Autophagy stimulation by rapamycin suppresses lung inflammation and infection by Burkholderia cenocepacia in a model of cystic fibrosis

Macrophages harboring the CFTR ΔF508 mutation support increased B. cepacia survival and produce more IL-1β than WT macrophages.

We examined whether B. cepacia had a survival advantage in primary murine macrophages expressing the CFTR protein harboring the ΔF508 mutation, which is the most common mutation in CF patients.^60–62 WT and CFTR ΔF508 (ΔF508) macrophages were infected with the B. cepacia clinical isolate K56-2 and colony-forming units (CFU) were determined from lysed infected macrophages at 30 min (Fig. S1) and at 24 h post-infection (Fig. 1A). We found that more B. cepacia was recovered from ΔF508 macrophages than WT cells after 24 h of infection (Fig. 1A), whereas, the initial uptake was similar in both cells (Fig. S1). We next examined the number of B. cepacia associated with WT and ΔF508 macrophages by confocal microscopy. WT and ΔF508 macrophages were infected with red fluorescent protein (mRFP)-expressing B. cepacia for 30 min and 2 h and the number of B. cepacia associated with 100 macrophages was evaluated. At an early time point (30 min post-infection) similar numbers of B. cepacia were associated with WT and ΔF508 macrophages (Fig. S2). In contrast, at 2 h there were <200 B. cepacia associated with 100 WT macrophages (Fig. 1B and C), whereas there were >300 B. cepacia associated with 100 ΔF508 macrophages (Fig. 1B and C). Thus, these data are consistent with the CFU data suggesting more growth of B. cepacia in ΔF508 macrophages than in WT macrophages.

Since IL-1β is an essential pro-inflammatory cytokine that affects CF patients,^50–58 we next determined the levels of active IL-1β in culture supernatants and found that ΔF508 macrophages produced significantly more IL-1β when infected with B. cepacia compared with WT cells (Fig. 1D). Yet, the mechanism is unclear. To rule out the role of macrophage survival in IL-1β production differences, the release of lactate dehydrogenase (LDH) was evaluated in WT and ΔF508 macrophages infected with B. cepacia. The ΔF508 mutation did not alter macrophage survival in response to B. cepacia (Fig. S3). Therefore, primary macrophages expressing the ΔF508 mutation support increased B. cepacia intracellular survival and bacterial replication and produce more IL-1β during B. cepacia infection.

The B. cepacia-containing intracellular vacuole acquires autophagy characteristics in WT macrophages but not in ΔF508 macrophages.

Since autophagy activity is compromised in CF epithelial cells we examined whether macrophages have the same defect by observing their autophagy response during the B. cepacia infection.^11,12 WT and ΔF508 macrophages were infected with mRFP-expressing B. cepacia for 2 h. The acquisition of endogenous LC3 by the B. cepacia-containing vacuole was assessed with specific antibodies by confocal microscopy. In WT macrophages, 20% B. cepacia-containing vacuoles were labeled with the specific autophagy marker LC3 within 2 h infection (Fig. 2A and C). Several LC3-labeled vacuoles (puncta) were identified in WT macrophages (Fig. 2A; white arrows). In contrast, B. cepacia-containing vacuoles in ΔF508 macrophages had rare LC3-labeled structures (puncta) in response to B. cepacia compared with WT macrophages and only 10% of mRFP-expressing B. cepacia-containing vacuoles showed co-localization with LC3 (Fig. 2B and C). Notably, as shown earlier, greater numbers of B. cepacia were visualized in ΔF508 macrophages than in WT cells at 2 h post-infection (Fig. 2A and B). To further confirm the identity of the B. cepacia-containing compartment, B. cepacia-infected WT and ΔF508 macrophages were examined by transmission electron microscopy. Infected WT macrophages contained few B. cepacia and they were surrounded by several multilamellar membranes (Fig. 2D; black arrows) similar to autophagosomes and showed signs of bacterial degradation. In contrast, in ΔF508 macrophages, several intact B. cepacia were associated with the macrophage (Fig. 2D; white arrow heads), and they lacked the autophagosome-like structure. Immunofluorescence quantification of puncta within infected macrophages also confirmed that the autophagy response is compromised in ΔF508 macrophages during B. cepacia infection (Fig. 2E). Together, these data indicate that the autophagy response of ΔF508 macrophages to B. cepacia is significantly less than that of WT macrophages and hence more B. cepacia are enclosed in autophagosome-like vacuoles within WT macrophages.

