Ethics statement

The study was approved by the Ethics Committee of the Hospital District of Helsinki and Uusimaa (HUS/853/2020, HUS/1238/2020). All volunteers gave a written informed consent, in accordance with the Declaration of Helsinki. For animal experiments, experimental procedures were approved by the Animal Experimental Board of Finland (license number ESAVI/28687/2020).

Patient population

Adult clinical patients with confirmed COVID-19 (RT-PCR positive for SARS-CoV-2) at Helsinki University Hospital (HUH) (hospitalized: n = 34; outpatients: n = 8) were enrolled in the present study. Blood samples were collected during hospitalization for the severe COVID-19 group, and after confirmation of diagnosis for the mild COVID-19 outpatient group. Samples for RNA sequencing were collected in 2020 and representing infections by the original and early SARS-CoV-2 variants, whereas samples for ex vivo culture experiments were collected in 2021–2022 likely representing infections by omicron subvariants of SARS-CoV-2. As controls, healthy blood donors were included for RNA sequencing (n = 7, age 57 ± 7, male/female 3/4) and ex vivo culturing experiments (n = 9, age 38 ± 14, male/female 4/5). For clinical correlation analysis, severe COVID-19 patients were further categorized by severity based on their need for hospitalization and oxygen supplementation, as described previously [7]. For each patient, medical history and clinical data were collected through retrospective patient record review and are presented for the severe COVID-19, hospitalized patients in Table 1 and as previously described [7]. Calprotectin was measured from serum (diluted 1:1000) by ELISA, according to the manufacturer’s protocol (calprotectin/S100A8 DuoSet kit, R&D systems).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Clinical parameters of hospitalized patients (n = 34).

https://doi.org/10.1371/journal.ppat.1012368.t001

The World Health Organization (WHO) Ordinal Scale for clinical improvement is a tool designed specifically to assess and measure the progression and clinical improvement of patients [67]. COVID-19 scoring: 1 = no limitations of activity, 2 = limitations of activity, 3 = no oxygen therapy, 4 = oxygen by mask or nasal cannulae, 5 = non-invasive ventilation or high-flow oxygen, 6 = invasive mechanical ventilation without other organ support, 7 = invasive mechanical ventilation with other organ support, 8 = dead. The baseline score represents the timepoint of the first laboratory sample taken, serving as a reference point for measuring improvement and establishing a starting point for comparison. In contrast, the worst score represents the most severe or critical state of the disease. None of the patients whose samples were used for RNA-seq underwent corticosteroid treatment.

Isolation of granulocytes from human blood

Blood samples from COVID-19 patients and healthy controls (HC) were collected in EDTA vacutainer tubes and transported to the laboratory. Peripheral blood mononuclear cells (PBMCs) or polymorphonuclear cells (PMNs) were isolated from whole blood by density gradient centrifugation using either Ficoll-Paque Plus (GE Healthcare) or Polymorphprep (Axis-Shield) respectively, following standard procedures including the use of 2 mM EDTA in PBS and red blood cell lysis with ACK lysis buffer (Lonza by Thermo Fisher). Subsequently, isolation of CD66⁺ granulocytes (low-density granulocytes, LDGs) from the PBMC fraction was performed using the CD66abce MicroBead Kit (Miltenyi Biotec, Germany) with an MS column, according to the manufacturer’s instructions. Both the positively selected CD66⁺ LDGs and the isolated PMNs were then washed and counted, using a TC20 Automated Cell Counter (Bio-Rad Laboratories, Inc.) with trypan blue staining for dead cell exclusion. All described procedures in this section were done at room temperature. An aliquot of cells was lysed in Trizol reagent (Thermo Fisher Scientific, USA) and stored at –80°C for later extraction of total RNA and subsequent RNA sequencing (RNA-seq) analysis.

Caspase1 activity

Caspase1 activity was assessed in isolated cells after 2 h of culture (1 million cells/ml) using the caspase-Glo1 inflammasome assay (Promega) according to the manufacturer’s protocol, with 2.5 μM nigericin (Invivogen) treatment as the activator. The resulting luminescence was measured by a Hidex Sense microplate reader (Hidex).

As another approach, caspase1 activity in isolated cells (1 million cells/ml) was measured by the fluorescent dye FAM-FLICA (Bio-Rad Laboratories). Cells were incubated with FAM-FLICA according to manufacturer’s recommendations for 30 min in culture medium at 37°C, after which cells were washed with PBS and analyzed by LSRII cytometer (BD Biosciences). Data was acquired with BD FACSDiva version 8.0.1 (BD Biosciences) software and further analysis was performed with the FlowJo software v10 (BD Biosciences).

Soluble factor stimulation assays

Isolated granulocytes from HC and COVID-19 patients were cultured at 2 million cells/ml in RPMI 1640 supplemented with 10% fetal bovine serum (R10) at 37°C. Cells were primed (1^st signal) with either LPS (20 ng/ml, Sigma Aldrich) or IFN-I (combination of 2.7*10⁴ IU/ml IFN-α and IFN-β, Immunotools) for 4 h, followed by activation (2^nd signal) by 2.5 μM nigericin or monosodium urate crystals (MSU, 100 μg/ml, Invivogen) for an additional 4 h. For the 24 h stimulation experiments, nigericin was added to the cultured cells, in the presence or absence of inflammasome inhibitors MCC995 (2 μg/ml) and Ac-YVAD-FMK (20 μg/ml, both from Invivogen). Cells were pelleted by centrifugation at 400 G for 5 min and stored in Trizol at –80°C for later RNA extraction whereas supernatants were used to measure IL-1β, IL-18, myeloperoxidase (MPO) and IL-8 by ELISAs according to the manufacturer’s protocols (DuoSet kits from R&D systems). LDH was measured in supernatants using Cyquant LDH cytotoxicity assay (ThermoFisher). Where indicated, priming and activation was performed in the presence of 100 μg/ml of anti-human IFNAR1 (Anifrolumab Biosimilar) or mouse IgG1 as control (both from Bio-X-Cell, New Hampshire, USA).

HL-60 cells (ATCC #CCL-240) were activated similarly to neutrophils after a 5-day differentiation period induced by 1% DMSO.

Virus propagation

The SARS-CoV-2 hCoV-19/Finland/THL-202117309/2021 (delta strain B.1.617.2) and the mouse-adapted strain MaVie [18] were propagated in VeroE6-TMPRSS2 cells (kidney epithelial cells expressing the transmembrane protease serine 2) [19] grown in DMEM supplemented with 10% inactivated FCS, 100 IU/mL Penicillin, 100 μg/mL Streptomycin and 2 mM L-glutamine at 37°C. The virus was purified from supernatants by ultracentrifugation (SW28 rotor, 27,000 rpm, 90 min, +4°C) through a 0.22 μm-filtered 30% ultra-pure sucrose cushion (in PBS), to obtain virus preparations free of cell culture contaminants. Virus titers were calculated by the median tissue culture infectious dose (TCID50) after assessing cytopathic effects by crystal violet staining of cell cultures infected for 5 days with serially diluted virus.

