF) from control donors, IPF, or COPD patients. We found that phLF from different disease-origins have a low baseline senescence. H 2 O 2 and bleomycin treatment induced a senescent phenotype in phLF, whereas TGF-β1 had primarily a pro-fibrotic effect. Notably, we did not observe any differences in susceptibility to senescence induction in phLF based on disease origin, while exposure to different stimuli resulted in distinct senescence programs in phLF. Moreover, senescent phLF reduced colony formation efficiency of distal alveolar epithelial progenitor cells in a stimuli-dependent manner. In conclusion, the senescent phenotype of phLF is mainly determined by the senescence inducer and impairs alveolar epithelial progenitor capacity in vitro .


Introduction
Chronic respiratory diseases are the third leading cause of death globally [1].Those include chronic obstructive pulmonary disease (COPD) and interstitial lung diseases such as idiopathic pulmonary brosis (IPF) [1].COPD is an in ammatory disease [2] characterized by small airway remodeling, emphysema, and chronic bronchitis [3].The main risk factors are cigarette smoking and age but exposure to air pollution or pathogens also contribute to COPD [3].IPF is a progressive brosing disease characterized by excessive matrix deposition [4,5].Higher age and exposure to cigarette smoke are the main risk factors for IPF [4,5].Familial cases of IPF have mostly been linked to mutations in genes encoding surfactant protein C and A2 (SFTPC, SFTPA2) and telomerases (TERT and TERC), which lead to telomere shortening, cellular senescence, and exhaustion of lung stem cells [4][5][6][7].The incidence rates for both COPD and IPF increase in the elderly population [8] and several cellular hallmarks of aging such as cellular senescence are increased in the lung tissue from both COPD and IPF patients [9][10][11][12].Senescent cells are characterized by an irreversible cell cycle arrest, resistance to apoptosis, altered metabolism and the secretion of growth factors and pro-in ammatory cytokines known as the senescence associated secretory phenotype (SASP) [13].
Fibroblasts are effector cells in both diseases, causing impaired tissue structure by aberrant deposition of extracellular matrix (ECM) on the one hand, and demonstrating hallmarks of senescence on the other hand [10,11,[14][15][16].Recent progress in single cell omics revealed the existence of disease-speci c cellular subtypes in the mesenchymal compartment [17][18][19][20][21].Although accumulation of senescent cells has been shown in both COPD and IPF, it remains unclear how senescence is induced in speci c cell types and whether certain subtypes are more prone to different senescence stimuli.Therefore, here we used well-known senescence inducers to study susceptibility and senescence programs in primary human lung broblasts (phLF) from control donors, IPF, and COPD patients.In this study, we show that phLF from different disease origins have a low baseline senescence in culture.Moreover, stimuli and not disease origin determines the senescence phenotype in phLF.Finally, senescent broblasts modulate progenitor cell capacity of alveolar epithelial progenitors in vitro.

Ethic statement
The study was approved by the local ethics committee of the Ludwig-Maximilians University of Munich, Germany (Ethic vote .Written informed consent was obtained for all study participants.

Cell culture
Primary human lung broblasts (phLF) from age-matched control, COPD, and IPF patients (Table 1) were obtained from the CPC-M bioArchive at the Comprehensive Pneumology Center (CPC Munich, Germany).PhLF were isolated [22] and cultivated in DMEM/F-12 (Life Technologies, USA) with 1% penicillin/streptomycin (10.000U/ml,Life Technologies, USA), and 10% Fetal Bovine Serum (PAN Biotech, Germany) and medium was changed every second day.For qPCR, Western Blot, and ELISA experiments, phLF were seeded on 6-well plates at a density of 4.5 x 10 4 cells per well.After 24h of incubation, treatment solutions were applied and changed every 48h.After day3 and day7 of treatment, supernatants of cells were collected, centrifuged, and frozen at -80°C.For RNA isolation, treatment solution was removed, the wells were washed twice with DPBS and cells were frozen directly at -80°C.

