Genetic ablation of histone deacetylase 2 leads to lung cellular senescence and lymphoid follicle formation in COPD/emphysema

Abstract

Histone deacetylase 2 (HDAC2), a critical determinant of chromatin remodeling, is reduced as a consequence of oxidative stress-mediated DNA damage and impaired repair. Cigarette smoke (CS) exposure causes DNA damage and cellular senescence. However, no information is available on the role of HDAC2 in CS-induced DNA damage, stress-induced premature senescence (SIPS), and senescence-associated secretory phenotype (SASP) during the pathogenesis of chronic obstructive pulmonary disease (COPD)/emphysema. We hypothesized that CS causes persistent DNA damage and cellular senescence via HDAC2-dependent mechanisms. We used HDAC2 global knockout (KO) and HDAC2 lung epithelial cell-specific KO [Clara cell-specific HDAC2 deletion (HDAC2 CreCC10)] mice to determine whether HDAC2 is a major player in CS-induced oxidative stress, SIPS, and SASP. HDAC2 KO mice exposed to CS show exaggerated DNA damage, inflammatory response, and decline in lung function leading to airspace enlargement. Chronic CS exposure augments lung senescence-associated β-galactosidase activity in HDAC2 KO, but not in HDAC2 CreCC10 mice. HDAC2 lung epithelial cell-specific KO did not further augment CS-induced inflammatory response and airspace enlargement but instead caused an increase in lymphoid aggregate formation. Our study reveals that HDAC2 is a key player regulating CS-induced DNA damage, inflammatory response, and cellular senescence leading to COPD/emphysema.—Sundar, I. K., Rashid, K., Gerloff, J., Rangel-Moreno, J., Li, D., Rahman, I. Genetic ablation of histone deacetylase 2 leads to lung cellular senescence and lymphoid follicle formation in COPD/emphysema.

Cigarette smoke (CS) is the major etiological factor that causes chronic obstructive pulmonary disease (COPD). COPD is a chronic inflammatory, debilitating disease, which is driven by 2 main signaling events (inflammation and cellular senescence and a state of irreversible growth arrest) that complement each other, culminating in accelerated or premature lung aging. Histone deacetylase 2 (HDAC2) is a class I HDAC that has been implicated in regulation of a variety of cellular processes and functions (cell cycle, proliferation, differentiation, development, and glucocorticoid response) (1). HDAC2, along with HDAC1, is involved in regulation of DNA damage response (DDR) and replicative senescence (2, 3), including inflammation and steroid resistance (4–6). We (7–9) and others (6, 10–13) have shown that HDAC2 level/activity is decreased in lung tissues, bronchial biopsies, alveolar macrophages, peripheral blood monocytes, and skeletal muscle from patients with COPD, as well as in lung inflammatory and structural cells (macrophages and epithelial cells), lung tissues, and skeletal muscle of CS-exposed mice. During chronic inflammation, corticosteroids suppress inflammation via recruitment of HDAC2 onto the promoter of NF-κB-dependent proinflammatory genes, thereby repressing their gene expression. Moreover, HDAC2-mediated deacetylation of the glucocorticoid receptor allows binding of the glucocorticoid receptor to the NF-κB complex, leading to suppression of NF-κB-driven proinflammatory gene transcription (4). Our previous report shows that HDAC2-deficient mice were nonresponsive to budesonide-mediated inhibition of LPS-induced lung inflammation (5). These studies highlight the critical role of HDAC2 in the pathogenesis of COPD/emphysema (4–6).

Cells undergoing senescence are primed to release proinflammatory mediators, a phenomenon termed senescence-associated secretory phenotype (SASP). It remains unclear whether HDAC2 directly plays a role in this positive feedback loop that could reinforce cellular senescence (14, 15). Furthermore, whether HDAC2 regulates CS-mediated DNA damage, proinflammatory response, and cellular senescence is not known. We hypothesize that CS exposure causes persistent DNA damage-induced, impaired repair [nonhomologous end joining (NHEJ)] via an HDAC2-dependent mechanism, thereby leading to augmented stress-induced premature senescence (SIPS), SASP, altered lung function, airspace enlargement, and premature lung aging. We used HDAC2 global knockout (KO) and HDAC2 lung epithelial cell-specific KO mice to determine the role of HDAC2 in CS-mediated SIPS and SASP, DDR, and lung inflammaging during the development of COPD/emphysema.

MATERIALS AND METHODS

Ethics statement and scientific rigor/reproducibility

All experiments for animal studies were performed in accordance with the standards established by the U.S. Animal Welfare Act, as set forth by the NIH guidelines. All animal protocols described in this study were approved by the University Committee on Animal Research Committee of the University of Rochester.

We used a rigorous/robust and unbiased approach throughout the experimental plans (e.g., in vivo mouse models) and during the analysis of the results to ensure that our data are reproducible as well as full and detailed reporting of both methods and raw/analyzed data. All of the key biologic and/or chemical resources that are used in this study were validated and authenticated (methods and resources) and are of scientific standard from commercial sources. Our results adhere to NIH standards of reproducibility and rigor.

Unless otherwise stated, all biochemical reagents used in this study were purchased from MilliporeSigma (St. Louis, MO, USA). The antibodies listed for human and mouse samples were of commercial grade and validated by manufacturers based on their data sheets.

HDAC2 global KO and HDAC2 lung epithelial cell-specific conditional KO mice

HDAC2 mutant (HDAC2 KO on C57BL/6J background) mice were kindly provided by Dr. J. A. Epstein (University of Pennsylvania School of Medicine, Philadelphia, PA, USA). These mice express a truncated and catalytically inactive form of HDAC2, with exons 9–14 replaced by a LacZ fusion gene created via a gene-trap method (16). We have used HDAC2 KO mice previously for acute CS exposure studies (5). Clara cell-specific HDAC2 deletion (HDAC2 CreCC10) mice were generated by crossing HDAC2 flox/flox (fl/fl) mice (17) [HDAC2 flox/flox mice were provided by Dr. Li-Huei Tsai (Massachusetts Institute of Technology, Cambridge, MA, USA)] with mice expressing the Cre recombinase transgene under the control of the CC10 promoter (CreCC10 mice; obtained from Dr. T. J. Mariani, University of Rochester). C57BL/6J mice (male and female) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and maintained separately in our mouse colony and used as wild-type (WT) controls, along with WT littermates. Our chronic exposure (air and CS for 6 mo) was conducted, along with other mouse strains that are not part of this study (18). Hence, we included WT-air and WT-CS groups from a previous study, along with additional groups for comparison (18). All mice were housed in the vivarium facility of the University of Rochester with a 12 h light/12 h dark cycle (lights on at 6:00 am).

CS exposures

Acute and subchronic CS exposure 

Two- to 4-mo-old (male and female) HDAC2 heterozygous (Het)/KO, WT littermates, HDAC2 fl/fl, and HDAC2 CreCC10 mice were bred and maintained with a 12 h light/dark cycle in the vivarium facility of the University of Rochester. For studies involving 3 d (acute) or 1 mo (subchronic) exposures, research-grade cigarettes (3R4F) were used to generate CS, and mice were exposed to CS, according to the Federal Trade Commission protocol (1 puff/min of 2 s duration and 35 ml vol) with a Baumgartner-Jaeger CSM2072i automatic CS-generating machine (CH Technologies, Westwood, NJ, USA; as previously described) (19, 20). The mainstream smoke concentration was set at a value of ∼250–300 mg/m³ total particulate matter (TPM) by adjusting the flow rate of the diluted medical air, and the level of carbon monoxide in the chamber was ∼350 ppm (19). Mice received 2, 1 h exposures (1 h interval/apart) daily for 3 consecutive d, and for 1 mo, exposure mice were exposed to CS for 5 h daily during 5 consecutive d/wk and were euthanized at 24 h after the last exposure. We used the following transgenic mouse strains for acute/subchronic CS exposures: HDAC2 lung epithelial cell-specific KO (HDAC2 fl/fl and HDAC2 CreCC10; for acute 3 d) and WT littermates, HDAC2 Het, and HDAC2 KO (homozygous) global KO (for subchronic 1 mo).

Chronic CS exposure 

Five- to 8-mo-old (male and female) WT littermates (C57BL/6J background), HDAC2 Het/KO, and HDAC2 lung epithelial cell-specific conditional KO (HDAC2 fl/fl and HDAC2 CreCC10) mice were exposed to chronic CS using research-grade cigarettes (3R4F), according to the Federal Trade Commission protocol (1 puff/min of 2 s duration and 35 ml vol) with a Baumgartner-Jaeger CSM2072i automatic CS-generating machine (CH Technologies), as previously described (19, 20). The mainstream smoke concentration was set at a value of ∼250–300 mg/m³ TPM by adjusting the flow rate of the diluted medical air, and the level of carbon monoxide in the chamber was ∼350 ppm (19). Mice were exposed to CS for 5 h daily during 5 consecutive d/wk and were euthanized at 24 h after the last CS exposure. Control mice were exposed to filtered air in an identical chamber according to the same protocol described for CS exposure.

