Macrophages control pathological interferon responses during viral respiratory infection

Abstract

Antiviral immune mediators, including interferons and their downstream effectors, are critical for host defense yet can become detrimental when uncontrolled. Here, we identify a macrophage-mediated anti-inflammatory mechanism that limits type I interferon (IFN-I) responses. Specifically, we found that cellular stress and pathogen recognition induce Oncostatin M (OSM) production by macrophages. OSM-deficient mice succumbed to challenge with influenza or a viral mimic due to heightened IFN-I activation. Macrophage-derived OSM restricted excessive IFN-I production by lung epithelial cells following viral stimulation. Furthermore, reconstitution of OSM in the respiratory tract was sufficient to protect mice lacking macrophage-derived OSM against morbidity, indicating the importance of local OSM production. This work reveals a host strategy to dampen inflammation in the lung through the negative regulation of IFN-I by macrophages.

Fig. 1.. OSM is induced by viral infection and is required for host survival.

(A) Illustration of overall hypothesis. (B) Bone-marrow derived-macrophages (BMDMs) were treated with indicated PAMPs or stressors, supernatant was collected at indicated time points, and protein levels were assessed by ELISA (n = 3). (C and D) Mice were infected intranasally with 225 PFU of A/WSN/1933(H1N1), and RNA from whole lung (C) and protein from BALF and plasma (D) were harvested at indicated time points (n = 3). (E and F) Percent initial body weight (E) and survival curve after infection (F) with 225 PFU of A/WSN/1933(H1N1), (n = 3 – 4). (G and H) (dpi) (n = 4 – 6). Relative expression calculated as (2^−ΔCt)*1000. All data are representative of at least two independent experiments. Error bars indicate standard deviation (SD), ****p ≤ 0.0001, Plaque forming units (PFU) detectable in BALF, lung, blood, and brain at 6 days post infection ns = not significant, (unpaired Student’s t test for (B) and (G), two-way ANOVA up to day 8 for (E), log-rank Mantel-Cox test for (F). LOD = level of detection.

Fig. 2.. OSM deficiency leads to elevated levels of type I IFN during IAV infection.

(A) Schematic of experimental design for (B) and (C). Mice were infected with 225pfu of A/WSN/1933(H1N1) and whole lung RNA was harvested three or seven days post-infection for bulk RNA-sequencing. (B and C) Transcriptional analysis of whole-lung gene expression (n = 2). Pathway enrichment analysis performed using Enrichr on genes that were at least 0.5 log2 fold change upregulated in both Osm^+/+ and Osm^−/− mice at 3dpi versus uninfected, and 0.25 log2 fold change more differentially upregulated in IAV-infected Osm^−/− mice as compared to Osm^+/+ mice. Gene expression pathways were retrieved from the Reactome Database (B). Volcano plot illustrating expression of MSigDB hallmark interferon alpha response signature genes between Osm^+/+ and Osm^−/− mice at 3dpi and 7dpi (C). (D and E) RNA from whole lung was harvested at indicated time points. (F) Protein from BALF harvested at 2dpi (n = 3 – 4; representative of two independent experiments). (G) Total protein in BALF at 2dpi (n = 4 – 6). IFN-α ELISA for IFN-α2 subtype. Relative expression calculated as (2^−ΔCt)*1000. Error bars indicate SD, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001, (unpaired Student’s t test for (D), (E), and (G); Mann-Whitney U test for (F)).

Fig. 3.. Macrophage-derived OSM is required for survival during poly(I:C) challenge.

(A) Schematic of experimental design. Mice were treated intratracheally (i.t.) with poly(I:C) (50μg) daily for five consecutive days and monitored daily. (B) and (C) Whole lung RNA and BALF were collected at indicated time points (n = 4 – 5). (D – G) Mice were treated i.t with poly(I:C) and monitored for body weight (D) and (F) and survival (E) and (G) (n = 3 – 7). (H – J) Whole lung RNA (H) and (I) and BALF (J) were collected at day 7 after the first poly(I:C) treatment (n = 3 – 5). (K) Lung neutrophils and monocytes were quantified by flow cytometry at baseline and at day 7 after the first poly(I:C) treatment (n = 5 – 11; pooled from three independent experiments for poly(I:C) treated mice). (L) Lung barrier permeability Evans blue dye (EBD) assay of BALF at day 7 after the first poly(I:C) treatment (n = 3 – 5). A combination of female and male mice was used for experiments in this figure. Relative expression calculated as (2^−ΔCt)*1000. (B), (C), (D), (E), (F), (G), (H), (I), (J) and (K) are representative of at least two independent experiments. Error bars indicate SD, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001, ns = not significant (two-way ANOVA up to day 7 for (D) and (F), log-rank Mantel-Cox test for (E) and (G), unpaired Student’s t test or Mann-Whitney U test for (H), (I), (J), (K), and (L).

Fig. 4.. Type I IFN signaling blockade restores survival in poly(I:C)-treated mice.

(A) Mice were treated intravenously (i.v.) with either 200μg of anti-mouse IFNAR-1 or isotype control antibody diluted in 100μl PBS on day −1, day 0, day 1, day 3, and day 4 after the first poly(I:C) treatment. Mice were weighed daily (n = 3 – 4). (B) Mice were treated i.t. with 1μg of murine rOSM in addition to 50μg of poly(I:C) daily for five consecutive days, and weighed daily (n = 4 – 5). (C) Cells isolated from digested lungs from WT mice were stimulated ex vivo with rOSM (100ng/mL) for 30 minutes and stained for surface markers and intracellular pSTAT3. Summary representation of pSTAT3 geometric mean fluorescence intensity (gMFI) in indicated populations following stimulation. pSTAT3-PE fluorescence signal was normalized per population by subtracting the PE signal in pSTAT3 fluorescence minus one controls (FMO) from the stained sample gMFI values. Epithelial cells were defined as EpCAM⁺ CD45⁻ CD31⁻, endothelial cells as CD31⁺ EpCAM⁻ CD45⁻, stromal as CD45⁻ CD31⁻ CD45⁻, and macrophages as CD45⁺ CD64⁺, all gated on viable singlets (n = 5; representative of two independent experiments). (D) RT-qPCR for Ifnb in fluorescence activated cell sorting (FACS)-sorted epithelial cells at day 7 after the first poly (I:C) treatment (n = 7 – 8). Data pooled from two individual experiments. (E) Model of OSM-mediated inhibition of IFN-I production in epithelial cells. (A) and (B) are representative of two independent experiments. Relative expression calculated as (2^-ΔCt)*1000. Error bars indicate SD, *p ≤ 0.05, ***p ≤ 0.001, ****p ≤ 0.0001, ns = not significant, (unpaired Student’s t test for (C), Mann-Whitney U test for (D), Two-way ANOVA up to day 6 for (A) and up to day 15 for (B).