Nucleotide-oligomerizing domain-1 (NOD1) receptor activation induces pro-inflammatory responses and autophagy in human alveolar macrophages

The subjects for this study were healthy nonsmokers with a median age of 24 (range 20–40) years, 25% female, 75% male, seronegative for HIV-1, and no history of pulmonary or cardiac disease or recent infections. These subjects were studied after giving a signed informed consent, according to the Declaration of Helsinki, for bronchoalveolar lavage and venipuncture. The National Institute for Respiratory Diseases (INER) Institutional Review Board in Mexico City approved this protocol.

Human bronchoalveolar cells were obtained by bronchoscopy, as previously described [12]. Cells were collected in sterile saline solutions and centrifuged at 400 × g for 15 minutes at 4°C. The pellets from the bronchoalveolar cells were suspended in culture medium, and the viability of the bronchoalveolar cells was assessed by Trypan blue exclusion (>98% in all cases). Bronchoalveolar cells were found to be 94.3 ± 2.8% alveolar macrophages according to flow cytometric analysis using a gate based on size and granularity. Therefore, we will refer to these cells as alveolar macrophages (AMs) in this study.

Monocyte-derived macrophages (MDMs) were obtained from peripheral blood mononuclear cells (PBMCs) that were prepared by centrifugation of whole heparinized venous blood diluted 1/1 with RPMI 1640 (Lonza, Walkersville, MD, USA) over a Lymphocyte separation solution (Lonza) gradient. PBMCs were plated in polystyrene dishes and incubated for 1 h at 37°C, in 5% CO₂. After discarding non-adherent cells and extensive washings, the monocytes were recovered using a cell lifter. The viability of the monocytes was assessed by Trypan blue exclusion and was greater than 98% in all cases. MN concentrations were adjusted to 10⁶ cells/ml and were incubated on 24-well plates at 37°C, in 5% CO₂ for 7 days, which allowed them to differentiate into MDMs that adhered to plastic [13]. MNs and MDMs were assayed for NOD1 activation. In all of the experiments, the culture medium consisted of RPMI 1640 supplemented with 50 μg/mL gentamycin sulfate, 200 mM L-glutamine, and 10% heat-inactivated pooled human serum.

To assess ligand-induced responses, 10⁶ AMs, MDMs, and MNs were cultured in a final volume of 1 mL in an ultralow attachment polystyrene 24-well plate (Corning Inc., NY, USA). The cells were stimulated using 5 μg/mL of synthetic L-Ala-γ-D-Glu-meso-diaminopimelic acid (Tri-DAP) (InvivoGen, San Diego, CA, USA) for 24 h. Next, the supernatants were collected and kept frozen until the cytokine assessment; the cells were harvested and prepared for protein or mRNA extraction. Culture medium alone was used as a negative control, and LPS was used as a positive control, as indicated. In selected experiments, 10 μM of Rip2/p38 inhibitor SB203580 (Promega, Madison, WI, USA), 20 μM of PI3K inhibitor LY294002 (Promega) and 20 μM of pan-caspases inhibitor Z-VAD-fmk (Calbiochem, La Joya, CA, USA) were added to the cells 30 minutes before Tri-DAP stimulation to block NOD1 signaling.

Total RNA was extracted and reverse transcribed as previously reported [13]. The cDNA was subjected to quantitative real-time PCR (qRT-PCR, TaqMan) to determine the NOD1, LC3 and IRGM mRNA expression levels using the comparative threshold cycle (ΔΔCt) as described previously [14]. Real-time PCR reactions were performed in duplicate wells according to the manufacturer’s protocol for Taqman predesigned gene assays; NOD1 (Hs01036717_m1), LC3 (Hs00171082_m1 and IRGM (Hs01013699_s1) were purchased from Applied Biosystems (Carlsbad, CA, USA). The Ct values for each gene were normalized to the endogenous control gene 18S rRNA (4319413E).

Supernatants of 24-h cultures were assayed for the release of IL1β, IL6, IL8, IL10, IL12p70, IFNα2, and TNFα cytokines using the Milliplex human cytokine detection kit (Millipore, Billerica, MA, USA) according to the manufacturer’s protocol.

Proteins extracted from cytoplasmic lysates were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes, as previously described [14]. Briefly, membranes were blocked and incubated with the following antibodies: anti-human NOD1 (AdB serotec, Raleigh, NC, USA), IRGM (Abcam, Cambridge, MA), Atg9, and LC3 (Novus Biologicals, Littleton, CA, USA) or α-tubulin (Sigma-Aldrich, St. Louis, MO, USA) for 2 h followed by an incubation with peroxidase-conjugated anti-rabbit or anti-mouse IgG antibody (Sigma-Aldrich) for 1 h at room temperature. Specific bands were detected with the chemiluminescence SuperSignal system (Thermo, Rockford, IL, USA) and were revealed using autoradiographic films. Densitometry was performed using ImageJ 1.44o (National Institutes of Health, USA).