Measurement of lung mechanics

Lung mechanical properties, including lung compliance, resistance, and elastance, were determined as previously described (21, 22). In brief, the mouse was weighed, deeply anesthetized by an intraperitoneal injection of pentobarbital (90 mg/kg), and tracheostomized. The trachea was cannulated, and the cannula was connected to a computer-controlled, small animal ventilator (FlexiVent; Scireq, Montreal, QC, Canada).

Bronchoalveolar lavage

Mice were injected with 100 mg/kg (body weight) of pentobarbiturate (Abbott Laboratories, Chicago, IL, USA) intraperitoneally and euthanized by exsanguination. The heart and lungs were removed en bloc, and the lungs were lavaged 3 times with 0.6 ml of 0.9% sodium chloride, as previously described (21, 22). The lavage fluid was centrifuged, and the cell-free supernatants were frozen at −80°C for later analysis. The bronchoalveolar lavage (BAL) cell pellet was resuspended in 1 ml of 0.9% sodium chloride and stained by acridine orange/propidium iodide (AO/PI) stain to determine the total cell counts per milliliter using a cellometer. For acute 3 d, subchronic 1 mo, and chronic 3 mo exposures, differential cell counts (minimum of 500 cells per slide) were counted on cytospin-prepared slides stained with Diff-Quik (Siemens, Newark, DE, USA) to determine the total number of macrophages, neutrophils, and total cells in BAL fluid (BALF).

Labeling of BAL cells for flow cytometry 

Flow cytometric analysis of immune inflammatory cells was performed using cell type-specific mAb in chronic (6 mo) air- and CS-exposed mice. In brief, 2.0 × 10⁵–4.0 × 10⁵ cells were stained in 1× PBS using cell type-specific markers for 30 min and then washed and resuspended in 0.1 ml of 1× PBS for analysis. Markers, such as LY6B.2 Alexa Fluor 488-conjugated antibody for neutrophils (Cat. #NBP213077AF488; Novus Biologicals, Littleton, CO, USA), F4/80 phycoerythrin-conjugated antibody for macrophages (Cat. #123109; BioLegend, San Diego, CA, USA), CD8a phycoerythrin-Cy5-conjugated antibody for T-lymphocytes (553034; BD Biosciences, San Jose, CA, USA), and CD45 allophycocyanin-conjugated antibody for leukocytes (Cat. #559864; BD Biosciences), were used. Flow cytometry data acquisition was performed on a BD Accuri flow cytometer (BD Accuri C6 software; BD Biosciences) and analyzed using FlowJo software (Ashland, OR, USA).

Protein extraction from lung tissues and quantification

One lobe of the lung tissue (∼50 mg) was homogenized (Pro 200 homogenizer; at maximum speed, fifth gear for 40 s) in 0.5 ml ice-cold RIPA buffer containing complete protease inhibitor cocktail (MilliporeSigma). The tissue homogenate was then incubated on ice for 45 min to allow total cell lysis. The homogenate was then centrifuged at 13,000 g for 5 min at 4°C to separate the protein fraction from the cell/tissue debris. The supernatant-containing protein was aliquoted and stored at −80°C for Western blotting. This fraction was taken for protein analysis by bicinchoninic acid colorimetric assay (Thermo Fisher Scientific, Waltham, MA, USA) using bovine serum albumin as a standard.

Western blot analysis

Western blotting was performed as described earlier (21, 22). Primary antibody incubations were performed overnight with 1:1000 dilutions [anti-p16 actin-related protein 2/3 complex subunit 5 (ab51243; Abcam, Cambridge, United Kingdom), anti-p21 (sc-6246; Santa Cruz Biotechnology, Dallas, TX, USA), anti-γH2A histone X (γH2AX) phospho-S139 (ab2893; Abcam), and anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH), -actin (sc-365062 and sc-1616; Santa Cruz Biotechnology), or -histone H3 (#9715; Cell Signaling Technology, Danvers, MA, USA)], followed by secondary anti-rabbit, -mouse, or -goat antibody incubation for 1 h at room temperature (1:5000 dilutions). ECL was used for detection, and images were taken with the ChemiDoc imaging system (Bio-Rad Laboratories, Hercules, CA, USA).

Measurement of SA-β-gal activity

Senescence-associated β-galactosidase (SA-β-gal) activity was quantitatively measured by the rate of conversion of 4-methylumbelliferyl-β-d-galactopyranoside to the fluorescent hydrolysis product 4-methylumbelliferone at pH 6.0, as previously described (23), using the cellular senescence activity assay kit (Enzo Life Sciences, Farmingdale, NY, USA), according to the manufacturer’s protocol. Normalized SA-β-gal activity was expressed as observed fluorescence divided by milligrams protein per milliliter.

SA-β-gal staining in lung tissues

Agarose-inflated lung tissues were fixed in 1% paraformaldehyde solution in 1× PBS for 1 h at 4°C with gentle rocking. After fixing, tissues were thoroughly washed 3 times with 10 ml 1× PBS at 4°C for 15–30 min. Then, the tissues were transferred to snap-cap tubes containing freshly prepared staining solution (5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 2 mM MgCl₂, 0.01% sodium deoxycholate, 0.02% Nonidet P-40, and 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside, 1 mg/ml in 1× PBS). The tubes containing tissues along with staining solution were incubated overnight at 30°C with gentle rocking. Finally, the stained tissues were washed in 1× PBS before the images were taken. SA-β-gal staining in lung tissues from chronic (3 mo) air- and CS-exposed HDAC2 KO mice was scored semiquantitatively in a blinded manner based on the intensity of staining in different lobes of lung tissues.

SA-β-gal staining in lung frozen sections

SA-β-gal staining was performed in lung frozen sections using the SA-β-gal staining kit, according to the manufacturer’s instructions (#9860; Cell Signaling Technology). The larger lobe of the mouse lung (left lobe) was inflated with 50% optimal cutting temperature (OCT) compound in 1 × PBS, embedded in optimal cutting temperature, and snap frozen at −80°C. Ten micrometers of thick, frozen sections was stored at −80°C until sections were ready for staining. In brief, lung frozen sections were thawed at room temperature, fixed with 2% formaldehyde containing 0.2% glutaraldehyde for 15 min, and washed with 1× PBS twice. Subsequently, lung sections were incubated with staining solution at 37°C overnight. Finally, slides were washed and counterstained with nuclear fast red and analyzed under light microscope for image capture using Spot Advanced software (SPOT Imaging, Sterling Heights, MA, USA).

Lung morphometry

Mouse lungs were inflated with 1% low melting agarose at a pressure of 25 cm H₂O and then fixed with 4% neutral-buffered paraformaldehyde (21, 22). Fixed lung was dehydrated and embedded in paraffin. Four micrometer sections were obtained with a rotary microtome (Microm International, Walldorf, Germany). Hematoxylin and eosin (H&E) staining was performed on the lung midsagittal sections to determine mean linear intercept (Lm) of airspace using the MetaMorph software (Molecular Devices, Sunnyvale, CA, USA). Eight to 10 randomly selected ×200 fields per slide were photographed in a blinded manner, and the images were manually thresholded for analysis (21).

Immunofluorescence staining and histology of BALT

Deparaffinized and rehydrated lung sections were boiled in Dako (Carpinteria, CA, USA) antigen retrieval solution for unmasking antigens crosslinked during formalin fixation. Nonspecific binding was blocked by incubating sections with 5% normal donkey serum and Fc block (10 μg/ml) in buffered saline containing detergent (0.1% Triton X-100 and 0.1% Tween-20). Endogenous biotin was blocked with a sequential avidin-biotin incubation step (MilliporeSigma). After blockade, the slides were incubated overnight at 25°C with biotin rat anti-CD45R (B220, Cat. #553092; BD Pharmingen, San Diego, CA, USA) to visualize B cells and goat anti-CD3ε to detect T cells (sc-1127; Santa Cruz Biotechnology). Additionally, we also stained lung sections with goat anti-mouse CXCL13 (AF470; R&D Systems, Minneapolis, MN, USA) and biotin rat anti-mouse CD21-CD35 (Cat. #123406; BioLegend), in combination with rat anti-mouse follicular dendritic cell (FDC; FDC-M1, Cat. #551320; BD Pharmingen) to detect FDCs and CXCL13-producing cells. Finally, the slides were incubated for 2 h at room temperature with Alexa Fluor 568 donkey anti-goat IgG (Thermo Fisher Scientific) and streptavidin, conjugated to either Alexa Fluor 594 or Alexa Fluor 488, and counterstained with ProLong Gold anti-fade with DAPI (Thermo Fisher Scientific). All sections were viewed, and images were taken with a Zeiss Axioplan 2 microscope using a Zeiss AxioCam HR digital camera (Carl Zeiss, Jena, Germany). Quantitative measurement in a formalin-fixed, paraffin-embedded saggital section was performed in H&E-stained lung sections. Stained slides were analyzed in a blinded fashion for the presence of mononuclear cell infiltrates and aggregates found adjacent to airways or blood vessels.

