The SARS-CoV-2 spike protein primes inammasome-mediated interleukin-1- beta secretion in COVID-19 patient-derived macrophages

Innate immunity triggers responsible for viral control or hyperinammation in COVID- 19 are largely unknown. Here we show that the SARS-CoV-2 spike protein primes inammasome activation and interleukin 1-beta (IL-1β) secretion in macrophages derived from COVID-19 patients but not in macrophages from healthy SARS-CoV-2 naïve controls. Chemical NLRP3 inhibition blocks spike protein-induced IL-1β secretion ex vivo. These ndings can accelerate research on COVID-19 vaccine design and drug treatment. or S-protein. statistical Two-way tukey post hoc test used. e IL-1β gene expression of macrophages from COVID- 19 patients (n = 5; red bars) or healthy individuals (n = 4; blue bars) stimulated with/without S-Protein were determined by qRT-PCR. Data are normalized to β-actin. Statistical signicance was analyzed using the Kolmogorov-Smirnov test. Graphs show mean ± SEM. *p<0.05; ***p<0.001.


Main Text
Since December 2019, coronavirus disease 2019 (COVID-19) has affected more than 3 million people globally 1. The disease is caused by infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a novel coronavirus. In COVID-19, little is known about protective or detrimental immune responses making rational therapeutic interventions di cult to assess. A subset of patients fails to control initial viral replication some of which present with severe pneumonia, signs of hyperin ammation and excessive release of cytokines in a second phase of the disease 2. This second phase immune response represents a putative target for host directed therapeutic intervention and several approaches such as blockade of the interleukin-6 receptor or inhibition of Janus kinases are being tested in clinical trials 2. However, knowledge on triggers of the SARS-CoV-2 speci c in ammatory response and key cytokines that are involved is scarce. Our data and those of others show that the major pro-in ammatory cytokine Interleukin-1-beta (IL-1β) is elevated in plasma from hospitalized COVID-19 patients and its associated signaling pathway seems to drive SARS-CoV-2 pathogenicity 3, 4 ( Figure S1A).
IL-1β secretion is primarily initiated by in ammasomes that represent multiprotein signaling platforms responsible for the coordination of the early antimicrobial host defense 5. In ammasomes are assembled by pattern-recognition-receptors such as the NOD-, LRR-and pyrin domain-containing protein 3 (NLRP3) following the detection of pathogenic microorganisms or danger signals in the cytosol of host cells. Upon activation, these receptors initiate the oligomerization of the adaptor protein ASC, serving as an activation platform for caspase 1. Active caspase 1 in turn cleaves pro-IL-1β yielding the mature active IL-1β which can be subsequently secreted. NLRP3 in ammasome activation is a two-step process. In a priming step, cellular receptors recognize conserved pathogen-associated molecular patterns (PAMPs) leading to pro-IL-1β, and pro-IL-18 expression. The activation step required for in ammasome assembly and secretion of mature IL-1β is triggered by a range of intrinsic or pathogen derived stimuli such as ATP, microbial toxins (e.g. nigericin), nucleic acids or vaccine adjuvants 6, 7, 8. For SARS-CoV-2, priming and activation triggers are unknown. We speculated that the ACE-2 receptor-binding spike glycoprotein (S-protein), a major SARS-CoV-2 antigen and focus of therapeutic strategies and vaccine design, may function as a PAMP leading to IL-1β secretion in patient derived macrophages. To test this hypothesis, we a nity puri ed the SARS-CoV-2 S-protein lacking the transmembrane domain (Fig. 1a) 9. The puri ed S-protein speci cally bound COVID-19 patient-derived IgG but not IgG from SARS-CoV-2 naïve controls con rming selective reactivity with patient derived antibodies (Fig. 1a). Next, we isolated peripheral blood mononuclear cells (PBMC) of six hospitalized COVID-19 patients and six SARS-CoV-2 naïve healthy controls followed by positive selection of CD14+ monocytes which were differentiated to macrophages by incubation with M-CSF (patient characteristics are provided in the supplementary Tables 1a and 1b and supplementary Fig. S1a). Overall monocyte counts and phenotypes were similar in both patients and controls (Supplementary Fig. 2a and 2b). In a second step, isolated macrophages were stimulated with SARS-CoV-2 S-protein followed by addition of nigericin as the in ammasome activating signal (Fig. 1c).
We show that the S-protein potently triggers secretion of IL-1β into the cell supernatants of patient derived macrophages after sequential incubation with nigericin (Fig. 1d). Intriguingly, cells from SARS-CoV-2 naïve healthy controls were non-reactive towards the S-protein (Fig. 1D). In contrast to S-protein, lipopolysacharide (LPS), a classical PAMP for priming in ammasome activation, induced IL-1β secretion in both groups indicating functional in ammasome signaling pathways (Fig. 1d). Gene expression analysis by quantitative real-time PCR revealed an S-protein induced increase of IL-1β mRNA levels in macrophages derived from COVID-19 patients and, to a much lesser extent, from SARS-CoV-2-naïve controls (Fig. 1e). IL-1β mRNA levels were signi cantly higher in COVID-19 patient-derived cells indicating that differential regulation of IL-1β secretion in patients versus SARS-CoV-2 naïve controls occurs on the transcriptional level (Fig. 1e). Macrophage treatment with LPS, nigericin or S-protein alone had no effect on IL-1β secretion showing that the S-protein solely functions as an NLRP3-in ammasome priming signal requiring a second stimulus for IL-1β secretion (Fig. 1d).
