Elimination of p19ARF‐expressing cells protects against pulmonary emphysema in mice

Abstract Senescent cells accumulate in tissues during aging and are considered to underlie several aging‐associated phenotypes and diseases. We recently reported that the elimination of p19ARF‐expressing senescent cells from lung tissue restored tissue function and gene expression in middle‐aged (12‐month‐old) mice. The aging of lung tissue increases the risk of pulmonary diseases such as emphysema, and cellular senescence is accelerated in emphysema patients. However, there is currently no direct evidence to show that cellular senescence promotes the pathology of emphysema, and the involvement of senescence in the development of this disease has yet to be clarified. We herein demonstrated that p19ARF facilitated the development of pulmonary emphysema in mice. The elimination of p19ARF‐expressing cells prevented lung tissue from elastase‐induced lung dysfunction. These effects appeared to depend on reduced pulmonary inflammation, which is enhanced after elastase stimulation. Furthermore, the administration of a senolytic drug that selectively kills senescent cells attenuated emphysema‐associated pathologies. These results strongly suggest the potential of senescent cells as therapeutic/preventive targets for pulmonary emphysema.

and are considered to contribute to tissue aging and aging-associated disorders through their cell nonautonomous functions (Freund, Orjalo, Desprez, & Campisi, 2010;Watanabe, Kawamoto, Ohtani, & Hara, 2017). Recent studies using transgenic mice designed to eliminate senescent cells from tissues by sensitizing them to specific drugs have more clearly elucidated the roles of senescent cells in tissue aging and disease. The targeting of senescent cells through a semi-genetic approach extends the health span by ameliorating the aging-associated phenotypes of kidney, eye, heart, bone, and lung tissues in wild-type or progeria model mice (Baar et al., 2017;Baker et al., 2016;Farr et al., 2017;Hashimoto et al., 2016). In addition, senescent cell elimination alleviates pathologies in disease models, which include those of atherosclerosis, idiopathic pulmonary fibrosis (IPF), hepatic steatosis, and osteoarthritis Jeon et al., 2017;Ogrodnik et al., 2017;Schafer et al., 2017). As the removal of senescent cells has beneficial effects on the maintenance of tissue homeostasis and the prevention of diseases in mice, senescent cells are expected to have potential as therapeutic targets for these diseases. One possible approach to apply the findings obtained from mouse studies to humans is the pharmacological targeting of senescent cells. Senescence shows resistance to cytotoxic stress due to an enhanced pro-survival pathway (Wang, 1995); however, certain drugs have the ability to induce cell death in senescent cells by targeting specific signaling pathways (Zhu et al., 2015). The efficacy of drugs that induce senescent cell death, namely senolytic drugs, has been demonstrated in several mouse disease models. The combination of dasatinib and quercetin, which inhibits pro-survival kinase pathways, ameliorates cardiovascular and vasomotor functions (Roos et al., 2016;Zhu et al., 2015), hepatic steatosis (Ogrodnik et al., 2017), and IPF (Schafer et al., 2017), similar to the semi-genetic elimination of senescent cells. The anti-apoptotic bcl-2 family protein inhibitors ABT-263/737 are effective in atherosclerosis and osteoarthritis models Jeon et al., 2017) and improve stem cell functions and other aging-associated phenotypes (Chang et al., 2016;Yosef et al., 2016).
Pulmonary emphysema is characterized by the destruction of alveolar walls, leading to permanent enlargement of the airspace in the lungs, and is a major component of chronic obstructive pulmonary disease (COPD), which is currently one of the leading causes of death worldwide. Pulmonary emphysema is also associated with the infiltration of inflammatory cells, which are believed to cause the accumulation of proteinases and lead to alveolar destruction (Barnes, 2004). Macrophages are predominant inflammatory cells in lung tissues, and their number increases in emphysema (Rodriguez, White, Senior, & Levine, 1977). Matrix metalloproteinase (MMP)-12, also known as macrophage elastase, is required for the development of cigarette smoke-induced emphysema in mice (Hautamaki, Kobayashi, Senior, & Shapiro, 1997), and the inhibition of macrophage recruitment or induction of apoptosis in alveolar macrophages was found to suppress alveolar collapse in an elastase-induced emphysema model (Houghton et al., 2006;Ueno et al., 2015). Neutrophils are also known to contribute to the pathology of emphysema, and the inactivation of neutrophil elastase was shown to confer resistance in a mouse emphysema model (Shapiro et al., 2003). Emphysema is accompanied by an increase in cellular senescence (Tsuji, Aoshiba, & Nagai, 2006). However, it currently remains unclear whether and how senescent cells are involved in the development of pulmonary emphysema or whether they are a consequence of this disease.
ARF-DTR mice express the diphtheria toxin receptor (DTR) and luciferase genes under the control of the CDKN2A promoter/enhancer and have enabled us to eliminate p19 ARF -expressing cells from lung tissues (Hashimoto et al., 2016). p19 ARF accumulates in mesenchymal cells of the lung parenchyma in adult mice, and the removal of p19 ARF -expressing cells was found to restore lung function with concomitant changes in the expression of senescence-associated genes. Therefore, p19 ARF -expressing senescent cells contribute, at least in part, to pulmonary hypofunction, which increases the risk of emphysema. Using ARF-DTR mice, we herein demonstrate that p19 ARF -expressing cells facilitate emphysema-associated lung dysfunction. The elimination of p19 ARF -expressing cells by a toxin-mediated cell knockout system (Saito et al., 2001) from lung tissue protected against elastase-induced emphysema. Alveolar wall destruction was reduced, and pulmonary function was maintained, in the absence of p19 ARF -expressing cells. The accumulation of inflammatory cells was also suppressed after the elimination of p19 ARF -expressing cells, as exemplified by the reduced infiltration of macrophages and other cells after elastase challenge. Moreover, the administration of a senolytic drug conferred resistance to elastase-induced pulmonary dysfunction with a concomitant reduction in senescent cells in lung tissues. Collectively, the present results imply that p19 ARF -expressing cells exacerbate pulmonary emphysema and have potential as a therapeutic/preventive target for emphysema.