Proinflammatory cytokine analysis in BALF 

The level of proinflammatory mediators [monocyte chemotactic protein 1 (MCP-1), chemokine keratinocyte chemoattractant (KC), IL-6, and macrophage inflammatory protein 2 (MIP-2)] in BALF or lung homogenates was measured by ELISA using respective dual antibody kits (R&D Systems), according to the manufacturer’s instructions or using the Luminex Bioplex assay kit (Bio-Rad Laboratories). The results were expressed as picograms per milliliter for cytokines measured in BALF or picograms per milligram protein for cytokines measured using lung homogenates.

RNA extraction and real-time PCR array

Total RNA was prepared using the RNeasy miniprep kit (Qiagen, Valencia, CA, USA), according to the manufacturer’s instructions. RNA samples were quantified by the Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific). Mouse DNA damage-response genes (custom-designed array: 13 genes and 1 housekeeping control) and the mouse cellular senescence RT² Profiler PCR array were obtained from SABiosciences (Frederick, MD, USA). One microgram of total RNA was used for reverse transcription using the RT² first strand kit, and the entire cDNA reaction was diluted and distributed among the 96 wells of the PCR array plates. All of the real-time PCR reactions were performed with RT² SYBR Green/ROX PCR Master Mix using the CFX96 real-time system (Bio-Rad Laboratories). The results were analyzed using software provided by the manufacturer (Qiagen, Germantown, MD, USA; http://saweb2.sabiosciences.com/pcr/arrayanalysis.php), as previously described (24). Differentially regulated genes between WT- and HDAC2 Het-air- or -CS-exposed mice were identified by statistical significance and fold change >1.5 for the DNA damage-response array and fold change >2.0 for the cellular senescence array (Qiagen; Cat. #PAMM-050ZD). The P values are calculated based on a Student’s t test of the replicate 2^Δ comparative threshold cycle values for each gene in the air and CS groups, and P < 0.01 is considered statistically significant. For the DNA damage-response gene array and cellular senescence array data, GAPDH and β-glucuronidase, respectively, were used as housekeeping controls. Primer sequences for genes involved in formation and organization of ectopic lymphoid follicles (LFs) in experimental mice [Tnfa, lymphotoxin β (Ltb), Cxcl12, Cxcl13, Ccl19, and Ccl21b] were obtained from PrimerBank (http://pga.mgh.harvard.edu/primerbank/), as previously reported, and 18S rRNA was used as housekeeping controls (25).

Cellular senescence gene-expression panel by NanoString

RNA, prepared from lung tissues of chronic air- and CS-exposed WT and HDAC2 Het/KO mice, using an RNeasy Mini Kit (Qiagen), was processed through the NanoString nCounter system (NanoString Technologies, Seattle, WA, USA). Up to 150–300 ng RNA was submitted to NanoString Technologies for analysis. The code set was custom designed for us by the company specific for 41 different mouse cellular senescence genes, including 5 housekeeping genes (Gapdh, hypoxanthine phosphoribosyltransferase 1, ribosomal protein L19, TATA-box binding protein, and tubulin, β 5 class I). NanoString mRNA counts were normalized and log2 transformed for differential analysis using linear models in the limma package (R; Bioconductor; https://www.bioconductor.org/). Comparisons among different experimental groups were performed using a linear contrast model; moderated t statistics was used to determine the differences in gene-expression levels with empirical the Bayes approach. The Benjamini-Hochberg procedure was further used to adjust the P values to control the false discovery rate at 5%. Normalized NanoString mRNA counts were used to represent selected genes from the cellular senescence panel that was differentially expressed among WT and HDAC2 Het/KO chronic air- and CS-exposed mouse lungs. Details of sequences for the custom probe design used in this study for the mouse cellular senescence panel using NanoString technology are provided (Supplemental Table S1).

Statistical analysis

Statistical analysis of significance was calculated using unpaired Student’s t test. Probability of significance compared with control was based on 2-tailed t tests. Statistical analysis of significance was calculated using 1-way ANOVA for multigroup comparisons (Tukey’s multiple comparison test) using GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA). Results are shown as means ± sem.

RESULTS

HDAC2 deficiency exacerbates inflammation, cellular senescence, and DDR, compromising lung function during subchronic/chronic CS exposure

WT, HDAC2 Het, and HDAC2 global KO mice were exposed to subchronic and chronic CS to determine the role of HDAC2 in the DNA damage-induced lung inflammatory response, cellular senescence, and altered lung mechanics. CS-induced inflammatory cellular influx into BALF was assessed by Diff-Quik. Subchronic (1 mo) CS exposure resulted in a modest increase in total inflammatory cells, macrophages, and neutrophils in BALF of WT and HDAC2 Het/KO mice (Supplemental Fig. S1A–C). HDAC2 Het mice exposed to chronic CS (3 mo) showed a significant increase in total cell counts compared with WT-air. Furthermore, the macrophage counts were significantly increased compared with WT-CS-exposed mice. Both WT- and HDAC2 Het-CS-exposed mice showed a significant increase in neutrophil counts compared with respective air-exposed mice (Supplemental Fig. S2A–C).

We measured SA-β-gal activity in lung tissues of subchronic air- and CS-exposed WT and HDAC2 KO. CS exposure in HDAC2 KO mice showed a trend toward an increase in SA-β-gal activity measured in lung homogenates (Fig. 1A). Additionally, we measured the markers of cellular senescence (p16 and p21), including DNA damage (phosphorylated γH2AX), in lung homogenates of air, CS-exposed WT, and HDAC2 Het mice. HDAC2 Het mice exposed to CS showed augmented p16 levels compared with WT-air-exposed mice (Fig. 1B, C). However, we did not observe any significant differences in abundance of p21 and γH2AX between air- and CS-exposed WT and HDAC2 Het mice (Fig. 1B, C).

Figure 1.

Increased SA-β-gal activity, expression of cellular senescence, and DNA damage markers in HDAC2 KO mice exposed to subchronic CS. WT and HDAC2 KO mice were exposed to subchronic CS (∼250–300 mg/m³ TPM; 1 mo). A) SA-β-gal activity was determined in lung homogenates of 1 mo air- and CS-exposed WT and HDAC2 KO mice (after 1 h). B) Abundance of p16, p21 (cellular senescence), and γH2AX (S139) phosphorylation (DNA damage) was assessed in the whole-lung homogenate/nuclear extract from air- and CS-exposed WT and HDAC2 Het mice by immunoblot analysis. GAPDH/total histone H3 was used as a loading control. C) The band intensity was measured by densitometry, and data were shown as fold change relative to GAPDH/total histone H3 loading control. Data are shown as means ± sem from n = 4–6/group. #P < 0.05 significant compared with WT-air.

Lung mechanical properties in subchronic and chronic air- and CS-exposed WT and HDAC2 Het/KO mice were measured using the Flexivent system. Lung compliance was significantly reduced in CS-exposed WT and HDAC2 Het/KO mice that correlated with a significant increase in resistance and elastance compared with respective air-exposed mice (Supplemental Fig. S3A–C). Subchronic CS-exposure (1 mo) altered lung function was associated with a modest increase in airspace enlargement in CS-exposed HDAC2 KO mice compared with WT-air (P < 0.05). WT-air- and WT-CS-exposed mice did not show any significant difference in Lm, but the HDAC2 Het/KO-air showed relatively higher baseline Lm compared with WT mice (data not shown). Likewise, when we measured lung mechanics in WT and HDAC2 Het mice exposed to chronic CS (3 mo), lung compliance was significantly increased in both HDAC2 Het-air- and -CS-exposed mice (Supplemental Fig. S3D). HDAC2 Het-air- and -CS-exposed mice showed a significant decrease in elastance compared with WT-air (Supplemental Fig. S3E). Resistance remained unaltered in both WT- and HDAC2 Het-air- and -CS-exposed mice (Supplemental Fig. S3F). These results suggest that subchronic CS exposure in HDAC2 Het augments inflammatory cellular influx in BALF; SA-β-gal activity in the lungs, associated with increased cellular senescence marker (p16); altered lung mechanics; and airspace enlargement.

HDAC2 deficiency enhances chronic CS-induced lung cellular senescence

To determine the role of HDAC2 in chronic (3 and 6 mo) CS-mediated cellular senescence, SA-β-gal activity staining was performed in lung tissues. We measured cellular senescence in agarose-inflated lung tissues of 3 mo air- and CS-exposed HDAC2 KO mice. CS exposure after 3 mo in HDAC2 KO mice showed a significant increase in SA-β-gal staining compared with HDAC2 KO-air (Fig. 2A). Likewise, chronic (6 mo) CS exposure significantly augmented SA-β-gal activity in lung homogenates from WT-CS, HDAC2 Het-air, and HDAC2 Het-CS compared with WT-air-exposed mice (Fig. 2B). These data suggest that HDAC2 partial (i.e., Het mice) and global deletion increased cellular senescence measured by SA-β-gal activity/staining in lung tissues/homogenates. The cellular senescence phenotype was further augmented by chronic CS exposure in lungs of WT and HDAC2 Het/KO mice.