Organoid assay
Murine distal lung epithelial cells (CD45-/CD31-/EpCAM+) were obtained after enzymatic and mechanic digestion of mouse lungs [23] and MACS sorting.phLF were treated as previously described and after 7 days, control and senescent phLF were treated with Mitomycin C (10 µg/ml) for 2h, at 37°C and 5% CO 2 to stop proliferation.Then, phLF were washed with 1X DPBS and kept in fresh medium for at least 1h at 37°C and 5%CO 2 .phLF and murine distal lung epithelial cells were mixed in a 1:1 ratio (10.000 cells each) in Matrigel (Corning, USA) and seeded on 96-well plates as previously described [23].Organoid medium supplemented with Rock inhibitor (Ri) was added only for the rst 48 h and after this, medium was changed every 2-3 days [23].After 14 days, organoids were imaged using a LifeCellImager Observer Z1 (Zeiss, Germany).Maximum projections were generated on Zen software (Zeiss, Germany).Organoid size and number were determined using the Napari organoid counter [24,25].Data and plots were generated and analyzed in GraphPad Prism 9.5.1.

RT-qPCR
RNA was isolated using the peqGOLD Total RNA Kit (VWR) according to manufacturer's instructions.Senescence associated β-Galactosidase staining phLF were seeded (8.0x10 3 cells/well) and the Senescence β-galactosidase staining Kit from Cell Signaling (9860) was used according to manufacturer's instructions to determine senescence induction after day3 and day7 of treatment.Immuno uorescence staining Organoids were xed with ice-cold methanol and 2D cell cultured cells were xed with 4%PFA.Then, samples were blocked with 5% donkey normal serum in 0.1% PBST for 1h at RT, and then incubated with primary antibodies (Table 3) diluted in 1% donkey normal serum at 4°C overnight.Samples were washed 3X for 20 min (organoids) or 5 min (cells) with 0.1% PBST and then incubated for 2h at RT with secondary antibodies (Table 3) plus DAPI.Cells were then washed again 3 times and mounted with uorescence mounting medium (Dako).Mean uorescence intensity of images taken with a LSM 710 Confocal microscope (Zeiss, Germany) were quanti ed using ImageJ (Fiji).

Data collection and analysis
Primary human broblasts from Donor, COPD, and IPF were used for different experiments in this study (Table 1).For titration (Fig. 2) and organoid assays (Fig. 6) phLF only from control donors were used and single points represent different biological or technical replicates as indicated in gure legends.To analyze the capacity of the triggers used to induce senescence (Fig. 3) we pooled together the data collected using all the different samples listed in Table 1.Here, single points represent biological replicates and points shape indicate the disease origin.Finally, to study differences linked to the background disease we separated the samples listed in Table 1 into three different groups and used the data collected with these same samples for downstream analysis.Here, single points represent biological replicates and points shape indicate the disease origin.

Results
Primary human lung broblast isolated from different lung diseases exhibit similar senescence phenotypes at baseline The phLF derived from Donor, COPD, or IPF patients were characterized at baseline for multiple wellaccepted senescence markers after short (Day3) and prolonged (Day7) culture (Fig. 1A).First, we evaluated the senescence-associated-β (SA-β)-galactosidase activity and observed a low percentage of SA-β-galactosidase + cells in all three groups with no signi cant differences among them (Fig. 1B).Similarly, we did not observe any signi cant difference in the proliferation rate as determined by cell counts ratio (Day7/Day3) (Suppl.Figure 1A).The expression of the cyclin dependent kinase inhibitor 1A (CDKN1A/P21) and 2A (CDKN2A/P16) and the tumor suppressor protein 53 (TP53) was determined.For all conditions, we observed an upregulation of CDKN1A/P21 and CDKN2A/P16 with no statistical difference among the different origins (Fig. 1C).Finally, we found no signi cant differences in the secretion of the SASP components: Plasminogen Activator Inhibitor 1 (PAI-1), Matrix Metalloproteinase-3 (MMP-3), Growth Differentiation Factor 15 (GDF-15), or Interleukin 6 (IL-6) by phLF from Donor, COPD, or IPF tissue, respectively (Fig. 1D) but we did observe an increase in GDF-15 and IL-6 secretion for all origins from day3 to day 7.In conclusion, we observed a low level of senescence-related markers in cultured phLF irrespective of their disease/non-disease background.