B. cepacia downregulates autophagy genes during infection of WT and ΔF508 macrophages.

B. cepacia resides within LC3 labeled compartments reminiscent of autophagosomes in WT macrophages with the presence of puncta in several macrophages (Fig. 2). However, compared with WT macrophages infected with other organisms such as Salmonella, there were many fewer puncta in WT macrophages infected with B. cepacia (Fig. 2E and data not shown).^13,14,63 Nevertheless, significantly fewer puncta were detected in ΔF508 macrophages when compared with WT macrophages infected with B. cepacia (Fig. 2E). The cause of this observation is unknown. To examine the effect of B. cepacia on the autophagy pathway, we performed an array analysis for a part of autophagy genes in B. cepacia infected WT and ΔF508 macrophages. B. cepacia infection led to a significant downregulation of several autophagy genes such as Atg9b, Atg5, Atg12 (Fig. 3A) and Atg8 (Fig. 3B) in ΔF508 and in WT macrophages (Fig. S4A and B). Quantitative PCR (qPCR) and protein gel blot analysis confirmed the downregulation of autophagy genes and their corresponding protein in both cell types (Figs. 3A, C–E and 2E and data not shown). The low number of puncta in ΔF508 macrophages compared with WT macrophages further confirmed that B. cepacia downregulates autophagy and that this effect is more profound in ΔF508 macrophages (Fig. 2E). On the other hand, other autophagyinteracting bacteria such as Salmonella induce the expression of several autophagy molecules during infection (data not shown). Therefore, it appears that the downregulation of essential autophagy molecules is a strategy employed by B. cepacia to delay the maturation of autophagosomes and hence avoid clearance by the host cell. The B. cepacia-mediated downregulation of autophagy genes is similar in WT and ΔF508 macrophages but the effect is more pronounced in the latter cells suggesting that they harbor an inefficient autophagy system as described in ΔF508 epithelial cells.^11,12

Depletion of LC3 from WT macrophages allows for more B. cepacia recovery and increased IL-1β production.

B. cepacia infection downregulates autophagy genes in both WT and ΔF508 macrophages, but the effect on autophagy activity is more pronounced in ΔF508 macrophages, suggesting a possible mechanism allowing for increased B. cepacia survival. This could be due to defective autophagy activity in macrophages harboring mutant CFTR protein similar to epithelial cells expressing the same mutation.^11,12 Therefore, we examined if compromising autophagy in WT macrophages would result in a similar phenotype to ΔF508 macrophages. We reduced autophagy activity in WT macrophages by depleting LC3, an essential autophagy molecule. The depletion of LC3 was achieved with specific small interfering RNA (siRNA) to LC3 and confirmed by protein gel blots using specific antibody that recognizes LC3I and LC3II forms (Fig. S5A). Initial uptake of B. cepacia was similar in WT and ΔF508 macrophages pretreated with siRNA-LC3 or siRNA-control (CT) (data not shown). Depletion of LC3 in WT macrophages by specific siRNA was accompanied by an increased recovery of B. cepacia (Fig. 4A). Depletion of LC3 in ΔF508 macrophages was accompanied by an increase in B. cepacia recovery although less significant (Fig. 4B). The decrease in LC3 expression was associated with increased secretion of the active form of IL-1β from B. cepacia-infected WT and ΔF508 macrophages (Fig. 4C and D). Therefore, compromising autophagy in WT macrophages with specific siRNA produced a ΔF508-like phenotype, such as increased B. cepacia recovery and enhanced IL-1β release.

Once activated, caspase-1 cleaves pro-IL-1β to its active form, which is then released into culture supernatants. Therefore, to understand why IL-1β release is increased in ΔF508 macrophages during B. cepacia infection, we compared the levels of IL-1β pro-form in ΔF508 and WT macrophages upon B. cepacia infection. The depletion of LC3 with specific siRNA in WT and ΔF508 macrophages increased the levels of pro-IL-1β in macrophages in response to B. cepacia infection (Fig. S5B) and led to more active IL-1β released into culture supernatants (Fig. 4C and D). Therefore, decreasing autophagy activity resulted in increased pro-IL-1β accumulation, activation and release of active IL-1β in response to B. cepacia infection.

Stimulation of autophagy by rapamycin decreases B. cepacia recovery and IL-1β production in vitro.