Figure 2.

Chronic CS exposure induces lung cellular senescence measured by SA-β-gal activity in HDAC2-deficient mouse lungs. WT and HDAC2 Het/KO mice were exposed to chronic CS (∼250–300 mg/m³ TPM; 3 and 6 mo). A) Representative images of agarose-inflated lungs fixed in 1% paraformaldehyde in 1× PBS, washed, and stained overnight with freshly prepared staining solution to demonstrate SA-β-gal staining in lung tissue. B) Lung SA-β-gal activity was determined in lung homogenates of chronic (6 mo) air- and CS-exposed WT and HDAC2 Het mice (after 1 and 3 h). Data are shown as means ± sem [n = 3/group for HDAC2 KO-air and -CS (3 mo); n = 6–12/group for WT- and HDAC2 Het-air and -CS (6 mo)]. ***P < 0.001 vs. respective air group; #P < 0.05, ##P < 0.01, ###P < 0.001 vs. WT-Air.

HDAC2 deficiency enhances proinflammatory cytokine production in mouse lungs after subchronic exposure to CS

To determine the role of HDAC2 in a CS-induced lung proinflammatory response, levels of SASP mediators in BALF and lung homogenates were measured. Subchronic and chronic CS exposure in WT and HDAC2 Het mice showed a differential response to SASP cytokines measured by ELISA/Luminex in BALF and lung homogenates. Subchronic CS exposure (1 mo) caused a significant increase in IL-6 and KC measured in BALF from WT-CS and HDAC2 Het-CS compared with respective air-exposed mice (Fig. 3A, B). Likewise, chronic CS exposure (3 mo) in HDAC2 Het mice showed a significant increase in MIP-2 levels in lung homogenates. Although other cytokines, such as MCP-1 and KC, showed a modest increase, differences were not significant when compared with respective air-exposed mice (Fig. 3C–E). WT-CS-exposed mice showed a significant increase in MCP-1 and MIP-2 and a modest increase in KC compared with WT-air (Fig. 3C–E). IL-6 levels remain unaffected in lung homogenates from air- and CS-exposed WT and HDAC2 Het mice (Fig. 3F). Additionally, we measured the levels of cytokines in BALF from chronic 3 mo air- and CS-exposed WT and HDAC2 Het mice by the Luminex multiplex assay (Supplemental Fig. S2D–H). HDAC2 Het-CS-exposed mice showed an increase in KC compared with HDAC2 Het-air (Supplemental Fig. S2F). WT-CS-exposed mice showed an increase in IP-10, IL-6, and KC levels compared with WT-air (Supplemental Fig. S2D–F). IL-1α was significantly decreased in both WT- and HDAC2 Het-CS-exposed mice compared with respective air-exposed control (Supplemental Fig. S2G). No significant change was observed in the release of IL-10 in BALF of air- and CS-exposed WT and HDAC2 Het mice (Supplemental Fig. S2H). These findings summarize that subchronic and chronic CS exposure preferentially augments SASP (proinflammatory) cytokines in BALF and lung homogenates from WT and HDAC2 Het mice.

Figure 3.

Subchronic and chronic CS exposure increased proinflammatory cytokines in BALF of HDAC2 KO mice. WT and HDAC2 Het mice were exposed to subchronic (1 mo) and chronic (3 mo) CS exposures (∼250–300 mg/m³ TPM). Mice were euthanized, and proinflammatory mediators in the BALF/lung homogenates were measured by ELISA. A, B) IL-6 (A) and KC (B) release in BALF of 1 mo air- and CS-exposed WT and HDAC2 Het mice. C–F) MCP-1 (C), MIP-2 (D), KC (E), and IL-6 (F), measured in lung homogenates of 3 mo air- and CS-exposed WT and HDAC2 Het mice. Data are shown as means ± sem (n = 3–4/group). *P < 0.05, **P < 0.01, ***P < 0.001, significant compared with respective air group.

HDAC2 deficiency differentially affects DNA damage response and cellular senescence pathway genes

To determine the role of HDAC2 in CS-mediated DDR and cellular senescence in the lungs, we performed custom- as well as pathway-specific quantitative PCR (qPCR) arrays. For DNA damage-response gene targets, we chose only 13 genes, along with 1 housekeeping control (GAPDH), for the analysis. Among the 13 genes, most of the genes were significantly downregulated >1.5-fold during 3 mo chronic CS exposure in HDAC2 Het and WT mice (Supplemental Fig. S4A–C). WT-CS-exposed mice showed a significant downregulation of 8 different DNA damage-response genes, which includes ataxia-telangiectasia mutated (Atm), ataxia telangiectasia and Rad3-related protein (Atr), Mre11a, CtIP, Xrcc5, Prdc, Terf2, and Rad54l compared with WT-air (Supplemental Fig. S4A). Likewise, when we compared HDAC2 Het-CS-exposed mice with HDAC2 Het-air, results revealed a significant downregulation of Atr, Mre11a, and Prdkc and upregulation of H2afx and Rad51 (Supplemental Fig. S4B). Furthermore, when we compared HDAC2 Het-CS with WT-air, Atm, Atr, Mre11a, CtIP, and Prkdc were significantly downregulated, whereas H2afx and Rad51 were upregulated during chronic CS exposure (Supplemental Fig. S4C). These data suggest that chronic CS exposure affects several different DNA damage/repair response genes in the HDAC2 Het and WT mouse lungs.

Next, we focused on mouse cellular senescence pathway genes differentially expressed in response to chronic (3 mo) CS exposure in HDAC2 Het and WT mouse lungs. The results showed downregulation of most of the senescence pathway genes >2.0-fold in chronic CS-exposed HDAC2 Het and WT mice (Supplemental Fig. S5A–D). In particular, when we compared WT-CS with WT-air-exposed mice, only Cdkn2a (p16^INK4a) was significantly upregulated, and all the other genes were significantly downregulated, e.g., genes for the senescence pathway [Atm, Bmi1, Cdk2, Cdk4, Cdk6, Cdkn2d, Chek1, Chek2, E2f3, Ets1, Ets2, retinoblastoma protein 1 (Rb1), Rbl2, Trp53, and Twist1]. Senescence initiators include p53/pRb signaling (cell cycle): Abl1, Cdkn1c, Gsk3b, Id1, Morc3, Rbl1, Serpinb2 (PAI-2), Serpine 1 (PAI-1), and Sirt1; IFN and IGF related: Egr1, Irf3, Igf1r, Igfbp3, and Igfbp5; MAPK signaling: Map2k1, Mapk2k6, and Mapk14; oxidative stress: Nox4 and Prkcd; and DNA damage: Pcna, Terf2, Tert, and Trp53bp1. Additional genes that belong to the category senescence responses include the p21 effector: Calr; cytoskeleton related/others: Fn1, Pik3ca, Thbs1, and Pten were also significantly downregulated (Supplemental Fig. S5A). Subsequently, we compared HDAC2 Het-CS with HDAC2 Het-air; Cdkn2a alone was significantly upregulated, and all of the other genes that belong to the above-mentioned senescence pathway were significantly downregulated (Atm, Cdk6, Cdkn2d, Ets1, Rbl2, Sirt1, Egr1, Mapk14, Pcna, Trp53bp1, Fn1, Pik3ca, and Thbs1; Supplemental Fig. S5B) .We found only 2 genes that were differentially expressed (Cdkn2a significantly upregulated and Serpinb2 significantly downregulated) in HDAC2 Het-air- vs. WT-air-exposed mice (Supplemental Fig. S5C). Finally, we compared HDAC2 Het-CS-with WT-air-exposed mice, which showed significant upregulation of 1 gene (Cdkn2a) and downregulation of other senescence pathway/initiator/response target genes, such as Atm, Ccne1, Cdk4, Cdk6, Cdkn1a, Cdkn1c, Cdkn2d, Chek1, Chek2, E2f3, Ets1, Ets2, Rb1, Rbl1, Rbl2, Cdc25c, Id1, Morc3, Serpinb2, Egr1, Igf1r, Mapk2k1, Mapk2k6, Mapk14, Gadd45a, Pcna, Terf2, Trp53bp1, Fn1, and Thbs1 (Supplemental Fig. S5D). These data corroborate our finding that expression of the cellular senescence marker (Cdkn2a) and SA-β-gal activity was increased in HDAC2 Het-air- and -CS-exposed mouse lungs. Chronic CS exposure affected most of the cellular senescence pathway target genes by downregulating their expression, thereby inducing the senescence program in the lungs.