Induction of senescence in primary human broblasts with different stimuli
Aging and exposure to cigarette smoke are the main risk factors for IPF and COPD and have been linked to stress-induced senescence.Therefore, we exposed phLF to hydrogen peroxide (H 2 O 2 ) or bleomycin, since they induce the release of reactive oxygen species (ROS) as well as genomic DNA-damage as observed in the lungs of smokers [27].For bleomycin and H 2 O 2 , we observed a dose-dependent induction of SA-β-galactosidase activity (Suppl.Figure 1A, B).A dose of 3.3mU/ml for bleomycin and 180uM for H 2 O 2 was used for further experiments, since these doses induced a high percentage of senescent cells (58.8% and 61.5%, respectively) with a signi cant reduction in cell proliferation, both well-known characteristics of senescent cells (Suppl.Figure 1C).Next, we tested whether single (S) or repetitive (R) treatment would induce different responses in senescence-related markers.Here, we did not nd any signi cant difference between both treatment regimes, but the repetitive treatment showed higher induction of CDKN1A/P21 (Suppl.Figure 1D) and therefore, we continued with this treatment scheme for further experiments.
Next, we characterized the senescence phenotype of phLF derived from donor, IPF, or COPD tissue (Table 1) upon exposure to the treatment regimen for 3 and 7 days (Fig. 2A).Moreover, we included transforming growth factor beta 1 (TGF-β1), a well-known pro brotic mediator, as the third stimuli, given that previous studies showed that TGF-β1 not only promotes broblast activation but also senescence in vitro [27].First, we stained for senescence-related markers: the cell cycle inhibitor CDKN1A/P21 and the DNA damage response-related protein yH2Ax (gamma-H2A histone family member X).Here, we observed that only phLF treated with H 2 O 2 and bleomycin displayed higher levels of these proteins compared to untreated controls (Fig. 2B).Accordingly, phLF treated with H 2 O 2 and bleomycin showed signi cantly increased SA-β-galactosidase activity and decreased cell proliferation, consistent with a senescent phenotype (Fig. 2C and Suppl.Figure 2).Notably, TGF-β1 treatment increased classical brotic markers (Suppl. Figure 3) but did not affect the expression of CDKN1A/P21 and yH2Ax, SA-β-galactosidase activity or cell proliferation (Fig. 2B, C and Suppl.Figure 2).The gene expression of CDKN1A/P21 was increased again only by H 2 O 2 and bleomycin whereas CDKN2A/P16 was only signi cantly induced by H 2 O 2 treatment (Fig. 2D).Finally, we characterized the secretion of different SASP factors.Here, we found that bleomycin signi cantly induced the secretion of GDF-15 while H 2 O 2 signi cantly inducedMMP-3 secretion (Fig. 2E).Conversely, TGF-β1 treatment signi cantly induced IL-6 and PAI-1 secretion (Fig. 2E).
In conclusion, H 2 O 2 and bleomycin induced a robust senescent phenotype in all phLF characterized by cell cycle inhibition, reduced proliferation, increased SA-β-galactosidase activity, and secretion of SASPrelated proteins.On the other hand, TGF-β1 treatment led to increased brosis-related genes and the secretion of well-known downstream mediators of the TGF-β1 signaling pathway like PAI-1 and IL-6 but did not induce a senescent phenotype in phLF.
Senescence induction in phLF is not impacted by origin, but stimuli-speci c Next, we further analyzed the dataset we generated and asked the following two questions: 1) Do phLF from different disease background exhibits differences in susceptibility to cellular senescence? and 2) Do the senescent phenotypes induced by different stimuli differ among phLF derived from distinct disease background?To this end, we rst compared the level of senescence induction in donor-, IPF, or COPD-derived phLF, respectively.Here, we did not nd any signi cant differences in SA-β-galactosidase-induction among nondisease/disease origins (Fig. 3A, B).Similarly, we did not observe any signi cant difference in gene expression of three cell cycle regulators of CDKN1A/P21, CDKN2A/P16, or TP53 (Fig. 3C).
Next, we compared the senescence signatures of phLF induced by TGF-β1, bleomycin, or H 2 O 2 after strati cation by origin of the tissue.Here, we observed that for all treatments the SA-β-galactosidase activity increased from day 3 to day 7 (Fig. 4A).Moreover, bleomycin induced the highest level of SA-βgalactosidase in phLF derived from donor and COPD tissue, while H 2 O 2 had a higher effect in SA-βgalactosidase activity induction in IPF-derived phLF (Fig. 4A).As observed previously, TGF-β1 did not signi cantly induce SA-β-galactosidase activity in any of the phLF (Fig. 4A).Interestingly, both H 2 O 2 and bleomycin induced CDKN1A/P21 expression, whereas CDKN2A/P16 was mainly induced after H 2 O 2 treatment (Fig. 4B).Moreover, H 2 O 2 and bleomycin induced the expression of PAI-1, ACTA2, and Fibronectin 1 (FN-1) after 7 days of culture (Fig. 4B).Conversely, TGF-β1 did not increase the evaluated senescence-related genes but predominantly induced the expression of the pro-brotic markers: ACTA2, FN-1, and Collagen 1 (COL1A1) (Fig. 4B).Finally, we characterize the secretion of selected SASP factors over the culture time.Here, we found that H 2 O 2 and bleomycin strongly induced GDF-15 secretion on day 3, whereas IL-6 was strongly induced by TGF-β1 and bleomycin treatment (Fig. 4C).PAI-1 secretion was mainly induced by TGF-β1 and sustained over culture time (Fig. 4C).Conversely, H 2 O 2 had the stronger effect on MMP-3 secretion after 7 days in comparison to TGF-β1 and bleomycin (Fig. 4C).In general, SASP pro les were similar between H 2 O 2 and bleomycin as shown by an initial induction of IL-6, later induction of MMP-3 secretion, and sustained GDF-15 secretion (Fig. 4C).Finally, supporting a mainly probrotic effect, TGF-β1 only induced the secretion of PAI-1 and IL-6 (Fig. 4C).In conclusion, we observed speci c senescence programs mostly depending on the senescence trigger.