RNA sequencing

Neutrophils isolated from different cohorts comprised three PMN groups (severe COVID-19, mild COVID-19, and healthy controls), and one LDG group (given that these cells were rare in mild COVID-19 patients and HC, only LDGs from patients with severe COVID-19 were included).

cDNA synthesis from total RNA was performed according to Takara SMARTseq v4 Ultra-low input RNA kit for Sequencing user manual (Takara Bio, Mountain View, CA, USA) followed by Illumina Nextera XT Library preparation according to Illumina Nextera XT Reference Guide (Illumina, San Diego, CA, USA). UDI index setup was used for the Nextera XT libraries. Library quality check was performed using LabChip GX Touch HT High Sensitivity assay (PerkinElmer, USA) and libraries were pooled based on the concentrations acquired from the assay. The pooled libraries were quantified for sequencing using KAPA Library Quantification Kit (KAPA Biosystems, Wilmington, MA, USA) and sequenced on the Illumina NovaSeq6000 system for 200 cycles using S1 flow cell (Illumina, San Diego, CA, USA). Read length for the paired-end run was 2x101 bp. The human RNA-seq data are deposited at the GEO (accession number GSE#272381).

RNA data analysis

Principal Component Analysis (PCA) and enrichment analyses were obtained using ExpressAnalyst [20]. Briefly, PCA was performed to identify patterns in the data and reduce the dimensionality of the dataset, where the top principal components were selected based on the percentage of variance explained. For enrichment analyses, Gene Set Enrichment Analysis (GSEA) and Over-Representation Analysis (ORA) were performed on the top 5000 DE genes identified by DESeq2 (adjusted P value < 0.05, log2FC >1) [20]. GSEA was used to identify enriched signaling pathways using the Reactome database, while ORA was used to identify enriched pathways using the KEGG database. The resulting p-values were corrected for multiple testing using the Benjamini-Hochberg method, and pathways with a corrected p-value <0.05 were considered significant.

To visualize the expression patterns of the DE genes, the data was analyzed using the AltAnalyze software [21], which selected the top 118 genes based on correlation and determined the heatmap clustering, using the Euclidean distance metric and the complete linkage method. Then, the obtained heatmap was re-generated using heatmapper.ca [22] for better visualization.

CIBERSORTx, a machine learning algorithm that infers cell type proportions using a reference gene expression matrix of known cell types was used to perform RNA-seq deconvolution on the gene expression data to estimate the abundance of immune cell types in the samples [23]. The signature matrix used was taken from Lasalle et al. [8]. This reference matrix made use of a published whole-blood single-cell dataset [9], and included the main immune cell types: monocytes, NK cells, T lymphocytes, B lymphocytes, plasmablasts and neutrophils, the latter subclassified into mature and immature. The smaller subsets of granulocytes (eosinophils and basophils) are not considered separately and are most likely categorized as neutrophils in the bulk data deconvolution. Nonetheless, the resulting cell type proportions were used to compare the immune cell composition between groups.

Additionally, the determination of sample purity (>65% identified as neutrophils) served as a limiting parameter for the visualization of differentially expressed inflammasome related genes from the RNA sequencing results, which were selected and graphed in a heatmap using heatmapper.ca [22], clustered by complete linkage and ordered by Spearman’s rank.

For the GSEA of the reanalyzed RNA-seq data from LaSalle et al. [8], we used the fgsea R package using MSigDB pathway sets, as specified in S1 Table.

Volcano plots

To visualize differentially expressed (DE) genes between groups from human and mice RNA-seq results previously identified by DESeq2, a volcano plot was generated using GraphPad Prism. Genes with a P-adjusted value (padj or FDR) <0.05 were considered significant. Similarly, RNA sequencing data from GSE93996 [24] was reanalyzed, and all DE genes in ATRA-differentiated HL-60 cells were visualized in a volcano plot.

Single cell transcriptomics data analysis

This study made use of the “COVID-19 Immune Atlas: integration of 5 public COVID-19 PBMC single-cell datasets” available online [25]. This standardized data collection contains cells from different assays (10x 3’ v2, 10x 3’ v3, 10x technology and Seq-Well) and consists of a total of 239,696 cells from the peripheral blood, 3,693 of which are neutrophils. These neutrophils were further subclassified as mature (59%) and immature (41%), based on the immune atlas predetermined cell classes. This was confirmed by a CD16b expression in mature neutrophils, and a higher CD66b expression in the immature population. This data was obtained from and analyzed in the Chan Zuckerberg CELLxGENE platform [25].

Reverse transcription and quantitative PCR (RT-qPCR) for human selected human genes

Total RNA was extracted from unstimulated or ex vivo stimulated PMNs using the Trizol reagent (Invitrogen, USA) according to the manufacturer’s protocol. Subsequently, cDNA synthesis was performed using the RevertAid RT Reverse Transcription Kit (Thermo Scientific, USA) as per the manufacturer’s instructions. Quantitative PCR (qPCR) was performed using the Stratagene model (Agilent Technologies) and SYBR Green/ROX master mix (Thermo Scientific, USA). The primer sequences for qPCR are presented in S2 Table.

Primer specificity was confirmed using melting curve analysis and dissociation curves. The relative expression levels of the genes of interest were calculated using the 2-ΔΔCT method and normalized to the expression of the housekeeping gene GAPDH. Baseline gene expressions of unstimulated samples were statistically assessed using the Mann-Whitney test, while the two-way ANOVA Tukey’s multiple comparisons test was performed for the ex vivo stimulated samples.

Mouse infections

Female BALB/c mice (Envigo, Indianapolis, IN, USA; 7 to 8 weeks, n = 36 in total) were transferred to the University of Helsinki biosafety level-3 (BSL-3) facility and acclimatized to individually ventilated biocontainment cages (ISOcage; Scanbur, Karl Sloanestran, Denmark) for 7 days with ad libitum water and food (rodent pellets). For subsequent experimental infection, the mice were placed under isoflurane anesthesia and inoculated intranasally with 50 μL of SARS-CoV-2 MaVie strain (5*10⁵ TCID50/animal) or PBS (mock-infected control). Daily weighting of all mice was performed, and their well-being was carefully monitored for signs of illness (e.g., changes in posture or behavior, rough coat, apathy, ataxia). Euthanasia was performed by cervical dislocation under terminal isoflurane anesthesia. All animals were dissected immediately after euthanasia, and the lungs were sampled for multiple downstream analyses. The infections were performed as 4 separate experiments (exp): 1) Exp 1 included 8 mice infected with MaVie and 4 mock infected mice. At 2 days post infection (dpi), 4 infected and the mock infected mice were euthanized; the remaining infected mice were euthanized at 4 dpi. The right lung was sampled for virus-specific RT-qPCR (1/5) and neutrophil isolation (4/5), the left lung was fixed for histological and immunohistochemical examination. 2) Exp 2 included 8 infected and 4 mock infected mice of which half were euthanized at 2 dpi and 4 dpi, respectively. From these mice, both lung lobes were subjected to neutrophil isolation. 3) Exp 3 included 8 mice that were infected and immediately inoculated intraperitoneally with 250 μg of anti-mouse IFNAR-1 (n = 4) or mouse IgG1 control (n = 4) (Bio-X-Cell, New Hampshire, USA), and 4 mock-infected animals. All mice were euthanized at 2 dpi. Each 1/5 of the left lobe was processed for virus-specific RT-qPCR and histology/immunohistochemistry, respectively. The remaining 4/5 of the lungs served for neutrophil isolation. 4) Exp 4 included 4 animals in the following 5 groups: PBS-inoculated controls, infected animals euthanized at 2 dpi and at 4 dpi respectively, infected animals euthanized at 2 dpi with intraperitoneal injection of control IgG or anti-mouse IFNAR-1. Each 1/5 of the left lobe was processed for virus-specific RT-qPCR and histology/immunohistochemistry, respectively. The remaining lung tissue served to prepare single cell suspensions. Each 1 million cells were subjected to neutrophil quantification or caspase1 activity measurement by flow cytometry, the rest (approx. 8 million cells) to neutrophil isolation.