Macrophages harboring ΔF508 mutation show increased intracellular survival and growth of B. cepacia (Fig. 1A and B). To investigate whether this phenotype is due to reduced autophagy activity in ΔF508 macrophages, we examined the effect of autophagy stimulation by rapamycin on B. cepacia-infected ΔF508 macrophages. Rapamycin is an inhibitor of the mammalian target of rapamycin (mTor) that induces autophagy.^64,65 Macrophages were treated with rapamycin or with the drug vehicle DMSO before and after B. cepacia infection. The number of B. cepacia recovered was evaluated at 2, 4 and 6 h post infection and expressed as CFUs/ml. Rapamycin significantly decreased the number B. cepacia recovered from WT and ΔF508 macrophages (Fig. 5A and B). However, the effect was observed in ΔF508 macrophages as early as 2 h post-infection despite equal initial uptake (Fig. 5B and data not shown). Adding rapamycin directly to bacterial cultures in the absence of macrophages did not alter B. cepacia survival (data not shown). By confocal microscopy, the number of B. cepacia associated with 100 macrophages at 2 h post infection was significantly lower in ΔF508 macrophages in the presence of rapamycin, thus confirming the CFU data (Fig. 5C). In contrast, the reduction in B. cepacia numbers recovered from infected WT macrophages was evident at later stages of infection (Fig. 5A).

We next investigated if rapamycin affects the levels of active IL-1β released in culture supernatants.⁶⁶ In contrast to DMSO, rapamycin treatment of ΔF508 macrophages markedly decreased IL-1β production in response to B. cepacia infection (Figs. 5D and S6A). Rapamycin also decreased IL-6 production (Fig. S6B and C). The levels of other cytokines such as TNFα were comparable in the presence or absence of rapamycin at 24 h post-infection (Fig. S6D and E). Therefore, stimulation of autophagy by rapamycin decreases the production of IL-1β and not all pro-inflammatory cytokines in response to B. cepacia infection in vitro.

Rapamycin treatment decreases the bacterial burden and reduces the signs of inflammation in the lungs of ΔF508 mice.

Since B. cepacia infects the lungs of CF patients we used the ΔF508 mouse model for CF to examine if rapamycin treatment decreases the bacterial burden in vivo and if it alleviates inflammatory findings in the lungs of infected mice. WT and ΔF508 mice were pretreated with two doses of rapamycin or DMSO by intra-peritoneal injections. Mice were then infected intra-tracheally with B. cepacia followed by an additional dose of rapamycin or DMSO. On the second day post-infection, mice were sacrificed, lungs were collected and the bacterial burden was determined. Rapamycin treatment did not affect B. cepacia recovery from WT lungs but reduced it drastically from ΔF508 lungs (Fig. 6A and B). Therefore, rapamycin treatment stimulates autophagy and decreases the B. cepacia burden in vivo in ΔF508 mice.

Next, we determined the effect of rapamycin treatment on the inflammatory response of lung tissue of B. cepacia-infected mice. Histological examination of H&E stained lung sections of WT mice infected with B. cepacia revealed the preservation of most of the lung tissue with patchy areas of accumulation of inflammatory cells (Fig. 6C and left upper part). Lung sections of ΔF508 mice demonstrated significantly more pronounced and generalized recruitment of inflammatory cells (arrow head) throughout the lung tissue with peribronchiolar and perivascular infiltrates and hemorrhage (arrow) (Fig. 6C and left middle part). Higher magnification of ΔF508 lung sections closely showed that B. cepacia infection also led to the accumulation of inflammatory cells and exudates in the alveolar spaces of ΔF508 lungs (Fig. 6C and left lower part). To determine whether rapamycin treatment rescues the appearance of the ΔF508 lung tissue during B. cepacia infection, mice were treated with rapamycin and infected intra-tracheally with B. cepacia as described above. H&E stained lung sections of rapamycin-treated mice showed that rapamycin treatment drastically reduced the recruitment of inflammatory cells throughout the CF lung, prevented peribronchiolar and perivascular hemorrhage and preserved the alveolar space of infected ΔF508 mice (Fig. 6C and right middle and lower parts). Therefore, rapamycin treatment decreases inflammation and thus rescues the lung tissue in a mouse model of CF following B. cepacia infection.

Rapamycin treatment decreases B. cepacia recovery from ΔF508 macrophages by enhancing their delivery to lysosomes.