Senescent broblasts disrupt progenitor potential of distal alveolar epithelial cells
Senescent broblasts accumulate in COPD and IPF lungs.In the alveolar niche the epithelial and mesenchymal compartments closely interact but the paracrine effects of senescent cells in this environment are still understudied.To explore this, we co-cultured control or senescent phLF with distal alveolar epithelial progenitor cells in an organoid assay (Fig. 5A).Here, we also evaluated whether the modulation of stem cell function by senescent phLF depends on trigger.After 14 days, we observed formation of alveolar (small and dark and surfactant protein C (SP-C) positive), bronchiolar organoids (big with a lumen and positive for acetylated Tubulin (ACT)) and bronchoalveolar organoids (SP-C+/ACT+) in all conditions (Fig. 5C).However, phLF treated with bleomycin or H 2 O 2 signi cantly reduced colony formation e ciency (CFE) of the distal lung epithelial progenitors (Fig. 5B).Moreover, bleomycintreated phLF signi cantly reduced the organoid size (Fig. 5B).Finally, to characterize the cellular composition of the formed organoids, we stained them for SP-C, a marker for alveolar type 2 (AT2) cells, Keratin-8 (Krt8), a transdifferentiation marker for AT2 cells, and ACT, a marker for airway epithelium as shown in Fig. 5C.Here, we observed a similar cellular composition for H 2 O 2 and control-derived organoids.Notably, organoids derived from the co-culture with bleomycin-treated phLF showed lower Krt8 with a higher expression of SP-C in comparison with controls and other senescence inducers, suggesting impaired AT2 activation/differentiation (Fig. 5C).In conclusion, the co-culture with senescent broblasts reduced stem cell capacity of alveolar epithelial progenitor cells in vitro.