Virus titration from mouse lungs

Supernatants of single cell suspensions from Exp 4 were used for virus titration by fluorescent focus forming unit (FFU)-based assay, in which Vero E6 cells were incubated with serially diluted supernatants for 24 hours at 37°C in growth medium. Cells were fixed with 4% PFA for 10 min, blocked and permeabilized with PBS containing 3% BSA and 0.2% TritonX-100 for 15 min and incubated with rabbit anti- SARS-CoV-2 receptor binding domain (RBD) [19] for 1 hr, followed by anti-rabbit AlexaFluor488 conjugated secondary antibody (Thermo Scientific) for 1 hr. Fluorescence was observed with Zoe fluorescence imager (Bio-Rad laboratories) and the highest dilution not showing any RBD positive cells considered as the virus titer.

Neutrophil quantification and caspase1 activity measurement from mouse lungs by flow cytometry

Single cell suspensions from mouse lungs of Exp 4 were subjected to flow cytometric quantification of Ly6G+ neutrophils. Cells were initially incubated with BV605-conjugated yellow live/dead dye (Thermo Scientific) for 15 min before addition of 1% FBS and a cocktail of antibodies recognizing CD3+ and CD19+ lymphocytes (FITC-conjugated clones 145-2C11 and 1D3, respectively, from Immunotools), Ly6G+ neutrophils (PE-Cy7-conjugated clone 1A8 from BD biosciences) and CD11b (APC-conjugated clone M1/70.15 from Immunotools). After incubation for 30 min at RT, cells were fixed with 2% paraformaldehyde for 30 min and washed with PBS. In parallel, single cell suspensions were stained with FAM-FLICA together with Ly6G antibody and 30 min incubations were performed in R10 at 37°C before fixation. Finally, all cells were subjected to flow cytometric analysis with a three-laser (Blue/Red/Violet lasers) 14-color Fortessa LSRII cytometer (BD Biosciences). Data was acquired with BD FACSDiva version 8.0.1 (BD Biosciences) software and further analysis was performed with the FlowJo software v10 (BD Biosciences).

Neutrophil isolation from mouse lungs

Neutrophil isolation was performed from the lungs of all mice. The dissected lung tissue was chopped into small pieces using scissors and enzymatically digested with a cocktail of Liberase (50 ug/ml; Roche #05401020001 from Merck) and DnaseI (100 ug/ml; Roche #11284932001 from Merck) in RPMI-1640 for 30 min at 37°C. The resulting homogenate was diluted 10-fold in R10 and passed through a 70 μm Cell strainer (Pluriselect) to obtain a single-cell suspension. Neutrophils were isolated by positive selection using Ly6G-binding magnetic beads and MS columns according to the manufacturer’s recommendations (Miltenyi Biotec). Neutrophils were isolated with a purity exceeding 95% based on flow cytometry analysis of Ly6G expression. Where indicated, isolated neutrophils were attached to glass slides through cytospin (800 g, 5 min), fixed with 4% PFA for 10 min and the nuclei stained with Hoechst33342.

RNA sequencing of mouse neutrophils

Mouse neutrophils from Exp 1 were isolated, lysed in Trizol (Thermo Scientific) and the RNA extracted in the liquid phase using chloroform. RNA isolation was carried out using the Rneasy micro kit (Qiagen). Isolated RNA (1 ng) underwent whole transcriptome sequencing with ribodepletion. Briefly, RNA sequencing was performed using the Illumina Stranded with RiboZero library preparation method. Sample quality and integrity were assessed using TapeStation RNA analysis. Sequencing was conducted on the Illumina NextSeq platform, followed by standard bioinformatics analysis for gene expression quantification.

The service was provided by the Biomedicum Functional Genomics Unit at the Helsinki Institute of Life Science and Biocenter Finland at the University of Helsinki. The mouse RNA-seq data are deposited at the GEO (GSE271808).

RT-qPCR of mouse samples

RNA was extracted from dissected lung samples (1/10 of the whole lung) as well as isolated neutrophils of mice in Exp 1, 3 and 4 using Trizol (Thermo Scientific) following the manufacturers’ instructions. The isolated RNA was directly subjected to one-step RT-qPCR analysis based on a previously described protocol using primer-probe sets detecting the viral genome encoding for the RNA-dependent RNA polymerase (RdRp [26], subgenomic E [27] as well as mouse OasL2, caspase1, IL1b and GAPDH (Applied biosystems #Mm01336189_m1, #Mm00438023_m1, #Mm00434228_m1 and #Mm99999915_g1 respectively, Thermo Scientific). The PCRs were performed with TaqPath 1-step master mix (Thermo Scientific) using AriaMx instrumentation (Agilent, Santa Clara, CA, USA).

Histology and immunohistochemistry

From animals in Exp 1, 3 and 4 the whole left lung (Exp 1) or 1/5 of the left lung (Exp 3 and 4) were trimmed for histological examination and routinely paraffin wax embedded. Consecutive sections (3 μm) were prepared and routinely stained with hematoxylin-eosin (HE) or subjected to immunohistochemistry (IHC) for the detection of SARS-CoV-2 nucleoprotein (NP) [28] and Ly6G (neutrophil marker); for Exp 3, a further section of the infected lungs was stained for histone H3 (NET marker) [29]. All stains followed previously published protocols [30].

Morphometric analyses

For quantification of SARS-CoV-2 antigen expression and the extent of neutrophil influx into the lungs, a morphometric analysis was undertaken on the slides stained for SARS-CoV-2 NP and Ly6G, respectively. The stained slides were scanned using NanoZoomer 2.0-HT (Hamamatsu, Hamamatsu City, Japan), and several sections of the lung of each animal were quantitatively analysed using the Visiopharm 2022.01.3.12053 software (Visiopharm, Hoersholm, Denmark). The average total tissue area used for quantification was 19.5 ± 6 mm². The morphometric analysis served to quantify the area, in all lung sections of an animal, that showed immunostaining for viral NP and Ly6G, respectively. In Visiopharm, for each section, the lung was manually outlined and annotated as a Region Of Interest (ROI), manually excluding artifactually altered areas. The manual tissue selection was further refined with an Analysis Protocol Package (APP) based on a Decision Forest classifier, with the pixels from the ROI being ultimately classified as either “Tissue” or “Background”. This new “Tissue” ROI, regrouping the different lung samples analysed for each animal, was further quantified by executing two APPs successively. The first APP was based on a Threshold classifier and served to detect and outline areas with immunostaining. The second APP then measured both the surface of the immunostained area (μm²) and the surface of the “Tissue” ROI (μm²). The percentage of immunostained area (%), expressed as the ratio between the immunostained area and the total area, was obtained for each animal in Excel (Microsoft Office 2019; Microsoft, Redmond, Washington, United States), according to the following formula: ([positive area (μm2)]/ [total area (μm2)]) x 100.