| RESULTS
To investigate the roles of p19 ARF -expressing cells in the development of pulmonary emphysema, we employed a porcine pancreatic elastase (PPE)-induced emphysema model that is closely related to human panlobular emphysema. Five-month-old female wild-type or ARF-DTR mice pretreated with DT or PBS were administered with PPE or PBS for 3 weeks (Figure 1a). ARF-DTR mice carry extra CDKN2A alleles as a transgene, in which the first exon of the ARF gene is replaced with DTR (human HB-EGF I117V/L148V) and firefly luciferase genes, thereby enabling the elimination and detection of p19 ARF -expressing cells by the administration of DT and in vivo imaging, respectively (Hashimoto et al., 2016). An in vivo imaging analysis of lung luciferase activity before and after treatment revealed that the PPE treatment did not affect luciferase activity, Pulmonary emphysema is characterized by an enlarged airspace as a result of irreversible alveolar collapse. We examined the impact of p19 ARF -expressing cell elimination on emphysema-associated morphology in these mice. Lung tissues were inflated with fixative solutions under constant pressure before embedding and sectioning. The PPE treatment caused massive alveolar collapse, and the pre-elimination of p19 ARF -expressing cells attenuated this effect (Figure 2a,b).
The results of a morphometric analysis indicated that the alveolar mean linear intercept, which reflects the mean alveolar size, increased threefold in PPE-treated animals due to alveolar collapse, and this result was suppressed by approximately 50% in DT-treated animals ( Figure 2c). The effects of DT are attributed to the elimination of p19 ARF -expressing cells because DT had no effect on PPEtreated wild-type mice. DT had no significant effects on the PPE-untreated ARF-DTR mice at this age (5 months old) but restored alveolar walls and their size in older animals (Hashimoto et al., 2016).
We then attempted to clarify whether the physiological function of lung tissues is retained after PPE treatment by eliminating p19 ARF -expressing cells. We performed pulmonary function tests using spirometry. The PPE treatment significantly increased both static and dynamic lung tissue compliance (Cst and Crs, respectively) as well as total lung capacity (inspiratory capacity; IC) due to alveolar collapse ( Figure 3). However, the effects of DT on lung function were diminished in mice pretreated with DT, and PPE exerted weaker effects on Cst, Crs, and IC in DT-treated ARF-DTR mice than in wild-type mice. Collectively, these results strongly suggest that the elimination of p19 ARF -expressing cells has protective effects against PPE-induced lung dysfunction. Nevertheless, DT did not affect all of the observed PPE-induced changes in the parameters examined ( Figure S1), suggesting that the effects of p19 ARF -expressing cell elimination are limited in the PPE-induced lung injury model. Similar results were obtained in male ARF-DTR mice (data not shown).
We also tested the effects of p19 ARF -expressing cell elimination in older (13-14 months old) female animals. Lung compliance was decreased by DT treatment in the absence of PPE treatment in those mice ( Figure S2). PPE increased the lung compliance (Crs), which was ameliorated by DT treatment in ARF-DTR but not in wild-type animals. These results likely reflect our previous observation that lung function can be restored by eliminating the p19 ARF -expressing cells in aged (12 months old or older) animals (Hashimoto et al., 2016). To focus on the effects of p19 ARF -expressing cells on PPE-induced lung pathology, we used 5-month-old mice for further analysis.
To gain further insight into the role of p19 ARF -expressing cells in PPE-induced emphysema, we analyzed cells in bronchoalveolar lavage fluid (BALF). Although a significant change was observed in lung morphology and function 3 weeks after the administration of PPE (Figures 2 and 3), no significant difference was found in the number or composition of inflammatory cells in BALF among PPE-and/or DT-treated samples at this time point ( Figure S3). This result was expected because a single shot of PPE only has temporal effects on BALF cells in the C57BL/6J strain (Limjunyawong, Craig, Lagassé, Scott, & Mitzner, 2015;Ueno et al., 2015). Therefore, we analyzed BALF cells at an earlier time point after the administration of PPE.
