Hypercapnia Inhibits Autophagy and Bacterial Killing in Human Macrophages by Increasing Expression of Bcl-2 and Bcl-xL

Abstract

Hypercapnia, the elevation of CO₂ in blood and tissue, commonly develops in patients with advanced lung disease and severe pulmonary infections, and is associated with high mortality. We previously reported that hypercapnia alters expression of host defense genes, inhibits phagocytosis, and increases the mortality of Pseudomonas pneumonia in mice. However, the effect of hypercapnia on autophagy, a conserved process by which cells sequester and degrade proteins and damaged organelles that also plays a key role in antimicrobial host defense and pathogen clearance, has not previously been examined.

In the present study we show that hypercapnia inhibits autophagy induced by starvation, rapamycin, LPS, heat-killed and live bacteria in the human macrophage. Inhibition of autophagy by elevated CO₂ was not attributable to acidosis. Hypercapnia also reduced macrophage killing of Pseudomonas aeruginosa. Moreover, elevated CO₂ induced the expression of Bcl-2 and Bcl-xL, anti-apoptotic factors that negatively regulate autophagy by blocking Beclin 1, an essential component of the autophagy initiation complex. Furthermore, siRNA targeting Bcl-2 and Bcl-xL and the small molecule Z36, which blocks Bcl-2 and Bcl-xL binding to Beclin 1, prevented hypercapnic inhibition of autophagy and bacterial killing. These results suggest that targeting the Bcl-2/Bcl-xL-Beclin 1 interaction may hold promise for ameliorating hypercapnia-induced immunosuppression and improving resistance to infection in patients with advanced lung disease and hypercapnia.

INTRODUCTION

Hypercapnia, elevation of CO₂ in blood and tissue, commonly develops over time in patients with advanced chronic lung disorders, such as chronic obstructive pulmonary disease (COPD), cystic fibrosis and neuromuscular syndromes with respiratory muscle weakness (1–4). Such individuals are also at risk for acute respiratory decompensations accompanied by marked acute or acute-on-chronic hypercapnia. Additionally, patients with respiratory failure due to the acute respiratory distress syndrome (ARDS) and status asthmaticus are often hypercapnic during the course of their acute illness (5).

Hypercapnia has long been recognized as a risk factor for increased morbidity and mortality in patients with acute exacerbations of COPD (1, 2, 6, 7). Of note, COPD exacerbations are most commonly triggered by bacterial or viral respiratory infections (8–10). Hypercapnia is also an independent risk factor for mortality in hospitalized patients with community-acquired pneumonia (11, 12), children with lower respiratory tract adenovirus infection (13), and cystic fibrosis patients awaiting lung transplantation (3). The association between hypercapnia and mortality in patients with respiratory infections and acute or chronic lung disease suggests the possibility that elevations in CO₂ may play a causal role in poor clinical outcomes by adversely affecting pulmonary host defense.

Consistent with this possibility, we (14) and others (15, 16) have shown that elevated levels of CO₂ inhibit macrophage expression of TNF and IL-6, cytokines that play critical roles in antibacterial host defense (17–19). We showed that the inhibitory effect of elevated CO₂ on macrophage cytokine synthesis was concentration-dependent, reversible and not due to extracellular or intracellular acidosis (14). We also found that elevated CO₂ inhibits phagocytosis of bacteria by macrophages in vitro (14). Moreover, we have shown that hypercapnia increases the mortality of Pseudomonas pneumonia in mice, also in an acidosis-independent manner (20). In the latter study, hypercapnia inhibited bacterial phagocytosis and reactive oxygen species (ROS) generation, and decreased pulmonary clearance of Pseudomonas. Thus, elevated CO₂ inhibits multiple phagocyte antimicrobial functions in vitro and in vivo, and it increases mortality in a clinically relevant model of bacterial pneumonia in mice.

Autophagy, a conserved eukaryotic stress-response pathway in which cells sequester damaged or surplus proteins and organelles in double-membrane vesicles and deliver them to lysosomes for degradation, also plays a seminal role in antimicrobial host defense. Classic triggers of autophagy include nutrient deprivation and inhibitors of the metabolic regulator mTOR, such as rapamycin (21). Autophagy is also strongly induced by LPS and other TLR ligands, as well as by bacteria, viruses and fungal organisms that ultimately are enclosed within autophagosomes and digested within the autophagolysosomes (21–23).

Regardless of the initial stimulus, autophagy is orchestrated by autophagy-related proteins including Beclin 1, which binds the class III PI3 kinase Vps34 and other components to initiate formation of autophagosomes, which mature by incorporating additional autophagy-related proteins, including the lipidated form of the microtubule-associated protein light chain 3 (LC3 II) (24–27). The importance of autophagy in mammalian host defense has been demonstrated by studies showing that expression of Beclin 1 and other autophagy genes is essential for protection of mice against infection with various bacterial and viral pathogens (23).

