FGF receptors mediate cellular senescence in the cystic fibrosis airway epithelium

The number of adults living with cystic fibrosis (CF) has already increased significantly because of drastic improvements in life expectancy attributable to advances in treatment, including the development of highly effective modulator therapy. Chronic airway inflammation in CF contributes to morbidity and mortality, and aging processes like inflammaging and cell senescence influence CF pathology. Our results show that single-cell RNA sequencing data, human primary bronchial epithelial cells from non-CF and CF donors, a CF bronchial epithelial cell line, and Cftr-knockout (Cftr–/–) rats all demonstrated increased cell senescence markers in the CF bronchial epithelium. This was associated with upregulation of fibroblast growth factor receptors (FGFRs) and mitogen-activated protein kinase (MAPK) p38. Inhibition of FGFRs, specifically FGFR4 and to some extent FGFR1, attenuated cell senescence and improved mucociliary clearance, which was associated with MAPK p38 signaling. Mucociliary dysfunction could also be improved using a combination of senolytics in a CF ex vivo model. In summary, FGFR/MAPK p38 signaling contributes to cell senescence in CF airways, which is associated with impaired mucociliary clearance. Therefore, attenuation of cell senescence in the CF airways might be a future therapeutic strategy improving mucociliary dysfunction and lung disease in an aging population with CF.


Fibroblast Growth Factor Receptors Mediate Cellular Senescence in the Cystic Fibrosis Airway Epithelium
Introduction Cystic Fibrosis (CF) is the most common autosomal recessive disorder, affecting more than 70,000 people worldwide (1).Respiratory failure is the leading cause of morbidity and mortality in people with CF (pwCF) (2).The emergence of highly effective modulator therapies [HEMT] led to a significant decrease in disease burden and increased life expectancy, but chronic airway inflammation continues to persist thereby affecting many cellular processes leading to accelerated aging and lung function decline (3).Investigations of the aging biology in chronic lung diseases have advanced and several cellular processes, termed "the hallmarks of aging", have been used to characterize and study accelerated aging processes in lung diseases (4).Although, little is known about the aging processes in the CF lung.
Fibroblast growth factor receptors (FGFRs) encompass a subfamily of receptor tyrosine kinases that consists of four family members (FGFR1, 2, 3 and 4) with diverse functions (11,12).FGFR1 and 4 are increased in CF and COPD airways and regulate airway inflammation (13,14).FGFR1 signaling contributes to airway inflammation in CF by activating the extracellular signal-regulated kinase/mitogen-activated protein kinase (ERK-MAPK) signaling cascade (13).FGFR4 induces airway inflammation through phospholipase C γ (PLCγ)/calcineurin and nuclear factor of activated T-cells (NFAT) signaling (13,14).FGFR signaling plays a complex role in cellular senescence with both pro-and anti-senescence qualities (15)(16)(17).However, no studies to date have examined the consequences of cellular senescence and accelerated aging in the CF bronchial epithelium.To our knowledge, this will be the first study to characterize cellular senescence in both in vitro and in vivo models of CF lung disease.We show that MAPK p38 signaling regulates cellular senescence in the CF bronchial epithelium and seems to involve FGFR4 and partially FGFR1 making FGFR blockade a novel and potential future amenable therapeutic target for senolytic therapies targeting the CF lung, which is independent of CFTR function.

Single-cell RNA sequencing data from primary CF airway epithelial cells demonstrated evidence of cellular senescence
To characterize cell senescence in the CF bronchial epithelium in a large, representative CF patient cell cohort, single-cell RNA sequencing (scRNA-seq) data of airway epithelial cells from the previously published GSE150674 dataset was used (18).This dataset includes donors with end stage CF lung disease and healthy controls and ages ranging 6-60 years in CF (n=17,590 cells and n=19 donors) and 18-63 years in healthy donors (n=23119 cells and n=19 donors).Using uniform manifold approximation and projections (UMAPs), the dataset contains a variety of cell types (Suppl.Figure 1A) with the majority consisting of basal, ciliated and secretory cells in both control and CF donors (Suppl.Figure 1B).For our analysis, we filtered the CF cell population to cells homozygous for the DF508 mutation (n= 10,131 cells and n=8 donors with age range of 25-60 years) and created 3 senescence signature scores visualized via UMAPs and violin plot from three gene sets: CellAge senescence inducing genes (408 genes) database (19), SenoMayo (124 genes) (20) and a commonly used cellular senescence marker panel (CDKN1A, CDKN2A, BCL2, BCL2L1, IL6, IL1B and GLB1).
UMAPs and violin plots examining the senescence score from the CellAge database showed a significant increase in the CF cohort compared to controls (Figure 1A, B).Furthermore, the SenoMayo score mean was significantly increased in the CF group compared to controls with noticeable differences in the UMAP between CF and control groups (Figure 1C, D).Moreover, we created a cellular senescence score using the following cellular senescence markers: CDKN1A, CDKN2A, BCL2, BCL2L1, IL6, IL1B and GLB1, which demonstrated significant differences in the combined score visually and via violin plot in the CF epithelial cell group compared to control (Figure 1E, F).
In order to identify the specific cell types involved, we subclassified the airway epithelial cells into the major representing groups and found significant differences in cellular senescence scores based on cell type.Using the CellAge database senescence score, basal and secretory epithelial cells from CF donors had a significantly increased score, when compared to non-CF controls, whereas there was no difference between ciliated cells from CF and control donors (Figure 1G).Similar results were seen using the SenoMayo score (ciliated CF cells showed a decrease in senescence markers) (Figure 1G) and the 8 gene senescence score (Figure 1G).Overall, these data support the evidence of cellular senescence in the CF airway epithelium using multiple senescence scores with large datasets of genes associated with cellular senescence, which seems to be primarily localized to basal and secretory airway cells.
Single-cell RNA sequencing data revealed an increase in cell senescence in CF airway epithelial cells homozygous for the DF508 mutation.
To further characterize cellular senescence in GSE150674, we filtered the dataset to compare senescence scores in homozygous ΔF508, CF airway cells harboring other CF mutations, and control cells.Senescence scores in ΔF508 CF airway cells and Non-D508 CF airway cells were increased as shown via UMAPs, when compared to control (Suppl.Figure 2A-C).Violin plots comparing the three groups showed a marked increase in all three senescence scores just in the ΔF508 CF airway cells (Suppl.Figure 2D).Overall, these data suggest that the CF airway cells from donors homozygous for the ΔF508 mutation exhibit increased senescence markers.