In line with this, we found that MCC950, a selective NLRP3 inhibitor fully blocked IL-1β secretion in Sprotein stimulated patient cells indicating that in ammasome inhibitors may provide valuable therapeutic tools for COVID-19 patients by preventing hyperin ammatory syndromes (Fig. 2a) 10. In addition, hydroxychloroquine, a drug known for its immune-modulatory effects showed similar suppressive effects on S-protein primed patient cells and, to a much lesser extent, on LPS stimulated cells ( Fig. 2A, Fig. S2C). Hydroxychloroquine has an ill-de ned, most likely host cell-directed inhibitory effect on viral replication in vitro 11. Several clinical trials are currently investigating this drug for treatment and prevention of COVID-19.
Furthermore, we show that the S-protein also triggers release of NLRP3-independent pro-in ammatory cytokines such as IL-8, IL-6 and tumor necrosis factor α (TNFα) (Fig. 2b). In contrast to IL-1β, these cytokines were secreted by both COVID-19 patient-derived cells and by cells isolated from SARS-CoV-2 naïve controls in an LPS-like manner (Fig. 2b). This indicates that in macrophages, the S-protein primed NLRP3 in ammasome is differentially regulated depending on previous exposure to SARS-CoV-2 whereas other pro-in ammatory signaling pathways are activated non-selectively.
We revisited our nding of SARS-CoV-2 naïve macrophages being non-reactive towards the S-protein and tested cells derived from COVID-19 convalescent individuals, which had only mild disease manifestations (patient characteristics can be found in the supplementary Table 1c). Interestingly, macrophages from these individuals showed elevated IL-1β secretion upon stimulation similar to hospitalized patients (Fig.  2c). Two convalescent individuals were tested sequentially, 7 days after initial sampling. Here, S-protein dependent IL-1β detected in cell supernatants remained high, but levels had declined in both patients within the 7-day period while IL-1β levels of LPS treated cells were stable (Fig. 2d). Assuming that the Sprotein functions as a classical PAMP, there seems to be a certain degree of trained innate immunity in individuals having survived COVID-19 and only little or no S-protein driven in ammasome activation in cells derived from SARS-CoV-2 naïve individuals. The latter may be a surrogate for failing early viral control in host tissue which is mainly driven by the in ammasome and IL-1β as rst lines of defense 8, 12, 13. However, once patients are infected, macrophages become highly reactive, secreting large amounts of IL-1β (Fig. 1C). Here, the NLRP3 in ammasome may contribute to pathophysiology and exuberant in ammation as shown for in uenza A virus infection or acute respiratory distress syndrome (ARDS) 14, 15.
In conclusion, we provide rst ex vivo evidence for a SARS-CoV-2 structural component being a PAMP and driver of pro-in ammatory cytokine secretion. The S-protein as the major antigen of most vaccine constructs currently under investigation seems to have a dual role in both adaptive and innate immunity. In ammasome formation is crucial for vaccine immunogenicity and for mounting an effective humoral immune response 7. Intriguingly, and possibly relevant for vaccine development, S-protein driven in ammasome activation seems to require prior SARS-CoV-2 in vivo priming since naïve individuals failed to secrete IL-1β when their macrophages were exposed to S-protein ex vivo. Pathogen or vaccine exposure-dependent in ammasome activation is known in the context of trained immunity and epigenetic reprogramming of monocytes, for example after vaccination with BCG 16, 17. However, it has not been shown yet for a viral infection and a correlating PAMP.
Our data also indicate that patients with severe, SARS-CoV-2-induced hyperin ammatory syndrome may bene t from treatment with IL-1 receptor antagonists or small molecules targeting in ammasomes.
Thus, our ndings are highly relevant for further research on SARS-CoV-2 derived triggers of innate immunity pathways required for rational designs of urgently needed therapeutic and preventive measures.
IL-1β ELISA IL-1β ELISA (BioLegend, San Diego, CA, USA) was performed according to manufactures manual. Brie y, supernatant was diluted 1:50 in IL-1β ELISA Kit diluent and incubated for 2h on previously coated 96-well ELISA plates (Thermo Fisher Scienti c). All samples were measured in technical duplicates (inhibition assay was performed in technical duplicates) and concentration was determined with a corresponding standard curve provided by IL-1β ELISA kit. OD was determined with Hidex Sense microplate reader (Hidex, Turku, Finland)

Statistical analysis
Statistical analysis was performed with GraphPad Prism 8.0.2 software (GraphPad, San Diego, CA, USA). Statistical parameters (value of n, statistical calculation etc.) are stated in the gure legend. P-values less than or equal to 0.05 were considered statistically signi cant.

Data availability statement
The data that support the ndings of this study are available from the corresponding author upon reasonable request. The authors declare that the data supporting the ndings of this study are available within the paper and its supplementary information les.
Declarations (2018  were stimulated with LPS (red triangles) or S-protein (red circles) and IL-1β was quanti ed by ELISA.