(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Autophagy prevents early proinflammatory responses and neutrophil recruitment during Mycobacterium tuberculosis infection without affecting pathogen burden in macrophages [1]

['Rachel L. Kinsella', 'Department Of Molecular Microbiology', 'Center For Women S Infectious Disease Research', 'Washington University School Of Medicine', 'St. Louis', 'Missouri', 'United States Of America', 'Jacqueline M. Kimmey', 'Asya Smirnov', 'Reilly Woodson']

Date: 2023-06

The immune response to Mycobacterium tuberculosis infection determines tuberculosis disease outcomes, yet we have an incomplete understanding of what immune factors contribute to a protective immune response. Neutrophilic inflammation has been associated with poor disease prognosis in humans and in animal models during M. tuberculosis infection and, therefore, must be tightly regulated. ATG5 is an essential autophagy protein that is required in innate immune cells to control neutrophil-dominated inflammation and promote survival during M. tuberculosis infection; however, the mechanistic basis for how ATG5 regulates neutrophil recruitment is unknown. To interrogate what innate immune cells require ATG5 to control neutrophil recruitment during M. tuberculosis infection, we used different mouse strains that conditionally delete Atg5 in specific cell types. We found that ATG5 is required in CD11c + cells (lung macrophages and dendritic cells) to control the production of proinflammatory cytokines and chemokines during M. tuberculosis infection, which would otherwise promote neutrophil recruitment. This role for ATG5 is autophagy dependent, but independent of mitophagy, LC3-associated phagocytosis, and inflammasome activation, which are the most well-characterized ways that autophagy proteins regulate inflammation. In addition to the increased proinflammatory cytokine production from macrophages during M. tuberculosis infection, loss of ATG5 in innate immune cells also results in an early induction of T H 17 responses. Despite prior published in vitro cell culture experiments supporting a role for autophagy in controlling M. tuberculosis replication in macrophages, the effects of autophagy on inflammatory responses occur without changes in M. tuberculosis burden in macrophages. These findings reveal new roles for autophagy proteins in lung resident macrophages and dendritic cells that are required to suppress inflammatory responses that are associated with poor control of M. tuberculosis infection.

Funding: This work was supported by NIH grants R01 AI132697 and U19 AI142784, a Burroughs Wellcome Fund Investigators in the Pathogenesis of Infectious Disease Award ( 1016714), and the Philip and Sima Needleman Center for Autophagy Therapeutics and Research to C.L.S., a Potts Memorial Foundation postdoctoral fellowship to R.L.K., and a National Science Foundation Graduate Research Fellowship DGE-1143954 and the NIGMS Cell and Molecular Biology Training Grant GM007067 to J.M.K. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. None of the authors received a salary from any of the funders.

In this manuscript, we dissect the role for ATG5 in regulating neutrophil recruitment and accumulation during M. tuberculosis infection in vivo. We find that ATG5 functions with other autophagy proteins specifically in CD11c + lung macrophages and DCs to limit the production of cytokines and chemokines that otherwise promote neutrophil influx to the lung early in M. tuberculosis infection. We demonstrate that loss of autophagy in macrophages and DCs does not affect M. tuberculosis burden in these cell types in vivo and instead changes the inflammatory response to the infection. In addition, ATG5 is required in lung macrophages and DCs to limit IL-17A production from CD4 + T cells. Together, our studies reveal new roles for ATG5 and other autophagy proteins in regulating inflammatory responses during infection, which with further dissection could provide insight into pathways that may be targeted to effectively promote protective immune responses during TB.

ATG5 is required for the intracellular pathway of autophagy, a process by which cytoplasmic contents are targeted to the lysosome for degradation [ 18 , 19 ]. Initiation of autophagy involves phagophore formation from the endoplasmic reticulum, which is mediated by the ULK1 complex (ULK1/ULK2, ATG13, FIP200, and ATG101) and the PI3 kinase complex (ATG14L, BECLIN1, VPS15, and VPS34) [ 20 , 21 ]. Elongation of the autophagosomal double membrane depends on 2 ubiquitin-like conjugation systems. In the first system, ATG12 is activated by ATG7, transferred to ATG10, and covalently attached to ATG5. The second ubiquitin-like component is LC3 (microtubule-associated protein 1 light chain 3), which is conjugated to phosphatidylethanolamine, generating the membrane bound form called LC3-II through the actions of ATG7 and ATG3. ATG5-ATG12 facilitates LC3 lipidation through its interactions with ATG3, while ATG16L1 specifies the localization of LC3 conjugation to the autophagosome membrane [ 18 , 19 , 21 , 22 ]. The autophagosome membrane is then completed and targeted for fusion with the lysosome where the autophagosome cargo are degraded. In addition, ATG5 also functions outside of autophagy, including during M. tuberculosis infection [ 3 ], although these activities remain poorly understood. Recent work supports an autophagy-dependent role for ATG5 in LysM + innate immune cells in suppressing neutrophil recruitment to the lungs during M. tuberculosis infection [ 23 ]. Using in vitro cell culture experiments, multiple groups have reported that macrophages require autophagy to control M. tuberculosis replication by targeting the pathogen to the lysosome (xenophagy) as well as to prevent necrosis following days of infection in culture [ 5 , 23 – 29 ]. However, to date there is no evidence that xenophagy functions in this capacity in vivo. Therefore, the mechanistic basis for how loss of ATG5 results in early and exaggerated recruitment of neutrophils during M. tuberculosis infection in vivo remains unknown.