Primary human bronchial epithelial cells from CF donors expressed increased cellular senescence markers
To validate the findings from the RNAseq dataset, primary human bronchial epithelial cells, cultured at the air liquid interface (ALI) from CF donors, homozygous for the ΔF508 mutation, and non-CF controls were assessed for the expression of an established set of cell senescence markers (7).
Similarly, to the cells from the RNAseq dataset, the differentiated primary CF ALI cultures also mainly consisted of basal cells, which were a higher percentage in non-CF donors (70%) versus CF donors (53%) with similar abundance of ciliated cells (18% non-CF vs 22% CF) and more secretory cells in the CF donor lungs, compared to the non-CF controls (25% vs 11%) (Suppl.Figure 3A, B).
Using western blot imaging and densitometric analysis, we observed increases in p16, p21 and BCL-xL protein levels in the CF ALI cultures (Figure 2A).SA β-gal staining was increased in CF ΔF508 ALI cultures, when compared to non-CF controls (Figure 2B).Additionally, mRNA levels of the senescence associated secretory phenotype (SASP) (IL1B ,IL6, IL8) were also substantially increased in the CF ALI cultures (Figure 2C).Since we have previously shown an association between FGFR signaling and IL-8 secretion in the CF epithelium, including upregulation of FGFR1 (13), we compared the FGF receptor expression between control and CF ALI cultures.Interestingly, most FGFRs were markedly increased in CF ΔF508 primary human bronchial epithelial cells compared to non-CF control ALI cultures (Figure 2D).In summary, markers of cellular senescence and FGFR expression were increased in CF ΔF508 primary human bronchial epithelial cell ALI cultures, when compared to non-CF controls.
To further characterize cell type specific localization of cell senescence markers, we co-labeled ALI cultures with commonly used cell type specific markers and senescence markers which showed that secretory cells, which express uteroglobin (21), also showed expression of p16, p21 and BCL-xL (Suppl.Figure 4A and Suppl.Figure 9B).Co-labeling of basal cells with KRT5 (22), exhibited less co-localization with p16, but overlapped with p21 and BCL-xl as well (Suppl.Figure 4B and Suppl.Figure 9B).Furthermore, we attempted to validate these findings in human CF lung tissue sections and achieved labeling of secretory cells in sections of the bronchial epithelium, which co-labeled with p16 (Suppl.Figure 5).FGFR inhibition decreased cellular senescence markers and phosphorylation of p38 MAPK CF bronchial epithelial cells (CFBEs) were treated with different FGFR inhibitors, including the clinically used FGFR1-3 inhibitor AZD4547 (13,(23)(24)(25) and the FGFR4 inhibitor BLU9931 (26).
Expression of p16, p21 and BCL-xL showed a marked attenuation following treatment with each specific inhibitor (Figure 3A).Furthermore, the ratio of SA β-gal positive cells in the CFBEs treated with AZD4547 or BLU9931 were substantially lower compared to vehicle treated cells (Figure 3B).
Next, we examined the activation of the downstream signaling mediators of FGFRs including ERK, PLCγ and p38 MAPK (13,14,(27)(28)(29)(30). CFBEs, treated for 24 hours with AZD4547 or BLU9931, did not show any relevant differences in phosphorylation of PLCy or ERK (Figure 3C).However, there was a marked decrease in p38 MAPK phosphorylation in CFBEs treated with AZD4547 or BLU9931 (Figure 3C).Primary human airway CF ALI cultures exhibited a baseline increase in p38 MAPK phosphorylation compared to non-CF controls (Figure 3D).
BLU9931 is a potent, selective and irreversible FGFR4 inhibitor with over 50-fold selectivity over FGFRs 1-3 (16), whereas the clinically established and used inhibitor AZD4547 has selectivity for FGFR1, 2 and 3 at very similar concentrations.In addition, AZD4547 demonstrates weaker activity against FGFR4, VEFGR2 and p38.To further characterize the isoform specific FGFR responsible for the demonstrated "senolytic" effects, we treated CF ALI cultures with PD173074, a potent and more selective FGFR1 inhibitor (IC50 of 25 nM) at 25 and 50 nM (31,32).
Targeting FGFRs via siRNA knockdown, specifically in ALI cultures, is challenging and will not block residual kinase activity; therefore, inhibitors have been shown to work best for our in vitro studies (13,14).PD173074-treated CF ALI cultures showed a dose dependent reduction in p16 expression (Suppl.Figure 6A) but no changes in p21 and BCL-xL expression (Suppl.Figure 6B, C).
Assessment of the SASP markers IL-6 and IL-8 via ELISA from basolateral media also demonstrated a decrease in PD173074-treated cells, which was in a dose dependent manner for IL-8 (Suppl.Figure 6D).Interestingly, AZD4547 and BLU9931 treatment did not affect IL-6 and 8 levels in the basolateral media of CF ALI cultures (Suppl.Figure 7A,B).In summary, FGFR inhibition attenuated cellular senescence in the CFBEs, which seems to be mediated by FGFR signaling pointing to partial dependence on FGFR1 and a more comprehensive involvement of FGFR4 in vitro.
Inhibition of p38 MAPK decreased cellular senescence markers in CFBEs.
To further investigate whether p38/MAPK mediates cellular senescence in the CF bronchial epithelium, CFBEs were treated with a p38/MAPK inhibitor (SB203580) for 24 hours.A marked decrease in protein expression of p16, p21 and BCL-xL was observed following pharmacological blockade of p38/MAPK (Figure 4A).In addition, the SASP cytokines IL-6 and IL-8 were also attenuated by p38/MAPK inhibition in CFBEs (Figure 4B).These findings were also accompanied by a considerable decrease in the ratio of SA β-gal positive cells in CFBEs treated with the p38 MAPK inhibitor (Figure 4C).Together, these data suggest that multiple cellular senescence markers are regulated by p38 MAPK in the CF bronchial epithelium.
Cellular senescence markers were increased in lung tissue and the airway epithelium of Cftr -/-rats compared to littermate controls.