Mice were treated with DT and PPE, and BALF cells were collected 1 week after the administration of PPE (Figure 1a). In contrast to the 3-week treatment, a significant increase was noted in the total cell number in the BALF of PPE-treated lungs at this time point (Figure 4a). This increase largely accounted for the change in macrophage numbers (Figure 4b), although the number of other inflammatory cells was also slightly increased (Figure 4c-e). The DT treatment diminished these effects, and the increase in inflammatory cells by PPE was significantly suppressed in ARF-DTR but not in wild-type mice. Taken together, these results suggest that the presence of p19 ARF -expressing cells in the lungs facilitates PPE-induced inflammation.
Macrophages have been reported to express high levels of p16 INK4a and senescent-associated β-galactosidase activity (Hall et al., 2017). Although p19 ARF expression in macrophages has not been documented, and no luciferase activity was detected in the BALF cells of ARF-DTR mice (Hashimoto et al., 2016), it is still possible that DT acted through the macrophage depletion in PPE-treated ARF-DTR mice. We checked the expression of INK4a, ARF, and DTR/ Luc in BALF cells prepared from ARF-DTR mice treated with PPE for 1 week. Although very low levels of INK4a were detected, both ARF and DTR/Luc were barely detectable in these cells ( Figure S4). Moreover, the INK4a level was unchanged in the BALF cells of DT-treated mice. Thus, it is unlikely that the effects of DT on PPE-induced emphysema were attributed to the ablation of nonsenescent macrophages in ARF-DTR mice.
The results described above imply that the elimination of p19 ARF -expressing cells suppressed the accumulation of inflammatory cells in lung tissues, thereby protecting these tissues from PPEinduced emphysema. These inflammatory cells express enzymes that exhibit elastolytic activity. MMP-12, also known as macrophage elastase, is involved in the development of emphysema, and a targeted deletion of MMP-12 confers resistance to cigarette smoke-induced emphysema (Hautamaki et al., 1997). Furthermore, many pro-inflammatory cytokines and MMPs are incorporated in SASP in humans and mice (Freund et al., 2010) and are elevated in senescent cells as well as in aged tissue. Therefore, we analyzed the mRNA expression levels of SASP-related cytokines and MMPs in PPE-treated lung tissues. Only MMP-12 showed a significant increase 1 week after the PPE treatment, and this was inhibited by DT pretreatment (Figure 5a). In addition, we observed an increase in the tissue inhibitor of metalloproteinase-2 (TIMP-2) levels in DT-treated lungs, although PPE by itself had no effect. The effects of DT are attributed to the elimination of p19 ARF -expressing cells because DT had no significant effect on PPE-treated wild-type animals ( Figure S5a). Immunohistochemical analyses revealed that MMP-12 was expressed in alveolar macrophages and capillary endothelia, although TIMP-2 was predominantly observed in macrophages (Figure 5c,d). As alveolar fibroblasts MIKAWA ET AL. These results imply that the pre-elimination of p19 ARF -expressing cells from adult lung tissues has protective effects in a mouse emphysema model. However, the method we utilized to eliminate p19 ARF -expressing cells depends entirely on the transgene and is not directly applicable to humans. One alternative approach that may be applicable to humans is "senolytic" agents that selectively kill senescent cells by targeting senescent cell-specific pro-survival pathways . Several potential senolytic drugs have recently been discovered or developed, and we used ABT-263 (Navitoclax) in this study. Six-month-old female ARF-DTR mice were orally administered ABT-263 for 8 weeks, as shown in Figure 6a (Figures 6h and S6).
However, ABT-263 did not affect Rn, which appears to reflect the  (Hashimoto et al., 2016) or in a PPE-induced emphysema model ( Figure S1). Consistent with these results, a morphometric analysis of lung sections revealed that ABT-263 partly restored alveolar size (mean linear intercept) in these mice (Figure 6i,j). Collectively, these results imply that senolysis prevents aspects of chemically induced emphysema.

