APOE protects against severe infection with Mycobacterium tuberculosis by restraining production of neutrophil extracellular traps

Abstract

Mice lacking apolipoprotein E (APOE, Apoe-/- mice) on a high cholesterol (HC) diet are highly susceptible to infection with Mycobacterium tuberculosis (Mtb) but the underlying immune dysregulation has been unclear. While neutrophils are often the predominant cell type in the lungs of humans with severe tuberculosis (TB), they are relatively scarce in the lungs of some strains of mice that are used to study the disease. The neutrophil levels in the lungs of Mtb-infected Apoe-/- HC mice are very high, and thus studies in this model offer the opportunity to examine the role of specific neutrophil functions in the pathology of severe TB. We determined that depleting neutrophils, depleting plasmacytoid dendritic cells (pDCs), or blocking type I interferon signaling improved the outcome of TB in Apoe-/- HC mice. We also demonstrated that blocking the activation of peptidylarginine deiminase 4 (PAD4), an enzyme critical to NET formation, leads to fewer NETs in the lungs and dramatically improves the outcome of TB in Apoe-/- HC mice without affecting the number of neutrophils in the lung. We found that the transcriptional profile of neutrophils in Mtb-infected Apoe-/- HC mice is biased towards a state that resembles the "N2" phenotype that has been defined in cancer models and has been implicated in matrix degradation and tissue destruction. Our observations strongly suggest that the state of the neutrophil when it encounters the Mtb-infected lung is one of the main drivers of severe disease and implies that targeted interventions that alter specific states or functions, such as the production of NETs, may improve outcome while preserving sufficient capacity for host-defense.

Fig 1. Apoe^-/- HC mice are highly susceptible to infection with Mtb.

(A) Male mice of the indicated genotypes were fed either normal food or high-cholesterol food for two weeks and then infected with ~50 CFU Mtb H37Rv and maintained on their pre-infection diet. (n = 3-5 mice/group) (B) Serum cholesterol profiles at day 28 following infection of the indicated genotypes of mice fed HC food and infected with Mtb H37Rv as in (A). HDL = high-density lipoproteins, LDL = low-density lipoproteins. (n = 3 mice/group) (C) Bacterial burden in the lung measured by CFU counting for mice of the indicated genotypes at the indicated time points fed HC food and infected with Mtb H37Rv as in (A). (n = 5-7 mice/group) Bars/lines indicate mean; error bars indicate SEM. Significance analysis was performed using the two-sided Student’s t-test allowing for unequal variances (C).

Fig 2. T cell priming is intact in Apoe^-/- HC mice.

(A) Expansion of CFSE-labeled, CD8 (OT-I) or CD4 (OT-II) T cells specific for Ova peptides as measured by flow cytometry, shown as a percentage of cells dividing in the draining (mediastinal) lymph node, in mice of the indicated genotypes maintained on a HC diet at 4 days following intranasal inoculation with 2x10⁸ CFU BCG-Ova. (n = 3-4 mice/group) (B, C) Expansion of CFSE-labeled, ESAT-6 specific transgenic CD4 + T cells (C7) as measured by flow cytometry, shown as a percentage of cells dividing in the draining (cervical) lymph node, in mice of the indicated genotypes maintained on a normal (B) or HC (C) diet at the indicated days following inoculation with 10,000 CFU Mtb H37Rv in the dermis of the ear. (n = 3-5 mice/group) (D) Number of CD8 + TB10.4 + T cells in the lung parenchyma (defined by lack of labeling by an intravenous anti-CD45 antibody (IV-), see Methods) at day 19 following infection with ~50 CFU Mtb H37Rv in the indicated genotypes of mice maintained on a HC diet. (n = 7 mice/group) (E) Percentage CD8 + T cells in single-cell suspensions of lung tissue from mice in (D) producing both IFNG and TNF when restimulated with TB10.4 peptides assessed by intracellular staining and flow cytometry. (n = 7 mice/group) Bars indicate mean; error bars indicate SEM. Data are representative of 2-4 independent experiments. See S9 Fig for gating strategies.

Fig 3. Neutrophils, pDCs, and type I interferon contribute to the susceptibility of Apoe^-/- HC mice.

(A) The kinetics of neutrophil infiltration into the lungs of mice of the indicated genotypes maintained on a HC diet and infected with ~50 CFU H37Rv, as assessed by flow cytometry, expressed as the total number of neutrophils in the lung. (n = 5-7 mice/group) (B-J) Apoe^-/- or B6 mice were placed on a HC diet for two weeks, infected with ~50 CFU H37Rv via aerosol, and maintained on the diet for the entire experiment. Neutrophil depletion: (B) Pulmonary bacterial burden and (C) total neutrophil numbers in the lung at day 24 PI in the indicated treatments. IFNAR blockade: (D) Pulmonary bacterial burden, (E) expression of Ifnb1 mRNA, and (F) neutrophil fractions of CD45 + cells in the lung at day 21 PI in the indicated treatments. pDC depletion: (G) Total numbers of pDCs in the lung, (H) expression of Ifnb1 mRNA in the lung, (I) pulmonary bacterial burden at day 28 PI, and (J) total numbers of neutrophils in the lung. Bars/lines indicate mean; error bars indicate SEM. Data are representative of 2 independent experiments (n = 4-7 mice/group) (B-J). Significance analysis was performed using the two-sided Student’s t-test allowing for unequal variances (C, E-H, J) or the Wilcox rank-sum test (B,D,I). See S10 Fig for gating strategies.