Six-month old Cftr -/-rats, a model that exhibits CF-like airway disease (33), were used to validate our findings of increased cellular senescence in vivo.Lung tissue from 6-month-old Cftr -/-rats and controls were assessed via immunohistochemistry and showed increased staining of p16, p21 and BCL-xL in the bronchial epithelium (Figure 5A), when compared to littermate controls and secondary antibody control only (Suppl.Figure 8A, B).Furthermore, SA-β-gal staining was increased in the Cftr -/-rat lungs compared to controls (Figure 5B).mRNA levels for Cdkn2a (p16), Cdkn1a (p21), Bcl2l1 (BCL-xL), and SASP markers (Il-1b, Il-6, Cxcl2) from total lung tissue were also significantly increased in comparison to Cftr +/+ lungs (Figure 5C).To show bronchial localization of some of the senescence markers, we used immunofluorescence staining and co-labeled frozen sections of Cftr -/-rat lung tissue demonstrating that there is co-localization of p16 with uteroglobin, a secretory cell marker, and KRT5, a basal cell marker, validating our in vitro results (Figure 6A-B) which was also compared to secondary controls (Suppl.Figure 9A).In summary, there is evidence of cellular senescence in the lung and bronchial epithelium from a well-established in vivo model exhibiting CF airway disease.
Fibroblast growth factor receptor expression were increased in the Cftr -/-rat lung Formalin-fixed paraffin embedded sections of total lung tissue from 6 month-old Cftr -/-rats were stained with Alcian blue-periodic acid-Schiff (AB-PAS) to recapitulate the previously established muco-obstructive phenotype when compared to controls (Figure 7A) (33).Immunohistochemical analysis using an isoform specific antibody against FGFR4 revealed increased staining of the Cftr -/- bronchial epithelium, when compared to airways from wild type littermates (Figure 7B).FGFR4 protein expression, determined by Western blot analysis, was also substantially increased in Cftr -/-rat lungs when compared to controls (Figure 7C).A dearth of validated antibodies limited our ability to assess protein expression of FGFRs 1-3, but we have previously shown that FGFR protein expression correlated with mRNA levels; therefore, qRT-PCR was performed and confirmed a marked increase in mRNA levels of Fgfr1, Fgfr2 and Fgfr4 in Cftr -/-rat lung tissue (Figure 7D).Transcript levels of the FGFRs have been shown previously to corroborate with protein expression (14).In summary, cell senescence markers as well as FGFRs are upregulated in the lungs of 6-month-old Cftr-/-rats.
Treatment of Cftr -/-rats with the FGFR inhibitor AZD4547 attenuated cell senescence in the lungs and bronchial epithelium, which correlated with marked improvements in mucociliary clearance To investigate the effects of FGFR inhibition on "reversal" of cellular senescence in vivo, Cftr -/-rats were treated with AZD4547 via oral gavage daily for a total of 5 days.Pharmacological blockade of FGFRs with AZD4547 led to a decrease in p16 and p21 staining in the bronchial epithelium of Cftr -/- rats compared to sham treated Cftr -/-rats (Figure 8A).Furthermore, total lung protein expression of p21 and Bcl-xL from AZD4547-treated Cftr -/-rats was significantly decreased when compared to sham treated rats (Figure 8B).A reduction in phosphorylation of p38 MAPK expression (Figure 8C) and IL-8 secretion (Figure 8D) was also observed.Tracheae of AZD4547 treated Cftr -/-rats were analyzed via µOCT (Figure 8E,F) demonstrating marked improvements in air surface liquid [ASL] depth but no considerable differences in ciliary beat frequency (CBF) or periciliary liquid (PCL) depth, suggesting treatment with AZD4547 did not negatively affect the functional microanatomy of the lung epithelium (Figure 8E,F).Further, there were substantial improvements in mucociliary transport (MCT) in the AZD4547 treated Cftr -/-rat trachea when compared to the vehicle treatment (Figure 8F).Overall, these data validate our in vitro findings that there is a decrease in cell senescence markers after FGFR inhibition with functional consequences, leading to improved mucociliary clearance without affecting the microanatomy of the CF airway epithelium.
Isoform specific inhibition of FGFR4 in an ex-vivo Cftr -/-rat trachea model decreased cellular senescence and partially restored mucociliary clearance To validate our in vitro findings that FGFR4 inhibition attenuated cell senescence and assess functional outcomes, we utilized an ex vivo Cftr -/-rat trachea model (34) and treated with BLU9931(0.1 uM for 24h) (Figure 9).BLU9931 treatment decreased p16 and p21 levels shown through immunohistochemistry staining of ex vivo Cftr -/-rat tracheae compared to the vehicle and secondary only controls (Figure 9A and Suppl.Figure 8C).In addition, Cdkn2a (p16), Cdkn1a (p21), Bcl2, and Bcl2l1 (BCL-xL) (Figure 9B) and SASP markers (Il1b, Il6, Cxcl2 (IL-8)) (Figure 9C) mRNA levels were also decreased in BLU9931-treated Cftr -/-rat tracheae, when compared to vehicle treated Cftr -/-tracheae.Treatment with BLU993 also led to marked improvements in mucociliary clearance including restoration of MCT and increased ASL depth as well as improved ciliary beat frequency without changes in PCL depth (Figure 9D, E).In summary, these data imply that isoform specific inhibition of Fgfr4 in an ex vivo CF model led to restoration of mucociliary clearance, which was accompanied by attenuation of several cell senescence markers.
Treatment with Dasatinib and Quercetin decreased cellular senescence and improved mucociliary clearance in the ex vivo Cftr -/-rat trachea model To test whether cellular senescence itself contributes to mucociliary dysfunction, we utilized a widely used combination of senolytic drugs (Dasatinib and Quercetin: D+Q), which have been shown to attenuate cell senescence in other cell types including the lung (35,36) and treated Cftr -/-rat tracheae with D+Q for 24 hours.As shown previously by others in other systems, D+Q treatment caused a reduction in of p16 and p21 compared to vehicle and secondary only controls (Figure 10A and Suppl.Figure 8C).µOCT assessment of both trachea groups showed that D+Q treatment also led to restoration of mucociliary transport and ASL depth without affecting ciliary beat frequency and PCL depth (Figure 10B,C).Furthermore, Il1b, Il-6, Cxcl2 mRNA levels were also substantially decreased in D+Q treated Cftr -/-tracheae, when compared to vehicle-treated Cftr -/-tracheae (Figure 10D).Along with Cdkn2a (p16), Cdkn1a (p21), and Bcl2 mRNA levels which were also significantly down regulated with D+Q treatment (Figure 10E).In summary, targeting senescence in the CF airway ex vivo using senolytic therapy improves mucociliary clearance.