Atg5 fl/fl -LysM-Cre mice, which delete the Atg5 gene specifically in macrophages, inflammatory monocytes, some dendritic cells (DCs), and neutrophils, are severely susceptible to M. tuberculosis infection [ 3 – 5 ], highlighting ATG5 as a critical component of a protective immune response to M. tuberculosis. M. tuberculosis-infected Atg5 fl/fl -LysM-Cre mice fail to control bacterial replication and succumb to infection by 40 days postinfection (dpi) [ 3 – 5 ]. The uncontrolled M. tuberculosis replication is associated with an early (by 14 dpi) and sustained influx of neutrophils in the M. tuberculosis-infected Atg5 fl/fl -LysM-Cre mice. Depletion of neutrophils during M. tuberculosis infection in Atg5 fl/fl -LysM-Cre mice rescues the susceptibility and extends their survival [ 3 ], demonstrating that the neutrophil-dominated inflammation contributed to their susceptibility. In general, higher abundance of neutrophils during M. tuberculosis infection have been associated with worse disease outcomes in mice [ 6 – 13 ] and humans [ 12 , 14 – 17 ]. Therefore, understanding the regulatory mechanisms that govern neutrophil recruitment and accumulation during M. tuberculosis infection could be key for manipulating inflammatory responses to better control TB.

According to the World Health Organization, 10 million people fell ill with Mycobacterium tuberculosis infection and 1.5 million people died of tuberculosis (TB) in 2020, marking the first increase in TB-associated deaths in over a decade [ 1 ]. Whether a person controls the initial M. tuberculosis infection or develops active TB disease is directly impacted by the type of immune response elicited in the infected individual [ 2 ]. Therefore, better understanding of what constitutes a protective versus non-protective immune response to M. tuberculosis infection is critical for developing better therapies and prevention measures to fight this deadly disease. Genetic mouse models have provided invaluable insight into the immunological processes that are required for control of M. tuberculosis infection. Infection of mice through the aerosol route leads to phagocytosis of M. tuberculosis by alveolar macrophages, initiating an inflammatory response and recruitment of innate immune cells to the lung [ 2 ]. M. tuberculosis replicates within these innate immune cells until antigen specific T cells traffic to the lung where they activate the innate immune cells to restrain M. tuberculosis replication and suppress inflammation. M. tuberculosis establishes a chronic infection in wild-type (WT) mice, which survive for over a year with this infection.

Results

ATG5 is required in CD11c+ lung macrophages and DCs to control neutrophil recruitment and accumulation early during M. tuberculosis infection in vivo M. tuberculosis infection of Atg5fl/fl-LysM-Cre mice results in the recruitment of a higher number of neutrophils in the lungs at 14 dpi as compared to Atg5fl/fl controls, despite equivalent bacterial burdens at this time point [3]. There are also no differences in the abundance of non-neutrophil cell types in Atg5fl/fl-LysM-Cre and Atg5fl/fl mice at 14 dpi [3]. This indicates that specifically neutrophils are accumulating in M. tuberculosis-infected Atg5fl/fl-LysM-Cre mice due to a defect in inflammatory responses to infection and not due to higher burden. To determine which LysM+ cells required ATG5 to control the early influx of neutrophils into the lungs during M. tuberculosis infection, we compared bacterial burdens and neutrophil inflammation in Atg5fl/fl-LysM-Cre, Atg5fl/fl-Mrp8-Cre (deletion in neutrophils), Atg5fl/fl-Cd11c-Cre (deletion in lung macrophages and DCs), and Atg5fl/fl controls at 14 dpi. At 14 dpi, the Atg5fl/fl-LysM-Cre and Atg5fl/fl-Cd11c-Cre mice, but not Atg5fl/fl-Mrp8-Cre mice, had higher levels of neutrophil inflammation in the lungs as compared to Atg5fl/fl controls (Fig 1A and 1B). The degree of increased neutrophil frequency was similar in Atg5fl/fl-LysM-Cre and Atg5fl/fl-Cd11c-Cre mice, indicating that loss of Atg5 in CD11c+ cells, but not neutrophils, leads to the early influx of neutrophils into the lungs during M. tuberculosis infection. At 14 dpi, none of the mouse strains harbored increased M. tuberculosis burden in their lungs (Fig 1C), indicating that the increase in neutrophil abundance in Atg5fl/fl-Cd11c-Cre mice is not due to elevated bacterial burden and reflects a dysregulated inflammatory response to infection. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. ATG5 is required in CD11c+ cells to regulate the early influx of neutrophils during M. tuberculosis infection in vivo. (A) Representative flow cytometry plots of neutrophils (CD45+Ly6G+CD11b+) at 14 dpi from Atg5fl/fl, Atg5fl/fl-LysM-Cre, Atg5fl/fl-CD11c-Cre, and Atg5fl/fl-Mrp8-Cre mice. (B) Proportion of CD45+ cells that are neutrophils in the lung at 14 dpi in Mtb-GFP infected Atg5fl/fl (n = 15), Atg5fl/fl-LysM-Cre (n = 8), Atg5fl/fl-CD11c-Cre (n = 10), and Atg5fl/fl-Mrp8-Cre (n = 6) mice. Neutrophil frequency is reported as a ratio relative to the average neutrophil frequency in Atg5fl/fl control mice at 14 dpi within a given experiment. (C) Lung burden from the right lung at 14 dpi in Mtb-GFP infected Atg5fl/fl (n = 15), Atg5fl/fl-LysM-Cre (n = 8), Atg5fl/fl-CD11c-Cre (n = 10), and Atg5fl/fl-Mrp8-Cre (n = 6) mice. (D) Proportion of CD45+ cells that are neutrophils in the lung at 21 dpi in Mtb-GFP infected Atg5fl/fl (n = 19), Atg5fl/fl-LysM-Cre (n = 6), Atg5fl/fl-CD11c-Cre (n = 6), and Atg5fl/fl-Mrp8-Cre (n = 7) mice. Neutrophil frequency is reported as a ratio relative to the average neutrophil frequency in Atg5fl/fl control mice at 21 dpi within a given experiment. Susceptible and healthy Atg5fl/fl-Mrp8-Cre mice are defined as done previously where susceptible Atg5fl/fl-Mrp8-Cre mice have lost more than 5% of their pre-infection body weight by 20 dpi and healthy Atg5fl/fl-Mrp8-Cre mice have lost less than 5% of their pre-infection body weight at 20 dpi [3]. (E) Lung burden from the right lung at 21 dpi in Mtb-GFP infected Atg5fl/fl (n = 19), Atg5fl/fl-LysM-Cre (n = 6), Atg5fl/fl-CD11c-Cre (n = 6), and Atg5fl/fl-Mrp8-Cre (n = 7) mice. (F) Kaplan–Meier curve of survival proportions during Mtb-GFP infection of Atg5fl/fl (n = 9) and Atg5fl/fl-CD11c-Cre (n = 9) mice. Statistical differences were determined by a log-rank Mantel–Cox test (F) or one-way ANOVA and Šídák multiple comparison test (B–E). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Differences that are not statistically significant are designated as ns. Pooled data from at least 2 separate experiments is graphed where each data point is from 1 biological replicate. The individual numerical values used to generate the graphed data in Fig 1, the statistical analyses performed to analyze these data, and the p values from these statistical tests are in S1 Data. dpi, days postinfection. https://doi.org/10.1371/journal.pbio.3002159.g001 To determine if the higher levels of neutrophils in the lungs of M. tuberculosis-infected Atg5fl/fl-CD11c-Cre mice was due to elevated neutrophil abundance in circulation prior to or during infection, we monitored neutrophil frequency in the blood in uninfected and 14 dpi Atg5fl/fl and Atg5fl/fl-CD11c-Cre mice. There was no significant difference in the frequency of neutrophils in the blood of uninfected or 14 dpi Atg5fl/fl and Atg5fl/fl-CD11c-Cre mice (S1A Fig), suggesting that the accumulation of neutrophils in the lungs of M. tuberculosis-infected Atg5fl/fl-CD11c-Cre mice was due to specific recruitment of neutrophils to the site of infection or an inability to clear neutrophils from the lung. To begin to investigate this latter possibility, we monitored whether dead neutrophils were accumulating in the lungs of Atg5fl/fl-CD11c-Cre mice during M. tuberculosis infection by analyzing neutrophil viability at 14 dpi by flow cytometry (S1B Fig). We did not observe a significant difference in the frequency of viable neutrophils at 14 dpi in Atg5fl/fl and Atg5fl/fl-CD11c-Cre mice, indicating that increased neutrophil inflammation in the lungs of Atg5fl/fl-CD11c-Cre mice was not due to differences in neutrophil viability. The higher levels of neutrophils in the lungs of Atg5fl/fl-LysM-Cre and Atg5fl/fl-Cd11c-Cre mice were sustained through 21 dpi (Fig 1D). However, only Atg5fl/fl-LysM-Cre mice, and not Atg5fl/fl-Cd11c-Cre mice, had higher bacterial burdens in the lungs at 21 dpi (Fig 1E), similar to as previously reported [3]. Loss of Atg5 in neutrophils results in