Discussion
Aging is the main risk factor for CLDs such as COPD and IPF and previous studies have shown that senescent cells accumulate with age [28].Although the mechanism is not fully understood, senescent cells can evade immune clearance, thereby accumulating and promoting organ dysfunction [29][30][31][32].
Indeed, elevated levels of CDKN1A/P21, CDKN2A/P16, and SA-β-galactosidase were found in broblasts from both IPF and COPD lungs and therefore, have been linked to the disease pathobiology [16,28,[33][34][35].However, whether the senescent phenotype is different depending on the disease background is not well understood.Therefore, in this study we aimed to characterize the senescence of phLF from control, IPF, and COPD patients at baseline and after exposure to different senescence inducers.Finally, we used an organoid assay to study the crosstalk between epithelial and mesenchymal cells in vitro.
First, we characterized classical markers of senescence in phLF at baseline.Here, we found that phLF from control, IPF, and COPD had a very similar phenotype characterized by low expression of the different senescence-related markers: CDKN1A/P21, CDKN2A/P16, and TP53 gene expression, SA-β-galactosidase activity, and secretion of SASP-related components.However, all these markers increased with prolonged culture as described for replication-induced senescence [36].Previous studies showed that broblasts originated from COPD patients had an elevated senescence signature as judged by enhanced expression of P21/CDKN1A and CDKN2A/P16, increased SA-β-galactosidase activity [37,38], reduced proliferation rates [35,38], and higher secreted levels of proteins associated with the SASP [15].Moreover, COPDderived broblasts inhibit canonical WNT-β-catenin signaling in alveolar epithelial cells by secreting WNT-5A, leading to stem cell exhaustion and impaired lung repair [39].Similarly, phLF obtained from IPF lung tissue, showed decreased proliferation rates, increased expression of CDKN1A/P21, CDKN2A/P16, and TP53 as well as senescence-related morphological changes [16].However, in the present study we did not observe any major differences in senescence markers among disease origin, as reported previously [16, 35,37,38].This could be explained by the different isolation protocols: in this study phLF were isolated by enzymatic digestion contrary to the outgrowth from lung tissue pieces used in other studies [37].Furthermore, intrinsic characteristics like the smoking history or passage number, as well as different culturing conditions and media supplementation can in uence senescence readouts [40].Finally, the composition of the isolated phLF can also vary depending on different anatomical localizations such as airway [38,41] versus whole lung [16,35] and current isolation methods do not discriminate among the different broblast subtypes recently described for the lung.In the past decades, it was believed that ACTA2 + positive myo broblasts were the main contributor for ECM deposition in the IPF lung [19].However, recent single cell-based studies have revealed that IPF lungs have a higher heterogeneity in broblasts than control lungs and that these subpopulations coexist in the lung and might play distinct roles in the disease progression [17,19].Interestingly, the susceptibility to typical brotic and senescence inducers such as bleomycin, has been shown to differ among these broblast subpopulations in the mouse lung [19].Therefore, the response to the stimuli used in this and other studies might be de ned by the composition of the isolated and treated broblast population.Therefore, the development of isolation protocols for primary lung broblasts based on the newly described markers would help to nd speci c disease-relevant cellular responses of these subpopulations that could be therapeutically targeted.