Because autophagy plays such an important role in antimicrobial host defense, in the current study we explore the effects of elevated CO₂ on autophagy in the macrophage. We show that hypercapnia inhibits autophagy induced by multiple stimuli, including LPS and bacteria, and that the inhibitory effect of elevated CO₂ is unrelated to acidosis. We also find that hypercapnia increases expression of Bcl-2 and Bcl-xL, anti-apoptotic factors that inhibit autophagy by binding Beclin 1. Moreover, reducing expression of Bcl-2 and Bcl-xL or blocking their ability to bind Beclin 1 prevents inhibition of autophagy and defective macrophage bacterial killing caused by elevated CO₂. These previously unrecognized inhibitory effects of elevated CO₂ on regulation of autophagy may in part account for impaired resistance to infection and increased mortality in patients with severe lung disease and hypercapnia.

MATERIALS AND METHODS

Materials

All materials were purchased from Sigma Chemical unless otherwise specified.

Cells

Human monocytic leukemia THP-1 cells (ATCC) were cultured in RPMI 1640, supplemented with 10% FBS, 2 mM l-glutamine, 1 mM sodium pyruvate, 20 μM β-mercaptoethanol, 100 U/ml penicillin, and 100 μg/ml streptomycin and differentiated to a macrophage phenotype by exposure to 5 nM PMA for 48 h (14). Human alveolar macrophages were obtained by bronchoalveolar lavage from the contralateral lung of subjects undergoing bronchoscopy for clinical diagnosis of noninfectious focal lung lesions (14, 28) under a protocol approved by the Northwestern University Institutional Review Board. Alveolar macrophages, purified to ≥98% by adherence to plastic and removal of nonadherent cells, were cultured in RPMI 1640 as above (14) and rested for 24 h to avoid the transient pro-inflammatory profile of freshly isolated macrophages (29). HeLa cells stably expressing GFP-LC3 (30), generously provided by C. He, were cultured in DMEM with 10% FBS containing 100 U/ml penicillin, 100 μg/ml streptomycin, and 10 μg/ml G418.

Exposure of cells to normocapnia and hypercapnia

Normocapnia consisted of standard incubator atmosphere: humidified 5% CO₂ (PCO₂ 36 mmHg)/95% air, at 37°C. Hypercapnia consisted of 15% CO₂ (PCO₂ 108 mmHg)/21% O₂/69% N₂. In selected experiments, Tris base was added to RPMI 1640 and HBSS so that pH was maintained at 7.4 in 15% CO₂ or lowered to 7.2 in 5% CO₂. Cells were exposed to hypercapnia in an environmental chamber (C-174, Biospherix) contained within the same incubator where control cultures were simultaneously exposed to normocapnia. In all cases, cells were exposed to hypercapnia or maintained in normocapnia as control for 18 h prior to stimulation of autophagy. The pH of the culture media was measured with a pHOx Plus blood gas analyzer (Nova Biomedical). All media were pre-saturated with 5% or 15% CO₂ before addition to the cells.

Stimulation of autophagy

After incubation in normocapnia or hypercapnia for 18 h, autophagy was induced in PMA differentiated THP-1 macrophages, human alveolar macrophages, or GFP-LC3 HeLa cells by amino acid starvation or exposure to rapamycin, LPS, bacterial particles or live bacteria. Cells were maintained in normocapnia or hypercapnia during exposure to autophagic stimuli. Starvation was achieved by replacing culture media with serum-free HBSS for 1 h. Rapamycin or Ultra-Pure E. coli K12 LPS (InvivoGen) were added to cells at final concentrations of 25 μM and 10 ng/ml, respectively, for 18 h. In addition, cells were exposed during 4 h to 0.1 μg/ml pHrodo-E. coli or Alexa 488-S. aureus BioParticles (both from Life Technologies) or live P. aeruginosa strain PAO1 (MOI: 1:10) prepared as described (31).

Autophagy assays

To determine early autophagy events and autophagic activity, cells were immunostained with ATG12 (32) or LC3 II antibodies (Cell Signaling), respectively, and ATG12 and LC3 II puncta formation was imaged with an Axioplan 2 microscope (Zeiss). DAPI (1 ng/ml, Life Technologies) staining was used to visualize nuclei. Formation of GFP-positive LC3 puncta in GFP-LC3 HeLa cells was assessed by fluorescence microscopy. Formation of ATG12 and GFP-LC3 puncta was quantified using Image J as the intensity of the fluorescence signal associated with puncta minus background cytoplasmic fluorescence associated with dispersed ATG12 or GFP-LC3, normalized for each experimental condition to the normocapnia control. For each condition, at least three optical fields with at least ≥ 30 cells per experimental condition were analyzed in three independent experiments. Conversion of endogenous LC3 I to LC3 II was determined by immunoblot of whole cell lysates under reducing conditions as described (33), using LC3 II antibody (Cell Signaling). β-actin was also detected by immunoblot (antibody from Abcam) as protein loading control. HRP-conjugated secondary antibodies (Cell Signaling) were used, and chemiluminescence from SuperSignal West Dura substrate (Thermo Fisher Scientific) was detected using the Odyssey Fc imaging system (LI-COR).

