Hypoxia-induced deoxycytidine kinase expression contributes to apoptosis in chronic lung disease

Abstract

Chronic obstructive pulmonary disease (COPD) is characterized by persistent inflammation and tissue remodeling and is a leading cause of death in the United States. Increased apoptosis of pulmonary epithelial cells is thought to play a role in COPD development and progression. Identification of signaling pathways resulting in increased apoptosis in COPD can be used in the development of novel therapeutic interventions. Deoxyadenosine (dAdo) is a DNA breakdown product that amplifies lymphocyte apoptosis by being phosphorylated to deoxyadenosine triphosphate (dATP). dAdo is maintained at low levels by adenosine deaminase (ADA). This study demonstrated that mice lacking ADA developed COPD manifestations in association with elevated dAdo and dATP levels and increased apoptosis in the lung. Deoxycitidine kinase (DCK), a major enzyme for dAdo phosphorylation, was up-regulated in mouse and human airway epithelial cells in association with air-space enlargement. Hypoxia was identified as a novel regulator of DCK, and inhibition of DCK resulted in diminished dAdo-mediated apoptosis in the lungs. Our results suggest that activating the dAdo-DCK-dATP pathway directly results in increased apoptosis in the lungs of mice with air-space enlargement and suggests a novel therapeutic target for the treatment of COPD.—Weng, T., Karmouty-Quintana, H., Garcia-Morales, L. J., Molina, J. G., Pedroza, M., Bunge, R. R., Bruckner, B. A., Loebe, M., Seethamraju, H., and Blackburn, M. R. Hypoxia-induced deoxycytidine kinase expression contributes to apoptosis in chronic lung disease.

Chronic obstructive pulmonary disease (COPD) is a deadly lung disease characterized by persistent inflammation and tissue destruction that leads to loss of lung function. COPD is initiated by exposure to noxious particles or gases that induce inflammation and lead to chronic bronchitis-bronchiolitis and/or emphysema (1), causing airflow limitation. COPD significantly impairs human health and has become one of the leading causes of death in the United States (2–4). Moreover, the morbidity and mortality rates of this disease are anticipated to increase in the coming decades, and there are few therapeutic options. Therefore, a better understanding of the pathogenesis of COPD and the regulatory pathways involved is critical for the development of novel therapies for this disorder.

Apoptosis is one of the most important cellular processes associated with COPD pathogenesis (4). Increased apoptosis of pulmonary epithelial cells, endothelial cells, and mesenchymal cells are prominent in patients with COPD and is directly linked to lung tissue destruction and the development of emphysema (4). Molecular mechanisms implicated in the pathogenesis of COPD have been shown to modulate cell survival by regulating, impairing, or promoting the apoptosis of various cells, including epithelial, endothelial, and inflammatory cells. Although apoptosis appears to be a critical cellular process in the development and progression of COPD, the signaling pathways responsible for its increase are not fully understood. It is essential to gain further understanding of the mechanisms that regulate apoptosis in COPD in order to attenuate or reverse the progression of the disease.

We have shown that mice with adenosine deaminase (ADA) enzyme deficiency spontaneously develop emphysemic changes in the lung similar to what is observed in patients with COPD (5). This is in line with clinical observations where patients with COPD exhibit a down-regulation of ADA in the lungs (6). ADA is responsible for the deamination of adenosine (Ado) and deoxyadenosine (dAdo), molecules with potent biological activities. Ado binds to G-coupled Ado receptors and can elicit both tissue protective and/or destructive activities (7). dAdo is normally generated as a by-product of apoptosis from DNA breakdown and is highly cytotoxic to cells. If allowed to accumulate, dAdo leads to apoptosis by a pathway that involves its phosphorylation to deoxyadenosine triphosphate (dATP), which, in turn, promotes the induction of apoptosis (8). This is the mechanism by which T cells undergo apoptosis in ADA-deficient mice and humans, leading to the characteristic lymphopenia seen in this disorder (9). However, the effect of dAdo on apoptosis of stromal cells in the lung has not been examined.

Deoxycytidine kinase (DCK) and adenosine kinase (AK) are key enzymes that phosphorylate dAdo into dATP (10). DCK is widely expressed in nonlymphoid tissues at low levels and is elevated in certain lymphoid cells and in malignant cells (11). DCK-knockout mice are immunodeficient, with impaired T and B lymphocyte development, likely due to an imbalance between apoptosis and proliferation of certain lymphocytes at specific stages of development (12). DCK activation is normally correlated with increased dATP accumulation and cell apoptosis (13). In the context of COPD, increased levels of apoptosis are observed; however, whether DCK plays a role in the pathogenesis of COPD has not been investigated.

The purpose of this study was to investigate whether the dAdo-DCK pathway is activated in lungs exhibiting features of COPD and to determine whether this pathway can directly affect the pathogenesis of disease. To achieve this goal, we investigated the role of the dAdo-DCK pathway using Ada^−/− mice as a model of air-space enlargement. We found that levels of DCK were elevated in alveolar epithelial cells (AECs) in the lungs of Ada^−/− mice in conjunction with elevations in dATP and increased apoptosis. Moreover, in vivo and in vitro evidence suggest that hypoxia regulates DCK expression in airway epithelial cells, which provides a novel mechanism of DCK up-regulation in COPD. Clinical significance of these findings was shown by examination of DCK expression in patients with COPD in which a marked up-regulation in hyperplastic epithelial cells was observed. These results suggest that DCK up-regulation and dAdo accumulation are novel mediators of disease amplification in COPD.

