Baricitinib attenuates the proinflammatory phase of COVID-19 driven by lung-infiltrating monocytes

Abstract

SARS-CoV-2-infected subjects are generally asymptomatic during initial viral replication but may suffer severe immunopathology after the virus has receded and monocytes have infiltrated the airways. In bronchoalveolar lavage fluid from severe COVID-19 patients, monocytes express mRNA encoding inflammatory mediators and contain SARS-CoV-2 transcripts. We leverage a human small airway model of infection and inflammation, whereby primary blood monocytes transmigrate across SARS-CoV-2-infected lung epithelium to characterize viral burden, gene expression, and inflammatory mediator secretion by epithelial cells and monocytes. In this model, lung-infiltrating monocytes acquire SARS-CoV-2 from the epithelium and upregulate expression and secretion of inflammatory mediators, mirroring in vivo data. Combined use of baricitinib (Janus kinase inhibitor) and remdesivir (nucleoside analog) enhances antiviral signaling and viral clearance by SARS-CoV-2-positive monocytes while decreasing secretion of proneutrophilic mediators associated with acute respiratory distress syndrome. These findings highlight the role of lung-infiltrating monocytes in COVID-19 pathogenesis and their importance as a therapeutic target.

Keywords: CP: Immunology; CP: Microbiology; Janus kinase; RNA-seq; cytokine release syndrome; inflammation; interferon; therapy.

Graphical abstract

Figure 1

HLE-ALI cells grown at the air-liquid interface can be infected by OC43, SARS-CoV-2, and influenza and exhibit different transcriptional responses (A) Experimental scheme illustrating the infection of human lung epithelium (HLE) differentiated at the air-liquid interface (ALI) with influenza A strain A/PR8/1934 (IAV), betacoronaviruses OC43, or SARS-CoV-2, isolate USA-WA1/2020, for 24–72 h. (B) Extracellular ATP generated by HLE-ALI cells infected with IAV, OC43, or SARS-CoV-2 at an MOI of 0.1 for 24–72 h was measured using a luciferase assay. Plotted is the median with interquartile range. No statistics were calculated due to low sample number. (C) Multiplexed qRT-PCR of HLE-ALI cells infected with either no virus or influenza A strain A/PR8/1934, betacoronaviruses OC43, or SARS-CoV-2, isolate USA-WA1/2020, for 24–72 h at MOIs of either 0.01, 0.1, or 1. Data were normalized using the delta delta Ct method relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the no virus condition at time 0, and then Z score normalized. Values shown are the average of biological triplicates.

Figure 2

SARS-CoV-2 prevents induction of an antiviral response in differentiated human lung epithelium HLE-ALI cells were infected at an MOI of 0.1 with IAV, OC43, or SARS-CoV-2 for 48 h, after which total RNA was extracted and library prepared following the protocol of the TruSeq RNA Sample Preparation Kit. (A) Volcano plots of differentially expressed genes (DEGs) in HLE-ALI cells infected with IAV, OC43, and SARS-CoV-2 compared with uninfected cells. (B) Gene ontology enrichment profiles of DEGs in HLE-ALI cells infected with IAV, OC43, or SARS-CoV-2. Gene names associated with GO terms are in Table S2C. Heatmaps of targeted analysis of genes associated with the viral response or cellular process pathways. Gene names associated with the cellular process terms are in Table S2, row 9, and raw expression values in Table S1.

Figure 3

Lung-recruited primary human monocytes induce proinflammatory signaling after transmigration across a SARS-CoV-2-infected epithelium (A) Schematic diagram of the monocyte transmigration process: HLE-ALI cells were differentiated at ALI for two weeks and infected with SARS-CoV-2 at an MOI of 0.1 for 48 h, after which 10⁶ monocytes were transmigrated across the infected epithelium toward leukotriene B4 (LTB4) (100 nM) and C-C motif chemokine ligand 2 (CCL2) (250 pg/mL) for 24 h, in the absence or presence of remdesivir (1 μM) and/or baricitinib (1 μM). All conditions contained 0.1% v/v DMSO. (B) Heatmaps of selected genes from each treatment condition. Monocyte mRNA was analyzed by multiplexed qRT-PCR and first normalized to GAPDH using the delta delta Ct method and then Z score normalized.

Figure 4

scRNA-seq of BALF from patients hospitalized with mild or severe COVID-19 shows that monocytes harbor SARS-CoV-2 genomes and express CXCL8 and IL-1β among other chemokines (A) Combined Uniform Manifold Approximation and Projection (UMAP) plots of scRNA-seq from n = 3 mild patients and n = 3 severe patients. Seurat was used to normalize gene barcodes and generate UMAP clustering plots. Expression values for SARS-CoV-2, CD14, CXCL8, and IL-1β were overlaid onto the UMAP plot. (B) DEGs in cells identified as epithelial cells or monocytes between the severe and mild patient groups were investigated. Enriched pathways for each cell type were plotted based on whether they were up- or down-regulated (severe versus mild). The top 20 up- and downregulated genes were listed for both cell types and plotted as a heatmap.

Figure 5

Lung-infiltrating monocytes release inflammatory mediators and harbor replicative SARS-CoV-2 (A) Inflammatory mediators in the apical fluid following transmigration were quantified by a multiplexed electrochemiluminescent assay. The Z score for each mediator in each condition was calculated and plotted. (B) Transmigration efficiency of monocytes was calculated by dividing the number of monocytes in the apical fluid after 24 h by the input number of cells. (C–E) RNA was extracted from each component of the model (epithelium, lung-infiltrating monocytes, and extracellular fluid) and reverse transcribed. Total SARS-CoV-2 genome copies were calculated. (F) The sum of each of the components of the model was calculated to depict the total amount of virus remaining in the system after monocytes were allowed to transmigrate for 24 h. All statistics were calculated using the Mann-Whitney U-test in Prism between the “no drug” and each treatment group. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001. Shown are median and interquartile range.

Figure 6

Treatment with baricitinib and remdesivir increases the rate of viral clearance in lung-infiltrating monocytes (A) Schematic of the experimental setup to quantify replication of the virus in monocytes. HLE-ALI cells were infected with SARS-CoV-2 at an MOI 0.1 for 48 h, after which 10⁶ monocytes were transmigrated across the infected epithelium toward LTB4 (100 nM) and CCL2 (250 pg/mL) for 24 h, in the absence/presence of remdesivir (1 μM) and/or baricitinib (1 μM). Monocytes were washed and purified by negative depletion using anti-CD326 beads to remove contaminating epithelial cells and placed into medium with no drug, remdesivir (1 μM), and/or baricitinib (1 μM) for 0–72 h. All conditions contained 0.1% v/v DMSO. (B) Quantification of total SARS-CoV-2 genome copies in the monocytes. (C) Quantification of the amount of virus in the extracellular fluid at each time point. (D) Quantification of the N-subgenome in monocytes at each time point. Data were normalized to 18S rRNA in the untreated condition at 0 h using the delta delta Ct method. (E) The extracellular fluid was layered onto VeroE6 cells to perform a plaque assay from which a TCID50 was calculated. The positive control was the direct application of 2.5 × 10⁴ genome copies of SARS-CoV-2 to the cells. (F–H) Inflammatory mediators were measured using an electrochemiluminescent assay. All statistics were calculated with a two-way ANOVA, main effects model in Prism with Geisser-Greenhouse correction applied. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001. Shown are median and interquartile range.