Chronic CS exposure in HDAC2-deficient mice increases inflammatory cellular influx and proinflammatory cytokine release, which coincides with a decline in lung function and airspace enlargement

To determine if HDAC2 deletion can further augment chronic CS-induced inflammation, SASP, altered lung mechanics, airspace enlargement, and cellular senescence, we exposed HDAC2 Het/KO and WT mice to chronic (6 mo) CS exposure. First, we measured inflammatory cell influx in BALF using cell type-specific, fluorescently labeled antibodies by flow cytometry. CS-exposed HDAC2 Het and WT mice show a significant increase in the total cell counts and neutrophil percentage compared with respective air-exposed mice (Fig. 4A, C). These data correlate with significant reduction in the percentage of macrophages in HDAC2 Het-CS- and WT-CS-exposed mice compared with respective air-exposed mice (Fig. 4B). There was a modest increase in the percentage of lymphocytes in the HDAC2 Het-CS group compared with HDAC2 Het-air-exposed mice. There was no significant difference in the percentage of lymphocytes between WT-CS and WT-air (Fig. 4D).

Figure 4.

HDAC2-deficient mice show increased inflammatory cellular influx and proinflammatory cytokines in response to chronic CS exposure in BALF. WT and HDAC2 Het mice were exposed to chronic CS (∼250–300 mg/m³ TPM; 6 mo). After 24 h postlast exposure, mice were euthanized, and BALF was used for differential cell counts and proinflammatory mediator analysis. A) Total cell counts per milliliter in BALF were determined by AO/PI staining using a cellometer. B–D) Percentages of F4/80⁺ macrophages (B), LY6B.2⁺ neutrophils (C), and CD8a⁺ lymphocytes (D) were determined by flow cytometry. Proinflammatory mediators in the BALF were measured by the Luminex Bioplex assay kit (Bio-Rad Laboratories). E–I) MCP-1 (E), KC (F), MOP-1α (G), G-CSF (H), and IL-12p40 (I) release in BALF of 6 mo air- and CS-exposed WT and HDAC2 Het mice. Data are shown as means ± sem (n = 5–7/group). **P < 0.01, ***P < 0.001, significant compared with respective air group.

Next, we measured some of the SASP cytokines in BALF of chronic air- and CS-exposed HDAC2 Het and WT mice by the Luminex Bioplex assay (Bio-Rad Laboratories; Fig. 4E–I). Of the several proinflammatory cytokines measured, only 3 SASP mediators (KC, MIP-1α, and IL-12p40) showed a significant increase in both HDAC2 Het-CS- and WT-CS-exposed mice compared with respective air-exposed mice (Fig. 4F, G, I). MCP-1 levels were increased in WT-CS compared with WT-air-exposed mice (Fig. 4E). Likewise, we found that G-CSF levels were slightly increased in WT-CS and HDAC2 Het-CS compared with respective air-exposed mice (Fig. 4H). These data suggest an increased cell influx into the lung during chronic CS, which was associated with heightened release of specific SASP cytokines in BALF of WT and HDAC2 Het mice.

HDAC2 Het mice exposed to CS for 6 mo show altered lung mechanical properties compared with HDAC2 Het-air-exposed control. CS exposure significantly increased lung compliance in HDAC2 Het-CS and WT-CS compared with respective air-exposed mice (Fig. 5A). We did not observe any significant difference in lung resistance measured in HDAC2 Het- and WT-air- and -CS-exposed mice (Fig. 5B). As expected, WT-CS-exposed mice show an increased compliance that correlated with a significant decrease in elastance compared with WT-air, and HDAC2 Het-CS also showed a decrease in elastance compared with HDAC2 Het-air-exposed mice (Fig. 5C). We performed Lm analysis as a measure of alveolar destruction/airspace enlargement in the lungs following chronic CS exposure. HDAC2 KO mice exposed to chronic CS showed a significant increase in airspace enlargement compared with WT-air-exposed mice (Fig. 5D). Lm was increased, but it was not significant in WT-CS and HDAC2 KO-air-exposed mice compared with WT-air (Fig. 5D).

Figure 5.

Chronic CS exposure altered lung mechanics and increased airspace enlargement in HDAC2-deficient mice. WT and HDAC2 Het/KO mice were exposed to chronic CS exposure (∼250–300 mg/m³ TPM; 6 mo). A–C) Lung compliance (A), resistance (B), and elastance (C) were measured in chronic air- and CS-exposed WT and HDAC2 Het mice using a computer-controlled small animal ventilator (FlexiVent; Scireq). D) Lung morphometry analysis (Lm) was measured in paraffin-embedded lung sections stained with H&E by MetaMorph (Molecular Devices). Arrows indicate the airspace enlargement in the alveolar regions. Data are shown as representative of n = 3–6/group. *P < 0.05, significant compared with respective air group; #P < 0.05, significant compared with WT-air. Original magnification, ×200; original scale bars, 100 μm.

Additionally, we used frozen lung sections from chronic HDAC2 KO-CS- and WT-CS-exposed mice for SA-β-gal activity staining. We found increased SA-β-gal activity specifically in the bronchial lung epithelium (larger airways) of chronic CS-exposed HDAC2 KO and WT mice (Supplemental Fig. S6). For the first time, we show a greater degree of cellular senescence in the lungs caused by chronic CS, mainly in the airway epithelium (Clara cells/club cells), but not in alveolar type II cells that drive cell-senescent programming in the lung during COPD pathogenesis in vivo (Supplemental Fig. S6). Furthermore, when we analyzed Lm in HDAC2 Het/KO young (4–10 mo) vs. old (14–18 mo) mice, data revealed a significant increase in airspace enlargement in older HDAC2 Het/KO compared with young HDAC2 Het/KO mice (Supplemental Fig. S7A, B). These data are coherent with other parameters measured in HDAC2 Het/KO- and WT-CS-exposed mice. Our findings from a chronic CS model suggest that HDAC2 Het/KO mice show increased inflammatory cellular influx, proinflammatory cytokines (SASP) concomitant with a decline in lung function, airspace enlargement/emphysema, and greater cellular senescence response in the lungs.

Differential expression of cellular senescence genes by NanoString in HDAC2 KO mice by CS exposure

Our data from HDAC2 Het/KO mice showed a preferential role of the cellular senescence pathway in chronic CS-induced COPD/emphysema. Therefore, we performed the validation of our findings that CS induced cellular senescence in the lung using a novel transcript analysis for selected cellular senescence pathway genes by the NanoString technology. Based on the literature and current knowledge from animal models of cellular senescence, we designed a panel of 41 different mouse cellular senescence panels for differential expression analysis by the NanoString nCounter (Supplemental Table S1). A boxplot was generated using the boxplot function in R that shows the distribution of normalized transcript levels from different experimental groups (Supplemental Fig. S8A). A heatmap was generated using the heatmap.2 function in the g plot package in R. We have included the normalized, differentially expressed target genes and sample groups (WT-air, WT-CS, HDAC2 KO-air, HDAC2 Het-air, and HDAC2 KO-CS) for cluster analysis using the hierarchical clustering method. The heatmap showed normalized transcript levels from 41 different cellular senescence genes analyzed by the NanoString nCounter (Supplemental Fig. S8B). WT-CS-exposed mice showed a significant increase in mRNA transcripts for key cellular senescence genes, such as Mmp12, Ccl2, Cdkn2a, Tert, and Bub1b compared with WT-air (Supplemental Fig. S9A). HDAC2 KO mice exposed to CS significantly upregulate 2 genes (Mmp12 and Ccl2) and downregulate 4 genes (Terf2, H2afx, Ager, and Kl) compared with HDAC2 KO-air (Supplemental Fig. S9B). Air-exposed HDAC2 KO mice significantly upregulated mRNA transcripts of Tert, Cdkn2a, Kl, and Rgn genes compared with WT-air (Supplemental Fig. S9C). Other target genes, such as Ager, Hmgb1, Pik3ca, S100a8, Mmp9, and Hdac2, were significantly downregulated in HDAC2 KO-air vs. WT-air (Supplemental Fig. S9C). Likewise, when we analyzed HDAC2 KO-CS vs. WT-CS, only Serpinb2 was significantly increased, and other genes—H2afx, Pik3ca, and Hdac2—were significantly decreased (Supplemental Fig. S9D). Additionally, HDAC2 Het-air-exposed mice showed upregulation of Cdkn2a, Tert, and Camp and downregulation of Lmnb1 compared with WT-air (Supplemental Fig. S9E). Finally, HDAC2 KO-air-exposed mice showed a significant upregulation of Bub1b and downregulation of HDAC2 mRNA compared with HDAC2 Het-air (Supplemental Fig. S9F). We have additionally included scatter plots data for selected cellular senescence genes in lung tissues identified by NanoString during chronic CS-induced cellular senescence (Supplemental Fig. S10A–E). From the NanoString data, it is evident that chronic CS exposure differentially up- or downregulates cellular senescence pathway genes (SASP/prosenescence genes) in HDAC2 Het/KO and WT mice.