MATERIALS AND METHODS

ADA-deficient (Ada^−/−) mice and enzyme replacement

Ada^−/− mice were generated and genotyped as previously reported (14). Once Ada^−/− mice were identified at birth, ADA enzyme therapy was conducted every 4 d from postnatal day 2 (P2) to P21. At 3 wk after cessation of enzyme therapy (P42), mice were euthanized, and the lungs were collected for further analysis (15).

Nucleoside and nucleotide measurement

Nucleosides and nucleotides were measured as described previously (14). Briefly, to measure the nucleoside levels in bronchoalveolar lavage (BAL) fluid, mice were anesthetized with 2.5% avertin, and the lungs were lavaged 4 times with 0.3 ml PBS containing 2 μM dipyridamole (Sigma-Aldrich, St. Louis, MO, USA) and 5 μM ADA inhibitor deoxycoformycin (dCF; R&D Systems, Minneapolis, MN, USA). BAL fluid was then centrifuged to remove cells and debris. To measure nucleoside levels, 200 μl BAL supernatant was analyzed by reverse-phase HPLC, as described previously (17). Representative peaks were identified and quantitated using external standard curves.

To measure nucleotide levels in lung tissue, whole mouse lungs were homogenized in 800 μl of 60% methanol, vigorously vortexed, and stored at −20°C overnight (17). To measure nucleotide levels in MLE12 cells, cells were scraped, washed, vigorously vortexed in 60% methanol, and stored at −20°C overnight. The lysate was centrifuged, and supernatants were collected and air dried to remove the methanol. Pellets were dissolved in 400 μl mobile phase (0.02 M NH₄H₂PO₄, pH 6.5) and analyzed by HPLC, as described previously (14).

Western blot analysis

MLE12 cells were collected and lysed in RIPA lysis buffer (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; and 1% Nonidet P-40) containing a protease inhibitor cocktail (Thermo Fisher Scientific, Fair Lawn, NJ, USA). Equal amounts of protein were separated on SDS-PAGE and transferred to nitrocellular membranes. The membranes were then blocked with 5% (w/v) nonfat milk, washed with Tris-buffered saline–Tween-20 (TBST), and incubated with primary rabbit anti-DCK antibodies, rabbit anti-AK antibodies (Abcam, Cambridge, MA, USA), mouse anti-Hif-1α antibodies (BD Biosciences, San Jose, CA, USA), mouse anti-α-actin antibodies (Sigma-Aldrich), or rabbit anti-SPC antibodies (Millipore, Billerica, MA, USA) overnight at 4°C. Membranes were then rinsed, incubated with corresponding secondary antibodies conjugated to horseradish peroxidase (Jackson ImmunoResearch, West Grove, PA, USA) for 1 h at room temperature, and developed with Pierce ECL Western blotting substrate (Thermo Fisher Scientific).

Primary AEC type II (AEC II) cell isolation

Primary AEC II cells were isolated from 10-wk-old female C57BL/6 mice, as described previously (15). Briefly, mouse lungs were isolated and perfused with 0.9% NaCl through the right ventricle of the heart. The lungs were injected with 1 to 2 ml Dispase (BD Biosciences, San Jose, CA, USA), filled with 0.45 ml 1% low-melting gel (prewarmed at 45°C), and immediately covered with ice. Next, the lungs were immersed in 2 ml of Dispase at room temperature for 45 min and then diced into 1-mm cubes and filtered through 100-, 40-, and 25-μm nylon mesh to separate cells. The filtered cells were incubated in an IgG plate for 45 min at 37°C to remove macrophages. To remove lymphocytes, cells were incubated with biotinylated anti-CD45 (1.5 μg/10⁶ cells) and biotinylated anti-CD32 (0.65 μg/10⁶ cells) antibodies (BD Biosciences) for 30 min at 37°C. The biotinylated antibodies were then removed by magnetic particles using the magnetic tube separator (Promega, Madison, WI, USA), and cells were centrifuged and resuspended in culture medium (DMEM containing 10% FBS and antibiotics). Finally, cell numbers were counted, and viabilities were determined using trypan blue staining.

Cell culture, hypoxia exposure, and transient transfection

MLE12 cells were cultured in RPMI 1640 medium supplemented with 10% antibiotics at 37°C in a humidified 5% carbon dioxide atmosphere. For hypoxia exposure, cells were placed in a sealed hypoxia chamber and incubated with 2% oxygen supplied with 95% N₂ and 5% CO₂ for a variety of time periods up to 24 h.

For siRNA studies, MLE12 cells were transfected with DCK siRNA or a scrambled control siRNA (Sigma-Aldrich) at a final concentration of 100 nM using the RNAimax (Invitrogen, Carlsbad, CA, USA), as per the manufacturer's instructions. At 48 h after transfection, the cells were incubated with 2% oxygen or normal air for 6 h and then incubated with 25 μM 2′-chloridedeoxyadenosine (2′-cdA) for an additional 24 h. Finally, cells were harvested for further analysis.