increased susceptibility to M. tuberculosis infection in some, but not all, Atg5fl/fl-Mrp8-Cre mice [3]. The susceptible Atg5fl/fl-Mrp8-Cre mice accumulate higher neutrophil numbers and bacterial burdens in their lungs at 21 dpi (Fig 1D and 1E) [3]. Therefore, loss of Atg5 in neutrophils is likely contributing to the higher burdens in the lungs of Atg5fl/fl-LysM-Cre mice at 21 dpi. These data indicate that ATG5 has a role in CD11c+ lung macrophages and DCs to regulate early recruitment of neutrophils, but not the control of M. tuberculosis replication during M. tuberculosis infection at 14 and 21 dpi. To determine how the loss of Atg5 in CD11c+ cells and the resulting early influx of neutrophils into the lungs affected host susceptibility, we monitored survival in M. tuberculosis-infected Atg5fl/fl-Cd11c-Cre mice as compared to Atg5fl/fl controls. Atg5fl/fl-Cd11c-Cre mice succumbed to M. tuberculosis infection between 100 and 150 dpi, which was significantly earlier than Atg5fl/fl controls (median survival time of 259 dpi) (Fig 1F), but not as early as Atg5fl/fl-LysM-Cre mice (succumb 30 to 40 dpi [3]). These data demonstrate that ATG5 is required in CD11c+ lung macrophages and DCs to control early neutrophil recruitment and promote survival following M. tuberculosis infection.

The role for ATG5 in lung macrophages and DCs in regulating neutrophil recruitment is dependent on other autophagy proteins Deletion of multiple different autophagy genes in all LysM+ innate immune cells can result in increased neutrophil recruitment to the lung during M. tuberculosis infection [23]. However, we previously showed that at least 1 role for ATG5 in LysM+ innate immune cells in controlling M. tuberculosis infection is autophagy independent [3]. Therefore, it is not known if the role for ATG5 specifically in CD11c+ lung macrophages and DCs is the same as described when broadly deleting Atg5 in all LysM+ cells. To determine whether the regulation of neutrophil recruitment by ATG5 in CD11c+ lung macrophages and DCs was dependent on other autophagy proteins or represented the autophagy-independent role for ATG5 during M. tuberculosis infection, we monitored neutrophil abundance in the lungs of mice lacking expression of another essential autophagy protein, BECLIN 1, in CD11c+ cells (Becn1fl/fl-Cd11c-Cre) at 14 dpi by flow cytometry. Similar to Atg5fl/fl-Cd11c-Cre mice, Becn1fl/fl-Cd11c-Cre mice also exhibited elevated neutrophil frequency in the lung at 14 dpi relative to Becn1fl/fl control mice (Fig 2A), despite no difference in bacterial burden (Fig 2B). In addition, analysis of M. tuberculosis-infected Atg16l1fl/fl-LysM-Cre and Becn1fl/fl-LysM-Cre mice also revealed higher levels of neutrophils in the lungs at 14 dpi relative to controls, without higher bacterial burdens (Fig 2C and 2D). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. The role for ATG5 in lung macrophages and DCs in regulating neutrophil recruitment is dependent on other autophagy proteins but does not involve control of pathogen infection or burden. (A) Proportion of CD45+ cells that are neutrophils (CD45+Ly6G+CD11b+) in the lung at 14 dpi in Mtb-GFP infected Becn1fl/fl (n = 8) or Becn1fl/fl-CD11c-Cre (n = 11) mice. Neutrophil frequency is reported as a ratio relative to the average neutrophil frequency in Becn1fl/fl control mice at 14 dpi. (B) Lung burden from the right lobes of the lung at 14 dpi in Mtb-GFP infected Becn1fl/fl (n = 8) or Becn1fl/fl-CD11c-Cre (n = 11) mice. (C) Neutrophil frequency of CD45+ cells reported as a ratio to the average neutrophil frequency in floxed control mice at 14 dpi in Mtb-GFP infected Atg5fl/fl (n = 6), Atg5fl/fl-LysM-Cre (n = 5), Becn1fl/fl (n = 6), Becn1fl/fl-LysM-Cre (n = 6), Atg16l1fl/fl (n = 7), and Atg16l1fl/fl-LysM-Cre (n = 7) mice. Neutrophil frequencies at 14 dpi in Mtb-GFP infected Rubicon-/- (n = 6) mice were compared to WT C57BL/6J (n = 6) mice. (D) Lung burden from right lobes of the lung at 14 dpi in Mtb-GFP infected Atg5fl/fl (n = 6), Atg5fl/fl-LysM-Cre (n = 5), Becn1fl/fl (n = 6), Becn1fl/fl-LysM-Cre (n = 6), Atg16l1fl/fl (n = 7), Atg16l1fl/fl-LysM-Cre (n = 7), WT C57BL/6J (n = 6), and Rubicon-/- (n = 6) mice. (E, F) The proportion of M. tuberculosis infected (Mtb-GFP+) and the median