HDAC2 epithelial cell-specific conditional KO mice exposed to acute CS exposure show an increased lung inflammatory cellular influx and proinflammatory cytokine production

It has been shown that HDAC2 expression is significantly reduced in the lung epithelium during chronic CS in mouse lungs. To determine the lung epithelium-specific role of HDAC2 in a lung inflammatory response, we exposed HDAC2 fl/fl and HDAC2 CreCC10 mice to acute CS exposure (3 d). HDAC2 CreCC10 and HDAC2 fl/fl mice exposed to acute CS showed a significant increase in neutrophil counts in BALF compared with respective air-exposed mice (Fig. 6C). The total inflammatory cell counts were increased in acute CS-exposed HDAC2 fl/fl and HDAC2 CreCC10 mice, but not significantly compared with respective air-exposed mice (Fig. 6A). Macrophage counts were not significantly increased in acute air- and CS-exposed HDAC2 fl/fl and HDAC2 CreCC10 mice (Fig. 6B). Acute CS exposure-induced inflammatory cellular influx was associated with increased proinflammatory cytokine release (SASP: MCP-1 and KC) in BALF of HDAC2 CreCC10 and HDAC2 fl/fl mice compared with respective air-exposed mice. HDAC2 CreCC10 exposed to CS showed a significant increase in MCP-1 and KC levels compared with HDAC2 CreCC10 air- and HDAC2 fl/fl CS-exposed mice (Fig. 6D, E). These data suggest that lung epithelial cell-specific deletion of HDAC2 augments acute CS-induced inflammatory response (inflammatory cellular influx and proinflammatory cytokine release) in the mouse lungs.

Figure 6.

Lung epithelial cell-specific deletion of HDAC2 augments acute CS-induced neutrophil influx in the lungs and proinflammatory cytokine release in BALF. HDAC2 fl/fl and HDAC2 CreCC10 mice were exposed to acute CS (∼250–300 mg/m³ TPM; 3 d). After 24 h, postlast 3 d exposure, air- and CS-exposed mice were euthanized, and BALF was collected to perform differential cell counts. Total cell counts per milliliter in BALF was determined by AO/PI staining using a cellometer. A–C) At least 500 cells in BALF were counted with a hemocytometer to determine the number of total cells (A), macrophages (B), and neutrophils (C) on cytospin slides stained with Diff-Quik. D, E) Levels of proinflammatory cytokines MCP-1 (D) and KC (E) were measured in BALF after 3 d acute air or CS exposure. F, G) Levels of proinflammatory cytokines MCP-1 (F) and KC (G) were measured in lung homogenates of chronic air- or CS-exposed mice. Data are shown as means ± sem (n = 5–14/group). *P < 0.05, **P < 0.01, ***P < 0.001, significant compared with respective air-exposed group; #P < 0.05, significant compared with HDAC2 fl/fl CS; ##P < 0.01, significant compared with HDAC2 fl/fl air.

HDAC2 conditional KO did not affect chronic CS-induced lung proinflammatory response and airspace enlargement, but promotes BALT formation

To determine the impact of HDAC2 deletion in epithelial cells on the lung microenvironment after chronic exposure to CS, we assessed SASP mediators, airspace enlargement, cellular senescence, and bronchus-associated lymphoid tissue (BALT) formation and organization (LFs) in lungs of our experimental mice. We did not have sufficient numbers of mice to perform analyses for all the parameters in chronic air- and CS-exposed HDAC2 CreCC10 and HDAC2 fl/fl mice. Hence, we were unable to perform differential cell counts and lung-function measurements. We focused our interest to determine the local production of SASP mediators in lung homogenates, and we were interested in evaluating changes in the composition of pulmonary cell infiltrates and morphometric (airspace enlargement) and histologic (immunofluorescence for lymphoid aggregates) analyses in the lungs of HDAC2 fl/fl and HDAC2 CreCC10 mice that were exposed to air or CS. Chronic CS exposure increased proinflammatory cytokine release (SASP: MCP-1 and KC) in lungs of HDAC2 CreCC10 compared with air-exposed mice (Fig. 6F, G). HDAC2 fl/fl exposed to CS showed a significant increase in MCP-1 compared with air-exposed control. KC levels in HDAC2 fl/fl mice remained elevated without a significant difference between air- and CS-exposed groups (Fig. 6F, G). Chronic CS exposure did not show any significant increase in Lm measured by morphometry, both from HDAC2 fl/fl and CreCC10 mice (data not shown). HDAC2 CreCC10 air- and CS-exposed mice showed a modest increase in Lm, but it was not significant compared with HDAC2 fl/fl air- and CS-exposed mice (data not shown). Although HDAC2 fl/fl and CreCC10 mice did not show apparent alveolar damage, they developed BALT as early as 10 mo of age. Surprisingly, we did not detect LFs in 14- to 18-mo-old HDAC2 Het/KO mice. To determine the cellular composition of BALT-like structures in the HDAC2 fl/fl and CreCC10 mouse lung, we performed immunofluorescence staining in lung sections for detecting CD3⁺ T-lymphocytes, CD45R⁺ B-lymphocytes, and CD21⁺CD35⁺ FDCs. Our data showed distinctive BALT areas containing B and T cells, as well as the presence of organized FDC networks, which are indicative of ongoing B cell activation, in defined regions of the lungs of air- and CS-exposed HDAC2 fl/fl and HDAC2 CreCC10 mice (Fig. 7). Additionally, we performed staining in the lung sections for CXCL13, a chemokine implicated in the attraction of CXCR5⁺ B and T cells, to areas of BALT formation. Our data show the presence of BALT-like structures with FDC networks containing activated B cells that were localized with stromal cells expressing CXCL13 in the lung (Fig. 7). Histologic analysis preferentially showed the presence of multiple BALT-like structures in close proximity to the larger airways in HDAC2 fl/fl and HDAC2 CreCC10 mice exposed to air (Fig. 8A). In the lungs of mice chronically exposed to CS, BALT-like structures were distributed predominantly around blood vessels (perivascular) and near the larger airways (peribronchial; Fig. 8A). A blinded, semiquantitative analysis revealed a significant increase in the area covered by BALT among HDAC2 CreCC10 CS compared with HDAC2 CreCC10 air-exposed mice. We found a slight increase in the BALT score in HDAC2 fl/fl CS compared with HDAC2 fl/fl air mice (Fig. 8A, B). We measured mRNA levels for genes that are crucial in the formation and organization of ectopic lymphoid structures in the periphery using total lung RNA isolated from 10- to 12-mo-old HDAC2 fl/fl and HDAC2 CreCC10 mice exposed to air or CS. The results revealed a significant increase in expression of Cxcl13 and a modest increase in Tnfa, Ltb, Cxcl12, Ccl19, and Ccl21b in CS-exposed HDAC2 CreCC10 mice compared with respective air-exposed controls (Fig. 8C). Except for Cxcl13 and Ccl19, all of the other genes involved in ectopic lymphoid structure formation were significantly increased in CS-exposed HDAC2 fl/fl mice compared with air-exposed HDAC2 fl/fl mice at 10–12 mo of age (Fig. 8C). Overall, lymphoid aggregates observed in CS-exposed HDAC2 CreCC10 point to a potential role for epithelial cell HDAC2 in modulating the recruitment and organization of immune inflammatory cells into the lungs during the progression and pathogenesis of COPD/emphysema.

Figure 7.

BALT formation in chronic CS-exposed HDAC2 fl/fl and HDAC2 CreCC10 mouse lungs. HDAC2 fl/fl and HDAC2 CreCC10 mice were exposed to chronic CS (∼250–300 mg/m³ TPM; 6 mo). Cellular composition in pulmonary LFs of HDAC2 fl/fl and HDAC2 CreCC10 mice was determined by immunofluorescence staining. Lymphocytic aggregates were rich in T- and B-lymphocytes (CD3, red; B220, green) and contained FDC networks (FDC-M1/CD21-CD35, green) and stromal cells expressing CXCL13 (red). Representative images from n = 4–5/group. Original magnification, ×200. Original scale bars, 100 μm.

Figure 8.

Lung histology shows increased BALT formation and differential expression of genes involved in ectopic LF formation/organization in chronic CS-exposed HDAC2 fl/fl and CreCC10 mouse lungs. HDAC2 fl/fl and HDAC2 CreCC10 mice were exposed to chronic CS (∼250–300 mg/m³ TPM; 6 mo). Mononuclear cell aggregates in lungs of chronic air- and CS-exposed HDAC2 fl/fl and HDAC2 CreCC10 mice. A) Representative images from H&E-stained lung tissues from HDAC2 fl/fl and HDAC2 CreCC10 mice after chronic CS exposure (10–12 mo of age) are shown. We found LFs (arrow) adjacent to airways (AW; peribronchial) and blood vessels (V; perivascular). Original magnification, ×100 and ×200. Original scale bars, 100 μm. B) The number of BALT structures was significantly increased in chronic CS-exposed HDAC2 fl/fl and HDAC2 CreCC10 mice. Assessment of BALT coverage in lung tissues was performed semiquantitatively in a blinded manner. C) Expression of genes involved in the formation of ectopic LFs. The transcription levels of genes encoding Tnfa, Ltb, Cxcl12, Cxcl13, Ccl19, and Ccl21b were examined by qPCR using the 2^Δ comparative threshold cycle method. Bar graphs represent the mean normalized expression of respective air- vs. CS-exposed mouse lungs. Data are shown as means ± sem (n = 4–7/group). *P < 0.05, **P < 0.01, ***P < 0.001, significant compared with respective air-exposed group.