Since autophagy is a dynamic process involving autophagosome synthesis, autophagosome fusion with the lysosome, followed by lysosomal degradation of autophagic substrates at the autophagosome, induction of ATG12 and LC3 II puncta formation and LC3 II accumulation may reflect either an increase in autophagy or defective lysosomal degradation of autophagic markers. To differentiate between these alternatives, we assessed autophagic flux in the absence and presence of bafilomycin A (BA, 10 nM), an inhibitor of autophagosome-lysosome fusion (27, 34).

Quantitative real-time PCR

RNA was extracted using RNeasy Mini Kit (Qiagen) and reverse-transcribed to cDNA using iScript cDNA synthesis Kit (Bio-Rad). PCR amplification was performed using CFX Connect^™ Real-Time System (Bio-Rad) and the TaqMan® Gene Expression Assay with FAM^™ labeled probes (Applied Biosystems). The following primer/probe sets were utilized: Bcl-2 (Hs00608023_m1), Bcl-xL (Hs00236329_m1), and Beclin-1 (Hs00186838-m1). Samples were normalized using the housekeeping gene GAPDH (Hs99999905_m1). Relative expression was calculated by the comparative CT method (ΔΔCT) (35).

Bcl-2 and Bcl-xL immunoblotting and immunocytochemistry

THP-1 macrophages lysates were immunoblotted using mouse anti-Bcl-2 (Abcam) and rabbit anti-Bcl-xL (Cell Signaling), followed by appropriate HRP-secondary antibodies. Chemiluminescence was detected as above. In addition, THP-1 macrophages were fixed and immunostained with anti-Bcl-2 or anti-Bcl-xL antibodies, followed by Alexa Fluor 488 donkey anti-mouse or Alexa Fluor 555 donkey anti-rabbit (Life Technologies), respectively. Nuclei were stained with DAPI. Cells were imaged using fluorescence microscopy, and fluorescence intensity was quantified using NIH ImageJ software. These data are presented as corrected total cell fluorescence (CTCF), the integrated density after subtraction of background fluorescence.

Bcl-2 and Bcl-xL co-immunoprecipitation with Beclin 1

THP-1 macrophages were lysed with a nonionic detergent (Nonidet P-40) to preserve protein-protein binding (36) and incubated with either agarose-conjugated Bcl-2 antibody (N-19, Santa Cruz Biotechnology), rabbit polyclonal anti-Bcl-xL antibody plus Dynabeads (Life Technologies), or nonimmune rabbit IgG. Immunoprecipitates were immunoblotted using rabbit anti-Beclin 1 antibody conjugated with HRP (Novus Biologicals), and chemiluminescence was assessed as indicated above. Beclin-1 was not detectable in samples immunoprecipitated with rabbit IgG (results not shown).

siRNA transfection

THP-1 macrophages were transfected with 25 pmol ON-TARGETplus SMARTpool Bcl-2 siRNA, Bcl-xL siRNA, or nontargeting (NT) negative-control siRNA (Thermo Fisher Scientific) using Lipofectamine® RNAiMAX transfection reagent (Life Technologies) following the manufacturer’s instructions. Knockdown efficiency was measured by qPCR and immunofluorescence. Using this protocol, typical transfection efficiencies were 70 to 80%. Transfected cells were then exposed to normocapnia or hypercapnia overnight prior to stimulation of autophagy.

Tetrazolium dye reduction assay of bacterial killing

Killing of P. aeruginosa by THP-1 macrophages was quantified using a tetrazolium dye reduction assay, as described (37, 38). Briefly, P. aeruginosa was added to THP-1 macrophages (MOI 10:1) in duplicate 96-well plates and incubated for 30 min at 37°C. Next, cells were washed and placed at 4°C (T0) or 37°C (T90) for 90 min, lysed with 0.5% saponin in tryptic soy broth, then incubated at 37°C for 2.5 h. MTT (5 mg/ml) was added to each plate for 30 min. Absorbance was read at 595 nm. Results were expressed as surviving bacteria (T90/T0), determined as the ratio between A595 at T90 and T0.

Statistical analysis

Data are presented as means ± SE. Differences between two groups were assessed using Student’s t-test. Differences between multiple groups were assessed by ANOVA followed by the Tukey Kramer honestly significant difference test. Levene’s test was used to analyze the homogeneity of variances. Significance was accepted at p < 0.05.