HEMT did not substantially decrease cellular senescence markers in CF airways
To investigate whether highly effective modulator therapy [HEMT] could affect cellular senescence, we treated CF primary human bronchial epithelial cells on ALI with VX-661/VX-445/VX-770 [ETI] and vehicle.First, we demonstrated ETI corrected CFTR dysfunction by µOCT analysis of ASL depth in CF primary bronchial epithelial cells showing ASL depth restoration (Figure 11A).However, we did not find any difference in protein expression of the cellular senescence markers p16, p21, and BCL-xL (Figure 11B).To further define the effect of ETI on cellular senescence we examined relative expression of SASPs markers and found that there was no difference in relative expression between treated and vehicle-treated groups either (Figure 11C).Additionally, we did not see differences in CDKN2a (p16), CDKN1a (p21), or BCL2 with ETI (Suppl.Figure 10A-C) nor did the addition of ETI + FGFR inhibition improve these markers (Suppl. Figure 10A-F).FGFR expression was also not different between groups (Suppl.Figure 11A).Moreover, we assessed cellular senescence in a hG551D rat model, which is receptive to ivacaftor (VX-770) treatment (37).
Immunoblots from hG551D rat lungs treated with VX-770 demonstrated no considerable changes in protein levels of the cellular senescence markers p21 and Bcl-xL (Figure 11D) or SASPs markers (Figure 11E).There was also no difference noted in expression of the different Fgfr isoforms (Suppl.Figure 11B).Furthermore, the hG551D rat model itself did not exhibit any increase in cell senescence in its lung, when compared to wild type littermates (Suppl.Figure 12A-E).Taken together, these results show that CFTR correction in CF did not attenuate cellular senescence in our in vitro and in vivo models.