| DISCUSSION
Pulmonary emphysema is a progressive lung disease characterized by the permanent enlargement of airspaces and is one of the most common conditions of COPD. Emphysema accompanies cellular senescence, and the incidence of this disease increases with age (Karrasch, Holz, & Jörres, 2008;Tsuji et al., 2006). clearance; however, this action may also result in adverse effects by inducing local inflammation (Coppé, Desprez, Krtolica, & Campisi, 2010). SASP appears to elicit deleterious effects in several disease models (He & Sharpless, 2017;Watanabe et al., 2017), and its inhibition was shown to be sufficient for preventing senescence-induced bone loss (Farr et al., 2017). Thus, SASP appears to accelerate alveolar macrophage accumulation upon PPE challenge and the release of tissue-destructive enzymes, including MMP-12, which ultimately leads to alveolar collapse.
We used a semi-genetic method for the elimination of targeted cells; however, as this method is not feasible in humans, the pharmacological targeting of senescent cells is being investigated . The senolytic activities and efficacies of some drugs have already been confirmed in mouse disease models Farr et al., 2017;Jeon et al., 2017;Ogrodnik et al., 2017;Roos et al., 2016;Schafer et al., 2017). We used the antiapoptotic Bcl-2 family inhibitor, ABT-263, in the PPE-induced emphysema model. ABT-263, or its related molecule ABT-737, has been shown to alleviate senescence-associated pathologies (Chang et al., 2016;Childs et al., 2016;Demaria et al., 2017;Yosef et al., 2016;Zhu et al., 2015Zhu et al., , 2016, but it may not be effective in all cell types (Schafer et al., 2017). In the PPE-induced emphysema model, The PPE-induced model is highly reproducible but does not capture all features of human emphysema (Antunes & Rocco, 2011). We also observed that cigarette smoking-induced pulmonary dysfunction was ameliorated by the elimination of p19 ARF -expressing cells (data not shown). Hence, our results support the targeting of senescent cells or a certain function of these cells, such as pro-inflammatory SASPs, as a promising approach for the prevention and suppression of the progression of emphysema. However, it currently remains unclear whether this approach is effective as a treatment for this disease.
Alveolar collapse is considered to be irreversible, and, once alveoli are broken, they may not be reconstructed in adult lung tissues. Nevertheless, it may still be possible to partially restore pulmonary function by reinforcing the remaining alveolar walls. It is important to note that lung tissue compliance was almost fully restored in the absence of