RESULTS

Elevated CO₂ inhibits autophagy induced by starvation and rapamycin

Starvation is a potent trigger of autophagy, a process that allows the cell to meet its energy needs when exogenous nutrients are scarce by degrading nonessential components for use as fuel (39). Therefore, to study the effects of hypercapnia on autophagy, THP-1 macrophages were exposed to 5% CO₂ (normocapnia) or 15% CO₂ (hypercapnia) and subjected to amino acid starvation. Under normocapnic conditions, starvation induced formation of ATG12 and LC3 II-positive puncta (Fig 1A–B) and accumulation of LC3 II protein (Fig 1C). These changes were greatly augmented in the presence of bafilomycin A, confirming that in THP-1 macrophages starvation increased autophagic flux, i.e. an increase of autophagic markers by autophagy induction per se, rather than by blockade of degradation at the lysosome. Notably, hypercapnia inhibited starvation-induced autophagy, as indicated by reduced ATG12 and LC3 II puncta formation and LC3 II accumulation, both in the absence or the presence of bafilomycin A (Fig 1A–C). Similarly, hypercapnia blocked starvation-induced autophagy in HeLa cells expressing GFP-LC3 (Fig 1D). To determine whether the inhibitory effect of hypercapnia was unique to starvation-induced autophagy, we stimulated cells with the mTOR inhibitor rapamycin, another well-known autophagy trigger (40, 41). Like starvation, rapamycin induced ATG12 puncta formation and LC3 II protein accumulation in THP-1 macrophages (Fig S1A–B) as well as GFP-LC3 redistribution in HeLa cells (Fig S1C), all of which were further increased by bafilomycin A. And like with starvation, hypercapnia inhibited autophagy induced by rapamycin in both THP-1 macrophages and in HeLa cells (Fig S1A–C), indicating that the inhibitory effect of elevated CO₂ was not specific to a single autophagy stimulus.

Figure 1. Hypercapnia inhibits starvation-induced autophagy independently of extracellular acidosis.

PMA-differentiated THP-1 macrophages and GFP-LC3 expressing HeLa cells were exposed to 5% CO₂ (normocapnia, NC) or 15% CO₂ (hypercapnia, HC) for 18 h, then starved for 1 h by incubation in HBSS in the absence or presence of bafilomycin A (BA, 10 nM) in normocapnia or hypercapnia, respectively. Formation of autophagosomes in THP-1 cells, detected as ATG12-positive (A) and LC3 II-positive (B) puncta, was assessed by immunofluorescence microscopy; nuclei were stained with DAPI (blue). LC3 II accumulation in THP-1 cells was quantified by immunoblot with β-actin as loading control (C). Formation of LC3 puncta (small bright green spots) in GFP-LC3 HeLa cells was assessed by fluorescence microscopy (D). The quantity of fluorescence associated with ATG12-positive and LC3-GFP autophagic puncta, normalized to the normocapnia control, is given in the upper left of each panel in A and D, respectively. To control for pH effects of elevated CO₂, experiments were performed with THP-1 cells using both unbuffered and pH-buffered media, and autophagy was assessed by LC3 II puncta formation (E) and immunoblot (F). Bars represent means ± SE, n≥3, *p<0.01 vs. NC control, **p<0.05 vs. NC starved.

Hypercapnia inhibits autophagy independently of extracellular acidosis

Since CO₂ levels above 5% reduce the pH of normal culture media (14), we performed experiments to distinguish whether inhibition of autophagy resulted from the elevated concentration of CO₂ or from acidosis. To do this, we buffered the media so that pH was maintained at 7.4 in 15% CO₂ or lowered to 7.2 in 5% CO₂ while THP-1 macrophages were starved to trigger autophagy. As shown in Figure 1, hypercapnia inhibited starvation-induced LC3 II puncta formation (Fig 1E) and LC3 II accumulation (Fig 1F) at both pH 7.2 and 7.4. Furthermore, normocapnic acidosis (5% CO₂, pH 7.2) did not cause any reduction in starvation-induced autophagy (Fig 1E–F). Similar results were obtained when autophagy was induced by other triggers, including P. aeruginosa (results not shown). Therefore inhibition of autophagy by hypercapnia is not due to extracellular acidosis, but the result of elevated CO₂ itself.