Discussion
In this study, to our knowledge, we seem to be the first to demonstrate evidence of increased cellular senescence in both in vitro and in vivo CF models, which was paralleled by an increase in FGFR expression.Pharmacological inhibition of FGFRs led to a decrease in cell senescence, which seemed to be at least partially mediated by MAPK p38.The direct role of the FGFR-MAPK p38 signaling axis on cell senescence was also validated in an in vivo CF rat model.Furthermore, pharmacological inhibition of the FGFRs altered mucociliary clearance and air surface liquid volume, two important functional outcomes of CF airway disease severity (38).Additionally, using scRNA sequencing data (GSE150674), we were able to validate cellular senescence as a feature of CF airway epithelial cells, (Figure 1), mainly of basal and secretory airway cells, including more comprehensive cell senescence marker panels, which has proven to be useful in studying pathobiology in other lung diseases such as IPF (39).
Cellular senescence was first discovered by Hayflick in the 1960s and has since been described as a hallmark of aging and a common feature found in disease and age associated diseases (40).Several studies have shown that removal of senescent cells via senolytic treatments can reduce the number of senescent cells and significantly decrease disease burden in other chronic lung diseases (10,41).Our study suggests that FGFR inhibition may attenuate expression of several cell senescence markers, which could be transiently expressed in the CF bronchial epithelium or be a sign of dysfunctional repair as it has been shown in the murine alveolar epithelium, when injured (42).Furthermore, our in vitro data do not point to highly effective CFTR modulator therapy attenuating cell senescence markers in primary ALI homozygous for the D508 mutation, which is also supported by no changes in the hg551D rats, that were treated with Ivacaftor (Fig. 11 and suppl.Fig. 10-12).Nevertheless, those treatments were short term and there might be a benefit in patients, who have been on HEMT long-term, which needs future investigation.In addition, cell senescence markers were predominantly upregulated in the D508 CFBEs, but we did not analyze a sufficient number of G551D donors or donors with nonsense mutations to draw conclusions.Although, hg551D rats did not show increased senescence where cell senescence was observed in Cftr -/-rat airways.
FGFR signaling plays a complex role in cellular senescence, and there is evidence for both pro-and anti-senescence effects.For example, FGFR signaling has been shown to induce cellular senescence in pancreatic cancer cells (7,16); however, FGFRs have been shown to delay or prevent senescence in stem cells, fibroblasts and neurons (15,17,43).Additionally, FGFR signaling has been shown to regulate telomerase activity, which prevents telomere shortening and reduces cellular senescence (44).
FGF receptor signaling has been studied extensively by our lab and others in multiple organs including the lung, kidney, heart, and the parathyroid glands (13,14,(27)(28)(29)(30). FGFR1 signaling can occur via binding to FGF23 and α-klotho, leading to activation of ERK.FGFR4 has been shown to be abundantly expressed in the lung and the bronchial epithelium and shows activation and downstream signaling via PLCγ phosphorylation in the COPD lung (14).In our study, all FGFRs were expressed in the bronchial epithelium, which mainly consisted of basal epithelial cells and secretory cells.Those cell types also showed consistent expression of cell senescence markers, which were increased in the CF donors.Interestingly, the RNAseq dataset also showed upregulation of FGFRs in the ΔF508 CF dataset, which was also shown when combined with the senescence scores (Suppl.Figure 13, 14) but there was relatively low expression of FGFR4, which was abundantly expressed and upregulated in the CF ALI cultures and PLCγ phosphorylation was not increased in CFBEs (Figure 3C).Furthermore, FGFR3 was not consistently upregulated in the CF rat lung (Figure 7D).In previous reports, FGFR3 expression was associated with lung cancer (45), whereas FGFR2 played an important role in alveolar epithelial cell homeostasis and survival following injury (46,47).
A previous study found that FGFR1 plays a role in Cftr maturation (36) however, to date there is no data suggesting FGFR regulation of cell senescence in the CF lung.
Our results suggest that FGFR inhibition leads to reduced cell senescence markers, which seems to be partially mediated by MAPK p38.Given that cellular senescence is defined by various markers and involves intricate crosstalk of signaling pathways, FGF receptor signaling may overlap with upstream/downstream signaling molecules causing FGFR compensatory mechanisms (48)(49)(50).In addition, tools to target isoform specific FGFRs are limited with FGFR1 knockout mice not being viable and siRNA mediated FGFR knockdown not efficient due to residual tyrosine kinase activity.
Therefore, we used a pharmacological approach against FGFR isoforms that has limitations including potential lack of isoform specificity and off target effects .We have shown previously that a-klotho in the CF airways exhibits an anti-inflammatory action and can attenuate FGF23 and TGF-bmediated IL-8 secretion (13).However, other studies have shown that FGFR inhibition can regulate downstream mediators without a ligand (51).In order to specify which FGFR isoform mainly contributes to mediation of cell senescence, we used isoform specific FGFR inhibitors, which are partially used in clinical studies.Targeting FGFR1 and FGFR4 as specifically as possible, we could demonstrate that both receptors seem to be involved in the regulation of cell senescence markers, but FGFR4 inhibition showed attenuation of cell senescence in a more comprehensive manner (Figure 9).Interestingly, when using AZD4547, which can not only inhibit FGFR1 but also FGFR2 and 3 and other tyrosine kinases, we saw a greater response when compared to a more specific FGFR1 inhibitor (Figure 3 and Suppl.Figure . 6).Those results show that FGFR signaling is complex and receptor isoforms might be able to compensate for each other, which makes validation of those inhibitors also quite challenging in in vitro systems that do express all four FGFRs.In addition, many of the cell senescence markers are target genes of NF-kB signaling, but we neither observed any phosphorylation of p65 in vivo and in vitro nor did FGFR inhibition affect phosphorylation of p65 (Suppl.Figure 15A-C).Further studies are needed to investigate signaling pathways involved as well as upstream signaling regarding which FGF ligands might contribute.The Fgfr4 knockout mouse is not lethal, though we have shown that the adult mouse develops airway inflammation along with changes consistent with emphysema, but the murine lung is not a great model to study airway biology, especially cystic fibrosis (27,52).In addition, FGFR4 inhibition has been studied in cancer though whether FGFR4 is a suitable target in cancer therapy is still controversially discussed (53).
Future studies could include neutralizing isoform specific FGFR antibodies to avoid off target effects.Furthermore, whether cell senescence is mediating functional outcomes in the CF bronchial epithelium is of interest and our studies potentially show an indirect link with FGFR inhibition improving mucociliary transport and ASL depth.We have shown previously that TGF-b signaling, which is implicated in aging processes in other lung diseases (54,55), leads to a decrease in ASL volume and Pirfenidone, a therapy for idiopathic pulmonary fibrosis, can restore ASL volume (56).
In this manuscript, we show that senolytic therapy with Dasatanib and Quercetin can improve mucociliary transport in an ex vivo CF rat model to further support a potential relation between senescence and mucociliary dysfunction (Figure 10).It is not clear how exactly D+Q achieves the effect in our model system; we did not see changes in cell number in vitro but further studies are needed to define whether this effect could be also mediated via MAPK p38.Cell senescence is one hallmark of aging and several other hallmarks will be worth investigating in future studies including other potential aging related contributing pathomechanisms such as the length-associated transcriptome imbalance (57).There are several additional limitations of our study.Most marked differences were seen in CF donors harboring the D508 mutation and the hg551D rat model did not replicate those findings, but the Cftr -/-rat airways exhibited an increase in our studied cell senescence markers and both ex vivo and in vitro models showed attenuation with FGFR inhibition pointing to the role of FGFRs but not excluding additional contributing pathways such as inflammation itself or ER stress.Furthermore, mainly basal and secretory cells exhibited the senescent state and our studies assessing mucociliary function were done ex vivo and not in the primary cell cultures.Regulation of mucociliary transport is complex and with secretory cells not showing marked senescence features, we assume that senescence of secretory cells could contribute to the decrease in ASL volume and unfavorable mucus composition impeding on ciliary beat frequency.Therefore, FGFRs could potentially represent a novel target for senolytic therapies, which are applicable for all people with CF independent of their mutations and may become increasingly relevant in a time of lengthening patient lifespan.