Statistical analyses

Statistical analysis was performed using GraphPad Prism 8.3 software (GraphPad Software, San Diego, CA, USA) and R software v3.6.3 (R core team). Statistically significant correlations between parameters were assessed by calculating Spearman’s correlation coefficients, and differences between groups were assessed with Mann-Whitney, Kruskall-Wallis or ordinary one-way or 2-way ANOVA tests, depending on sample distribution and the number of groups analyzed. To elaborate, nonparametric tests like Mann-Whitney and Kruskall-Wallis were employed when the data violated assumptions of normality, while ANOVA tests were applied when the data met parametric assumptions.

Unsupervised RNA-seq analysis reveals an antiviral gene expression signature of circulating neutrophils in COVID-19 that is strongly influenced by maturity

Different neutrophil subsets have gained a lot of attention as modulators of COVID-19 pathogenesis. We recently found increased frequencies of one neutrophil subset, referred to as low-density granulocytes (LDGs, isolated from the PBMC fraction), during COVID-19 [7]. In the current study we sought to understand in more detail the transcriptomic profile of different neutrophil subsets, including LDGs and their “normal” density counterpart, the circulating polymorphonuclear cells (PMNs) [31], which typically consists of mainly mature neutrophils. Neutrophils isolated from different cohorts comprised three PMN groups (severe COVID-19, mild COVID-19, and healthy controls), and one LDG group. Initial deconvolution of the RNA sequencing (RNA-seq) data allowed us to gain a comprehensive understanding of the cellular composition within PMN and LDG fractions and verified that most cells present in the samples were neutrophils (S1A Fig). This analysis also demonstrated that cells in the LDG fraction were predominantly immature neutrophils, meanwhile PMNs were composed of mainly mature neutrophils.

The samples with predominant neutrophil cell populations were selected for subsequent gene expression analysis (neutrophils ≥ 65%). The high variance in gene expression between PMNs and LDGs was confirmed by principal component analysis (PCA) (Fig 1A), which revealed that the gene expression patterns of COVID-19 LDGs differed from those of all PMNs regardless of the patients’ disease state. Functional enrichment analyses through gene overrepresentation (ORA) and gene-set enrichment analyses (GSEA) (Fig 1B) compared PMNs with LDGs from severe COVID-19 patients. The most statistically significant result was an overrepresentation of the NOD-like receptor signaling pathway in PMNs in contrast with LDGs, highlighting that the different neutrophil fractions have a distinct inflammatory profile. This was supported by GSEA, where the most obvious increases in fold changes were the enrichment of the interferon signaling pathways. Another relevant difference was the cell cycle and DNA replication pathways, identified by both ORA and GSEA, which supported our previous findings suggesting LDGs to be predominantly immature cells [7]. Furthermore, a heatmap of selected type I IFN (IFN-I) related genes confirmed a robust IFN-I gene signature in severe COVID-19 PMNs, while LDGs from severe COVID-19 distinctively lacked this signature (Fig 1C). Unsupervised clustering analysis, namely Iterative Clustering and Guide Gene Selection (ICGS) using the AltAnalyze software, supported these findings by identifying the top 118 differentially expressed (DE) genes, including several IFN-related genes (S1B Fig). Similarly to the selected samples included in Fig 1, this analysis classified the samples into two major clusters: a first one containing all isolated LDG samples, and a second one comprising all isolated PMN samples. The former cluster consisted of neutrophil antimicrobial and granule marker genes (e.g. DEFA3, DEFA4, SERPINB10, CTSG), while in the latter cluster the most significantly upregulated genes in the PMNs from severe COVID-19 subgroup were mainly interferon inducible (e.g. IFI44L, IFI6, GBP3, IRF7). These differences were supported by a detailed gene analysis (S2A Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Increased IFN-I related gene expression in mature COVID-19 neutrophils.

The analysis was reduced to include only the samples with the highest purity (cell fraction over 0.65 of neutrophils), as identified by CIBERSORTx. (A) Principal component analysis (PCA) of the RNA-seq samples (n = 7 PMNs from HC, n = 10 PMNs from severe COVID-19, n = 8 PMNs from mild COVID-19, and n = 6 LDGs from severe COVID-19). (B) Ridgeline diagrams depicting the top 20 enriched signal pathways from the genes differentially expressed by PMNs versus LDGs during severe COVID-19: overrepresentation analysis (ORA) using KEGG database and gene-set enrichment analysis (GSEA) according to Reactome database. Both enrichment analyses were made using ExpressAnalyst and are sorted by P-value, obtained from Welch’s t-test. (C) Heatmap of differentially expressed IFN-related genes in COVID-19 PMNs and LDGs as compared to HC PMNs. RNA sequencing was performed on purified PMNs from healthy controls, mild COVID-19 and severe COVID-19, as well as LDGs from severe COVID-19. The heatmap was clustered by complete linkage and ordered by Spearman’s rank. FC = fold change.

https://doi.org/10.1371/journal.ppat.1012368.g001

Increased expression of inflammasome related genes in severe COVID-19 PMNs

In addition to the strong IFN-I signature, PMNs of severe COVID-19 upregulated several genes involved in inflammatory processes, such as the formation of inflammasomes. We further analyzed the differential expression of selected inflammasome related genes across all samples using RNA-seq (Fig 2A). PMNs of severe COVID-19 displayed higher levels of inflammasome genes such as NLRP3 and caspases 1, 4 and 5. LDGs did not display similar upregulation of inflammasome genes, with the notable exception IL-18 and NLRC4, which were not upregulated by PMNs. These findings prompted us to look more closely into PMN fractions between different disease states. Pathway analyses identified the inflammasome related NOD-like and RIG-like receptor signaling pathways among the most significantly overrepresented pathways, differentially expressed in severe COVID-19 PMNs versus HC PMNs (Figs 2B and S2B and S2C) or mild COVID-19 PMNs (Figs 2C and S2D and S2E). However, mild COVID-19 PMNs did not significantly differ from HC PMNs in their inflammatory profile (S2F Fig). The increased expression of selected IFN-I (OAS1, OAS2, and IFIT1) and inflammasome related genes (CASP1, CASP5, NLRC5 and NAIP) between COVID-19 and HC PMNs was confirmed by RT-qPCR. However, some inflammasome related genes (IL-1β, NLRP3 and NLRC4) were seemingly downregulated, although not statistically significant (S3 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Inflammasome related gene expression.