MDMs were seeded in 8-well chamber slides (Thermo) at 5 × 10⁵ cells/well. The cells were transduced with 15 viral particles per cell for ectopical expression of p62-RFP (Premo Autophagy Sensor p62 kit, Molecular Probes, Carlsbad, CA). After 18 h, the cells were stimulated with 5 μg/ml of Tri-DAP and incubated for 5 or 24 h at 37°C and 5% CO₂. Chloroquine 60 μM and medium were used as controls. We used Lysotracker (Molecular Probes) to stain lysosomes and Hoescht 33342 (Enzo Life Sciences, Farmingdale, NY) to stain nuclei following the manufacturer’s instructions. Cells were visualized in an AxioScope.A1 microscope (Carl Zeiss, Oberkochen, Germany) with the appropriate fluorescence filters. Images were acquired and analyzed with ZEN Pro software (Carl Zeiss).

M. tuberculosis (Mtb) strain H37Rv (ATCC 25618) was grown as previously described [12]. AMs (10⁵) were infected with Mtb in RPMI with 30% non-heat-inactivated, pooled human AB serum at an infection ratio of 1–2 bacteria/20 cells in 96-well polystyrene plates; they were incubated for 1 h followed by three washes to remove any non-phagocytized bacteria. The cells were then cultured for another hour in RPMI supplemented with 10% heat-inactivated pooled human serum with or without 5 μg/ml of Tri-DAP. The infected macrophages were incubated after Tri-DAP treatment for 1 h (Day 0), 24 h (Day 1) and 96 h (Day 4) to evaluate the effects of macrophages on mycobacterial intracellular growth by quantifying the colony-forming units (CFUs). The intracellular growth index was calculated as the ratio of CFUs at Day 1 or 4 relative to the CFUs at Day 0.

AMs were infected with a multiplicity of infection (MOI) of 5 using 4 × 10⁶ cells cultured in polypropylene tubes under the same conditions described above and analyzed by transmission electron microscopy. Cells were treated with Tri-DAP with or without prior inhibition of Rip2/p38 and PI3K for 24 h post-infection and fixed in preparation for electron microscopy detection and subcellular localization of proteins.

The subcellular localization of IRGM and LC3 proteins was performed using transmission electron microscopy (TEM), as previously described [15]. Briefly, cells were fixed in 4% paraformaldehyde in 0.2 M Sörensen buffer; the samples were dehydrated with increasing concentrations of ethylic alcohol and infiltrated with LR-White hydrosoluble resin (London Resin Co., Hampshire, United Kingdom). Sections that were 60- to 80-nm-thick were placed on nickel grids. The grids were incubated overnight at 4°C with specific polyclonal rabbit anti-IRGM (Abcam) and anti-LC3 (Novus Biologicals) antibodies followed by a 2 h incubation at room temperature with goat anti-rabbit IgG (Sigma-Aldrich) conjugated to 5-nm gold particles (Sigma-Aldrich) and diluted 1:20 in PBS. The grids were contrasted with uranyl acetate (Electron Microscopy Sciences, Fort Washington, PA) and examined with an M-10 Zeiss electron microscope (Karl Zeiss, Jena, Germany). To quantitatively assess autophagy protein recruitment to the mycobacteria-containing vesicle, we performed morphometric analyses by counting gold particles that colocalized with bacteria in 10 randomly selected cells from each condition (10–18 bacteria/condition) using ImageJ software.

Differences in Toll-like receptor patterns have been previously described between differentiated macrophages and monocytes [13]; therefore, we investigated whether NOD1 was expressed in unstimulated AMs, MDMs, and MNs. We observed a basal protein expression in all cell populations studied (Figure 1A) and densitometry revealed that NOD1 was more abundant in AMs (Figure 1B). We did not observe significant differences in the relative abundance of NOD1 mRNA in unstimulated cells (Figure 1C). However, after stimulation with Tri-DAP, AMs significantly up-regulated their NOD1 mRNA levels relative to the unstimulated cells, while MDMs barely modulated their gene expression and MNs showed down-regulated NOD1 levels (p < 0.05 vs. AMs, Figure 1D).