Hematoxylin and eosin (H&E) staining and immunohistochemistry

Mouse lungs were collected and fixed in 10% formaldehyde, while human lungs were collected and fixed in 4% formaldehyde, for ≥24 h. Lungs were then dehydrated, paraffin embedded, and sectioned (5 μm). Sections were rehydrated and stained with H&E (Sigma-Aldrich), according to manufacturer's instructions. For DCK immunostaining, sections were quenched with 3% hydrogen peroxide, incubated in citric buffer (Vector Laboratories, Burlingame, CA, USA) for antigen retrieval, and blocked with an avidin/biotin blocking system (Vector Laboratories). Slides were then blocked with 5% normal goat serum and incubated with primary antibodies for DCK (1:250, rabbit polyclonal, 4°C overnight; Abcam). To detect the hyperoxia in situ, mice were injected with hypoxyprobe (60 mg/kg body wt; HPI, Burlington, MA, USA) 90 min before euthanasia. Lungs were then collected, sectioned, and stained with rabbit anti-hypoxyprobe antibody (1:500; HPI) at 4°C overnight. Slides were incubated with appropriate secondary antibodies (1:1000; Vector Laboratories) for 1 h, and ABC Elite streptavidin reagents (Vector Laboratories) for 30 min at room temperature. Finally, slides were developed with 3,3-diaminobenzidine (Sigma-Aldrich) and counterstained with methyl green.

For TUNEL assays, sections were stained with the In Situ Cell Death Detection Kit (Roche, Indianapolis, IN, USA) as instructed. Briefly, sections were dewaxed, rehydrated, treated with 1% hydrogen peroxide in methanol, and then incubated with TUNEL reaction mixture for 1 h at 37°C. Finally, slides were mounted with Vectashield antifade medium with DAPI (Vector Laboratories).

Apoptosis assay

MLE12 and A549 cells were seeded in 96-well plates, grown to 70% confluence, and then preincubated with normal air or 2% oxygen for 6 h. Cells were then treated with different doses of 2′-cdA with or without 50 mM 2′-chloridedeoxycytidine (2′-cdyC). At 24 h after treatment, 100 μl Caspase-Glo 3/7 assay reagents (Promega, Madison, WI, USA) were added to each well, and luminescence intensity was measured after 40 min.

To determine the caspase-3/7 activities in mouse lungs, the lungs were homogenized using a Dounce homogenizer (Sartorius, Goettingen, Germany) in a hypotonic extraction buffer (25 mM HEPES, pH 7.5; 5 mM MgCl₂; and 1× protease inhibitor cocktail). The lung lysate was cleared by centrifugation at 4°C, and diluted to 400 μg/ml protein concentrations. To perform caspase assays, 50 μl Caspase-Glo 3/7 assay reagents were mixed with 50 μl of diluted cell extract in each well of a 96-well plate. After 40 min incubation at room temperature, the luminescence intensities were measured, and data were analyzed.

Arterial oxygen saturation measurement

Arterial oxygen saturation was measured using MouseOx oximeter (Starr Life Sciences Corp., Oakmont, PA, USA) as instructed. The necks of the mice were shaved and a CollarClip sensor (Starr Life Sciences) was used to continually monitor oxygen saturation.

Intranasal administration of 2′-cdA

P42 Ada^−/− or Ada⁺ mice were anesthetized by inhalation of isoflurane. After the mice were deeply anesthetized, they were gradually injected intranasally with 20 μl sterile PBS containing 0.5 mg 2′-cdA or 0.5 mg 2′-cdA plus 1 mg 2′-cdyC. Mice injected with PBS alone were used as negative controls. The rate of administration was carefully adjusted to allow the mice to inhale the liquid without forming bubbles. After administration, mice were held upright for additional minutes until their breathing returned to normal. After 4.5 h, mice were euthanized; the left lungs were collected for TUNEL staining, and the right lungs were collected for apoptosis assay.

Quantitative RT-PCR

Total RNA was isolated from MLE12 cells or frozen human patients with COPD using TRIzol reagent (Invitrogen), treated with DNase, and reverse-transcripted into cDNA using Superscript II reverse transcriptase (Invitrogen). Equal amounts of cDNA were subjected to quantitative real-time RT-PCR. 18S rRNA was used as internal control. The primer sequences used in this experiment included human DCK: forward, 5′-CCATCGAAGGGAACATCGCT-3′, and reverse, 5′-GGTAAAAGACCATCGTTCAGGT-3′; mouse DCK: forward, 5′-TCTCCACGGTCTGCCCAAT-3′, and reverse, 5′-CAGAACCTCTTAGGTGGGGTG-3′; and 18s: forward, 5′-GTAACCCGTTGAACCCCATT, and reverse, 5′-CCATCCAATCGGTAGTAGCG. The data were quantified using the comparative C_t method and presented as mean ratio to 18sRNA.

Human sample collection

The explanted lung tissues from patients with COPD used in this study were provided in a deidentified manner from the Methodist Hospital J. C. Walter Jr. Transplant Center. Methods for the collection and distribution of these tissues for research have been approved by the Methodist Hospital Institutional Review Board. Lung tissues from stage 0 lobes of the same patients with COPD or from patients without COPD were used as controls. Clinical readouts, including pulmonary function tests (PFTs), were performed at the time the patients were listed for transplantation as a part of a routine evaluation. The data were deidentified prior to use in the current study.