fluorescence intensity of Mtb-GFP in infected alveolar macrophage (Alv. Mac.), eosinophils (Eos.), neutrophils (Neut.), inflammatory monocytes (Inf. Mono.), CD11b+ DC, CD103+ DC, and non-alveolar macrophages in the lung in Atg5fl/fl (n = 7) and Atg5fl/fl-LysM-Cre (n = 7) mice at 14 dpi. (G) The proportion of viable (Zombie-) alveolar macrophages, eosinophils, neutrophils, inflammatory monocytes, CD11b+ DC, CD103+ DC, and non-alveolar macrophages in the lung in Atg5fl/fl (n = 8) and Atg5fl/fl-CD11c-Cre (n = 7) mice at 14 dpi. Statistical differences were determined by Student t test to compare the LysM-Cre or CD11c-Cre mice to their respective floxed control and Rubicon-/- mice to WT C57BL/6J mice (A–G).* P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Differences that are not statistically significant are designated as ns. Pooled data from at least 2 separate experiments is graphed where each data point is from 1 biological replicate. The individual numerical values used to generate the graphed data in Fig 2, the statistical analyses performed to analyze these data, and the p values from these statistical tests are in S2 Data. DC, dendritic cell; dpi, days postinfection; WT, wild-type. https://doi.org/10.1371/journal.pbio.3002159.g002 In addition to their role in canonical autophagy, the proteins ATG5, BECLIN 1, and ATG16L1 are also required for the process of LC3-associated phagocytosis (LAP), where LC3 is recruited to the phagosome, resulting in LC3+ single membrane vesicles that traffic to the lysosome for degradation. LAP can dampen inflammatory responses through efferocytosis, pathogen removal, stimulating inhibitory immune-receptor signaling, and reducing auto-antigen levels [30–33]. In contrast to canonical autophagy, LAP uses RUBICON and UVRAG instead of ATG14 in the PI3K complex and does not depend on ULK1 [33,34]. To distinguish between whether ATG5, BECLIN 1, and ATG16L1 were functioning through autophagy or LAP to regulate neutrophil recruitment during M. tuberculosis infection, we infected mice lacking RUBICON expression, a protein specifically required for LAP. Rubicon-/- mice had no difference in neutrophil accumulation or bacterial burdens as compared to WT controls following M. tuberculosis infection (Fig 2C and 2D), indicating that LAP is not required to control neutrophil inflammation during M. tuberculosis infection. Importantly, BECLIN 1 and ATG5 function at different steps of autophagy. Therefore, the requirement of both BECLIN 1 and ATG5 suggests that both the initiation and elongation steps of autophagy are required in CD11c+ cells to control neutrophil recruitment early during M. tuberculosis infection. Targeting of pathogens to the lysosome via autophagy is termed xenophagy, and multiple studies have reported roles for xenophagy in controlling M. tuberculosis replication in macrophages and DCs in cell culture in vitro [5,23–29]. However, in addition to there being no differences in bacterial burden in the lungs at 14 dpi (Figs 1C and 2D) [3], there was no significant difference in the proportion of macrophages, eosinophils, neutrophils, inflammatory monocytes, or DCs that were infected with M. tuberculosis (Fig 2E) and no difference in the M. tuberculosis burden in autophagy-deficient cells in the lungs of Atg5fl/fl-LysM-Cre and Atg5fl/fl mice at 14 dpi (Fig 2F). Loss of xenophagy in bone marrow-derived macrophages (BMDMs) in vitro has also been associated with increased necrosis during M. tuberculosis infection [23]. However, we did not detect a difference in the viability of macrophages, inflammatory monocytes, eosinophils, neutrophils, or DCs in the lungs of Atg5fl/fl and Atg5fl/fl-Cd11c-Cre mice at 14 dpi (Figs 2G and S1B), although we cannot rule out effects on the balance of different cell death pathways. These data support that the role for autophagy in CD11c+ lung macrophages and DCs early during M. tuberculosis infection in vivo is independent of xenophagy regulating M. tuberculosis replication. In addition, these data show that the elevated neutrophil abundance in M. tuberculosis-infected Atg5fl/fl-LysM-Cre mice is not due to differences in the cell viability or the cell types infected with M. tuberculosis but instead is driven by an imbalanced inflammatory response.