Fig 4. Restraining NET formation protects Apoe^-/- HC mice against severe tuberculosis.

(A) Representative images of lung sections from Mtb H37Rv infected Apoe^-/- HC mice treated with GSK484 or vehicle daily from days 7-28 PI. Sections were labeled with anti-Cit-H3 antibody (orange) and imaged with confocal microscopy. Red outline indicates approximate lesion extent. Scale bar is 100 µm. (B) Quantification of the mean fluorescent signal of Cit-H3 labeling for 6 lesions from 3 mice from each condition in (A). (n = 6 lesions/group) (C) Expression of Ifnb1 mRNA in the lung, (D) total number of neutrophils in the lung, (E) and bacterial burden at day 28 PI in mice treated as indicated. (n = 4-7 mice/group) (F) Apoe^-/- HC mice were infected as in (A) and treated with GSK484 or vehicle daily starting at either day 7 or day 14 PI. The fraction of mice surviving to day 40 is plotted. (n = 5-6 mice/group) (G) The expression of MHCII on MDM expressed as MFI assessed by flow cytometry from mice treated as in (A). (n = 7 mice/group) (H) C3H mice were infected with ~50 CFU Mtb SA161 treated with GSK484 or vehicle daily starting at day 7 PI. Bacterial burden in the lung was measured by CFU at day 28 PI. (n = 6-7 mice/group) (I) B6 mice were infected with ~50 CFU Mtb H37Rv and treated with GSK484 or vehicle daily starting at day 7 PI. Bacterial burden in the lung was measured by CFU at day 28 PI. (n = 6-7 mice/group) (J) The number of neutrophils and monocyte-derived macrophages in the lung as measured by flow cytometry at day 28 PI in mice described in (I). Bars/lines indicate mean; error bars indicate SEM. Data are representative of two independent experiments (C-E, G). Significance analysis was performed using the two-sided Student’s t-test allowing for unequal variances (B,C,D,G,J), the Wilcox rank-sum test (E,H,I), or the Mantel-Haenszel test (F). See S10 Fig for gating strategies.

Fig 5. Neutrophils in Apoe^-/- HC mice have a distinct polarization state.

(A) UMAP plot of expression measurements from single-cell RNA-seq analysis of pulmonary immune cells isolated from B6, Apoe^-/-, and Ldlr^-/- HC mice at day 14 PI with ~50 CFU Mtb H37Rv (See Methods.). The neutrophil population, identified by comparison with the ImmGen database of transcriptional profiles (https://www.immgen.org) and confirmed by examining expression of Ly6g and S100a8, is shown in orange and circled. (B) UMAP plot of re-clustered expression measurements for the neutrophil population shown in (A). (C) Heatmap of row-normalized expression measures for the top 100 genes that distinguish the clusters labeled N1 and N2 in (B). (D) UMAP plot of neutrophils from Apoe^-/- and Ldlr^-/- HC mice at day 14 PI. The relative sizes of the N1 and N2 clusters are Apoe^-/- log₂(N2/N1) = 2.1 ± 1.5 and Ldlr^-/- log₂(N2/N1) = -1.8 ± 0.3 (mean ± SEM for 3 replicates). See S10 Fig for gating strategy.

Fig 6. Transcriptional analysis of macrophages and neutrophils isolated from B6:Apoe^-/- mixed bone marrow chimeric mice.

(A) Volcano plot depicting differential expression between Apoe^-/- and B6 bystander (uninfected) neutrophils isolated from B6:Apoe^-/- mixed bone marrow chimeric mice, maintained on a normal diet, 28 days following infection with ~50 CFU H37Rv. Genes that are most characteristic of N1 and N2 neutrophils in Mtb-infected mice on a HC diet as determined by single-cell RNA-seq analysis are colored (See Fig 5C). Dashed line indicates FDR = 0.05. (n = 3 mice/group) (B) Multidimensional scaling (MDS) plot⁶⁵ of gene expression in alveolar macrophages (AM), monocyte-derived macrophages (MDM), and neutrophils isolated by cell sorting from B6:Apoe^-/- mixed bone marrow chimeric mice at Day 28 following infection with ~50 CFU of Mtb H37Rv expressing mCherry⁶⁶. The top 500 genes with the largest standard deviations across samples were used to generate the plot. Open (filled) shapes indicate uninfected (infected) cells, shapes indicate cell types (△ = MDM, □ = Neutrophil), and colors indicate genotype (red = Apoe^-/-, black = B6). Distances on the plot represent the leading log2-fold-changes, which are defined as the root-mean-square average of the top largest log2-fold-changes between each pair of samples. (n = 3 mice/group) See S10 Fig for gating strategy.