RESULTS

dAdo and dATP levels are elevated in the lungs of Ada^−/− mice

ADA is an enzyme that catalyzes the deamination of Ado and dAdo, and down-regulation or absence of this enzyme results in the accumulation of these two nucleosides (9). dAdo is normally released from DNA breakdown of dead cells. Ada^−/− mice exhibit increased inflammation that causes abnormal cell death and may contribute to the release of dAdo. To determine whether dAdo levels increased in the lungs of Ada^−/− mice, we collected BAL from the lungs of P42 Ada⁺ or Ada^−/− mice and measured the levels of dAdo in BAL using HPLC. This time point was chosen on the basis of previous studies that have shown air-space enlargement to be prominent at this stage (15). As shown in Fig. 1A, dAdo levels were not detectable in the BAL fluid of Ada⁺ mice, but increased dramatically in the BAL fluid of Ada^−/− mice. The levels of an unrelated nucleoside, adenosine monophosphate (AMP), were not affected (Fig. 1A). dAdo can be phosphorylated into dATP, a cytotoxic purine, which may result in apoptosis. To analyze whether dATP levels were increased, we collected lungs from Ada⁺ and Ada ^−/− mice, extracted nucleotides, and measured dATP levels by HPLC. dATP levels were found to be elevated >2-fold in the lungs of Ada^−/− mice compared to Ada⁺ mice (Fig. 1B). These findings suggest that dAdo and dATP levels are elevated in the lungs of mice with air-space enlargement. However, which cells have high dATP accumulation is still not known.

Figure 1.

Detection of the elevation of dAdo and dATP in Ada^−/− mice. dAdo and dATP levels were quantitated on P42. A) dAdo levels in BAL were measured using HPLC. Left panel: representative HPLC chromatograph for dAdo in Ada⁺ and Ada^−/− mice. Middle panel: dAdo concentration was quantitated. Right panel: concentration of AMP was also quantitated as a control nucleoside. B) dATP levels in the whole-lung lysates were measured using HPLC and quantitated (n = 4). *P < 0.05 vs. Ada⁺.

dAdo-DCK pathway is activated in the lungs of Ada^−/− mice

The phosphorylation of dAdo to dATP requires two key enzymes: DCK and AK (12, 16). To determine whether these pathways were activated in the lungs of Ada^−/− mice, the expression of DCK and AK was examined by Western blot and immunostaining on P42, a stage when air-space enlargement is prominent (5). DCK protein expression was significantly increased in the lungs of Ada^−/− mice (Fig. 2A), where increased expression was localized to airway epithelial cells (Fig. 2B, top right panel, arrows) and lymphocytes (Fig. 2B, top right panel, arrows). AK expression was decreased (Fig. 2A) and was also localized to airway epithelial cells (Fig. 2B, bottom right panel, arrows). To further demonstrate whether AEC II cells overexpress DCK, we colocalized DCK with surfactant protein C (SP-C), a marker of AEC II cells. As observed in Fig. 2C, some AEC II cells have DCK overexpression (arrows), but DCK overexpression is not limited to AEC II cells. Since DCK is the rate-limiting enzyme for dAdo phosphorylation, cells with high DCK expression may have high dATP accumulation and, thus, be more sensitive to apoptosis.

Figure 2.

AK and DCK expression in Ada^−/− mice. Age-matched lungs were collected from P42 Ada⁺ and Ada^−/− mice and subjected to Western blot and immunostaining analysis. A) Western blots were used to check the expression of AK and DCK in the lung lysates (left panel). Densitometry of the Western blot bands was quantitated and normalized to β-actin (right panel). *P < 0.05 vs. Ada⁺. B) Immunostaining showed the cellular localization of AK (bottom panels, arrows) and DCK (top panels, arrows) in the lungs of Ada⁺ (left panels) and Ada^−/− (right panels) mice. C) Lung sections of Ada^−/− mice were coimmunostained using antibodies to DCK (red, arrows) and surfactant protein C (SP-C, green, arrows). Scale bars = 100 μM (B); 50 μM (C).

Ada^−/− mice exhibit an increase in apoptosis

Apoptosis of epithelial cells is an important mechanism that leads to air-space enlargement in COPD (4). Previously, we have shown that Ada^−/− mice develop air-space enlargement (9, 17). However, whether apoptosis of epithelial cells plays a role in this model is not known. To determine the levels of apoptosis in the lungs of Ada^−/− mice, we carried out H&E staining to examine the morphology of the lung and TUNEL staining to detect apoptotic cells. The lungs of Ada^−/− mice displayed air-space enlargement, accompanied by infiltration of various inflammatory cells (Fig. 3A), and a marked increase in apoptosis in the alveolar epithelium (Fig. 3A, B). The elevation of apoptosis in the lungs of Ada^−/− mice could also be demonstrated by the increased activity of caspase-3/7 (Fig. 3C). Increased apoptosis was also observed in AEC II cells, as viewed by coimmunostaining for TUNEL and SP-C (Fig. 3D). These results demonstrated that Ada^−/− mice exhibit an increase in apoptosis.

Figure 3.

Increased apoptosis in the lungs of Ada^−/− mice. A) Lungs were isolated for Ada⁺ and Ada ^−/− mice on P42 and subjected to H&E and TUNEL staining. B) TUNEL-positive cells were quantified (n=5). *P < 0.001 vs. Ada⁺. C) Caspase-3/7 activity was measured using the whole-lung lysate (n=5). *P < 0.01 vs. Ada⁺. D) Lung sections of Ada^−/− mice were coimmunostained using antibodies to SP-C (red) and TUNEL (green). Scale bars = 400 μM (A); 50 μM (D).