We further determined whether cell-specific deletion of HDAC2 augments chronic CS-induced cellular senescence. Chronic CS exposure did not show changes in SA-β-gal activity measured in lung homogenates from HDAC2 fl/fl and HDAC2 CreCC10 mice compared with respective air-exposed controls (data not shown). Lung epithelial cell-specific removal of HDAC2 did not augment CS-induced cellular senescence response in the lungs measured as SA-β-gal activity. Overall, these data suggest that lung epithelial cell-specific deletion of HDAC2 causes CS-induced inflammatory response (proinflammatory cytokine release), moderate changes in airspace enlargement, and enhanced BALT formation in the lungs but did not show an augmented cellular senescence response.

DISCUSSION

We and others have previously shown that the levels/activity of HDAC2 are substantially reduced in lungs of patients with COPD/emphysema, as well as in lungs of CS-exposed mice and macrophages, human bronchial epithelial, and small airway epithelial cells in vitro (4–7, 9, 26). CS is known to cause lung cellular senescence, and HDAC2 is implicated in regulation of DNA damage-initiated cellular senescence (1). However, the role of endogenous HDAC2 in the development of airspace enlargement is not known. In this study, we exposed HDAC2 global KO and HDAC2 lung epithelial cell-specific (HDAC2 CreCC10 mice) conditional KO mice to chronic CS to determine the role of HDAC2 in persistent DNA damage and cellular senescence in the pathogenesis of COPD/emphysema.

Our findings indicate that HDAC2 deficiency augmented CS-induced DDR, SIPS/SASP, and cellular senescence, leading to a decline in lung function and airspace enlargement, which are the key characteristic features of COPD/emphysema. Furthermore, chronic CS exposure in HDAC2 lung epithelial cell-specific conditional KO showed a modest inflammatory response, airspace enlargement, and enhanced BALT formation but not cellular senescence. Collectively, these data suggest that HDAC2 exhibits a novel role in protection against DNA damage and cellular senescence in premature lung aging/COPD. HDAC2 partial deficiency or KO mice developed spontaneous airspace enlargement (COPD-like phenotype) in an age-dependent manner but did not develop lymphoid aggregate formation. Additionally, global HDAC2 Het/KO mice exposed to chronic 3–6 mo of CS exposure caused increased inflammatory response, SIPS and SASP phenotype associated with increased DDR, cellular senescence, lung-function decline, and airspace enlargement. This study supports that HDAC2 reduction by CS can augment DNA damage, proinflammatory responses (SIPS/SASP), and cellular senescence, eventually leading to COPD/emphysema.

DNA double-strand breaks (DSBs) are among the major types of DNA damage caused by CS exposure (27, 28). HDACs and histone acetyltransferases play an important role during DSBs (29). HDAC1 and HDAC2 act as gatekeepers, thus selectively recruit proteins to the DSB to initiate repair during DDR (29, 30). Depletion of HDAC2 resulted in cell hypersensitivity to DNA damaging agents and sustained DDR, during which NHEJ levels are reduced (3). The additional role of HDAC1 and HDAC2 on deacetylation of histone H3K56 and H4K16 was provided as H3K56ac is perturbed at sites of DNA damage (3). A previous report from our laboratory has shown increased H3K56ac and H4K12ac by CS in mouse lungs associated with HDAC2 reduction (24). Our data from HDAC2 Het exposed to chronic CS show an increased expression of H2afx and Rad51 in the lungs, which complements the role of HDAC2 in DDR via NHEJ. Gene expression of all of the other DDR genes (Atr, Mre11a, and Prkdc) was significantly reduced in HDAC2 Het-CS- compared with air-exposed control. Chronic CS-induced reduction in HDAC2 and HDAC2 KO exposed to CS led to augmented DNA damage/impaired DNA repair (reduced Ku70/Ku80), as measured by expression of DDR and cellular senescence pathway genes, which is in concurrence with the previous reports demonstrating the HDAC2 role in DDR and genomic stability (3, 31, 32). Cell-cycle arrest is a phenomenon that is caused by the activation of DDR, initiated by ATM and ATR kinases, via activation of tumor suppressor protein p53 and upregulation of cyclin-dependent kinase inhibitors (p16^INK4a and p21^CIP1) (2, 33–35). The hallmark of senescent cells includes proliferation arrest (no cell division), very high metabolic activity, and enlarged cell phenotype associated with secretory phenotype (release of cytokines, chemokines, growth factors, and matrix metalloproteases), which are collectively referred to as SASP (36). A recent study shows the HDAC2 role in suppression of IL-17A-mediated airway remodeling in a human and mouse model of COPD (37). Our data from HDAC2 Het/KO mice exposed to CS show augmented cellular senescence markers (p16 and SA-β-gal activity) in the lungs associated with proinflammatory SASP mediators (IL-6, KC, MCP-1, and MIP-2) in BALF or lung homogenates.

Our NanoString mRNA transcript analysis demonstrated the role of cell-cycle inhibitor p16^INK4a, serine protease inhibitor Serpinb2 (PAI2), matrix metalloprotease Mmp12, and cytokine Ccl2 genes in CS-induced DNA damage, cellular senescence, and COPD/emphysema. We found that chronic CS exposure in HDAC2 KO mice significantly induced SASP mediators Mmp12 and Ccl2 gene (MCP-1) at the mRNA level in lung tissues. WT-CS-exposed mice showed increased expression of Tert, whereas the HDAC2 KO-CS-exposed mice showed reduced expression of Terf2, suggesting that telomere length-related genes play an important role in the regulation of CS-induced DNA damage and cellular senescence. The Serpinb2 gene was significantly upregulated in HDAC2 KO-CS-exposed mice compared with WT-CS. Recently, SerpinB2 has been shown as a direct downstream target of p53 that is activated by the DNA damage-response pathway and binds to stabilize p21 levels in senescent cells (38). These findings are in agreement with a recent report that SerpinB2 has a critical role in mediating cellular senescence (38). Our data on SASP mediators and cellular senescence biomarkers identified in lungs of CS-exposed mice are comparable with SASP factors and senescence genes measured in senescent cells or in age-associated pathologies (39, 40).

It has been shown that p16^INK4a and p21^CIP1/WAF1 (cyclin-dependent kinase inhibitors) abundance is increased by CS-induced DNA damage and cellular senescence in lung cells in vitro and mouse lungs in vivo (22, 28, 41–43). Additional evidence from patients with COPD shows an increase in SASP, determined by the percentage of p16 and phosphorylated NF-κB-positive alveolar type II cells in the lungs (44). It is known that HDAC2 regulates prosenescence gene expression (p16 and p21) at the level of transcription, which in turn, may delay cellular senescence (2, 33, 34). Previous findings from our laboratory showed the deletion of prosenescence gene p21 protected against CS-induced inflammation, airspace enlargement, and cellular senescence (43, 45). Furthermore, pharmacological inhibition of HDAC by trichostatin A in rats has been shown to cause emphysema via an HDAC-dependent mechanism in the lungs (46). We have previously shown that CS extract reduced total HDAC activity and protein levels of HDACs 1–3 associated with post-translational modification of HDAC protein by nitrotyrosine and aldehyde-adduct formation in MonoMac6 cells (9). Additionally, we reported that phosphorylation of HDAC2 on serine/threonine residues by a protein kinase casein kinase II-dependent mechanism reduces HDAC2 activity in vitro in MonoMac6 and bronchial and airway epithelial cells and in vivo in acute and chronic CS-exposed mouse lungs as a result of increased ubiquitin-proteasome-dependent degradation (7), which may influence steroid resistance and cellular senescence mechanisms during the pathogenesis of COPD/emphysema. It is possible that various HDACs (HDACs 1–3) differentially affect the proinflammatory response and cellular senescence in a cell type-specific manner. A recent report shows the role of microRNA-21 in a steroid-insensitive experimental asthma by amplifying PI3K-mediated suppression of HDAC2 (47). The present study does not focus on the contribution of other epigenetic mechanisms, such as microRNAs and histone modifications. Future studies will further investigate the role of novel epigenetic mechanisms in CS-induced cellular senescence and COPD in HDAC2 global KO and HDAC2 lung epithelial cell-specific KO mice.