Animals
Sex as a biological variable: Our study examined male and female animals, and similar findings are reported for both sexes.
All experiments used male and female SD-Cftr tm1sage rats (Cftr -/-) rats or wild type littermate controls at 6 months old as previously described (33).For experiments involving treatment of Cftr -/-rats with AZD4547 (Selleck Chemicals; Houston, TX, USA), we divided Cftr -/-rats into two groups (n=8 each): treated with 12.5 mg/kg bodyweight AZD4547 dissolved in DMSO with 1% sodium carboxymethyl cellulose (Selleck Chemicals; Houston, TX, USA) or vehicle once daily for five days.
Ex-vivo Cftr -/-rat trachea culture Tracheae were isolated from 6-month old Cftr -/-rats and carefully explanted to culture cassettes according to methods previously published (34).Those cultures were incubated for 5 days with Pneumacult media and then treated with either BLU9931 (0.1 uM) or a combination of Dasatinib and Quercetin (D+Q) at 100nM and 2uM for 24 hours respectively.After treatment, the tracheae were imaged via µOCT and the epithelial cell layer was removed and used for RNA and protein isolation.

Cell culture
Both primary human bronchial epithelial cells and CFBEdelta508 (CFBEs) (60) were used for experiments and cultured on Snap well filters or plates in medium consisting of Minimum Essential Media (MEM) with L-glutamine, Phenol Red then supplemented with 10% heat-inactivated fetal bovine serum (Atlas Biologicals; Fort Collins, CO, USA), 1% L-glutamine, 1% penicillin/streptomycin and 0.2% plasmocin.Human bronchial epithelial cells from cystic fibrosis (ΔF508) and non-CF donors were provided by the Cell Culture Core of the UAB Cystic Fibrosis Research Center and cultured and differentiated on air liquid surface interface as previously described (13).Briefly, passages 1-2 of the primary cells were seeded and kept in submerged cultures for about one week to expand to the necessary cell numbers and then splitted on filters and differentiated at the ALI for 4-6 weeks until ciliated cells were observed and mucus was produced in addition to assessment of transepithelial electrical resistance (TEER) (13,14,56).In addition, we assessed proportions of cell types, including ciliated, basal and secretory cells at the time cultures were used for experiments (Suppl.Fig. 2).

Western Blot
Protein lysates were collected using 1x radio immunoprecipitation assay (RIPA) buffer with 1x Halt protease and phosphatase inhibitor (Thermo Fisher Scientific, Waltham, MA, USA).Proteins were separated on 4-20% precast Ready Gels (Bio-Rad Hercules, CA, USA) and transferred onto polyvinylidene difluoride (PVDF) membranes (Pierce, Thermo Fisher Scientific, Waltham, MA, USA).Membranes were blocked with either 5% BSA or 5% low-fat milk depending on antibody manufacturer recommendations for 30 mins then incubated overnight with the following primary antibodies: rabbit anti-p21, rabbit anti-BCL-xL, rabbit anti-FGFR4, rabbit total and phospho-anti-ERK1/2, rabbit total and phospho-anti-p38 mitogen-activated protein kinase (MAPK), rabbit total and phospho-anti-PLCγ1 (Cell Signaling Technologies, Danvers, MA, USA), mouse anti-β-actinperoxidase (Sigma, St. Louis, MO, USA) and rabbit anti-p16 (Proteintech, Rosemont, IL, USA) diluted according to the manufacturer's recommendations.After three washes with TBST, membranes were incubated with goat anti-rabbit peroxidase conjugated (Invitrogen, Carlsbad, CA, USA) at 1:6000 in either 5% low fat milk or 5% BSA depending on primary antibody manufacturer recommendations for one hour.After three washes in TBST, the membranes were imaged by chemiluminescence on a ChemiDoc XRS system (Bio-Rad Hercules, CA, USA) and acquired using Image Lab software (Bio-Rad Hercules, CA, USA).Image J (National Institutes of Health, Bethesda, MD, USA) was used to measure densitometry of positive signals on the membranes.

Immunohistochemistry
Lungs from control and Cftr-/-rats were collected and fixed in 10% neutral buffered formalin for 24 hours followed by dehydration in ethanol for 24 hours.The tissue was then embedded in paraffin and cut into 3-5 mm sections and mounted on slides.Lung tissue slides were deparaffinized and stained using rabbit anti-p16, rabbit anti-p21 (Proteintech, Rosemont, IL, USA), anti-rabbit FGFR4 antibody (sc-124; Santa Cruz Biotechnology, Dallas, TX, USA), rabbit anti-BCL-xL (Cell Signaling Technologies, Danvers, MA, USA) and developed using a rabbit specific HRP/DAB detection IHC kit (Abcam Cambridge, UK) then counter stained with haematoxylin.The lung sections were stained with Alcian blue-periodic acid-Schiff and haematoxylin and eosin by UAB Comparative pathology laboratory core.

Immunofluorescence
Epithelial subtype ratios and ALI co-localization: Primary CF and non-CF bronchial epithelial cells were obtained from the UAB Cell Model and Translation Core upon air liquid interface (ALI) differentiation.The media was aspirated and washed with PBS prior to fixation with 4% paraformaldehyde.Staining methods were adapted from the Cell Signaling Immunostaining Protocol.
Briefly, after fixation for at least 24 hours, cells were washed with PBS and permeabilized with 0.1% Triton for 15 mins.The specimen was then washed and blocked with 1% bovine serum albumin (BSA) for 60 mins prior to primary incubation overnight with the following antibodies: Krt5 (1:50), Uteroglobin (1:50), FoxJ1 (1:50), p16 (1:100), p21 (1:100).Secondary antibodies were added at 1:2000 (anti-mouse, anti-rat, anti-rabbit) for two hours in the dark.Cells were then stained with NucBlue (Invitrogen, R37606) for 5 minutes and concurrently mounted using Prolong Gold Mounting Media (Invitrogen, P36930) and sealed with coverslips.After mounting, microscope slides were protected from light and stored at 4°C until imaging.Images were obtained on a Nikon Eclipse Ts2 with red, green, and blue cubes at 20x and cells were counted and subsequently analyzed using ImageJ and Excel respectively.
Primary CF bronchial epithelial cells co-labeled with a cell type marker and a senescence marker were imaged using a Zeiss Axio Observer with fluorescent monochrome Hamamatsu ORCA-Flash 4.0 LT camera and software Zen Blue.Images were obtained at 40X magnification with fixed exposure, gain, and signal threshold settings for each target.Signal threshold was based on negative controls stained only the corresponding secondary antibodies.ImageJ 1.54d (61) and Fiji (62) software was used for image processing.
Rat lung tissue staining: Rat lung tissue sections were co-labeled with a cell type marker and a senescence marker were imaged using a Zeiss Axio Observer with fluorescent monochrome Hamamatsu ORCA-Flash 4.0 LT camera and software Zen Blue.Images were obtained at 40X magnification with fixed exposure, gain, and signal threshold settings for each target.Signal threshold was based on negative controls stained only the corresponding secondary antibodies.ImageJ 1.54d (61) and Fiji (62) software was used for image processing.
Human lung tissue staining: Formalin-fixed, paraffin-embedded (FFPE) CF and non-CF human lung tissue sections were obtained from the University of Alabama Tissue Biorepository (UAB-TBR).The IF protocol was adapted from prior IHC staining protocol (63) and the Abcam Immunofluorescence Protocol.Briefly, FFPE sections were melted for 50 minutes in an over set for 60℃.Slides were then de-paraffinized and rehydrated using Clear-Rite 3 (Epredid, 6901) and 100, 95, and 70% ethanol.
Enzyme-Linked Immunosorbent Assay (ELISA) ELISAs for the quantitative recognition for IL-6 and IL-8 (Invitrogen, Vienna, Austria) were performed using supernatant from CFBEs after treatment with the different inhibitors as outlined before.ELISA for IL-8 (Abcam Cambridge, UK) was performed on protein lysates normalized to 2mg/mL of total protein in each sample from SHAM and AZD treated Cftr-/-rats.