Hypoxia is present in the lungs of Ada^−/− mice and directly regulates DCK expression

Individuals with COPD are characterized with airway obstruction, shortness of breath, and inefficient air exchange, which may lead to hypoxia in the lungs in association with activation of hypoxia-inducible factor 1α (Hif-1α) expression (18, 19). Interestingly, our microarray data showed a significant up-regulation of Hif-1α expression in the lungs of Ada^−/− mice (data not shown), suggesting that this pathway might be activated in this model. To determine whether Ada^−/− mice exhibit hypoxia in the lung, we measured the arterial oxygen levels in Ada⁺ and Ada^−/− mice. The saturation of peripheral oxygen (S_PO₂) levels were significantly decreased in Ada^−/− mice compared to controls (Fig. 4A). In association with this, Hif-1α protein levels were elevated in the lungs of Ada^−/− mice (Fig. 4B). To visualize cell types associated with hypoxia, hypoxyprobe-1 was used to track hypoxic cells in the lung. As shown in Fig. 4C, there was an increase in hypoxia-positive cells in the lungs of Ada^−/− mice, and the cells were largely airway epithelial cells found in regions of tissue remodeling (Fig. 4C, arrows).

Figure 4.

Ada^−/− mice exhibit hypoxia and increase in Hif-1α in the lung. Lungs were isolated from P42 Ada⁺ and Ada^−/− mice. A) Arterial oxygen levels were measured (n=10). *P < 0.05 vs. Ada⁺. B) Western blot was carried out to detect Hif-1α protein levels. α-Tubulin was used as an internal loading control. C) Lung sections were stained with hydroxyprobe-1 antibody to detect regions of hypoxia (arrows). D) Immunohistochemistry was carried out on two adjacent sections from the lungs of Ada^−/− mice to visualize DCK and hypoxyprobe localization. Arrows denote hyperplastic airway epithelial cells. E) Primary AEC II cells were isolated, cultured for 3 d, and exposed to normal air (Nor) or 2% oxygen (Hypo) for 24 h. Western blots were carried out to determine SP-C and DCK expression. F) MLE12 cells were exposed to normal air or hypoxia for 6, 12, or 24 h, and real-time PCR was used to examine DCK transcript levels. β-Actin was used as an internal control (n=3). *P < 0.05 vs. corresponding control. Scale bars = 100 μM.

Next, we observed that DCK and hydroxyprobe-1 staining were colocalized in epithelial cells in the lungs of Ada^−/− mice (Fig. 4D, arrows), suggesting that hypoxia might be a potential regulator of DCK. To determine whether hypoxia directly induced DCK expression, primary AEC II cells were isolated and incubated in 2% oxygen for 24 h. As shown in Fig. 4E, hypoxia significantly increased DCK expression after treatment for 24 h. Figure 4F demonstrates that the transcript levels of DCK are gradually induced by hypoxia exposure, indicating that the regulation of hypoxia by DCK expression is at least at the transcriptional level. Taken together, these studies indicate that hypoxia is present in the lungs of Ada^−/− mice and directly regulates DCK expression.

Hypoxia promotes 2′-cdA-induced apoptosis by up-regulating DCK in airway epithelial cells

The role of dAdo in lymphocyte apoptosis has been well documented (9); however, its contribution to the progression of COPD by inducing epithelial cell apoptosis is not known. To determine whether dAdo induces apoptosis in cultured lung epithelial cells and whether hypoxia promotes this process by up-regulating DCK expression, MLE12 cells were preincubated in normal air or 2% oxygen for 6 h and treated for 24 h with different doses of 2′-cdA (a deaminase-resistant form of dAdo), and caspase-3/7 activity was measured as an indicator of apoptosis. As shown in Fig. 5A, 2′-cdA alone induced MLE12 apoptosis in a dose-dependent manner. Hypoxia exposure alone had no effect on caspase activity, but it significantly increased the sensitivity of MLE12 cells to 2′-cdA (Fig. 5A). Given that hypoxia exposure increased DCK expression (Fig. 5A), our data suggest that increases in DCK might be directly responsible for dAdo-induced apoptosis. Similar results were observed in MLE12 cells treated with 2′-dAdo and dCF (an ADA inhibitor), but with a lower induction of apoptosis (Fig. 5B). Consistent with the increased levels of apoptosis, dATP levels in cells treated with 2′-dAdo and dCF were highly up-regulated, and hypoxia further promoted the accumulation of dATP in cells (Fig. 5C). Next, to analyze whether 2′-cdA-induced apoptosis was directly dependent on the level and activity of DCK, we pretreated MLE12 cells with 2′-dyC, a DCK inhibitor. 2′-dyC completely blocked 2′-cdA-induced apoptosis in MLE12 cells (Fig. 5D). Similar results were observed in A549 cells (Fig. 5E) and primary AEC II cells (Fig. 5F). Whereas DCK expression was inhibited by DCK-specific siRNA in MLE12 cells, 2′-cdA-induced apoptosis was significantly repressed, even in cells preexposed to hypoxia (Fig. 5G). Taken together, these results demonstrate that hypoxia induces airway epithelial cell apoptosis through a pathway that involves up-regulation of DCK and elevation of dATP. These findings suggest that the activation of the dAdo-DCK-dATP pathway represents a novel mechanism for the up-regulation of apoptosis seen in an animal model of COPD.

Figure 5.