Rapamycin treatment significantly decreased the recovery of B. cepacia from ΔF508 mice and their derived macrophages, yet the mechanism is unclear. It is possible that rapamycin-mediated induction of autophagy enhanced the maturation and the fusion of the B. cepacia-enclosing vacuole with the lysosomal compartment. To explore the trafficking of B. cepacia within macrophages, the colocalization of mRFP-expressing B. cepacia with the autophagy marker LC3 was determined in the presence of rapamycin. Rapamycin increased the number of B. cepacia-containing vacuole that acquired LC3 in both WT and ΔF508 macrophages (Fig. 7A; white arrows and B). Since rapamycin promoted the acquisition of LC3, we examined if it enables the delivery of B. cepacia to the lysosome. Green lysotracker, which accumulates in lysosomes, was added to rapamycin or DMSO-treated WT and ΔF508 macrophages prior to infection with mRFP-expressing B. cepacia. Almost 40% and 20% of B. cepacia localized with the lysotracker in DMSO-treated WT and ΔF508 macrophages respectively within 2 h infection indicating that WT macrophages are more efficient in delivering B. cepacia to the lysosomes (Figs. 7A, B and S7). Rapamycin treatment slightly increased the number of B. cepacia that trafficked to the lysosome in WT macrophages. Nevertheless, rapamycin treatment significantly increased the delivery of B. cepacia to the lysosomes in ΔF508 macrophages up to 50% within 2 h (Figs. 7C, D and S7). Therefore, stimulation of autophagy by rapamycin drives the maturation of the B. cepacia-containing vacuole enabling it to fuse with the lysosome and thereby improving the clearance of the bacterium from ΔF508 macrophages.

Bone-marrow-derived macrophages.

All animal experiments were performed according to protocols approved by the Animal Care Use Committee of the Ohio State University College of Medicine. Wild-type (WT) C57BL/6, ΔF508 mice were obtained from Case Western University and housed in the OSU vivarium. Bone marrow-derived macrophages (BMDMs) were isolated from the femurs of 6- to 12-wk-old mice and were cultured in IMDM (GIBCO, 12440) containing 10% heat-inactivated FBS (GIBCO, 16000), 20% L cell-conditioned medium, 100 U/ml penicillin and 100 mg/ml streptomycin (GIBCO, 15140) at 37°C in a humidified atmosphere containing 5% CO₂. Macrophages were infected with B. cepacia K56-2 or the corresponding gentamicin sensitive strain MHK1 at a multiplicity of infection (MOI) of 10.

Bacterial strains and culture.

Burkholderia cepacia strain K56-2 was isolated from a CF patient. All bacterial strains were grown in Luria-Bertani (LB) broth at 37°C overnight with high amplitude shaking. The B. cepacia MHK1 strain has a mutation in an antibiotic efflux pump that confers gentamicin sensitivity but does not alter the trafficking of the mutant in macrophages.⁷⁹ To kill extracellular bacteria, Iscove's media (GIBCO, 12440) containing 10% heat-inactivated FBS (GIBCO, 16000) containing 50 µg/ml gentamicin (GIBCO, 3564) was added for 30 min as previously described in reference ⁷⁹. To enumerate intracellular bacteria, infected macrophages were lysed using ice cold PBS (GIBCO, 14190) and physical disruption. Macrophages from monolayers were scrapped and pipetted repeatedly against the walls and bottom of the well. Recovered bacteria were quantified by plating serial dilutions on LB agar plates and counting colonies using the Acolyte Colony Counter, 5710/SYN.

Mouse in vivo infection.

WT C57BL/6 and ΔF508 mice were infected intra-tracheally with 10 x 10⁶ WT bacteria with rapamycin (Sigma-Aldrich, R0395) or DMSO (Sigma-Aldrich, D2650) treatment (n = 3). Rapamycin was used in vivo at 4 mg/kg by intra-peritoneal injections. Mice were pretreated with two doses of rapamycin for 2 d (24 h interval), and then infected with B. cepacia 2 h after the second dose of rapamycin. On the third day, mice treated with a final dose of rapamycin. The mice were sacrificed 2 h later, and the number of bacteria in the lungs was determined at second day post-infection.¹⁴ All animal experiments were performed according to animal protocols approved by the Animal Care Use Committee of the Ohio State University College of Medicine.

In vitro rapamycin treatment.

Rapamycin (Sigma-Aldrich, R0395) was dissolved in DMSO (Sigma-Aldrich, D2650) at 1 mg/ml. Rapamycin was used at concentration 5 µg/ml, DMSO alone was used as a diluent control.

Immunoblotting.

