The type I NADH dehydrogenase of Mycobacterium tuberculosis counters phagosomal NOX2 activity to inhibit TNF-alpha-mediated host cell apoptosis

Abstract

The capacity of infected cells to undergo apoptosis upon insult with a pathogen is an ancient innate immune defense mechanism. Consequently, the ability of persisting, intracellular pathogens such as the human pathogen Mycobacterium tuberculosis (Mtb) to inhibit infection-induced apoptosis of macrophages is important for virulence. The nuoG gene of Mtb, which encodes the NuoG subunit of the type I NADH dehydrogenase, NDH-1, is important in Mtb-mediated inhibition of host macrophage apoptosis, but the molecular mechanism of this host pathogen interaction remains elusive. Here we show that the apoptogenic phenotype of MtbDeltanuoG was significantly reduced in human macrophages treated with caspase-3 and -8 inhibitors, TNF-alpha-neutralizing antibodies, and also after infection of murine TNF(-/-) macrophages. Interestingly, incubation of macrophages with inhibitors of reactive oxygen species (ROS) reduced not only the apoptosis induced by the nuoG mutant, but also its capacity to increase macrophage TNF-alpha secretion. The MtbDeltanuoG phagosomes showed increased ROS levels compared to Mtb phagosomes in primary murine and human alveolar macrophages. The increase in MtbDeltanuoG induced ROS and apoptosis was abolished in NOX-2 deficient (gp91(-/-)) macrophages. These results suggest that Mtb, via a NuoG-dependent mechanism, can neutralize NOX2-derived ROS in order to inhibit TNF-alpha-mediated host cell apoptosis. Consistently, an Mtb mutant deficient in secreted catalase induced increases in phagosomal ROS and host cell apoptosis, both of which were dependent upon macrophage NOX-2 activity. In conclusion, these results serendipitously reveal a novel connection between NOX2 activity, phagosomal ROS, and TNF-alpha signaling during infection-induced apoptosis in macrophages. Furthermore, our study reveals a novel function of NOX2 activity in innate immunity beyond the initial respiratory burst, which is the sensing of persistent intracellular pathogens and subsequent induction of host cell apoptosis as a second line of defense.

Figure 1. Mtb NuoG mediates inhibition of extrinsic but not intrinsic apoptosis pathways.

(A and B) THP-1 cells were either infected with Mtb or the nuoG mutant (ΔnuoG) for 4 h at an MOI of 5, or left uninfected (UI). (A) Cultures were either treated with a 20 µM of Caspase-3 inhibitor (C3I), an inactive analog of the inhibitor (C3I-A) or medium only (−), and analysis of TUNEL+ cells by FACS was performed after 5 days. (B) As in (A), THP-1 cells were infected or left untreated (UT), and cultured with specific inhibitors of Caspase-9 (C9I), Caspase-8 (C8I) or in medium alone (−) for 3 days, followed by analysis of TUNEL⁺ cells. Statistical analysis was performed on three independent experiments (ANOVA with Tukey post-test) and significance is indicated as follows: *, 0.01<p<0.05; **, 0.001<p<0.01; ***, p<0.001.

Figure 2. Mtb NuoG mediates inhibition of TNF-α-induced apoptosis and TNF-α secretion.

(A and B) TNF-α secretion was measured in human THP-1 (A) or primary murine (B) macrophages 3 days post infection with Mtb and nuoG knockout bacteria by ELISA. Results in (A) are a representative example. (C) Macrophages were infected with wild type (Mtb) and nuoG knockout (ΔnuoG) bacteria in the presence of 5 µg of TNF-α-neutralizing antibodies. Apoptotic cells were quantified by TUNEL staining 5 days post infection. (D) BMDM derived from C57B/6 (B6) and TNF-α knockout mice (B6 TNF^−/−) were infected with wild type and mutant bacteria and were assayed for apoptosis 5 days post infection.

Figure 3. Mtb NuoG mediates inhibition of ROS-dependent induction of apoptosis.