(A) Heatmap depicting selected differentially expressed inflammasome related genes from RNA sequencing performed in PMNs from HC, mild and severe COVID-19, as well as severe COVID-19 LDGs. Only the samples with the highest purity, determined by a cell fraction over 0.65 of neutrophils (identified by CIBERSORTx) are included. The heatmap was clustered by complete linkage and ordered by Spearman’s rank. (B-C) Ridgeline diagrams of overrepresentation analyses (ORA) according to KEGG database, depicting the top 10 enriched signaling pathways in PMNs during severe COVID-19 compared to (B) healthy controls and (C) mild COVID-19. (D) UMAP analysis of the COVID-19 Immune Atlas, which integrates 5 public COVID-19 PBMC single-cell transcriptomics datasets, created using CELLxGENE. (Top) UMAP showing the clustering of CD16+ cells (mature, FCGR3B expressing cells) and CD66b+ cells (immature, CEACAM8 expressing cells). Each dot represents a single cell colored according to the expression level of a selected gene. The color scale ranges from green (low expression) to purple (high expression). (Bottom) Pie chart summarizing the percentage of mature (black) and immature (blue) cells in the data. (E) The fraction of mature and immature neutrophils cells expressing inflammasome related genes identified in (D) are shown in bar graphs. For each gene, the proportion of expressing cells is shown in light blue, while the proportion of negative or not-expressing cells is shown in gray. Zoomed-in bar graph depicts the proportion of mature and immature cells expressing each gene.

https://doi.org/10.1371/journal.ppat.1012368.g002

Single cell sequencing data from the COVID-19 immune atlas, which integrates data from 5 independent studies analyzing COVID-19 PBMCs, confirmed our transcriptomic results (Fig 2D), from which a detailed gene by gene analysis of the most relevant inflammasome related genes is shown (Figs 2E and S4). Briefly, PYCARD gene coding for the ASC protein was expressed similarly in mature and immature neutrophils (S4 Fig), suggesting that both cell types may have ASC-dependent inflammasome forming capacity. However, most of the inflammasome gene expressions differed significantly and in the same manner as in our transcriptomic analysis.

Activation of neutrophil inflammasome related pathways during respiratory distress is not specific to COVID-19

We also reanalyzed the RNA-seq data generated by LaSalle et al. [8], focusing on neutrophil transcriptomics in patients with COVID-19 versus non-COVID-19 patients, and healthy controls. The non-COVID-19 patients exhibited acute respiratory distress and clinical suspicion for COVID-19. However, they tested negative for SARS-CoV-2 by PCR, unlike those classified as COVID-19 patients. Our analysis included IFN-α response, IL-1β production, TLR signaling, NLRP3 inflammasome, and pyroptosis pathways, using the Gene Ontology (GO) database; the NLR signaling pathway using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database; and inflammasome pathway using the REACTOME database (Fig 3). These pathways were significantly enriched in COVID-19 patients, supporting our findings. Importantly, the genes from the above-mentioned pathways were also induced in non-COVID-19 patients, suggesting that these pathways represent a general neutrophil response to inflammatory stimuli rather than a COVID-19 specific response.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Comparative neutrophil transcriptomics of COVID-19 and non-COVID-19 patients.

Bar graphs represent the activation levels of selected pathways and processes as identified by neutrophil transcriptomics. The analysis includes interpheron alpha (IFN-α) responses, interleukin (IL)-1β production, Toll-like receptor (TLR) signaling, NLRP3 inflammasome, and pyroptosis, as determined through the Gene Ontology (GO) database. The NOD-like receptor signaling pathway was investigated using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, and the inflammasome pathway was explored via the REACTOME database (more information in S1 Table). The graphs compare the activation levels of these pathways in healthy controls (HC), non-COVID patients with similar symptoms (COVID-19 negative), and COVID-19 positive individuals. Statistical significance is denoted as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. P values were calculated with Kruskall-Wallis test.

https://doi.org/10.1371/journal.ppat.1012368.g003

Inflammasomes are activated in severe COVID-19 PMNs, but not directly by SARS-CoV-2

Given the strong upregulation of many inflammasome related genes during severe COVID-19, we assessed whether PMNs exhibit active inflammasome formation in vivo. To evaluate spontaneous inflammasome mediated cytokine secretion, fresh PMNs isolated from severe COVID-19 patients and HC were cultured ex vivo for 24 hours. We measured the levels of IL-1β and IL-18 in the supernatant and found that IL-1β secretion was significantly increased in the supernatant of severe COVID-19 PMNs compared to HC PMNs (Fig 4A), whereas the IL-18 levels did not differ significantly (Fig 4B). Additionally, since SARS-CoV-2 viral particles were previously implicated to induce inflammasome formation in macrophages [17], the IL-1β and IL-18 levels after HC PMNs exposure to SARS-CoV-2 were also assessed but no significant effects in the secretion of these cytokines were observed (Fig 4C and 4D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Inflammasome activation in PMNs during severe COVID-19.

PMNs were freshly isolated from blood and cultured at 2 million cells/ml (A) IL-1β and (B) IL-18 levels in 24 h cell culture supernatants from COVID-19 (n = 11 for IL-1β and 9 for IL-18) and HC PMNs (n = 6 for both). (C) IL-1β and (D) IL-18 levels in 24 h cell culture supernatant from PMNs exposed or non-exposed to purified SARS-CoV-2 viral particles (10 virus particles / PMN) (n = 3). (E) Caspase1 activity in PMNs following a 2 h stimulation with nigericin or purified SARS-CoV-2 viral particles (10 virus particles / PMN). For HC PMNs, n = 9 for mock and nigericin and n = 6 for SARS-CoV-2 exposure. For COVID-19 PMNs, n = 12 for mock and nigericin and n = 9 for SARS-CoV-2 exposure. *p < 0.05 and **p < 0.01. Data presented as mean ± SD. Tukey’s multiple comparisons test for mixed-effect analysis was applied for (E), meanwhile P values for (A-D) were calculated with the Mann-Whitney U-test.

https://doi.org/10.1371/journal.ppat.1012368.g004

The spontaneous secretion of IL-1β by COVID-19 PMNs suggests that these cells are actively producing and releasing IL-1β through inflammasome formation which is dependent on caspase1 activity [32]. We assessed caspase1 activity in response to the second signal required for inflammasome activation, induced by nigericin, and observed increased caspase1 activity in severe COVID-19 PMNs compared to HC PMNs (Fig 4E). These findings suggest that severe COVID-19 PMNs have an increased capacity for inflammasome activation, potentially due to an existing priming signal during acute disease in vivo. However, no significant difference in caspase1 activity between non-exposed and virus-exposed PMNs were observed (Fig 4F), indicating that caspase1 activation in COVID-19 PMNs is not directly triggered by the virus.

Type I IFNs prime PMNs for inflammasome activation

Since PMNs from COVID-19 patients concomitantly display a strong IFN-I signature (Fig 1B and 1C) and an increased propensity for inflammasome activation, we hypothesized that IFN-I could act as the priming signal for PMN inflammasomes during COVID-19. Isolated HC PMNs were stimulated ex vivo with exogenous IFN-I and the well-described inflammasome priming (1^st signal) and activator (2^nd signal) agents LPS and nigericin, respectively [33,34]. After stimulation, both priming signals induced pro-IL-1β (31 kDa) in the cell lysates, followed by the release of active IL-1β (17 kDa) into the supernatant in response to nigericin (Fig 5A), confirming the ability of IFN-I to prime PMNs for inflammasome activation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. IFN-I primes inflammasome activation while COVID-19 PMNs show defective inflammasome responses ex vivo.