Hypercapnia impairs autophagy induced by LPS in THP-1 and human alveolar macrophages

Since LPS is another well-known initiator of autophagy (42, 43), we evaluated the effect of hypercapnia on LPS-induced autophagy. Like starvation and rapamycin, LPS effectively triggered ATG-12 and LC3 II puncta formation and LC3 II accumulation in normocapnia-exposed THP-1 macrophages (Fig. 2A–C), as well as GFP-LC3 redistribution in HeLa cells (Fig 2D). Hypercapnia blocked LPS-induced autophagy in both cell types. We also examined the effect of hypercapnia on autophagy in human alveolar macrophages obtained by bronchoalveolar lavage. We found that LPS triggered LC3 II accumulation in normocapnia, and that this was attenuated by hypercapnia (Fig 2E). Thus, in addition to blocking autophagy induced by metabolic stressors (starvation, rapamycin), hypercapnia inhibits autophagy triggered by LPS – a potent inflammatory stimulus – in macrophage and epithelial cell lines and in primary macrophages from the human lung.

Figure 2. Hypercapnia inhibits autophagy induced by LPS.

THP-1 macrophages and GFP-LC3 expressing HeLa cells in 5% CO₂ (NC) or 15% CO₂ (HC) were exposed 18 h to 1 ng/ml LPS with and without bafilomycin A (BA). In THP-1 cells, formation of ATG12 (A) and LC3 II (B) puncta was assessed by immunofluorescence microscopy, and LC3 II accumulation was quantified by immunoblot (C). GFP-LC3 redistribution in HeLa cells was assessed by fluorescence microscopy (D). The quantity of fluorescence associated with ATG12-positive and LC3-GFP autophagic puncta, normalized to the normocapnia control, is given in the upper left of each panel in A and D, respectively. Bars represent means ± SE, n≥3, *p<0.01 vs. NC control, **p<0.05 vs. NC LPS. In addition, human alveolar macrophages (AM) were stimulated with LPS in normocapnia or hypercapnia and LC3 II accumulation was assessed by immunoblot (E).

Hypercapnia inhibits bacteria-induced autophagy

We previously showed that elevated CO₂ decreases phagocytosis of bacteria in vitro and in vivo (14, 20) and further that hypercapnia reduced clearance of bacteria from the lungs and other organs in mice with Pseudomonas pneumonia (20). Because autophagy targets intracellular bacteria for lysosomal degradation (44, 45) and bacteria stimulate autophagy (22, 23), we next examined the effect of hypercapnia on bacteria-induced autophagy. We found that hypercapnia inhibited LC3 II puncta formation and protein accumulation induced by E. coli and S. aureus BioParticles in THP-1 macrophages (Fig 3A–C). Of note, LC3 II co-localized with many fluorescently-labeled intracellular E. coli and S. aureus BioParticles, indicating formation of autophagosomes around the bacterial fragments (Fig 3A). On the other hand, LC3 II was not co-localized with another proportion of the bacterial BioParticles (best seen with S. aureus) suggesting that these internalized bacterial fragments were being processed in phagosomes without engagement of the autophagic machinery. Interestingly, Fig 3A shows that while hypercapnia reduced the number of both LC3 II-positive autophagosomes containing BioParticles (yellow) and non-LC3 II-associated intracellular BioParticles (green), the inhibitory effect of elevated CO₂ on formation of BioParticle-containing LC3 II-positive autophagosomes appeared to be greater than on the uptake of BioParticles not associated with LC3 II. This suggests that hypercapnia may suppress autophagy to an even greater degree than it reduces phagocytosis. In addition, E. coli and S. aureus BioParticles induce LC3 II accumulation in the absence or presence of bafilomycin A in normocapnia-exposed THP-1 macrophages (Fig 3B–C), as well as GFP-LC3 redistribution in HeLa cells (results not shown). Hypercapnia blocked E. coli and S. aureus BioParticles -induced autophagy in both cell types (Fig 3B–C).

Figure 3. Hypercapnia inhibits autophagy triggered by heat-killed and live bacteria.

THP-1 macrophages were exposed to 5% CO₂ (NC) or 15% CO₂ (HC) for 18 h, then incubated with pHrodo- E. coli or Alexa 488-S. aureus BioParticles for 4 h in normocapnia or hypercapnia, respectively. Merged images of the autophagosome marker LC3 II (red) and bacteria (green) were captured by fluorescence microscopy, nuclei were labeled with DAPI (A). LC3 II accumulation induced by E. coli and S. aureus BioParticles in the absence or presence of bafilomycin A (BA) was quantified by immunoblot (B, C). Bars represent means ± SE, n≥3, *p<0.01 vs. NC control, **p<0.05 vs. NC with E. coli or NC with S. aureus. In addition, THP-1 cells in normocapnia or hypercapnia were incubated with live P. aeruginosa for 4 h, then LC3 II puncta formation was assessed by immunofluorescence microscopy (D), and LC3 II accumulation was quantified by immunoblot (E). In another set of experiments, cells were exposed continuously to 5% CO₂ (NC), or to 15% CO₂ for 18 h followed by 5% CO₂ for 24 h (HC → NC), prior to incubation with live P. aeruginosa for 4 h in 5% CO₂; LC3 II accumulation was quantified by immunoblot (F). Bars represent means ± SE, n≥3, *p<0.01 vs. NC control, **p<0.05 vs. NC with P. aeruginosa.

