Central role of TWIK2 in triggering ATP-induced macrophage training of bacterial killing.

A. Diagram of the ATP-induced AMφ training model in vivo. WT or KO mice were exposed to ATP (40 μL at 1mM, i.n.) and AMφ were isolated on day 7 to determine acquisition of training. For in vitro training, BMDMs or RAW cells were trained with or without ATP (1mM) and recovered for 7 days. B. AMφ subjected to training as described in A were incubated with live PA-GFP (MOI:1), and assessed for bacterial survival (CFU), n = 3. C. BMDMs subjected to in vitro training (1mM ATP for 7 days as described in A) were incubated with live PA-GFP (MOI:1) and were assayed for PA survival (CFU). D. Survival of control or ATP pretrained mice after receiving 40mL of 5×106 CFU of P. aeruginosa intranasally instilled, n = 7. E. Comparison of bacterial load clearing activity of AMφ trained with various DAMPs: ATP, or NAD. ATP showed the most dramatic response. F-H. Changes in concentrations of IL1-β (F), IL-6 (G), and IL-18 (H) in ATP-trained AMφ after PA infection. ∗, p<0.05, ∗∗, p<0.01, data are from three independent experiments (mean and s.e.m.).

Immune cell counting in the lungs before and after ATP training.

A-D. Immune cells counting before and after ATP training in lung tissue: neutrophils (A), eosinophils (B), monocytes (C) and AMφ (D), n = 3. Data are from three independent experiments (mean and s.e.m.).

Assessment of TWIK2 PM and phagosome translocation.

A. Images of RAW cells expressing TWIK2-GFP post-ATP challenge (1mM, PM stained in red). White arrowheads show PM translocated TWIK2. Bar, 1µm. B. Western blot analysis of ATP trained or untrained macrophage lysate after membrane isolation. Membrane and non-membranous fractions were compared to assess the relative abundance of TWIK2 and the Na-K ATPase with and without ATP training. C. Fluorescence Recovery After Photobleach of membrane-associated TWIK2-GFP in ATP- or control-treated cells. D. Ratiometric fluorescence intensity analysis of TWIK2-GFP signal within 1 micrometer of the plasmalemma to the corresponding TWIK2-GFP cytoplasmic intensity, highlighting enhanced and long-term membrane association after ATP treatment. Duration of PM-TWIK2 untreated or ATP or LPS treated cells is shown. *, P<0.05, compared with control. Data are representative of (A, B-left panel) or from (B-right graph, D) three independent experiments, mean and s.e.m. in B, D.

PM-TWIK2 re-internalizes into phagosomes upon phagocytosis.

A. Fluorescence images of TWIK2-GFP expressing RAW 264.7 cells phagocytosing AF594-conjugated zymosan (red) with or without ATP treatment. B. Quantification of TWIK2 localization frequency in A. We observed significantly more frequent phagosome-associated TWIK2 in ATP-trained cells vs. controls. Arrowheads point to phagosomes with TWIK2-GFP. ∗, p<0.01 compared with untreated control (n = 3). C, D, Representative images (C) and quantification (D) of K+ enrichment (indicated by a specific K+ dye, ION K+, green) in phagosomes relative to cytoplasm of control and ATP-trained BMDMs, ∗∗∗, p < 0.001. Data are representative of (A, C) or from (B, D) three independent experiments, mean and s.e.m. in B, D.

ATP training regulates transcriptional profile and metabolic reprogramming of macrophages.

A. PA killing ability (as indicated by the survival of PA, CFU) of BMDMs trained with or without ATP in the presence of metabolic inhibitors 2DG or protease inhibitors. B. RNA isolated from BMDMs trained with or without prior ATP treatment was mapped to the Mus musculus genome and subsequently sorted into GO terms in Galaxy. Shown are the top 10 GO categories overrepresented in ATP trained BMDMs. C. Basal glycolytic and mitochondrial metabolism in BMDMs trained with or without ATP seeded equivalently 2 days prior to assessment as exhibited by Seahorse assay. D, E. Western blot (D) and quantification (E) of NLRP3 assembly in BMDMs trained with or without 1mM ATP and challenged with 1.0 MoI PA for 2 hours. These cells were lysed with DSP crosslinker and assessed for NLRP3 oligomerization. Data are representative of (B, D) or from (A, C, E) three independent experiments, mean and s.e.m. in A, C, E.

ATP training rescues lung immunosuppression caused by pneumonia.

A. Diagram of the double-infection model of primary pneumonia (1° PN) with 1×106 E. coli and secondary pneumonia (2° PN) with 1×106 GFP-PA. B. Colony-forming units (CFU) per 5mg of lung homogenate with 1° or 2° pneumonia. ∗, p < 0.05. C. Relative bacterial killing ability in lung homogenate of WT or designated KO mice trained with or without ATP in the double exposure model, n = 5. Data are from (B, C) three independent experiments, mean and s.e.m. in B, C.

Negative controls for K+ enrichment as shown in Figure 4C.

Cells were incubated with fluorescently labeled zymosan 594 beads under identical conditions in Figure 4C, but without addition of Ion K+ Green dye. Images were captured using identical laser power and selected focal plane with most up taken AF594-Zymosan beads in frame. Scale bar, 10 µm.

Volcano plot displaying log2 fold change of differentially accessible chromatin in ATP treated BMDMs relative to that of vehicle-treated BMDMs 6 days after treatment.

The x-axis shows log2 (fold change), and the y-axis denotes -log10(Q-value). Peaks with Q-value (Benjamini-Hochberg FDR) are highlighted in red and labeled with the gene containing the nearest likely promoter region. No significant down-regulated peaks were noted in this comparison. The horizontal dashed line indicates the Q-value threshold of 0.05 (-1.3 on the -log10 scale). The ATAC-seq data has been submitted to the NIH SRA at SUB15765984.