Isolated HC or COVID-19 PMNs were non-stimulated or stimulated 4h with IFN-I (combination of 2.7*10⁴ IU/ml IFN-α and IFN-β) or 20 ng/ml LPS (1^st signal), followed by 4h with 2.5 μM nigericin or purified SARS-CoV-2 (10,1 virus/PMNs) (2^nd signal). Then, (A) western blot of pro-IL-1β (31 kD) and active IL-1β (17 kD) was performed from HC PMNs supernatant and cell lysates, (B) IL-1β (n = 5 HC PMN and 9 COVID-19 PMN) and (C) MPO (n = 5 HC PMN and 9 COVID-19 PMN) were measured from supernatants by ELISA. (D-E) Effect of inflammasome inhibitor MCC950 (2 μg/ml, added simultaneously with nigericin) on IL-1β secretion in (D) HC and (E) severe COVID-19 PMN supernatant (n = 3). (F) LDH and (G) IL-8 in HC and severe COVID-19 PMN supernatants (n = 3). (H-K) RT-qPCR of selected mRNAs in IFN-I or LPS-primed HC and COVID-19 PMNs (n = 6–8 HC PMN and 7–10 COVID-19 PMN). (L-M) HC PMNs were stimulated with high dose IFN-I (2.7*10⁵ IU/ml), normal dose IFN-I (2.7*10⁴ IU/ml) and 20 ng/ml LPS. After 4 hr stimulation caspase1 activity was measured using median fluorescence intensity (MFI) of FAM-FLICA by flow cytometry (L, n = 5, representative histogram of one donor shown) and after 24 hr stimulation IL-1β release was measured by ELISA (M, n = 5). P values calculated with Kruskall-Wallis test for the comparison between treatments by group (HC or COVID-19 PMNs), and Mann-Whitney test for the comparison between HC and COVID-19 PMNs by individual treatment for (B-G), and Two-way ANOVA Tukey’s multiple comparisons test for (B, H-K). The treatments in L-M were compared to mock by one-way ANOVA for repeated measures, corrected for multiple comparisons with the two-stage step-up method of Benjamini, Krieger and Yekutieli. *p < 0.05, **p < 0.01, ***p < 0.001, **** p < 0.0001. Data presented as mean ± SD.

https://doi.org/10.1371/journal.ppat.1012368.g005

To assess inflammasome formation in circulating neutrophils during COVID-19, PMNs from HC and COVID-19 patients underwent similar stimulation assays as above, followed by IL-1β measurement from supernatants by ELISA. In addition, to further assess the role of SARS-CoV-2 virus particles in neutrophil inflammasome activation, HC PMNs were cultured in the presence of purified viruses (10 infectious units/PMN). HC PMNs responded to both LPS and IFN-I by increasing their IL-1β secretion, which was exponentiated after exposure to nigericin (Fig 5A and 5B), confirming the ability of IFN-Is to prime for inflammasome assembly in PMNs, albeit less efficiently than LPS. Furthermore, as expected, the ability of IFN-I to prime for nigericin-mediated inflammasome activation was dependent on the IFN-I receptor IFNAR1 for both IFN-α and IFN-β (S5A Fig).

Interestingly, COVID-19 PMNs produced less IL-1β than HC PMNs upon exogenous inflammasome activation primed by either LPS or IFN-I, while SARS-CoV-2 particles did not have any effect on PMN inflammasome activation (Fig 5B). As with 24 h cultures (Fig 4B), we did not detect any significant changes in IL-18 secretion in either HC or COVID-19 PMNs (S5B Fig). However, the release of myeloperoxidase (MPO), used as a marker of degranulation and/or NETosis, in response to nigericin was similar between COVID-19 PMNs and HC PMNs (Fig 5C), and therefore the observed diminished IL-1β release by COVID-19 PMNs is not due to general cellular inertia but may be specific to the ex vivo induced inflammasome pathway. Furthermore, additional stimulation assays in the presence of the NLRP3 inhibitor MCC950 (Fig 5D and 5E) and caspase1 inhibitor YVAD (S5C Fig) confirmed that induced IL-1β secretion is dependent on canonical NLRP3 inflammasome activation. Unlike IL-1β (S5D Fig), increased IL-18 secretion was not detectable even after 24 h stimulation (S5E Fig). Furthermore, the observed residual IL-18 was not affected by inflammasome inhibitors, suggesting its secretion to be unrelated to inflammasome activity in PMNs.

We further assessed the specificity of inflammasome activation by measuring LDH and IL-8 levels in the supernatants from the same cells and under the same experimental conditions as shown in Fig 5D and 5E. The measurements of the former were done to assess inflammasome mediated cell death by pyroptosis in response to nigericin, while the latter was assessed to demonstrate the responsiveness of PMNs to an inflammasome unrelated inflammatory cascade. As with IL-1β secretion, COVID-19 PMNs were less responsive than HC PMNs to nigericin- and LPS-mediated LDH (Fig 5F) and IL-8 (Fig 5G) release, respectively. This suggests that COVID-19 PMNs are generally poorly responsive to inflammatory stimuli.

To examine this reduced responsiveness to external inflammatory priming, we evaluated the inflammasome related gene expression following ex vivo stimulation with IFN-I or LPS (Figs 5H–5K and S5F–S5I). OAS1 gene, an interferon stimulated gene (ISG), showed significant upregulation by IFN-I in COVID-19 PMNs as compared to HC PMNs (Fig 5H), while the inflammasome related genes IL-1β (Fig 5I), CASP1 (Fig 5J) and NLRC5 (Fig 5K) were more efficiently induced in HC PMNs than COVID-19 PMNs. This suggests that the inflammasome defect in COVID-19 PMNs is at the transcriptional level when using IFN-I as the priming factor, while high OAS1 gene expression indicates transcriptional defect is restricted to individual genes.

Next, we investigated whether IFN-I could also activate caspase1 directly without the 2^nd signal to boost inflammasome activation. In addition to treating HC PMNs with IFN-I and LPS as in previous experiments, we also included another group with a higher dose of IFN-I (ten-fold) and assessed caspase1 activity after 4 hours with FAM-FLICA, a fluorescent caspase1 reactive dye (Fig 5L) as well as IL-1β release after 24 hours by ELISA (Fig 5M). Results showed that similar to LPS, high-dose IFN-I induced significant caspase1 activity, which was not observed with the normal IFN-I concentration. Despite this, the normal IFN-I concentration still resulted in increased IL-1β levels in the supernatant. Taken together, these findings suggest that the 2^nd signal is not essential for inflammasome activation by IFN-I and that IL-1β release is a more sensitive method of detecting inflammasome activity as compared to caspase1 activity in our assay setup. Thus, these results can also explain our previous observation of increased spontaneous release of IL-1β by COVID-19 PMNs even though significant caspase1 activity is only detected after nigericin-mediated boosting in vitro (Fig 4A and 4E).

Association between ex vivo inflammasome activation and disease severity

Our analysis of the association between ex vivo inflammasome activation (caspase1 activity and IL-1β release) and clinical markers of disease severity, including neutrophil responses, revealed intriguing links. Calprotectin is a marker of neutrophil activation or death [35] but also potentially activates the inflammasome [36]. A significant positive correlation between calprotectin plasma levels and PMN caspase1 activity (Fig 6A and 6B) underscores this latter possibility and highlights the interplay between inflammation and inflammasome activation in PMNs of COVID-19 patients. Furthermore, the negative association of PMN IL-1β levels (after ex vivo stimulation with IFN and nigericin) with disease severity (WHO ordinal scale, Fig 6A) and patient neutrophil counts (Fig 6A and 6C) supports the exhaustion hypothesis, wherein PMNs from severe COVID-19 patients may be less responsive to stimuli due to prior in vivo activation. While these findings provide intriguing insights into the complex interplay between calprotectin release, caspase1 activity, and inflammasome activation in COVID-19, additional research is required to further elucidate these connections.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Correlation analysis between clinical parameters and ex vivo PMN inflammasome activation.