Macrophages were stimulated with B. cepacia K56-2 (MHK1) and the culture supernatant was removed. The cells were washed twice with PBS (GIBCO, 14190) and lysed in lysis buffer solution supplemented with a protease inhibitor mixture (Roche Applied Science, 10-519-978-001). The protein concentration was adjusted to 40 µg/ml. Proteins were separated by sodium dodecyl sulfate-15% PAGE and transferred to polyvinylidene difluoride (PVDF) membranes (Bio-Rad Laboratories, 1620117). Membranes were immunoblotted for pro-Interleukin-1β (IL-1β) antibody (kindly provided by Dr. Mark Wewers, Ohio State University) and Atg8 (LC3) (Sigma-Aldrich, L8918) protein bands were detected with secondary antibodies conjugated to horseradish peroxidase followed by enhanced chemiluminescence reagents (Amersham ECL protein gel blotting detection reagents GE Health Care-Life Sciences, RPN2106).

Enzyme-linked immuno sorbent assay (ELISA).

Macrophages were infected with B. cepacia K56-2 (MH1K) for different time points 6, 8, 24 h. Then, culture supernatants were collected, centrifuged and stored at -20°C until assayed for cytokine content. The amounts of IL-1β, IL-6 and TNFα in the supernatant were determined by specific sandwich ELISA following the manufacturer's protocol (R&D System Inc., DY201, DY406, DY410 respectively) and as previously described in reference ⁸⁰.

siRNA knockdown of LC3.

siRNA treatment was performed using siRNA against LC3 (Dharmacon, J-040989-09): CUA AUA AAG GCA CAA CGA A, GGA UAU AGC UCU AAG CCG G, CAU CCU AAG UUG CCA AUA A, ACU AUG GUG CGA UCA GUA A. siRNA was nucleofected into primary murine macrophages using Lonza Nucleofection kit (VPA-1009) and Amaxa equipment (AAD-10015) as we described previously in references 81 and 82.

Real time PCR.

Total RNA was isolated from cells were lysed in Trizol (Invitrogen Life Technologies, 15596-026) and submitted to SA Biosciences for autophagy array study. Gene expression was calculated as relative copy numbers (RCN), as described previously in references 81 and 83. Briefly, Ct values of every target gene were subtracted from the average Ct of five housekeeping genes, present on the autophagy array (Gusb, Hprt1, Hsp90ab1, Gapdh, Actb) and the resulted ΔCt was used in the equation: RCN = (2^-ΔCt) × 100. RCN for every gene represents its expression as number of copies relative to the 100 copies of average housekeeping genes.^81,83

Histopathological analysis.

Lungs were removed en bloc and fixed in (10% Formalin) at room temperature for 24 h then formalin was replaced by PBS (GIBCO, 14190), and processed for paraffin embedding. Formalin-preserved sections of the lungs we processed and embedded in paraffin by standard techniques. Sections of 5 µm thick were stained with hematoxylin and eosin (H&E) and examined. Longitudinal sections of 5 µm taken at regular intervals were obtained using a microtome from the proximal, medial and distal lung regions.

Confocal microscopy.

Immunofluorescence microscopy experiments were performed as previously described in reference ⁸⁴. B. cepacia expressing monomeric Red Fluorescent Protein (mRFP) was used. Localization of markers on B. cepacia phagosomes was performed as previously described in references 14 and 63. Antibodies used were rabbit anti-Atg8/LC3 (Abgent, AP1805a) followed by fluorescent secondary antibodies (Molecular Probes, A11008). Nuclei were stained with the nucleic acid dye 4′,6′-diamino-2-phenylindole and lysosomes were stained green with Lysotracker Green (Invitrogen, L-7526) as previously described in references 14, 63, 80 and 85. In each experiment at least 100 bacteria were scored. Experiments were performed at least three times. Samples were analyzed with (Olympus FV 1000 Spectral Confocal) at the Ohio State University Microscopy Core Facility.

Transmission electron microscopy.

WT and ΔF508 primary murine macrophages cultured on Permanox (Lab-Tek) chamber slides were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer with 0.1 M sucrose. The cells were postfixed with 1% osmium tetroxide in phosphate buffer and then en bloc stained with 2% uranyl acetate in 10% ethanol, dehydrated in a graded series of ethanol and embedded in Eponate 12 epoxy resin (Ted Pella Inc., 18012). Ultrathin sections were cut on a Leica EM UC6 ultramicrotome (Leica microsystems, EM FC7), collected on copper grids, and then stained with Lead citrate and rranyl acetate. Images were acquired with an FEI Technai G₂ Spirit transmission electron microscope (FEI), and Macrofire (Optronics) digital camera and AMT image capture software.

Statistical analysis.

All experiments were performed at least three independent times and yielded similar results. Comparisons of groups for statistical difference were done using Student's two-tailed t test. p value ≤ 0.05 was considered significant.