| Animals
All animal experiments were approved by and conducted in accordance with guidelines established by the National Center for Geriatrics and Gerontology Animal Ethics Committee. The animals were maintained under specific pathogen-free conditions, with a 12-hr light/dark cycle, constant temperature, and ad libitum access to food (CE-2; CLEA Japan) and water.
Female wild-type or ARF-DTR mice (Hashimoto et al., 2016) with the C57BL/6J background were randomly assigned to groups. Hemizygous ARF-DTR transgenic mice were used for the analysis. Five units of porcine pancreatic elastase (Elastin Products) in 100 μl phosphate-buffered saline (PBS) were intranasally administered using a standard pipette tip. For the DT treatment, the mice were intraperitoneally injected with 50 μg/kg body weight of DT (SIGMA) at 2week intervals. In the senolytic drug treatment, we used a protocol modified from that of Chang et al. (2016).  was dissolved in DMSO to obtain a 100 mM stock solution. The stock solution was diluted with vehicle containing 10% ethanol, 30% polyethylene glycol 400%, and 60% phosphatidylcholine and administered to mice via oral gavage at 25 mg/kg body weight per day.
The drug was administered for 5 days per cycle for 4 cycles, with a 1-week interval between cycles.

| Morphometry
All histopathological analyses were performed in a blinded manner.
The lungs were fixed with Mildform®20N (Wako Pure Chemicals Industries) at 25 cmH 2 O. Paraffin-embedded tissues were sectioned (5-μm-thick) and stained with hematoxylin and eosin. At least eight randomly selected fields per mouse were photographed. Test lines were randomly drawn on the images, and the intercepts with the tissue structure were counted for each line. Airway and vascular structures were eliminated from the analysis.

| In vivo imaging analysis
An in vivo luciferase imaging analysis was performed using the IVIS imaging system (Perkin Elmer). Mice were ventrally shaved and anesthetized with isoflurane (Wako Pure Chemicals Industries), and luciferin (VivoGlo; Promega) was intraperitoneally injected according to the manufacturer's instructions. Luciferase activity was monitored 10 min after the luciferin injection. Luminescence was quantified and analyzed using Living Image® software (Perkin Elmer).

| Pulmonary function tests (spirometry)
Pulmonary function tests were performed on a FlexiVent system (Scireq) as previously described (Hashimoto et al., 2016;Shalaby, Gold, Schuessler, Martin, & Robichaud, 2010). The mice were euthanized by an intraperitoneal injection of pentobarbital sodium (100 mg/kg of body weight) and connected to the FlexiVent system after tracheotomy. The mice were ventilated at a respiratory rate of 150 breaths/min with a tidal volume of 10 ml/kg against a positive Tissue elastance (H) and damping (G) were obtained from respiratory system impedance data using a constant phase model. All parameters were calculated using FlexiVent software.
In brief, BALF cells were prepared with 1 ml of PBS containing 5 mM EDTA, and the cells were collected from BALF by mild centrifugation. The collected cells were attached to glass slides using StatSpin Cytofuge (Beckman Coulter) and subjected to modified Giemsa staining using the Diff-Quick stain kit (Sysmex).

| Real-time PCR analysis
The total RNA was isolated from lung tissues using the PureLink®

| Immunohistochemistry
Paraffin-embedded tissue sections were deparaffinized, and antigens were retrieved for 5 min in a pressure cooker at 121°C in pH 9.0 antigen retrieval solution (Nichirei Bioscience). For MMP-12 staining, the sections were incubated with a rabbit monoclonal antibody (1:100 dilution, BS9869M, Bioworld Technology) at room temperature for 60 min. Sites of antibody binding were visualized with the Histofine simple stain MAX-PO(R) kit (Nichirei Bioscience). 3,3′-Diaminobenzidine tetrahydrochloride was used as a chromogen, and the sections were counterstained with hematoxylin. For TIMP-2 staining, a mouse monoclonal antibody (1:100 dilution, 3A4, Santa Cruz Biotechnology) was applied at room temperature for 60 min and visualized with the Histofine mouse staining kit (Nichirei Bioscience). For immunofluorescence, a rabbit polyclonal antibody against p19 ARF (1:300 dilution, ab80, Abcam) and a rat monoclonal antibody against fibroblasts (1:50 dilution, ER-TR7, Santa Cruz Biotechnology) were used. The sections were visualized with Alexa Fluor 488-conjugated anti-rat IgG and Alexa594-conjugated anti-rabbit IgG and were counterstained with DAPI.

| Statistical analysis
A one-way ANOVA was performed for the comparison of more than two sets of data. When the statistical model was proven to be significant, differences between combinations of the two groups were analyzed using a Tukey-Kramer or Steel-Dwass test. A two-tailed unpaired Student's t test was used for the comparison of two sets of experimental data. The data displayed a normal variance. Significance was represented by asterisks corresponding to *p < 0.05, **p < 0.01, and ***p < 0.001. No blinding was performed, except for histological quantifications. No statistical method was used to select the sample size.

ACKNOWLEDGMENTS
We thank Koichiro Kawaguchi, Noboru Ogiso, Ayumi Sugimoto, and Yuko Tottori for their technical assistance. We also thank Dr. Mitsuo Maruyama for his financial support and comments. This project was supported by grants from the National Center for Geriatrics and Gerontology

CONF LICT OF I NTEREST
The authors declare that they have no conflict of interests.

AUTHORS' CONTRIBUTI ON
RM performed the majority of the experiments. MS designed the experiments and wrote the manuscript, with contributions from TS.
YS and HB contributed to the establishment of the emphysema model. KK and KS performed the immunohistochemical studies.