Hypoxia increases the sensitivity of MLE12 cells to dAdo by promoting DCK expression. A, B) MLE12 cells were incubated in normal air or 2% oxygen for 6 h, and then treated with different doses of 2′-cdA (A) or 2′-dAdo with 5 μM dCF (B) for 24 h. DCK protein levels were verified using Western blot (A, inset). Caspase3/7 activities were examined to quantify levels of apoptosis (n=3). *P < 0.05 vs. normoxia with corresponding 2′-cdA or 2′-dAdo doses. C) HPLC was used to measure dATP levels in MLE12 cells treated with normoxia or hypoxia for 6 h, and then 2′-dAdo with dCF for 24 h (n=3). *P < 0.05 vs. control; ^#P < 0.05 vs. dAdo. D–F) MLE12 cells (D), A549 cells (E), or primary AEC II cells (F) were treated with normoxia or 2% oxygen for 6 h, and then incubated with 25 μM 2′-cdA, 50 μM cdyC, or both for 24 h. At the end of treatment, caspase-3/7 activities were examined (n≥3). *P < 0.05. G) MLE12 cells were transfected with control scrambled siRNA (sc-siRNA) or DCK-specific siRNA (DCK-siRNA) for 48 h, preexposed to normal air (Nox) or 2% oxygen (Hyp) for 6 h, and then incubated with 25 μM 2′-cdA for 24 h. Finally, cells were harvested to examine caspase-3/7 activity (top panel) and DCK protein levels (bottom panel). *P < 0.05 vs. sc-siRNA control; ^#P < 0.05 vs. sc-siRNA 2′-cdA; ^&P < 0.05 vs. sc-siRNA hypoxia + 2′-cdA.

Intranasal administration of 2′-cdA directly enhances apoptosis in the lungs

Our results demonstrated that dAdo can induce apoptosis in cultured lung epithelial cells. However, whether it is sufficient to induce apoptosis in the lung is not known. To examine this, we administered 0.5 mg of 2′-cdA into the lungs of P42 Ada⁺ and Ada^−/− mice via intranasal injection. At 4.5 h after injection, a significant increase in the number of apoptotic cells was observed in both Ada⁺ and Ada^−/− mice treated with 2′-cdA compared with controls treated with PBS, and these effects were completely blocked by the DCK inhibitor 2′-dyC (Fig. 6A, B). We also homogenized the lungs and measured caspase-3/7 activity and found that 2′-cdA significantly induced caspase-3/7 activity in both Ada⁺ and Ada^−/− mice, and 2′-dyC significantly suppressed these effects (Fig. 6C). Collectively, these studies suggest that intranasal administration of 2′-cdA is sufficient to enhance apoptosis in the lungs.

Figure 6.

Intranasal administration of 2′-cdA is sufficient to induce apoptosis in the lungs. A) P42 Ada⁺ and Ada^−/− mice were intranasally injected with PBS alone, 0.5 mg 2′-cdA, or 0.5 mg 2′-cdA plus 1 mg 2′-cdyC in 20 μl PBS for 4.5 h. The right lungs were collected for TUNEL staining. Scale bars = 200 μM. B) TUNEL-positive cells in each area were counted and quantitated for each treatment group. *P < 0.05. C) Caspase-3/7 activities were measured using the left lungs collected from the same mice used for TUNEL assay, and data were quantitated as mean luminance of each treatment group. *P < 0.05.

DCK levels are enhanced in patients with COPD

To examine whether this pathway is also activated in patients with COPD, we examined the expression of DCK in the lungs of patients with COPD. As shown in Fig. 7A, DCK protein levels are significantly increased in the lungs of patients with stage 4 COPD compared to patients with stage 0 COPD. Consistent with the results observed in mouse lungs, elevated DCK is located in airway epithelial cells in the lungs of patients with stage 4 COPD (Fig. 7B). To further determine the correlation of DCK levels to the severity of disease, we measured DCK mRNA levels and mapped them to different clinical parameters of COPD using linear regression analysis. DCK mRNA is not correlated with diffusing capacity of the lung for carbon monoxide (DLCO; data not shown). However, a significant negative correlation between DCK mRNA and forced expiratory volume in 1 s (FEV1) was observed (Fig. 7C). A significant negative correlation was also observed between DCK mRNA and the ratio of FEV1 to forced vital capacity (FEV1/FVC; Fig. 7D), as well as between DCK mRNA and forced expiratory flow (FEF; Fig. 7E). These clinical parameters are normally diminished in obstructive diseases due to increased airway resistance to expiratory flow. The significant correlation between DCK level and the severity of disease suggests that DCK may be used as a biomarker for clinical evaluation and prognosis of patients with COPD. Collectively, these studies demonstrate that DCK levels are increased in patients with COPD and correlated with disease severity.

Figure 7.

DCK expression in patients with COPD. Stage 0 and stage 4 COPD lungs were collected from patients and used to examine DCK protein level and cellular localization. A) Western blots were used to determine the expression of DCK in the lung lysates (left panel). Densitometry of the Western blot bands was quantitated and normalized to β-actin (right panel). B) Immunostaining showed DCK localization (arrows) in the hyperplastic airway epithelial cells in patients with stage 4 COPD. Scale bars = 100 μM. C–F) Transcript levels of DCK were measured in patients with COPD and mapped to the levels of forced expiratory volume in 1 s (FEV1; C), forced vital capacity (FEV1/FVC; D), and forced expiratory flow (FEF; E).