(A) THP-1 cells were infected with wild-type (Mtb), ΔnuoG, or complemented mutant (Comp) strains of Mtb or left uninfected (UI). Cultures were incubated in medium alone (−), or in medium containing 15 mM glutathione (GLU) or 10 µM DPI. Apoptotic cells were quantified via TUNEL staining 5 days post infection. (B) Supernatants of the cultures from (A) were harvested on indicated days, and levels of TNF-α were determined by ELISA. (C) Macrophages derived from wild type C57B/6 (B6) or NOX2 deficient gp91 knockout mice (B6 gp91 ^−/−) were infected with Mtb or MtbΔnuoG. Apoptosis was assayed after 5 days by TUNEL staining. (D) Supernatants of B6 gp91 ^−/− infected macrophages from experiment shown in (C) were harvested after indicated days and levels of TNF-α were determined by ELISA.

Figure 4. Mtb NuoG mediates inhibition of infection-induced phagosomal ROS production.

(A) Macrophages from C57B/6 (B6) or gp91 deficient mice (B6 gp91 ^−/−) were infected with Mtb or ΔnuoG and stained 24 hours post infection (hpi) with the ROS sensitive dyes DCFDA and DHE, which are more sensitive for H₂O₂ and O₂ ^- respectively. ROS production was measured by flow cytometry. (B and C) Quantification of the increase in ROS levels as detected by DCFDA (B) or DHE (C) fluorescence intensities. Net increases in individual fluorescence intensities were obtained by subtracting the fluorescence intensity distributions of untreated cells from the corresponding Mtb or ΔnuoG infected distribution. (B and C). C57/B6 macrophages were infected with wild type (Mtb) and NuoG deficient (ΔnuoG) bacteria labeled with the lipophilic dye DiI. Cells were visualized via fluorescence microscopy (scale bar  = 10 µm) at (D) 0 hpi, or stained with DCFDA and visualized at (E) 24 hpi. Arrows indicate the accumulation of ROS in the phagosome. Note that the DiI fluorescence is lost in the presence of ROS (E, lower row), but not in the absence of ROS (D and E, upper row). (F) Quantification of NO produced by infected macrophages. BMDMs from B6 mice were infected with Mtb or ΔnuoG and NO concentrations in the cell supernatant were determined by the Griess assay. Supernatants from macrophages primed with IFNγ and exposed to increasing concentrations of heat-killed M. smegmatis were used as positive controls. (A,D,E) Representative samples shown.

Figure 5. The pro-apoptotic phenotype of the nuoG mutant is conserved in primary human alveolar macrophages and is dependant on ROS.

(A) Fluorescence microscopy of human alveolar macrophages infected with wild type (Mtb) or NuoG deficient bacteria (ΔnuoG) and stained with TUNEL 3 days post infection (scale bar  = 20 µm)(representative sample). (B) Quantification of TUNEL positive macrophages (500 cells counted per condition; average of 5 donors). (C) Human macrophages were infected in the presence of the NOX2 inhibitor DPI (10 µM) and assayed for apoptosis 3 days post infection (500 cells counted per condition; average of two donors).

Figure 6. MtbΔnuoG induces phagosomal ROS production in infected primary human alveolar macrophages.

(A and B) Alveolar macrophages were infected with Mtb and nuoG knockout bacteria (ΔnuoG) and stained with the ROS sensitive dye DCFDA after 3 days. (A) Quantification is of cells containing one or more ROS positive phagosomes (two donors shown). (B) Fluorescence microscopy of DCFDA stained alveolar macrophages infected with DiI labeled bacteria (scale bar  = 10 µm)(representative sample). Arrows indicate phagosomal ROS accumulation.

Figure 7. Mtb KatG mediates neutralization of phagosomal ROS to inhibit host cell apoptosis.

(A) THP-1 cells were infected with wild type (Mtb) and KatG deficient Mtb (ΔkatG) and assayed for apoptosis via TUNEL staining 3 days post infection. (B) TNF-α concentrations in culture supernatants from (A) were assayed by ELISA. (C) BMDMs from C57B/6 (B6) or gp91 deficient mice (B6 gp91 ^−/−) were infected with Mtb and ΔkatG, stained with DHE at 24 hrs post infection, and quantified via flow cytometry. (D) Quantification of (C) using Overton cumulative histogram subtraction (FlowJo version 8.8.6 DMV). (E) Fluorescence microscopy of macrophages infected with DiI-labeled bacteria and stained with DCFDA (scale bar  = 10 µm). (C and E) representative samples shown. Arrows indicate phagosomal ROS accumulation.