NOD1 pathogen recognition is typically accompanied by proinflammatory cytokine production [16]. AMs are responsible for initiating inflammatory responses, which recruit large numbers of neutrophils into the alveolar spaces to clear bacteria from entering terminal airways [17]. In this study, we have determined that Tri-DAP stimulation of macrophages in 24 h cultured supernatants results in pro-inflammatory cytokine production. We found that AMs release significant amounts of IL1β, IL6, TNFα, and IL8 (p < 0.05, Figure 2A). Meanwhile, MDMs released significant quantities of IL1β, TNFα, and IL8 (p < 0.05), and remarkably, MNs only produce significant amounts of IL1β in response to Tri-DAP stimulation (p < 0.05). Interestingly, in all of the cell types tested, we did not find any anti-inflammatory production of IL10 after Tri-DAP exposure (data not shown). In addition, AMs, the maximum responders to Tri-DAP stimulation, were also assayed for IL12p70, IFNα, and IL17; we found that AMs do not produce these cytokines in response to NOD1 ligand binding (data not shown). To confirm the selectivity of NOD1’s proinflammatory response, we stimulated AMs, the highest responders, and MNs, the lowest responders, with 100 ng/ml of LPS and observed that significant amounts of TNFα, IL6, and IL10 were released from both cell types (p < 0.05, compared to medium), and the levels of cytokine production were similar in both cells (Figure 2B). Cytokine responses were confirmed by qPCR using cell lysate-derived cDNAs (data not shown).

NOD1 induces autophagy in epithelial cells [6], and we recently found that NOD2 activation induces autophagy in human AMs as part of their antibacterial pulmonary defense [14]. Thus, we evaluated the ability of NOD1 to induce autophagy in different macrophages by determining the expression of the autophagy proteins Atg9 and LC3 after Tri-DAP exposure. We found that the expression of Atg9 increased after NOD1 stimulation in AMs and MDMs (Figure 3A). Pretreatment of AMs and MDMs with the Rip2/p38 inhibitor SB203580 for 30 min prior to Tri-DAP stimulation blocked NOD1-mediated responses and blocked the increased expression of Atg9 (Figure 3A,B). NOD1 activation also induced an increment in LC3-I and II in AMs and MDMs (Figure 3A,C). In contrast, no increase in autophagy proteins was observed in MNs. The enzyme IRGM is a key autophagy component with antimicrobial functions. Thus, we analyzed IRGM expression levels and determined that IRGM gene expression was elevated in AMs after Tri-DAP exposure, slightly increased in MDMs, and decreased in MNs (Figure 3D). Consistent with the gene expression data, only the AMs overexpress IRGM protein after Tri-DAP stimulation (Figure 3E,F).

To confirm that autophagy was productive, we used ectopically expressed p62 in MDMs as an indicator of autophagic flux. The p62/SQSTM1 is a receptor for cargo destined to be degraded by autophagy [18]. Because the vector does not replicate within mammal cells, p62 is degraded when autophagy occurs and an accumulation of p62 positive vesicles indicate lack of autophagic flux. After 24 h of Tri-DAP treatment we observed cells with p62 puncta (Figure 4A). When we compared p62 puncta values obtained at 5 h to those of 24 h an evident reduction in p62 was observed confirming the development of productive autophagy (Figure 4B). In addition, most of the p62 puncta of Tri-DAP treated cells colocalyzed with lysotracker (Figure 4C). Chloroquine, as expected, induced autophagy but blocked the fusion autophagosomes-lysosome thus preventing the completion of the autophagic flux.

Our results reveal that NOD1 activation induces an antibacterial state in AMs, which may improve the control of intracellular pathogenic infections. Therefore, we evaluated the role of NOD1 stimulation during in vitro infections with Mtb, one of the most successful intracellular pathogens. Previously, we demonstrated that Tri-DAP-treated AMs develop autophagy (Figures 3 and 4). Therefore, we used immunoelectron microscopy to examine the recruitment of IRGM and LC3 to Mtb-containing vesicles in AMs infected with Mtb that were treated or non-treated with Tri-DAP. Our results demonstrate that when AMs are treated with Tri-DAP post-infection, the Mtb is enclosed in autophagic vesicles that are positive for autophagy proteins, such as IRGM and LC3, whereas a minimum recruitment of these proteins is observed in the untreated macrophages (Figure 5A-D). To confirm whether the recruitment of IRGM and LC3, as well as autophagy, were dependent on NOD1 signaling, we pre-incubated cells with Rip2/p38 (SB) and PI3K (LY) inhibitors prior to Tri-DAP stimulation (Additional file 1). The use of both inhibitors significantly diminished the recruitment of autophagy proteins to the Mtb-containing vesicle (Figure 5E, F). In absence of stimulation the expression of IRGM was very low, and LC3 was homogeneously distributed within the cytosol (Additional file 2).Next, we investigated whether Tri-DAP treatment improves the control of Mtb-intracellular growth. The intracellular growth index revealed that bacterial growth was lower inside AMs when they were treated with Tri-DAP post-infection (Figure 6A) compared with the infected AMs that did not receive Tri-DAP stimulation. Moreover, Tri-DAP stimulation induced the release of significant amounts of TNFα and the overexpression of LC3 mRNA in AMs infected with Mtb (Figure 6B,C) that were not observed in AMs when they were not treated with Tri-DAP.