DISCUSSION

Increased apoptosis is known to play an important role in the progression of COPD by inducing abnormal tissue damage, stimulating the release of inflammatory factors and initiating pulmonary fibrosis (4, 20). Although several studies have investigated the signaling pathways contributing to abnormal apoptosis during the pathogenesis of COPD, it remains unclear how apoptosis is continuously induced and amplified in lungs with air-space enlargement, and there is no effective way to prevent apoptosis and thus attenuate the progression of COPD. Therefore, it is important to identify new pathways to elucidate the progression of COPD and find new therapeutic targets. In this study, we discovered a novel feedback pathway between apoptosis and dAdo (Fig. 8). Initially, injury to epithelial cells (e.g., cigarette smoke or inflammation) induces epithelial cell apoptosis, resulting in DNA breakdown and increasing extracellular dAdo levels. Increased dAdo leads to cytoplasmic dATP accumulation and subsequent amplification of cell apoptosis. In Ada^−/− mice, or ADA-deficient patients, in which dAdo cannot be deaminated due to the absence of ADA activity (6), dAdo levels are increased and subsequent apoptosis is facilitated. In addition, we also determined for the first time that DCK, the key enzyme that phosphorylates dAdo, is elevated in the lungs of both mice and humans with air-space enlargement, thus potentially accelerating the accumulation of dATP and the amplification of apoptosis. Moreover, we found that DCK is elevated in hypoxic cells, and hypoxia directly induces DCK expression. Finally, by mapping DCK transcript levels with several clinical parameters of COPD, we demonstrated that DCK levels are significantly correlated with disease severity. Collectively, these studies suggest that the dAdo-DCK pathway might be a mechanism involved in the amplification of apoptosis, thus promoting the development and progression of COPD.

Figure 8.

Positive feedback loop between apoptosis and dAdo.

DCK is the key enzyme for the synthesis of all four nucleotides and is the rate-limiting step in the nucleotide salvage pathway (12, 13). The nucleotide salvage pathway is important for the phosphorylation of nucleosides from DNA degradation to nucleotides for DNA synthesis and repair. Therefore, the activation of this pathway has been considered to be a protective cellular response to DNA damaging agents (23). However, prolonged hyperactivation of DCK or overexposure to dAdo and/or its analogues can cause excess DNA damage in association with increased production of cytotoxic nucleoside triphosphate and its derivatives (12, 13). In light of this, dAdo and its analog-induced and DCK-mediated apoptosis have been studied in lymphocytes, whose high DCK activity renders them sensitive to dAdo-induced apoptosis. Moreover, certain cancer cells exhibit increased DCK expression (21–23), which facilitates the use of DCK substrate analogues as efficient chemotherapeutic agents (24–27). Similar to Ada^−/− mice, DCK-deficient mice are immunodeficient, characterized by abnormal T- and B-lymphocyte development. This impaired T- and B-lymphocyte maturation may be caused by a defect in proliferation or abnormal apoptosis during lymphocyte development, demonstrating an important role for DCK in regulating cell apoptosis. Lungs from patients and animals with air-space enlargement have dramatically increased apoptosis, which could be a source of dAdo. However, dAdo contribution to the apoptosis in COPD has not been reported until now. In this study, we characterized pathways of dAdo and demonstrated that DCK levels are elevated in airway epithelial cells of Ada^−/− mice and in patients with COPD. We also demonstrated that DCK levels strongly correlate with several disease parameters of COPD. In the presence of accumulated dAdo in COPD lungs, elevated DCK may accelerate the production of dATP and contribute to the amplification of apoptosis and the development of COPD.

In recent years, apoptosis has been listed as one of the major mechanisms of COPD (4). The effect of apoptosis on COPD development is supported by the evidence that blockage of apoptosis prevents the development of emphysema (28). Moreover, several studies successfully developed animal models of COPD by directly inducing apoptosis of alveolar cells (29) or suppressing VEGF/VEGFR signaling (28, 30, 31) without causing inflammatory cell infiltration, demonstrating that apoptosis of alveolar epithelial or endothelial cells is sufficient to cause pulmonary emphysema in animal models. Although it has not been reported whether dATP accumulates and induces apoptosis in the lungs of patients with COPD, it is well understood that elevated dATP is cytotoxic to the cells (13, 32). In response to increased dATP, mitochondria release cytochrome c together with apoptotic protease activating factor-1 (Apaf-1) and procaspase-9, which form the apoptosome complex and activate the apoptosis pathway (9). This is the underlining mechanism of apoptosis induced by dAdo analogs in lymphocytes and DCK-overexpressing cancer cells. In the lungs of Ada^−/− mice, we observed increased dAdo levels and elevated expression of DCK in airway epithelial cells, indicating that dAdo may induce apoptosis in these structural cells expressing high levels of DCK. We also detected a significant elevation of dATP in combination with a significant increase in the number of apoptotic cells in the lungs of Ada^−/− mice. Taken together, our data suggest that repeated exposure to dAdo may be one of the major mechanisms that damage airway epithelial cells, in turn, releasing more dAdo and forming a positive feed-forward loop to amplify apoptosis of the lung.

To further confirm that dAdo was able to induce airway epithelial apoptosis, we carried out in vitro studies and demonstrated that treatment of 2′-cdA, an ADA-resistant form of dAdo, induces apoptosis in airway epithelial cells, including human A549, mouse MLE12, and mouse primary AEC II cells, suggesting that this pathway is important to both systems. The effect of 2′-cdA was completely blocked by adding 2′-dyC, an inhibitor of DCK. This revealed that accumulated dAdo in ADA-deficient mice and humans may act as a death factor that continuously amplifies epithelial cell apoptosis and accelerates disease progression. Thus, blockage of the dAdo-DCK pathway may be a novel beneficial therapy to treat COPD.

