GM‐CSF suppresses antioxidant signaling and drives IL‐1β secretion through NRF2 downregulation

Abstract GM‐CSF is a potent inflammatory cytokine regulating myeloid cell differentiation, hematopoiesis, and various other functions. It is functionally associated with a number of inflammatory pathologies including rheumatoid arthritis and inflammatory bowel disease. GM‐CSF has been found to promote NLRP3‐dependent IL‐1β secretion, which may have a significant role in driving inflammatory pathologies. However, the molecular mechanisms remain unknown. Here, we show that GM‐CSF induces IL‐1β secretion through a ROS‐dependent pathway. TNF is required for reactive oxygen species (ROS) generation that strikingly does not promote NLRP3 activation, but instead drives ubiquitylation of IL‐1β, promoting its cleavage through basal NRLP3 activity. GM‐CSF regulates this pathway through suppression of antioxidant responses via preventing upregulation of NRF2. Thus, the pro‐inflammatory effect of GM‐CSF on IL‐1β is through suppression of antioxidant responses, which leads to ubiquitylation of IL‐1β and enhanced processing. This study highlights the role of metabolic regulation of inflammatory signaling and reveals a novel mechanism for GM‐CSF to promote inflammation.

▸ Figure EV2. Quantification of IL-1b mRNA and protein levels.
A HoxB8 macrophages were treated for 12 h as indicated. The RNA was extracted and analyzed by qPCR for IL-1b. B Densitometric quantification of pro-IL-1b protein level relative to LPS-treated samples. C HoxB8 macrophages were co-treated with 10 ng/ml recombinant M-CSF for 16 h as indicated. Media were analyzed for IL-1b secretion by ELISA. D HoxB8 macrophages were treated as indicated using 60 ng/ml UP-LPS and or 1% of GM-CSF supernatant. IL-1b secretion was analyzed in the supernatant by ELISA. E Densitometric quantification of pro-IL-1b protein level in the indicated HoxB8 genotypes treated for 16 h with LPS in the presence or not of GM-CSF. Shown are representative western blots of pro-IL-1b (right panel). Black vertical lines indicate non-relevant lanes that were removed during figure preparation. All samples were run on the same gel. F HoxB8 macrophages were treated with the indicated concentrations of either media from GM-CSF producing cells, E. coli-expressed recombinant GM-CSF, or HEK293expressed recombinant GM-CSF plus LPS. Supernatants were analyzed for IL-1 b by ELISA. G Densitometric quantification of pro-IL-1b protein level relative to LPS-treated samples. HoxB8 macrophages were treated as indicated for 16 h. Data correspond to Fig HoxB8 macrophages or BMDMM were treated for 16 h as indicated. The indicated cytokines were analyzed in supernatants by LEGENDplex antiviral response assay as indicated in the Materials and Methods. Shown are the concentrations (pg/ml) converted into log 10 in order to be following a normal Gaussian distribution for statistical analysis. Data information: n ≥ 3 biological replicates (every dot represents one biological replicate). For all panels, error bars are SEM. Significance was calculated using one-way ANOVA with multiple comparisons. P-values are indicated. Figure EV4. GM-CSF-induced IL-1b secretion is blocked by IL-10 but does not require IRG1 dysregulation.

A B C
A BMDM were generated as described in Materials and Methods and treated as indicated for 16 h. IL-1b levels were measured in media by ELISA. B HoxB8 macrophages were treated as indicated with or without recombinant IL-10 and secreted IL-1b levels were measured using ELISA. C Wild-type and Tnf À/À HoxB8 macrophages were treated as indicated for 12 h, and RNA was extracted. Levels of IRG1 were analyzed by qPCR.
Data information: n ≥ 3 biological replicates (every dot represents one biological replicate). For all panels, error bars are SEM. Significance was calculated using two-way ANOVA with multiple comparisons. P-values are shown.
A B D C Figure EV5. GM-CSF-regulated NRF2 stability independently of GSK3b.
A HoxB8 macrophages were treated for 12 h with LPS or LPS + GM-CSF and RNA was extracted and analyzed by qPCR for NRF2 mRNA. B HoxB8 macrophages were treated for 12 h AE GM-CSF. MG132 was added for 1 h prior to cells being lysed and analyzed for levels of NRF2 by western blot. C HoxB8 macrophages were treated for the indicated times with LPS or LPS + GM-CSF and proteins extracted and western blots made against KEAP1. High-molecularweight bands represent oxidized KEAP1. D HoxB8 macrophages were treated for 12 or 14 h with LPS or LPS + GM-CSF. Proteins were analyzed by western blot for levels of phospho-GSK3b.
Data information: n ≥ 3 biological replicates (every dot represents one biological replicate). Error bars are SEM and P-values were calculated using unpaired t-tests. Westerns are representative images from three biological replicates. Source data are available online for this figure.