A Neutrophil activation signature in Covid-19

Abstract Covid-19 is often related to hyperinflammation that drives lung or multi-organ damage and mortality. The immunopathological mechanisms that cause excessive inflammation following SARS-Cov-2 infection are under investigation while different approaches to limit hyperinflammation in affected patients are being proposed. Here, a computational network approach was used on recently available data to identify possible Covid-19 inflammatory mechanisms. First, network analysis of putative SARS-Cov-2 cellular receptors and their directly associated interacting proteins, led to the mining of a robust neutrophil-response signature and multiple relevant inflammatory response genes. Second, analysis of RNA-seq datasets of lung epithelial cells infected with SARS-Cov-2 found that infected cells specifically expressed neutrophil-attracting chemokines, further supporting the likely role of neutrophils in Covid-19 inflammation. The role of neutrophils in Covid-19 needs to be studied further. Different immunoregulatory molecules and pathways presented here (TNF Receptor, IL8, CXCR1, CXCR2, ADAM10, GPR84, MME-neprilysin, ANPEP, LAP3) are druggable and might be therapeutic targets in efforts to limit SARS-Cov-2 inflammation in severe clinical cases.


Introduction
New studies have highlighted that Covid-19 is often characterised by an extreme inflammatory response associated with lung and multi-organ injury and mortality and have suggested promising anti-inflammatory options (1). Other studies recommend caution with immunosuppression given that regulated inflammation is necessary for an effective anti-viral response (2). More ideas are needed to understand Covid-19 hyperinflammation and target relevant immunopathological mechanisms. The role and function of different myeloid cell types in Covid-19 lung injury, cytokine storm and sepsis are not well understood.

Methods
The data used for the analysis was from publicly available published sources as cited below. Lung epithelial cell RNA-seq data was downloaded from https://www.biorxiv.org/content/10.1101/2020.03.24.004655v1.supplementary-material (Supplemental Table 2). Only genes significant by adjusted p-value were included. Network analysis was performed using StringDB. For the 7 coronavirus receptors network inflation was performed on StringDB by adding 100 first-shell (direct first-degree protein interactions). For the lung epithelial cell RNA-seq data only the 115 significant proteins were used to make the protein-protein interaction network. Network cut-off was set at 0.4. Gene Ontology analysis was performed using GeneOntology and validated in BiNGO in Cytoscape Drug-Gene interactions were analysed and downloaded from DGIdb. Network analysis, comparisons and final designs were performed on Cytoscape.