We also determined the effect of high CO₂ on autophagy induced by live P. aeruginosa, an important cause of bacterial lung infections in COPD and other lung diseases (10, 46). In these experiments, hypercapnia markedly decreased P. aeruginosa-induced autophagy as assessed by LC3 II puncta formation (Fig 3D) and protein accumulation (Fig 3E). Additionally, we found that when THP-1 macrophages exposed to 15% CO₂ for 18 h were subsequently returned to 5% CO₂ for 24 h, P. aeruginosa-stimulated LC3 II accumulation returned to the same level as in cells not previously exposed to hypercapnia (Fig 3F). Thus, hypercapnic inhibition of P. aeruginosa-induced autophagy is reversible, indicating that it is a regulated phenomenon and not due to cytotoxicity.

Hypercapnia increases expression of Bcl-2 and Bcl-xL and their binding to Beclin 1

Alveolar macrophages from smokers have prolonged survival, which is associated with increased expression of the antiapoptotic Bcl-2 family member, Bcl-xL (47). Besides inhibiting apoptosis, Bcl-2 and Bcl-xL also function as negative regulators of autophagy by binding Beclin 1 at its BH3 domain and blocking formation of the autophagy initiation complex (24, 48–52). Thus, to investigate the role of the Bcl-2/Bcl-xL-Beclin 1 axis in hypercapnic inhibition of autophagy, we measured Bcl-2, Bcl-xL, and Beclin 1 mRNA expression in THP-1 macrophages exposed to 5% or 15% CO₂. We found that exposure to elevated CO₂ doubled Bcl-2 and Bcl-xL mRNA expression at 2 and 4 h, respectively (Fig 4A), with no effect on the level of Beclin 1 mRNA (results not shown). Similarly, hypercapnia increased Bcl-2 and Bcl-xL protein expression at 18 h, while Beclin 1 protein did not change (Fig 4B–D). Next, to explore the potential relevance of the increases in Bcl-2 and Bcl-xL expression to hypercapnia-induced inhibition of autophagy, we assessed Bcl-2 and Bcl-xL binding to Beclin 1 by co-immunoprecipitation. Figure 4D shows that hypercapnia increased Beclin 1 pull-down by Bcl-2 and Bcl-xL antibodies, despite equal Beclin 1 input from 5% and 15% CO₂-exposed cells. These results indicate that hypercapnia increases the binding of Bcl-2 and Bcl-xL to Beclin-1, suggesting these interactions as the basis for blockade of autophagy initiation in the presence of high CO₂.

Figure 4. Hypercapnia increases Bcl-2 and Bcl-xL expression and binding to Beclin 1.

After exposure of THP-1 macrophages to 5% CO₂ (NC) or 15% CO₂ (HC), Bcl-2 (2 h) and Bcl-xL (4 h) mRNA expression was evaluated using qPCR (A), and protein expression (18 h) was assessed by immunoblot (B) and immunofluorescence microscopy with measurement of corrected total cell fluorescence (CTCF, expressed in arbitrary units (AU)); nuclei were labeled with DAPI (C). Binding of Bcl-2 and Bcl-xL to Beclin 1 was assessed by immunoprecipitation with Bcl-2- and Bcl-xL-specific antibodies, followed by immunoblot for Beclin-1; β-actin was used as loading control (D). Bars represent means ± SE, n≥3, *p<0.05 vs. NC.

Bcl-2 and Bcl-xL knockdown prevents hypercapnic inhibition of bacteria-induced autophagy

To determine whether Bcl-2 and Bcl-xL are required for inhibition of autophagy by elevated CO₂, we used an siRNA knockdown approach. Bcl-2 siRNA strongly suppressed Bcl2 mRNA and protein expression in THP-1 macrophages exposed to normocapnia and blunted the hypercapnia-induced increase in protein expression to a level lower than that in normocapnic cells treated with a non-targeting (NT) siRNA (Fig. S2A–C). Likewise, Bcl-xL siRNA suppressed Bcl-xL mRNA expression to a very low level in normocapnia, and blocked protein expression almost completely, under both normocapnic and hypercapnic conditions (Fig. S2D–F). We then evaluated the effect of Bcl-2 and Bcl-xL knockdown on hypercapnic inhibition of autophagy induced by bacteria. Figure 5A and 5B show that Bcl-2 and Bcl-xL siRNAs each prevented hypercapnic inhibition of autophagy induced by heat-killed S. aureus or live P. aeruginosa, as assessed by LC3 II protein accumulation. These results confirm that hypercapnia-induced increases in expression of Bcl-2 and Bcl-xL are required for elevated CO₂ to inhibit autophagy triggered by bacteria.