Our data from a chronic CS exposure model demonstrate increased lung compliance and reduced elastance associated with increased cellular senescence and airspace enlargement in HDAC2 Het/KO mice. A previous report shows that p16 removal (p16^INK4a INK-ATTAC transgenic mouse model) delayed the onset of age-related pathologies in both WT and progeroid (mitotic checkpoint kinase budding uninhibited by benzimidazole-related 1) KO mice and increased the lifespan of WT mice (40, 48). Recently, the use of the INK-ATTAC transgenic mice for removal of senescent cells improved lung function and SASP in the mouse model of bleomycin-induced pulmonary fibrosis (49). We have made attempts to remove p16 in HDAC2 KO mice by backcrossing p16 KO with HDAC2 KO mice. Our preliminary findings from p16 × HDAC2 double-KO mice showed protection against an acute CS- and N-formylmethionyl-leucyl-phenylalanine-induced inflammatory response, suggesting that the selective removal of p16 in the lungs could protect against oxidative stress-mediated DNA damage, cellular senescence (SIPS/SASP), and development of COPD/emphysema (unpublished observations). As the genetic approaches for removal of senescent cells in patients with COPD may not be feasible, alternative methods using pharmacological agents (drivers of senescent cell apoptosis), referred to as senolytic drugs, offer a promising alternative (50). Senotylic drugs show specificity by inhibiting prosurvival and anti-apoptotic proteins that are significantly upregulated in senescent cells (50, 51). Recent reports show that senolytic drugs [dasatinib (a tyrosine kinase inhibitor) + quercetin (a flavonoid and inhibitor of PI3Ks)] target alveolar epithelial cell function and SASP factors and attenuate experimental lung fibrosis ex vivo and in vivo (49, 52). These studies suggest that senolytic drugs could be a viable treatment option that needs to be thoroughly explored in pathologies that involve cellular senescence mechanisms using mouse models of age-related chronic lung disease.

Mice with reduced HDAC1/2 activity (HDAC1 KO and a single HDAC2 allele) develop pathologic changes after 3 mo of age (neoplastic transformation of immature T cells in the thymus) (53), skeletal muscle alterations, tumorigenesis, and cardiac and neurologic changes (16, 17, 54, 55). Recently, it has been shown that chromatin-remodeling factors HDAC1/2 regulate sex-determining region Y-box 2⁺ progenitors in the lung endoderm differentially during lung development/postnatal homeostasis and regeneration via derepression of bone morphogenetic protein 4 and Rb1 in a stage-specific manner (56). All of the previous studies did not identify any lung abnormalities/phenotypes in HDAC2 global or conditional KO mice.

LFs are formed in lungs of patients with COPD, likely as a result of the ongoing chronic stimulation of stromal and immune cells by inflammatory mucus exudates, which results in the sustained attraction of more innate and adaptive inflammatory immune cells to the inflamed lung (57). LFs consist of compact and organized B cell follicles with central FDC networks and surrounded by T cells. LFs have been previously detected in the parenchyma and in bronchial walls of patients with COPD and chronic CS-exposed mice (58–60). In chronic CS-exposure models, clonal expansion and antigen-induced proliferation in B cell follicles were associated with progressive inflammation and an increase in alveolar airspace enlargement, as a result of local autoantibody production against CS components or other components of the extracellular matrix (61). In our study, HDAC2 deficiency in lung epithelial cells specifically showed an increased organization in the LFs (BALT), which contained numerous B- and T-lymphocytes and occasional FDC networks in the lungs of mice exposed to CS. Furthermore, we observed expression of CXCL13-positive cells in FDC networks of chronic CS-exposed mouse lungs by immunofluorescence and confirmed by qPCR analysis the significant expression of genes involved in the formation of ectopic LFs. It is known that CXCL13 is important in CS-induced development of tertiary lymphoid organs. Consistent with our findings, a previous report demonstrated the role of the B cell-attracting chemokine CXCL13 in the formation and organization of LFs in patients with COPD and in a mouse model of COPD (62). Based on the strategic localization of BALT structures, we assume that the production of proinflammatory cytokines by epithelial cells can create the ideal conditions for triggering formation of ectopic LFs. For example, IL-6, in combination with TGF, is critical in inducing the differentiation of T helper 17 cells (63, 64). Interestingly, IL-17 induces the expression of neutrophil-attracting chemokine KC by epithelial cells (65). Our data show increased levels of IL-6 and KC in the BALF/lungs and enhanced neutrophils in CS-exposed HDAC2-Het/KO and KC release in HDAC2 lung epithelial cell-specific KO mice. Hence, we speculate that HDAC2 deficiency in epithelial cells likely facilitates the local production of IL-17 that may have contributed to trigger BALT formation, as previously proposed in a neonatal model of LPS-induced pulmonary inflammation (66). Another possible explanation for BALT formation could be a result of an increase in autoantibodies against lung proteins (generation of citrullinated and carbonylated proteins). Thus, it is possible that citrullinated/carbonylated proteins or neoantigens induced by CS may induce the formation of ectopic LFs and the local production of pathogenic plasma cells that secrete antibodies against lung proteins (67, 68).

It is well known that the severity of COPD is associated with the formation and organization of LFs, which are likely supporting local B and T cell activation. A recent study supports that patients with COPD have increased steroid-resistant CD28null (senescent), proinflammatory T cell, and NK T-like cells (CD8⁺ subsets) in peripheral blood (69). Loss of HDAC2 in CD28nullCD8⁺ T and NK T-like cells is shown to be associated with lymphocyte senescence in patients with COPD. This was abrogated by upregulation of HDAC2 in these cells in the presence of theophylline and prednisolone or cyclosporine A by decreasing lymphocyte subset-specific proinflammatory cytokines (TNF-α and IFN-γ) (70). For the first time, we observed increased LF formation around the vessels in HDAC2 CreCC10 mice, possibly indicating lymphocyte senescence caused by CS, implicating the role of HDAC2 during steroid resistance in patients with severe COPD. Our findings indicate that HDAC2 plays an important role during CS-induced DDR, cellular senescence (SIPS and SASP), lung function decline, and airspace enlargement in the pathogenesis of COPD/emphysema.

Limitations of the study

We agree that our study has some limitations, such as the fact that acute, subchronic, and chronic CS exposures in vivo were conducted separately, in different batches, depending on the availability of HDAC2 global KO and lung epithelial cell-specific KO genotypes. Hence, we were unable to perform a direct comparison of our data among acute, subchronic, and chronic exposures, as well males and females for sex differences. We used both HDAC2 Het/homozygous KO mice for these studies and have explicitly mentioned the genotypes for parameters analyzed, as they develop similar lung pathology in acute vs. chronic CS exposures. We did not have a sufficient number of mice in HDAC2 fl/fl and HDAC2 CreCC10 mice to analyze all of the parameter, such as differential cell counts, BALF cytokines, and lung function. We performed acute CS and N-formylmethionyl-leucyl-phenylalanine exposures using p16× HDAC2 double-KO mice but were unable analyze all of the parameters as a result of increased tumor development in the spleen and lymphoid organs as they age (18). Our data from double-KO mice may provide confounding results as a result of systemic inflammation during tumor formation in other organs. Hence, we decided not to pursue use of double-KO mice in this study. It will be ideal to generate cell type-specific deletion of p16 × HDAC2, which is beyond the scope of the present study, to determine their role in CS-induced cellular senescence and COPD.

In summary, chronic CS exposure in HDAC2 deletion mice causes emphysema as a result of a persistent DNA damage-induced inflammatory response and cellular senescence (SIPS and SASP) mechanism. Furthermore, HDAC2 conditional deletion in lung epithelium showed an inflammatory response associated with mild airspace enlargement and enhanced lymphoid aggregate formation in the lungs. These findings highlight the crucial role of HDAC2 in CS-induced DNA damage/impaired repair response, cellular senescence phenotypes (SIPS and SASP), and LF formation and organization during the development of COPD/emphysema. Our data suggest that activation of HDAC2, using small molecules and/or novel senolytic drugs to maintain the balance between histone acetyltransferases and HDACs, may advance treatment options, where DNA damage, cellular senescence, and steroid resistance drive the pathogenesis in smoking-related chronic lung diseases, such as COPD.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

AO/PI

acridine orange/propidium iodide

ATM

ataxia-telangiectasia mutated

ATR

ataxia telangiectasia and Rad3-related protein

BAL

bronchoalveolar lavage

BALF

bronchoalveolar lavage fluid

BALT

bronchus-associated lymphoid tissue

COPD

chronic obstructive pulmonary disease

CreCC10

Cre recombinase Clara cell 10kD

CS

cigarette smoke

DDR

DNA damage response

DSB

double-strand break

FDC

follicular dendritic cell

fl/fl

flox/flox

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

H&E

hematoxylin and eosin

H2AX

histone H2AX

HDAC

histone deacetylase

HDAC2 CreCC10

Clara cell-specific deletion of histone deacetylase 2

Het

heterozygous

KC

keratinocyte chemoattractant

KO

knockout

LF

lymphoid follicle

Lm

mean linear intercept

Ltb

lymphotoxin β

MCP-1

monocyte chemotactic protein 1

MIP-2

macrophage inflammatory protein 2

NHEJ

nonhomologous end joining

qPCR

quantitative PCR

Rb1

retinoblastoma protein 1

SA-β-gal

senescence-associated β-galactosidase

SASP

senescence-associated secretory phenotype

SIPS

stress-induced premature senescence

TPM

total particulate matter

WT

wild type

AUTHOR CONTRIBUTIONS

I. K. Sundar, K. Rashid, J. Gerloff, and I. Rahman conceived of and designed the experiments; I. K. Sundar, K. Rashid, J. Gerloff, J. Rangel-Moreno, and D. Li performed the experiments and analyzed the data; and I. K. Sundar and I. Rahman wrote and edited the manuscript.