(A) Spearman’s correlation matrix depicting the relationships among clinical parameters and results of ex vivo experimentation. For the WHO ordinal scale, the baseline parameters were used. (B-C) Linear regression analysis demonstrating the associations between: (B) Positive association between PMN Caspase1 activity, measured after ex vivo nigericin stimulation, and the levels of Calprotectin in the matched patient’s peripheral blood; (C) Negative association between ex vivo stimulated PMN IL-1β levels (IFN+Nig) and the blood neutrophil count in matched patients at the time of sampling (n = 12). LOS = length of stay. WHO = World Health Organization. Min = minimum. Casp1 = caspase1. LPS or IFN + nig = lipopolysaccharide or type I interferon + nigericin ex vivo stimulation.

https://doi.org/10.1371/journal.ppat.1012368.g006

LDGs differ from PMNs in their ability to release IL-18 upon inflammasome activation

Transcriptomic analysis presented above revealed a distinct lack of IFN-I responsive and inflammasome related gene expression in LDGs as compared to PMNs of severe COVID-19 patients. This suggested that inflammasomes are not similarly regulated in LDGs as compared to PMNs during COVID-19. To assess the inflammasome forming capacity of LDGs, we conducted ex vivo stimulation assays using LDGs isolated from COVID-19 patients, similar to the approach used for PMNs described earlier. Like PMNs, IL-1β secretion by LDGs was elevated in the presence of a priming signal (IFN-I or LPS), which exponentially increased when the inflammasome activation signaling molecule nigericin was added (Fig 7A). Contrary to PMNs and in line with the transcriptomics data, an increased IL-18 secretion was detected (Fig 7B). Additionally, the secretion of both ILs by LDGs was inhibited in the presence of inflammasome specific inhibitors MCC950 and YVAD (Fig 7A and 7B).

These findings suggested that IFN-I can prime for inflammasome activation also in LDGs. Furthermore, the ability to release IL-18 upon neutrophil inflammasome activation varies based on cellular maturation state. To explore this further, we conducted in vitro stimulation studies using differentiated HL-60 cells, an immature neutrophil-like model [37]. Similar to LDGs from COVID-19 patients, HL-60 displayed comparable IL-18 secretion pattern upon LPS or IFN-I stimulation and nigericin-induced activation. Notably, their IL-1β release was only detected with LPS priming (Fig 7C and 7D). Furthermore, transcriptomic analysis revealed an upregulation of inflammasome related genes upon differentiation (Fig 7E). Overall, these findings suggest that neutrophils may lose the ability to secrete IL-18 in response to inflammasome activity during maturation, and increased release of neutrophil-derived IL-18 occurs primarily in disease states associated with extensive granulopoiesis and increased immature granulocyte counts in the blood, like COVID-19 [38].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Immature neutrophils express IL-18 in response to inflammasome activation.

(A-D) Isolated COVID-19 LDGs or HL-60 cells (differentiated for 5 days with 1% DMSO) were non-stimulated or stimulated 4h with IFN-I or LPS (1^st signal), followed by 4h with nigericin (2^nd signal) in the presence or absence of inflammasome inhibitors MCC950 or YVAD as previously. Secretion of (A, C) IL-1β and (B, D) IL-18 were measured from the supernatants by ELISA (n = 2 for LDGs and 3–5 for HL-60). *p < 0.05 and **p < 0.01. P values calculated with Kruskal-Wallis test. Data presented as mean ± SD. (E) Volcano plot of differentiated vs undifferentiated HL-60 cells gene expression from GSE93996, with inflammasome related genes marked in blue. Only significant DE genes are shown (adjusted p value < 0.05).

https://doi.org/10.1371/journal.ppat.1012368.g007

Neutrophils are recruited to the lungs in SARS-CoV-2 infected mice

Hamsters and human ACE2 expressing mice infected with SARS-CoV-2 develop pulmonary inflammation including neutrophil recruitment [39–41]. To further assess the role of neutrophils in COVID-19, we utilized a recently developed SARS-CoV-2 mouse model [18]. This model employs the MaVie strain, serially passaged in mouse lungs and causing pneumonia like human COVID-19 in wild-type BALB-C mice [18]. Infected mice started losing weight by day 2 post-infection, with some mice reaching the clinical endpoint of 20% weight loss by day 4 (S6A Fig, includes animals from 4 independent infection experiments, details of animal usage in S3 Table). The first experiments were performed to study infection kinetics, histopathology and Ly-6G+ neutrophil accumulation in lungs.

Viral loads, as assessed by RT-qPCR and titration of infectious virus, were significantly higher at 2 dpi than at 4 dpi (Fig 8A and 8B), and viral antigen expression, widespread at 2 dpi in bronchioles and alveoli, matched this pattern (Figs 8C and S6D). The extensive viral replication at 2 dpi was associated with degeneration of infected epithelial cells, most prominent in the respiratory epithelium, accompanied by neutrophil (Ly6G+) infiltration (S6D Fig). An increase in neutrophil numbers in the lungs of infected mice was observed as represented by increased Ly6G+ neutrophil/lymphocyte ratio and total Ly6G+ cell counts as assessed by flow cytometry of lung single cell suspension (Fig 8D and 8E; the gating strategy for Ly6G+ neutrophils is shown in S6B Fig). A significant increase in the number of neutrophils in the lungs of infected mice compared to PBS-inoculated mice was also confirmed in situ, by morphometrical quantification of lungs sections stained for Ly6G (Figs 8F and S6D). Detailed information on the histological and immunohistochemical features is provided in S3 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Neutrophil accumulation in the lungs correlates with viral loads in SARS-CoV-2 infected mice.

Female BALB/c mice were intranasally inoculated with 5*10⁵ TCID50 SARS-CoV-2 MaVie strain or PBS as control and euthanized at 2 dpi or 4 dpi. (A) RNA was isolated from lungs and subjected to RT-qPCR targeting viral subE and GAPDH as housekeeping gene. The relative expression of subE was measured using the comparative Ct method as compared to mock-infected control (in which subE was undetectable but set to 40 Ct) **** p < 0.0001. P values calculated with Welch’s t-test. (B) Infectious virus was calculated from supernatants of lung single cell suspensions of infected mice as fluorescence focus forming units (FFU) in Vero E6 cells. * p < 0.05. P values calculated with Mann-Whitney test. (C) Quantification based on morphometric analysis that determines the area of immunolabelling for SARS-CoV-2 nucleoprotein in relation to total tissue area. (D) Quantification of Ly6G neutrophil/lymphocyte ratio in lung single cell suspensions by flow cytometry. (E) Quantification of total Ly6G neutrophil counts in lung single cell suspensions by flow cytometry. (F) Quantification of Ly6G based on morphometric analysis that determines the area of immunolabelling for Ly6G in relation to total tissue area in mock-infected controls. (G) Quantification of median fluorescence intensity (MFI) of CD11b expression in ly6G neutrophils by flow cytometry. Representative histograms of CD11b expression in Ly6G+ neutrophils is shown. P values for C-G were calculated with ordinary one-way ANOVA using Tukey’s multiple comparison correction. Black line represents the mean. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. (H) Fluorescent nuclear staining of representative magnetic-bead isolated Ly6G neutrophils by Hoechst33342. Panels A, C and F are representative of two independent experiments.