Figure 5. siRNA knockdown of Bcl-2 and Bcl-xL and the inhibitor Z36 prevents hypercapnic inhibition of autophagy and impairment of bacterial killing.

THP-1 macrophages transfected with non-target (NT), Bcl-2, or Bcl-xL siRNA were exposed to 5% CO₂ (NC) or 15% CO₂ (HC) for 18 h, then incubated with S. aureus BioParticles (SA) (A) or live P. aeruginosa (PA) (B) for 4 h, in normocapnia or hypercapnia, respectively. LC3 II accumulation was quantified by immunoblot. Bars represent means ± SE, n≥3, *p < 0.05 vs. NC with NT siRNA; **p<0.05 vs. NC + SA or NC + PA with NT siRNA; ***p<0.05 vs. HC + SA or PA with NT siRNA. In other experiments, THP-1 macrophages cultured in normocapnia or hypercapnia were starved or exposed to LPS (1 ng/ml) in the absence or presence of Z36. LC3-II protein accumulation (C) and puncta formation (D) were assessed by immunoblot and immunofluorescence microscopy, respectively. Bars represent means ± SE, n≥3, *p<0.05 vs. NC, **p<0.05 vs. NC starved, and ##p<0.05 vs. HC starved. GFP-LC3 expressing HeLa cells were also cultured in normocapnia or hypercapnia and stimulated with LPS for 18 h in the absence or presence of Z36. GFP-LC3 redistribution was assessed by fluorescence microscopy (E). The quantity of fluorescence associated with ATG12-positive and LC3-GFP autophagic puncta, normalized to the normocapnia control, is given in the upper left of each panel in E. Finally, THP-1 macrophages cultured in NC or HC were incubated with live P. aeruginosa for 90 min in the absence or presence of Z36, after which the number of viable bacteria was quantified using the MTT reduction assay (F); results are expressed as means ± SE of the ratio of the number of surviving bacteria at 90 min relative to the number of bacteria initially added to the cells (T90/T0), n≥3. *p<0.05 vs. NC control, **p<0.05 vs. HC control.

The Bcl-2 and Bcl-xL inhibitor Z36 prevents inhibition of autophagy and defective bacterial killing in hypercapnia

The finding that Bcl-2 and Bcl-xL knockdown prevents hypercapnic inhibition of autophagy indicates that Bcl-2 and Bcl-xL are necessary for inhibition by elevated CO₂, but it does not reveal how they produce this effect. To address this, we used Z36, a novel small-molecule BH3 mimetic that competitively inhibits the binding of Bcl-2 and Bcl-xL to Beclin 1 (53). As shown, Z36 (0.1 μM) prevented hypercapnic inhibition of starvation-induced LC3 II accumulation (Fig 5C) and LPS-induced LC3 II puncta formation (Fig 5D) in THP-1 macrophages, as well as LPS-induced LC3 puncta formation in GFP-LC3 HeLa cells (Fig 5E). Taken together with the results of our Bcl-2/Bcl-xL-Beclin 1 co-immunoprecipitation (Fig. 4D) and Bcl-2 and Bcl-xL knockdown experiments (Fig. 5A and 5B), the fact that Z36 blocked hypercapnic inhibition of autophagy strongly suggests that Bcl-2 and Bcl-xL mediate the inhibitory effect of high CO₂ by binding the BH3 domain of Beclin 1, thereby preventing it from participating in formation of the autophagy initiation complex.

Finally, to determine whether inhibition of autophagy is of importance in hypercapnic suppression of antimicrobial host defense, we measured bacterial killing by THP-1 macrophages under normocapnic and hypercapnic conditions in the absence and presence of Z36. For these experiments, live P. aeruginosa was added to macrophages cultured in normocapnia or hypercapnia, and after 90 min, viable bacteria were quantified using the MTT reduction assay (37, 38). As shown, hypercapnia reduced macrophage bactericidal activity, leading to a three-fold increase in viable bacteria, while the hypercapnia-induced killing defect was completely blocked by Z36 (Fig. 5F). Thus, preventing hypercapnic inhibition of autophagy by blocking Bcl-2/Bcl-xL-Beclin 1 binding with Z36 maintained full macrophage bactericidal activity in the presence of elevated CO₂, indicating that autophagy is the major pathway for bacterial killing impacted by hypercapnia under the conditions of these experiments.

DISCUSSION

Multiple clinical studies have shown an association between hypercapnia and increased mortality in patients with COPD, pneumonia and cystic fibrosis (1–3, 6, 7, 11, 12, 54). Our previous studies (14, 20) and those of others (15, 16) demonstrate that hypercapnia suppresses multiple aspects of phagocyte antimicrobial function, including cytokine expression, reactive oxygen species generation and phagocytosis. Moreover, we have recently shown that hypercapnia increases the mortality of bacterial pneumonia in mice (20). These observations suggest that in addition to being a marker of advanced lung disease, hypercapnia may play a causal role in poor clinical outcomes by inhibiting lung host defense and increasing susceptibility to pulmonary infection.