Autophagy regulates proinflammatory responses in macrophages during M. tuberculosis infection We have previously shown that lungs of Atg5fl/fl-LysM-Cre mice at 14 dpi contain higher levels of G-CSF and IL-17A than control mice [3], cytokines that promote neutrophil development and recruitment. At this time point, the primary CD11c+ cell types that are infected by M. tuberculosis are the lung macrophages, encompassing alveolar and interstitial macrophages [35,36]. Therefore, we hypothesized that autophagy could be suppressing the production of these cytokines from M. tuberculosis-infected macrophages. We tested this hypothesis by culturing BMDMs from Atg5fl/fl, Atg5fl/fl-LysM-Cre, Atg16l1fl/fl, Atg16l1fl/fl-LysM-Cre, Becn1fl/fl, and Becn1fl/fl-LysM-Cre mice and infecting with M. tuberculosis in vitro before monitoring cytokine and chemokine production using a cytokine bead array (Bio-Rad) on the supernatants from infected cultures (Figs 3A–3F and S2). Of the 23 cytokines tested, we detected significantly higher levels of IL-1β, G-CSF, KC, TNF-α, and RANTES from the Atg5fl/fl-LysM-Cre macrophage cultures compared to controls at 24 hpi (Fig 3A–3E), despite no difference in bacterial burdens or BMDM viability at this time point (Figs 3F and S2A), indicating that the heightened proinflammatory response of Atg5-/- BMDMs was not in response to increased antigen or macrophage cell death. The levels of these cytokines were only different following M. tuberculosis infection and not in mock-infected cultures, indicating that the increased proinflammatory responses were infection induced. The higher levels of G-CSF and KC, both proinflammatory signals associated with neutrophil inflammation [37–39], produced from Atg5-deficient macrophages in response to M. tuberculosis infection were dose dependent and confirmed by ELISA (Fig 3H and 3I). Similar to Atg5fl/fl-LysM-Cre BMDMs, Atg16l1fl/fl-LysM-Cre BMDMs also produced higher levels of IL-1β, G-CSF, and TNF-α following M. tuberculosis infection in vitro (Fig 3A, 3B and 3D). Becn1fl/fl-LysM-Cre BMDMs also produced more IL-1β and G-CSF following M. tuberculosis infection in vitro compared to control cells (Figs 3A, 3B and S2) despite no difference in M. tuberculosis burden at this time point (Fig 3F). In addition, Becn1fl/fl-LysM-Cre BMDMs produced higher levels of IL-6, MIP-1α, MIP-1β, and MCP-1 following M. tuberculosis infection (Figs 3G and S2). IL-6 in particular is associated with neutrophil recruitment [40–43] and also trended higher in M. tuberculosis-infected Atg5fl/fl-LysM-Cre BMDMs (Fig 3G), so we further analyzed the levels of IL-6 produced by M. tuberculosis-infected Atg5fl/fl-LysM-Cre BMDMs using an ELISA (Fig 3J). These analyses confirmed higher levels of IL-6 secretion from M. tuberculosis-infected Atg5fl/fl-LysM-Cre BMDMs compared to controls. We were able to detect more cytokines and chemokines being differentially produced by M. tuberculosis-infected BMDMs than what was previously detected in the lungs of Atg5fl/fl-LysM-Cre mice at 14 dpi [3]. This is likely due to the dilution of signals produced by macrophages in the context of the total lung homogenate, decreasing our sensitivity to detect macrophage-specific responses. In addition, it is possible that the inflammatory responses of BMDMs differ from lung macrophages and DCs during M. tuberculosis infection. Nonetheless, loss of expression of the autophagy proteins ATG5, ATG16L1, or BECLIN 1 in macrophages both in vivo and in vitro results in higher levels of cytokines and chemokines that are associated with neutrophil recruitment and accumulation following M. tuberculosis infection relative to controls, indicating that canonical autophagy is required in macrophages to control proinflammatory responses during M. tuberculosis infection without affecting M. tuberculosis burden in macrophages. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Autophagy regulates proinflammatory responses in macrophages during M. tuberculosis infection. (A) Cytokine bead array data to quantify cytokines in culture supernatants from Atg5fl/fl, Atg5fl/fl-LysM-Cre-, Atg16l1fl/fl, Atg16l1fl/fl-LysM-Cre, Becn1fl/fl, or Becn1fl/fl-LysM-Cre BMDMs mock-treated or infected with Mtb-GFP at an MOI of 10 for 24 h. BMDMs generated from at least 3 mice were tested in duplicate to quantify cytokine production. (A) IL-1β, (B) G-CSF, (C) KC, (D) TNF-α, and (E) RANTES levels at 24 hpi are shown. (F) BMDM CFU counts at 24 hpi. (G) IL-6 levels at 24 hpi are shown. (H) KC, (I) G-CSF, and (J) IL-6 levels at 24 hpi in Atg5fl/fl (n = 8) and Atg5fl/fl-LysM-Cre (n = 8) BMDMs infected with M. tuberculosis at an MOI of 10 or 20 for 24 h determined by ELISA. Statistical differences were determined by Student t test to compare the autophagy gene-deficient cells to their respective floxed control cells (A–J). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Differences that are not statistically significant are designated as ns. Cytokine levels below detection limits are designated as dbl. Each data point is 1 biological replicate, and the samples were generated from at least 2 separate experiments. The individual numerical values used to generate the graphed data in Fig 3, the statistical analyses performed to analyze these data, and the p values from these statistical tests are in S3 Data. BMDM, bone marrow-derived macrophage; hpi, hours postinfection. https://doi.org/10.1371/journal.pbio.3002159.g003

Autophagy suppresses neutrophil recruitment early during M. tuberculosis infection independent of mitophagy and inflammasome activation Autophagy has been shown to suppress proinflammatory responses by negatively regulating inflammasome activation indirectly through regulation of NFκB signaling and directly by degrading pro-IL-1β and clearance of inflammasome components [44–46], which can otherwise promote proinflammatory responses, IL-1β secretion, and neutrophil recruitment [44,45,47–49]. Indeed, Atg5fl/fl-LysM-Cre, Atg16l1fl/fl-LysM-Cre, and Becn1fl/fl-LysM-Cre BMDMs produce significantly more IL-1β in response to M. tuberculosis infection in vitro at 24 hpi compared to control BMDMs (Fig 3A), supporting that loss of autophagy has resulted in increased inflammasome activation. The primary inflammasome activated during M. tuberculosis infection of macrophages is the NLRP3 inflammasome, which consists of the NOD-, LRR-, and pyrin-domain containing protein 3 (NLRP3) sensor, ASC adaptor, and CASPASE 1 [50–53]. TLR engagement and NFκB activation during M. tuberculosis infection constitute the priming step of inflammasome activation, resulting in increased expression of pro-IL-1β and NLRP3 [54,55]. Phagocytosis of M. tuberculosis and subsequent Esx-1-dependent plasma membrane damage leading to potassium efflux is the second signal promoting NLRP3 inflammasome formation, which mediates CASPASE 1 activation followed by IL-1β processing and secretion [26,29]. To determine whether the increased neutrophil inflammation following M. tuberculosis infection in autophagy-deficient mice results from increased inflammasome activation, we crossed Caspase1/11-/- mice to Atg5fl/fl-LysM-Cre and Becn1fl/fl-LysM-Cre mice and monitored neutrophil abundance in the lungs at 14 dpi. Caspase1/11-/-/Atg5fl/fl-LysM-Cre and Caspase1/11-/-/Becn1fl/fl-LysM-Cre mice had similar neutrophil abundances and bacterial burdens in the lungs at 14 dpi as Atg5fl/fl-LysM-Cre mice and Becn1fl/fl-LysM-Cre mice, respectively (Fig 4A and 4B), indicating that increased neutrophil recruitment in the absence of autophagy is occurring independent of CASPASE1/11. Caspase1/11 deletion also did not extend the survival of Atg5fl/fl-LysM-Cre mice, indicating that increased inflammasome activation does not contribute to the early susceptibility of these mice (Fig 4C). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Autophagy suppresses neutrophil recruitment independent of mitophagy and inflammasome activation during M. tuberculosis infection. (A) Proportion of CD45+ cells that are neutrophils (CD45+Ly6G+CD11b+) in the lung at 14 dpi in Mtb-GFP infected Becn1fl/fl (n = 5), Becn1fl/fl-LysM-Cre (n = 7), Caspase1/11-/-/Becn1fl/fl (n = 9), Caspase1/11-/-/Becn1fl/fl-LysM-Cre (n = 13), Atg5fl/fl (n = 5), Atg5fl/fl-LysM-Cre (n = 5), Caspase1/11-/-/Atg5fl/fl (n = 4), or Caspase1/11-/-/Atg5fl/fl-LysM-Cre (n = 5) mice reported as a ratio relative to the average neutrophil frequency in corresponding floxed control mice. (B) Lung burden at 14 dpi from right lobes of the lung in Mtb-GFP infected Becn1fl/fl (n = 5), Becn1fl/fl-LysM-Cre (n = 7), Caspase1/11-/-/Becn1fl/fl (n = 9), Caspase1/11-/-/Becn1fl/fl-LysM-Cre (n = 13), Atg5fl/fl (n = 5), Atg5fl/fl-LysM-Cre (n = 5), Caspase1/11-/-/Atg5fl/fl (n = 4), or Caspase1/11-/-/Atg5fl/fl-LysM-Cre (n = 5) mice. The legend in 4A should be used for 4B too. (C) Kaplan–Meier curve of survival proportions during Mtb-GFP infection of Atg5fl/fl, Atg5fl/fl-LysM-Cre, Caspase1/11-/-/Atg5fl/fl, and Caspase1/11-/-/Atg5fl/fl-LysM-Cre mice. (D) Proportion of CD45+ cells that are neutrophils (CD45+Ly6G+CD11b+) in the lung at 14 dpi in Mtb-GFP infected WT (n = 4), Parkin-/- (n = 3), or Pink1-/- (n = 3) mice reported as a ratio relative to the average neutrophil frequency in WT mice. (E) Lung burden from the right lobe of the lung at 14 dpi in Mtb-GFP infected WT (n = 4), Parkin-/- (n = 3), or Pink1-/- (n = 3) mice. Statistical differences were determined by log-rank Mantel–Cox test (C) and one-way ANOVA and Šídák multiple comparison test (A, B, D, and E). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Differences that are not statistically significant are designated as ns. Pooled data from at least 2 separate experiments is graphed where each data point is from 1 biological replicate. The individual numerical values used to generate the graphed data in Fig 4, the statistical analyses performed to analyze these data, and the p values from these statistical tests are in S4 Data. dpi, days postinfection; WT, wild-type. https://doi.org/10.1371/journal.pbio.3002159.g004 Autophagy has also been shown to suppress inflammatory responses via the process of mitophagy, where autophagy targets old and damaged mitochondria to the lysosome for degradation [56,57]. The build-up of damaged or dysfunctional mitochondria in the absence of autophagy results in loss of mitochondrial membrane potential and the release of reactive oxygen species (ROS), mitochondrial DNA, and ATP to the cytosol where it can lead to oxidative damage, inflammasome activation, and proinflammatory cytokine production [57–61]. Mitophagy requires the canonical autophagy proteins, including ATG5, ATG16L1, and BECLIN 1, as well as PARKIN and PTEN-induced kinase 1 (PINK1) [62]. PINK1 accumulates on damaged mitochondria and activates the mitochondrial E3 ubiquitin ligase, PARKIN, to ubiquitinylate damaged mitochondria [58,61]. Optineurin and NDP52 are the main mitophagy receptors that interact with the ubiquitinylated mitochondria and LC3, leading to autophagosome engulfment of the mitochondria [58,63]. To investigate whether loss of mitophagy could contribute to higher neutrophil accumulation in the lungs following M. tuberculosis infection, we measured neutrophil frequency in the lungs at 14 dpi by flow cytometry in Parkin-/- and Pink1-/- mice relative to WT controls. There was no difference in neutrophil abundance or bacterial burdens in M. tuberculosis-infected Parkin-/- or Pink1-/- mice relative to WT mice at 14 dpi (Fig 4D and 4E), indicating that mitophagy is not required to control neutrophil recruitment early during M. tuberculosis infection. To determine if mitophagy is required in macrophages to control proinflammatory cytokine and chemokine production during M. tuberculosis infection, we generated BMDMs from Parkin-/-, Pink1-/-, and WT mice and infected the macrophages with M. tuberculosis for 24 h. We measured cytokine and chemokine levels from mock and M. tuberculosis-infected cultures using the cytokine bead array (Bio-Rad). Unlike in autophagy-deficient BMDMs, there were no differences in IL-6, IL-1β, G-CSF, KC, TNF-α, or RANTES production by M. tuberculosis-infected Parkin-/- and Pink1-/- macrophages at 24 hpi compared to WT macrophages (S3 Fig), nor any differences in bacterial burden (S3 Fig). Therefore, loss of mitophagy in macrophages does not result in higher levels of inflammation early during M. tuberculosis infection.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002159

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.