Single-cell RNA sequencing data
Publicly available single-cell RNA sequencing dataset GSE150674 that was previously aligned, filtered, normalized and annotated was used to analyze senescence scores using three separate gene sets (18).From the total dataset of 19 controls and 19 CF donors with end-stage lung disease undergoing lung transplantation.For the analysis, we separated the CF group by CF cells that are ΔF508 homozygous (n=10,131 cells and n=8 donors) with a 34-36-year-old average age for the CF group and compared this CF ΔF508 homozygous group to the healthy donor group (n=23119 cells and n=19) with an average age of 46 years for the control group.Both groups include male and female donors.Using BBrowser 3 software (Bio Turning Inc, San Diego, CA, USA) and three different sets of gene list associated with cellular senescence, we created senescence scores visualized via UMAP and violin plots.The gene sets used in the senescence scores were: CellAge senescence genes database which was filtered for genes that induce cellular senescence (19), SenoMayo (20) and cellular senescence markers used in this study : CDKN1A, CDKN2A, BCL2, BCL2L1, IL6, IL1B and GLB1.A list of genes used from SenoMayo and CellAge senescence genes database is included as a supplemental table.Violin plots were made by extracting the signature score data from BBrowser3 and analysed in Prism9 (GraphPad, San Diego, CA, USA).
The Institutional Animal Care and Use Committee at the University of Alabama at Birmingham (UAB) approved all animal protocols.7: Fibroblast growth factor receptor expression is increased in the lungs of 6-month-old Cftr -/- rats.(A)AB-PAS staining indicating an increase of intercellular mucus staining in the bronchial epithelium of Cftr -/-rats, compared to control rats, arrows show areas stained in blue for intercellular mucus (magnification 20x) (B) Immunohistochemical analysis using an isoform specific and validated anti-FGFR4 showed increased staining in the bronchial epithelium of Cftr -/-rats, arrows highlight areas of airway epithelium stained for FGFR4 (scale bar = 100uM, magnification 40x).(C) Representative images of FGFR4 and b-actin with densitometric analysis demonstrating increased FGFR4 expression in Cftr -/-rat airways.(D) Relative mRNA levels of Fgfr1-4 from Cftr -/-total lung tissue, normalized to GAPDH expression.Statistical analysis was done using unpaired Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001 with n = 5-10 rats per group.-8)) from CF primary human bronchial epithelial cells demonstrates no significant change when compared to untreated CF primary human bronchial epithelial cells.(D) Immunoblots and densitometric analysis of p21 and BCL-xL in hG551D rats and hG551D rats treated with VX-770 for 14 days (hG551d= 3 rats, hG551d+VX-770= 5 rats).(E) Relative expression of SASPs markers Il1b, Il6 and Cxcl2 (IL-8) in hG551D rats and hG551D rats treated with VX-770 for 14 days (hG551d= 3 rats, hG551d+VX-770= 5 rats).Statistical analysis was done using Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001.

Figure 1 :
Figure 1: Single-cell RNA sequencing data reveals an increase in cell senescence markers in CF epithelial cells compared to control epithelial cells.Single cell RNA (scRNA) transcriptome of control and CF epithelial cells (GSE150674 Control (CO) and Cystic Fibrosis (CF) epithelial cell count CO n= 23119 and CF n= 17590) from 19 control and 19 CF donor lungs from donors with end stage CF lung disease and healthy controls, CF ΔF508 homozygous patients (n=10,131 cells and n=8 donors) and healthy donors (n=23119 cells and n=19) were separated from the whole dataset and analysed using BBrowser3 to generate UMAPs and violin plots showing senescence scores from three separate gene databases: (A,B) CellAge database of senescence inducing genes (416 genes), (C,D) SenMayo (124 genes), and (E,F) cellular senescence markers used in this study (CDKN1A, CDKN 2A, BCL2, BCL2L1, IL6, IL1B, IL6 and GLB1).(G) Senescence scores using the same three gene databases listed above but separated into violin plots looking at senescence scores based on major epithelial cell types: basal, ciliated and secretory.Statistical analysis was done using unpaired Student's t-test or one-way Anova shown with *p < 0.05, **p < 0.01, ***p < 0.001 and **** p<0.0001.

Figure 2 :
Figure 2: Cellular senescence markers are increased in CF primary human bronchial epithelial cells, cultured at the air liquid interface (ALI).(A) Representative immunoblot images and bar graphs showing densitometric analyses for p16, p21 and BCL-xL in CF ΔF508 and non-CF donors at ALI (N=5).(B) Representative images for senescence associated b-galactosidase (SA-b-gal) staining using brightfield imaging of the same CF ΔF508 and non-CF donor ALI cultures including quantification of b-gal staining using the ratio of SA-b-gal positive cells per brightfield by image J (N=3), arrows show b-gal positive cells (Scale bar = 100uM, Magnification 40x).(C) Bar graphs demonstrating relative mRNA levels of SASPs (IL1B, IL6, and IL8) markers normalized to GAPDH.(D) Bar graphs indicating relative mRNA levels of FGFR1-4 normalized to GAPDH in the same two groups.Statistical analysis was done using unpaired Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001 from 3-5 different donors per group with experiments repeated 3 times.