In this report, we show that hypercapnia suppresses autophagy, a critical pathway by which cells sequester and eliminate intracellular microbes. Hypercapnia inhibited autophagy induced by diverse stimuli, including amino acid starvation, rapamycin, LPS, Gram-positive and Gram-negative bacteria, in PMA-differentiated THP-1 cells and primary human alveolar macrophages, as well the epithelial carcinoma-derived HeLa cell line. The inhibitory effect of elevated CO₂ was not due to extracellular acidosis, and conversely, acidification of culture media without hypercapnia did not suppress autophagy. Starvation, rapamycin, LPS and Gram-positive and Gram-negative bacteria each signal through distinct upstream pathways, all of which converge to induce autophagy in a Beclin 1-dependent manner (55–57). While elevated CO₂ did not affect expression of Beclin-1, it increased expression of Bcl-2 and Bcl-xL, and this was accompanied by increased formation of Bcl-2-Beclin 1 and Bcl-xL-Beclin 1 protein complexes. Furthermore, siRNA knockdown of either Bcl-2 or Bcl-xL prevented hypercapnic inhibition of autophagy triggered by heat-killed S. aureus and live P. aeruginosa. In addition, Z36, a small molecule BH3 mimetic that competitively inhibits binding of Bcl-2 and Bcl-xL to Beclin 1 (53), fully blocked hypercapnic inhibition of starvation and LPS-induced autophagy. Importantly, Z36 also prevented the hypercapnia-induced defect in killing of P. aeruginosa by THP-1 macrophages. Thus, the mechanism by which hypercapnia interferes with autophagy and autophagy-dependent bacterial killing is consistent with prior observations that Bcl-2 and Bcl-xL inhibit starvation-induced autophagy by binding Beclin 1 at its BH3 domain, thereby preventing autophagy initiation (48, 58–60).

The fact that Z36 completely blocked the CO₂-induced defects in both autophagy and bacterial killing indicates that inhibition of autophagy is a major way in which hypercapnia suppresses macrophage antimicrobial function. These findings also underscore the importance of autophagy as pathway by which macrophages kill P. aeruginosa, a respiratory pathogen associated with high morbidity and mortality in patients with COPD, cystic fibrosis and other lung diseases (61–63). Recent studies have similarly demonstrated that autophagy plays a key role in clearance of P. aeruginosa by alveolar macrophages, mast cells and bronchial epithelial cells in vitro, as well as in a pneumonia model in mice (31, 64).

The finding that hypercapnic inhibition of autophagy was not related to extracellular acidosis coincides with previous observations that elevated CO₂ inhibits expression of inflammatory/host defense cytokines independently of effects on pH (14, 16). Likewise, hypercapnia inhibits epithelial ion transport (65, 66) and cell proliferation (67) in an acidosis-independent manner. Also of importance, we previously reported that elevated CO₂ inhibited expression of antimicrobial peptides in Drosophila, similarly unrelated to effects on pH (68). Moreover, as in murine P. aeruginosa pneumonia (20), hypercapnia increased the mortality of bacterial infections in Drosophila (68). These observations suggest that cells of diverse origin possess the ability to sense and respond to differing levels of molecular CO₂, and further, that the pathway(s) by which elevated CO₂ suppress innate immune responses may be evolutionarily conserved. While the molecular mechanisms of CO₂ sensing and upstream signaling that impact innate immunity are not yet defined, the recent identification of CO₂-sensitive neuronal receptors in Drosophila (69, 70), mosquitoes (71), and C. elegans (72) provides a paradigm for elucidating the components of such pathway(s) in future studies.

The major result of the present investigation – that hypercapnia inhibits initiation of autophagy and autophagy-mediated bacterial killing by macrophages – adds to the growing body of evidence that elevated levels of CO₂ suppress innate immune responses and interfere with host defense (14–16, 20, 73). Additionally, our finding that hypercapnia inhibits autophagy by increasing expression of Bcl-2 and Bcl-xL, which bind Beclin 1 and prevent autophagy initiation, has potential therapeutic implications. Pharmacologic BH3 mimetics that inhibit the antiapoptotic effects of Bcl-2 and Bcl-xL, currently in clinical trials for treatment of cancer (74–76), could also be effective in overcoming hypercapnic inhibition of autophagy. More specific BH3 mimetics, which like Z36 enhance autophagy without inducing apoptosis (53), might be of even greater benefit. These strategies, and others yet to be revealed by further investigation of CO₂ signaling pathways, hold promise for ameliorating hypercapnia-induced immunosuppression and improving resistance to infection in patients with advanced lung disease.