The regulation of DCK expression is not well studied. Identifying upstream factors of DCK will have essential clinical significance based on the important role of DCK in cancer chemotherapy and in regulating cell apoptosis in our model of COPD. DNA-damaging agents, including the DNA polymerase inhibitor aphidicolin, topoisomerase II inhibitor etoposide, γ-irradiation, and apoptosis inducer p53, are able to increase DCK activity (13, 33–35). However, none of these agents affect DCK protein or mRNA levels. To explore the mechanisms that induce DCK expression in COPD, we screened several agents, including those that induce apoptosis, such as hydrogen peroxide and 2′-cdA, or those that promote epithelial-mesenchymal transition, including TGF-β and hypoxia. The only agent that affected DCK expression was hypoxia. We found that hypoxia promoted DCK expression in mouse primary AEC II and MLE12 cells at both the mRNA and protein levels. We also observed colocalization of DCK and hypoxyprobe in airway epithelial cells, which further verified in vivo that hypoxia is a novel regulator of DCK. Furthermore, preexposing cells to hypoxia significantly increased 2′-cdA and dAdo-induced apoptosis. This demonstrates that DCK levels are important in determining the sensitivity of cells to dAdo, and hypoxia-induced DCK expression is directly responsible for the increased apoptosis induced by 2′-cdA or dAdo. Although we identified hypoxia as a novel regulator of DCK, the exact mechanism by which this regulation occurs is still not clear. Increased DCK transcript levels in response to hypoxia suggest that hypoxia regulates DCK expression at the transcriptional level. Transcription factors SP1 and USF have been reported to bind to the DCK promoter and promote DCK transcription in tumor cells (1). However, whether hypoxia-induced DCK expression is mediated by SP1 or USF is still not known. Hypoxia-induced DCK expression may also be mediated by other hypoxia-inducible transcription factors, including Hif-1a, p53, and NF-κB, all of which have binding sites at the DCK promoter, as predicted by PROMO 3.0. At the same time, hypoxia may also induce DCK expression by triggering posttranscriptional regulation pathways through RNA binding proteins. Further studies are needed to determine these underlying mechanisms.

Increased hypoxia may be present in the lungs of patients with COPD due to excessive airway remodeling and damage to the blood-air barrier, which normally results in insufficient air exchange and low oxygen levels in the blood (36, 37). Ada^−/− mice exhibit features of chronic lung diseases, including inflammation, emphysema, and fibrosis (38), and show a significant decrease in arterial oxygenation, suggesting the presence of hypoxia in these mice. In response to hypoxia, several inflammatory factors are activated, including nuclear factor-κB (NF-κB), a transcription factor that regulates multiple inflammatory and immune responses, tumor necrosis factor α (TNF-α), a proinflammatory cytokine, and Toll-like receptors (TLRs), which promote phagocytosis and immune responses (39). Hypoxia also induces anti-inflammatory signaling molecules by stabilizing Hif transcription factors. Hif-1α is a hypoxia-inducible factor responsible for the rapid expression of >60 genes in response to low oxygen levels, including genes involved in cell survival, adaptation, epithelial-mesenchymal transition, immune reaction, cytokine production, vascularization, and tissue homeostasis (40–42). Several studies have demonstrated that Hif-1α stimulates the generation and accumulation of the anti-inflammatory molecule Ado by directly inducing the expression of CD73 (36), an enzyme required for generation of extracellular Ado, or inhibiting the expression of equilibrative nucleoside transporters (ENTs), which uptake Ado and terminate Ado signaling (38). Hif also induces the expression of Ado receptors, thus further promoting Ado signaling (39, 43). Despite some studies demonstrating that Hif-1α transcript levels were decreased in severe COPD (43, 44), we observed a significant accumulation of Hif-1α in the lungs of Ada^−/− mice both at the protein and mRNA level. In addition, using hypoxyprobe, which can be reductively activated by low oxygen and form adducts that can be recognized by antibodies, we detected a significant increase in the number of hypoxic cells in the lungs of Ada^−/− mice, further confirming in vivo that hypoxia is present in the lungs of Ada^−/− mice. Hypoxia may have a direct effect on increased inflammation in these mice, and we have demonstrated a novel role of hypoxia on cell apoptosis by directly linking hypoxia to the dAdo-DCK pathway. Both the colocalization of hypoxia and DCK in the same group of cells in COPD and the regulation of DCK by hypoxia suggest that hypoxia may be directly responsible for the increased expression of DCK in COPD, as well as facilitate dAdo-induced apoptosis.

Apoptosis is important for the development and progression of COPD. Our study proposes a role for dAdo in the apoptosis of airway epithelial cells other than lymphocytes. In addition, we investigated the mechanisms of dAdo-DCK-dATP pathway activation and demonstrated the contribution of this pathway to epithelial cell apoptosis during disease progression using both in vitro and in vivo approaches. Further studies are needed to establish whether blocking components of the dAdo-DCK-dATP pathway will prevent or attenuate the development of COPD. However, our proof of concept experiments demonstrating the up-regulation of DCK in the airway epithelial cells of patients with COPD and the association of these increases with disease severity, suggest our findings are clinically important. Overall, our study may provide evidence for the development of novel approaches to treat COPD, as well as contribute to the treatment of other diseases associated with abnormal apoptosis, such as cancer and various inflammatory disorders (45, 46).