Figure 3 :
Figure 3: FGFR inhibition decreases cellular senescence markers and decreases phosphorylation of p38 MAPK.(A) Representative immunoblot images and densitometric analyses of the cellular senescence markers p16, p21, and BCL-xL from CFBEs, which were treated with AZD4547 0.1uM or BLU9931 0.1uM for 24 hours.(B) Representative images of SA-b-gal staining's in CFBEs treated with AZD4547 and BLU9931 and quantification by capturing three images from different regions of the cell culture plates and counting total cells and b-gal positive cells from the three images to make a ratio of b-gal positive cells to total cells, arrows indicate b-gal positive cells (scale bar= 100uM, Magnification 40x).(C) Representative immunoblots and densitometric analysis for p-ERK/ERK, p-PLC/PLCγ and p-p38/p38 MAPK in CFBEs treated with AZD4547 and BLU9931 for 24 hours with b-actin loading control (D) Representative immunoblot images and densitometric analyses from primary bronchial epithelial ALI cultures of CF (ΔF508) donors and non-CF control donors for p-p38/p38 MAPK expression.Statistical analysis was done using unpaired Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001; 3 independent experiments were done in triplicates.

Figure 4 :
Figure 4: Inhibition of p38 MAPK decreases cellular senescence markers in CFBEs.(A) Representative immunoblot images and densitometric analyses showing p16, p21, and BCL-xL expression of CFBEs treated with SB203580 at 20uM for 24 hours compared to controls (B) Bar graphs showing protein levels of IL-6 and IL-8 in CFBE supernatant after treatment with SB203580 for 24 hours.(C) Representative images of SA-b-gal staining in control and SB203580treated CFBEs including quantification, arrows indicate b-gal positive cells (scale bar = 100uM, Magnification 40x).Statistical analysis was done using unpaired Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001 with n =3-5 experiments.

Figure
Figure7: Fibroblast growth factor receptor expression is increased in the lungs of 6-month-old Cftr -/- rats.(A)AB-PAS staining indicating an increase of intercellular mucus staining in the bronchial epithelium of Cftr -/-rats, compared to control rats, arrows show areas stained in blue for intercellular mucus (magnification 20x) (B) Immunohistochemical analysis using an isoform specific and validated anti-FGFR4 showed increased staining in the bronchial epithelium of Cftr -/-rats, arrows highlight areas of airway epithelium stained for FGFR4 (scale bar = 100uM, magnification 40x).(C) Representative images of FGFR4 and b-actin with densitometric analysis demonstrating increased FGFR4 expression in Cftr -/-rat airways.(D) Relative mRNA levels of Fgfr1-4 from Cftr -/-total lung tissue, normalized to GAPDH expression.Statistical analysis was done using unpaired Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001 with n = 5-10 rats per group.

Figure 8 :
Figure 8: Systemic FGFR inhibition in Cftr -/-rats leads to decreased cellular senescence in the lung and improved mucociliary clearance.(A) Representative images of immunohistochemical staining for p16 and p21 in Cftr -/-and control rat lungs ± AZD4547 treatment (scale bar = 100uM, magnification 20x).(B) Representative immunoblot images and bar graphs demonstrating densitometric analyses of p21 and BCL-xL protein expression in Cftr -/-rat lungs ± AZD4547 treatment.(C) Representative immunoblot images of phosphorylated and total p38 MAPK and densitometric analysis.(D) IL-8 protein levels in Cftr -/-rat lung tissue ± AZD4547 treatment.(E) Representative images showing mucociliary transport (MCT) (cross-sectional arrow in blue indicates the velocity of the mucus particle via the slope).Along with, representative uOCT images of the trachea of Cftr-/-rats (yellow line representing airway surface liquid depth and the red line representing periciliary liquid depth, ep= epithelial layer, lp= lamina propria).(F) Bar graphs indicating analysis of uOCT images quantifying airway surface liquid [ASL], ciliary beat frequency (CBF), periciliary liquid depth (PCL), and mucociliary transport (MCT) from Cftr -/-rat trachea after treatment for 5 days with AZD4547 (12.5 mg/kg) or sham.Statistical analysis was done using unpaired Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001 with n = 3-4 rats per group.

Figure 11 :
Figure 11: HEMT does not significantly decrease cellular senescence markers or expression of FGFRs.(A) Representative µOCT images and bar graph of CF primary human bronchial epithelial cells (n=3 donors per group) treated with VX-661/VX-445/VX-770 [ETI] for 72 hours showing a significant increase in ASL depth in CF cells treated with ETI, labels: yellow line is ASL depth, ep: epithelial layer and F: filter.(B) Representative immunoblots and densitometric analysis showing no significant difference in protein expression of cellular senescence markers p16, p21 and BCL-xL in CF primary human bronchial epithelial cells treated with ETI for 72 hours when compared to untreated CF primary human bronchial epithelial cells (n= 4 donors per group).(C) Relative expression of SASPs markers (Il1b, Il6, and Cxcl2 (IL-8)) from CF primary human bronchial epithelial cells demonstrates no significant change when compared to untreated CF primary human bronchial epithelial cells.(D) Immunoblots and densitometric analysis of p21 and BCL-xL in hG551D rats and hG551D rats treated with VX-770 for 14 days (hG551d= 3 rats, hG551d+VX-770= 5 rats).(E) Relative expression of SASPs markers Il1b, Il6 and Cxcl2 (IL-8) in hG551D rats and hG551D rats treated with VX-770 for 14 days (hG551d= 3 rats, hG551d+VX-770= 5 rats).Statistical analysis was done using Student's t-test showing means ± SEM with *p < 0.05, **p < 0.01, and ***p < 0.001.