Results and Discussion
A recent study identified 7 putative SARS-Cov-2 receptors including the now well-known ACE2 (3) together with peptidases DPP4 and ANPEP as well as pathogen-binding proteins CD209, CLEC4G, CLEC4M and CEACAM1 (Fig.1A). These 7 proteins are expressed in different cell types and are implicated in the binding of coronaviruses on epithelial cells (4). To examine these 7 proteins further, I used freely-accessible network analysis (StringDB). The small signature (Fig.1A) was inflated to include up to 100 proteins that directly interact with the 7 input proteins (Fig.1B). Network inflation was performed to identify proteins that directly associate with the putative coronavirus receptors on human cells, thereby allowing mining of possible functional pathways. Surprisingly, the main ontology of this expanded network was "Neutrophil Degranulation" (GO:0043312) with 70 proteins in this category (Fig.1C). There were also 8 neutrophil-specific genes (Neutrophil-specific genes (5) including ANPEP, MME, MGAM, CD177, CEACAM1/3, FPR2 and CYSTM1 (Fig.1D). This finding suggested that coronavirus binding proteins (and their direct interactors) might be involved in neutrophil or other related classic inflammatory mechanisms. Two of the neutrophil-specific genes, ANPEP and CEACAM1 (Fig.1D) are coronavirus receptors (Fig.1A) indicating that neutrophils could be infected by SARS-Cov-2, as they are by Influenza A (6). At the time of writing this paper, interaction or infection of neutrophils with SARS-Cov-2 had not been reported.
The role of neutrophils in viral infections of the upper respiratory tract and their possible involvement in therapeutic strategies is not entirely clarified (7). They are involved in early anti-viral defence (8) but through degranulation and lysis, they can also be cytotoxic during ARDS and severe pneumonia, including from coronaviruses (9). Neutrophil hyperinflammation is also likely in other severe viral infections such as Hepatitis (7). In current Covid-19 literature, an increased peripheral neutrophil-to-lymphocyte ratio is observed in severe cases (10) but not much is known regarding neutrophils in Covid-19 affected lungs. Covid-19 lung injury in specific patients might involve dysregulated neutrophil activity.
Are neutrophils and related inflammatory mechanisms likely targets in Covid-19 complications? This is a difficult question given the complexity of acute innate immune responses and the importance of neutrophils in early anti-viral defence. Universal suppression of neutrophils or other myeloid cell types is not trivial and the clinical evidence on the use of steroids in Covid-19 inconclusive. Nevertheless, it might be possible to target specific inflammatory mechanisms before ARDS. The target-mining presented here includes multiple druggable proteins (Fig.1I-K) such as neutrophil-attracting chemokine signalling (Fig.1I), neutrophil-relevant inflammatory entities (Fig.1J) and SARS-Cov-2 receptors (Fig.1K). IL8 (Fig.1I) a cardinal neutrophil chemoattractant and product of activated neutrophils, can be blocked by neutralising antibodies (HuMax-IL8; FDA/EMA-approved). The neutrophil chemokine receptors CXCR1 and CXCR2 are also targeted by different experimental drugs including the approved CXCR2 blocker AZD5069 (Fig.1H). Other interesting inflammatory proteins include the monocyte and neutrophil-enriched genes TNF-receptor-2 (TNFR2 or TNFRSF1B), GPR84 and ADAM10 (Fig.1J). TNFR2 is clinically blocked by the receptorantagonist Enbrel. While TNF has a well-established role in neutrophil activation and prolongs neutrophil survival (13), the role of TNF in Covid-19 infection has not received attention. GPR84 and ADAM10 are targeted with experimental drugs including approved PBI-4050 (Setogepram) and the metzincin inhibitor Ilomastat respectively (Fig.1J). FPR2 (formyl peptide receptor), MME (neprilysin) and MGAM (maltase-glucoamylase) are directly involved in inflammatory neutrophil recruitment and activity and are neutrophil-specific genes (Fig.1D). ANPEP (Aminopeptidase-N), a likely coronavirus receptor (Fig.1A) and a neutrophil-specific gene (Fig.1D), is potentially blocked by approved drugs Ezetimibe and Icatibant, while there is experimental evidence that it interacts with Tosedostat and Ubenimex. It is however unlikely that binding of these drugs to the aminopeptidase will interfere with viral binding. Ubenimex is also an inhibitor of LAP3 (leucine aminopeptidase-3), the only gene shared between the coronavirus receptors inflated network (Fig.1A) and SARS-Cov-2 infected lung epithelial cells (Fig.1F). LAP3 was induced in lung epithelial cells following SARS-Cov-2 infection and it is a monocyte and neutrophil-enriched gene. Like other aminopeptidases (like ANPEP), LAP3 might be another coronavirus receptor.

Conclusions
Neutrophil recruitment and related inflammatory activity might be important components of severe Covid-19 immunopathology. In contrast to universal immunosuppression, specific inflammatory proteins and pathways (chemokine receptors, TNFR, neprilysin, FPR2, complement C3) could be considered in patients suffering from Covid-19 hyperinflammation. Moreover, the role of neutrophils in recognition of SARS-Cov-2 and the possibility that they are infected with the virus needs to be studied further. Finally, it is perhaps worth noting that some currently investigated Covid-19 therapeutic options include hydroxychloroquine, azithromycin and colchicine. All three have well-studied anti-neutrophil effects (14-16).   The main gene ontology of the inflated network is "Neutrophil Degranulation" (C) and included 8 Neutrophil-specific genes (D). E-H: 113 genes that were differentially regulated (RNA-seq) in lung epithelial cells following stimulation with SARS-Cov-2 (E). Following virus infection, most upregulated genes were related to inflammatory and virus responses (F). 6 classic neutrophil chemokines were upregulated (G). SARS-Cov-2 infected epithelial cells also upregulated C3 and related complement pathway genes (H). I-K: Putative druggable targets with neutrophil and inflammatory function derived from the analysis. Approved and experimental drugs with solid pharmacological evidence are presented as interaction networks. Protein-Drug associations were examined in DGiDB and were manually curated to exclude false-positive or non-experimentally validated interactions. All protein-protein interaction networks were performed in Cytoscape using interaction data derived from StringDB(v10).