https://doi.org/10.1371/journal.ppat.1012368.g008

Furthermore, flow cytometry revealed diminished surface expression of the maturation and activation marker CD11b in Ly6G+ neutrophils of infected mice both at 2 and 4 dpi as compared to PBS controls (Fig 8G). However, morphologically, neutrophils from infected mice were similar to those in the PBS control mice in displaying equally multilobed nuclei (DNA staining of isolated Ly6G+ neutrophils shown in Fig 8H; a representative flow cytometry histogram showing the purity of Ly6G+ neutrophils after isolation is provided in S6C Fig). This suggests that the reduced expression of CD11b on the surface of neutrophils from infected mice was not the consequence of an accumulation of CD11b-negative immature neutrophils but rather an activation-related phenomenon, such as shedding or internalization.

Neutrophils from SARS-CoV-2 infected mice display IFN-I dependent caspase1 activation

Next, we investigated whether neutrophils of SARS-CoV-2 infected mice show increased caspase1 activation. We utilized the fluorescent FAM-FLICA caspase1 reactive probe in conjunction with Ly6G+ neutrophil staining to detect caspase1 activity in neutrophils by flow cytometry. Indeed, we observed significantly an increased median fluorescent intensity of FAM-FLICA in Ly6G+ neutrophils at 4 dpi, but not at 2 dpi, compared to PBS-inoculated controls (Fig 9A). This finding suggests increased neutrophil caspase1 activity at latter stages of the infection, concomitant with decreased viral loads.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Neutrophils of SARS-CoV-2 infected mice display increased caspase1 activation ability that is dependent on IFN-I.

Female BALB/c mice were intranasally inoculated with 5* 10⁵ TCID50 SARS-CoV-2 MaVie strain or PBS as control. (A) Lungs were harvested at 2 and 4 dpi and single cell suspensions stained with FAM-FLICA and Ly6G antibody followed by flow cytometric analysis. FAM-FLICA median fluorescence intensity (MFI) was recorded in Ly6G+ neutrophils (n = 4, representative histogram image shown). * p < 0.05, ** p < 0.01. P values calculated with one-way ANOVA using Tukey’s multiple comparisons. (B-E) Ly-6G+ neutrophils isolated from lung single cell suspensions based on positive selection with magnetic beads. (B-D) RNA was isolated and subjected to transcriptomic analysis by RNA-seq. (B) Principal component analysis (PCA) of the PBS-inoculated control and SARS-CoV-2 infected mice lung neutrophil RNA-seq samples. (C) Heatmap of the top differentially expressed genes (DEGs). (D) Volcano plots of DEGs between neutrophils isolated from SARS-CoV-2 infected mice versus uninfected PBS-inoculated mice. Blue points represent significant terms (adjusted p-value < 0.05), while smaller gray points represent non-significant terms. Relevant inflammasome and interferon related genes are shown with larger and darker blue points. (E) Caspase1 activity in isolated mice neutrophils following a 2 h stimulation with nigericin was assessed by a bioluminescence method (Caspase-Glo 1 Inflammasome Assay). (F-I) Mice were intraperitoneally inoculated with 250 μg anti-IFNAR or IgG1 isotype control directly after infection with SARS-CoV-2 and lung neutrophils isolated at 2 dpi (including also intranasally PBS-inoculated control mice without intraperitoneal injection). (F) Caspase1 activity was assessed following a 2 h stimulation with nigericin by bioluminescence method. (G-I) RNA was isolated from isolated neutrophils and fold change mRNA expressions of (G) Oasl2, (H) Caspase1 (Casp1) and (I) IL-1β (Il1b) was assessed by RT-qPCR in isotype control and anti-IFNAR treated infected mice as compared to mock-infected control mice. *p < 0.05, **p < 0.01 and ***p < 0.001. P values for A, E and F panels were calculated with ordinary one-way ANOVA using Tukey’s multiple comparisons correction, while Welch’s t-test was used for panels G-I. Data presented as mean ± SD.

https://doi.org/10.1371/journal.ppat.1012368.g009

The potential role of IFN-I in mediating caspase1 activity in neutrophils of infected mice was initially assessed by isolating neutrophils from the lungs of infected mice at 2 and 4 dpi, as well as from non-infected mice, for transcriptomic analysis by RNA-seq. PCA showed differences between neutrophils from infected and non-infected mice, with slight variation between the 2 and 4 dpi time points (Fig 9B). These differences were reflected in many DEGs, including several IFN-I responsive and inflammasome related genes, which showed strong upregulation at 2 dpi with slightly lower but still significantly elevated levels at 4 dpi, compared to non-infected mice (highlighted in the DEG heatmap; Fig 9C). The volcano plot (Fig 9D) provided a comprehensive view of the DEG pattern between neutrophils from SARS-CoV-2 infected and mock-infected mice. In addition to confirming the upregulation of IFN-I responsive and inflammasome related genes observed in the heatmap, the plot revealed a broader transcriptional response to viral infection with several additional DEG.

Having established a robust IFN-I transcriptional signature in neutrophils from SARS-CoV-2 infected mice we wanted to assess whether they possess increased propensity for nigericin-induced caspase1 activation, similar to human PMNs after exogenous priming by IFN-I. Neutrophils from infected mice, harvested at 2 and 4 dpi, displayed increased caspase1 activity upon nigericin stimulation, compared to neutrophils from non-infected mice (as displayed by both bioluminescence and FAM-FLICA fluorescence assays; Figs 9E and S7A, respectively). To directly assess the role of IFN-I nigericin-induced caspase1 activity, we inoculated mice with an IFN-I blocking anti-IFNAR monoclonal antibody or an isotype control antibody post-infection. Remarkably, neutrophils from anti-IFNAR treated mice showed diminished nigericin-induced caspase1 activity (Figs 9F and S7A). Furthermore, as expected due to their typical IFN-responsiveness, oasl2, caspase1 and IL-1β gene expressions were lower in anti-IFNAR treated than in isotype treated mice (Fig 9G–9I). Taken together, the results indicate that IFN-I is responsible for the increased caspase1 activity in neutrophils of infected mice.

Blocking IFN-I signaling did not significantly alter virus replication, virus-induced pathological changes, neutrophil CD11b expression or neutrophil caspase1 activity without exogenous stimuli (S7B–S7D, S7H and S7I Fig and S3 Table). Neutrophil counts indicated a significant increase in neutrophil accumulation in the lungs of anti-IFNAR treated mice; however, this was not confirmed by neutrophil/lymphocyte ratio or morphometry (S7E-S7G Fig). Interestingly, regardless of treatment, some neutrophils in infected mice displayed degeneration and NETosis, evidenced by histone H3cit staining (Fig 7J; S3 Table).