Cellular senescence can be induced by several stimuli such as increased oxidative stress, caused by exposure to cigarette smoke, or genomic DNA damage, induced by chemotherapeutic agents.Therefore, we used H 2 O 2 and bleomycin to mimic these insults in vitro.Moreover, we also included TGF-β1, since it has been shown to induce a senescent phenotype in phLF [16,27].In our study, only H 2 O 2 and bleomycin induced cell cycle arrest as measured by CDKN1A/CDKN2A gene expression and other senescence-related markers such as reduced proliferation, increased SA-β-galactosidase activity, and increased secretion of GDF-15 and MMP-3 after 7 days.Notably, TGF-β1 only induced the expression of pro-brotic markers (ACTA2, FN-1, COL1A1, and PAI-1) as well as secretion of IL-6 and PAI-1 but did not induce a senescence phenotype de ned by canonical senescence markers.This could be explained by difference in the dosage, since in previous studies much higher doses were used to observe TGF-β1-induced senescence [27].Previous studies showed that aged individuals have around 3-4 ng/ml circulating TGF-β1 in plasma [42].Therefore, based on results using a more physiologically relevant dose (5ng/ml), we propose TGF-β1 as a pro-brotic rather than a senescence stimulus.
Next, we addressed whether the susceptibility to senescence was different among the different disease origins.Here, we found that IPF-derived phLF had a trend towards a reduced response to all stimuli in comparison to Donor-and COPD-derived phLF as previously described [43].However, as observed at baseline, we did not nd any signi cant difference in the senescence response among the different disease origins.This could be attributed to the chosen effective treatment regimens, which consistently induce a senescent phenotype overriding the cell origin.In conclusion, the senescence response of phLF is mainly de ned by the trigger, in this case DNA damage and oxidative stress, than by cellular predispositions.Interestingly, we observed similar gene expression and SASP pro les for bleomycin and H 2 O 2 that differ from the one induced after TGF-β1.However, we also observed differences in the effect size for the tested markers.For example, the gene expression of CDKN2A or the secretion of MMP-3 was more pronounced on H 2 O 2 -treated phLF.Therefore, more comprehensive analysis of gene expression changes and secreted factors might be useful to better understand differences among these two senescence inducers.
Senescent cells can modulate their microenvironment in a paracrine manner by their SASP or direct cellular interactions [28,44].Therefore, we assessed the secretion of proteins related to in ammation and ECM deposition after induction of senescence.Here, we found that bleomycin and H 2 O 2 induced the secretion of GDF-15, MMP-3, and PAI-1.On the other hand, TGF-β1 treatment induced the secretion only of proteins downstream of it signaling pathway: IL-6 and PAI-1.In IPF lungs, MMP-3 is secreted by different cell types, including broblasts, and has been linked to lung epithelium dysfunction and poor regenerative capacity as well as broblasts activation [45,46].GDF-15 and PAI-1 are well known SASP factors that also have been linked to in ammation and ECM remodeling in the diseased lung [47][48][49].Interestingly, as previously described for paraquat-induced cellular senescence of lung broblasts [37], here we also observed that bleomycin and H 2 O 2 decreased the expression of COL1A1 in phLF.Moreover, bleomycin and H 2 O 2 induced the gene expression of PAI-1 and FN-1.This suggests that senescent broblasts can contribute to ECM remodeling as seen in CLDs.
Given the fact that senescent phLF can modulate their microenvironment by secreting pro-in ammatory and ECM-related proteins, we used an organoid assay to evaluate whether co-culture with them would alter stem cell function of distal alveolar progenitor cells.Here, we found that both bleomycin and H 2 O 2induced senescence signi cantly reduced progenitor cell capacity as assessed by colony forming e ciency.However, only bleomycin signi cantly altered the size of the formed organoids.These could be attributed to differences in the SASP and ECM-related genes expression between bleomycin-and H 2 O 2induced senescence programs in phLF.Further studies focusing on a comprehensive characterization of these factors could provide insights in the speci c phLF-derived factors modulating the alveolar progenitor function in vitro.
In conclusion, this study provides novel insights into the senescence phenotype of primary human lung broblasts exposed to disease-relevant insults.Moreover, in vitro organoid assays revealed that senescent phLFs modulate the regenerative capacity of the lung progenitors.Further characterization of these phenotypes using state-of-art techniques such as single cell sequencing could help elucidate the underlying mechanism that de nes these senescence programs.

Figure 1 Primary
Figure 1

Figure 3 No
Figure 3

Figure 4 Different
Figure 4