key: cord-0716791-xco2yfbt
authors: MacDonald, Lucy; Otto, Thomas D.; Elmesmari, Aziza; Tolusso, Barbara; Somma, Domenico; McSharry, Charles; Gremese, Elisa; McInnes, Iain B.; Alivernini, Stefano; Kurowska-Stolarska, Mariola
title: COVID-19 and Rheumatoid Arthritis share myeloid pathogenic and resolving pathways
date: 2020-07-26
journal: bioRxiv
DOI: 10.1101/2020.07.26.221572
sha: 39063d36bd6e3116f06b6eb92546800fb2aad3a2
doc_id: 716791
cord_uid: xco2yfbt

Background We recently delineated the functional biology of pathogenic and inflammation resolving synovial tissue macrophage clusters in rheumatoid arthritis (RA). Whilst RA is not a viral respiratory syndrome, it represents a pro-inflammatory cytokine-driven chronic articular condition often accompanied by cardiovascular and lung pathologies. We hypothesised that functionally equivalent macrophage clusters in the lung might govern inflammation and resolution of COVID-19 pneumonitis. Methods To provide insight into the targetable functions of COVID-19 bronchoalveolar lavage (BALF) macrophage clusters, a comparative analysis of BALF macrophage single cell transcriptomics (scRNA-seq) with synovial tissue (ST) macrophage scRNA-seq and functional biology was performed. The function of shared BALF and ST MerTK inflammation-resolving pathway was confirmed with inhibitor in primary macrophage-synovial fibroblast co-cultures. Results. Distinct BALF FCNpos and FCNposSPP1pos macrophage clusters emerging in severe COVID-19 patients were closely related to ST CD48highS100A12pos and CD48posSPP1pos clusters driving synovitis in active RA. They shared transcriptomic profile and pathogenic mechanisms. Healthy lung resident alveolar FABP4pos macrophages shared a regulatory transcriptomic profile, including TAM (Tyro, Axl, MerTK) receptors pathway with synovial tissue TREM2pos macrophages that govern RA remission. This pathway was substantially altered in BALF macrophages of severe COVID-19. In vitro dexamethasone inhibited tissue inflammation via macrophages’ MerTK function. Conclusion Pathogenesis and resolution of COVID-19 pneumonitis and RA synovitis might be driven by similar macrophage clusters and pathways. The MerTK-dependent anti-inflammatory mechanisms of dexamethasone, and the homeostatic function of TAM pathways that maintain RA in remission advocate the therapeutic MerTK agonism to ameliorate the cytokine storm and pneumonitis of severe COVID-19.

The clinical, social and economic disruption caused by SARS-CoV-2 infection and its unpredictable progression to COVID-19 and Acute Respiratory Distress Syndrome (ARDS)

represents a global emergency. The severity of COVID-19 is attributable to immune dysregulation, abnormal blood clotting and tissue disruption, particularly implicating proinflammatory innate immunity [1] [2] [3] . Interstitial lung disease and alveolitis is a co-morbidity of rheumatoid arthritis (RA) 4 , in which articular inflammation and disease remission are driven by distinct synovial tissue macrophage clusters 5 . Despite suggestive early data 2 6-10 there is a knowledge gap of how myeloid cell pathways mechanistically regulate severity or contribute to the resolution of COVID-19, hindering development of effective treatments 11-14 . Single-cell profiling and fate-mapping indicate spatial and functional macrophage heterogeneity that maintains lung homeostasis [15] [16] [17] [18] [19] . The alveolar macrophages (AM) expressing Fatty Acid Binding Protein 4 (FABP4) localise to the alveolar epithelial surface and recycle surfactants with type-2 epithelial cells to maintain compliance and efficient gas exchange 20 . This function is compromised in severe COVID-19 21 . Recent scRNAseq analysis 22 5 . Whilst RA is not a viral respiratory syndrome, it represents a proinflammatory cytokine-driven chronic articular condition often accompanied by cardiovascular and lung pathologies 4 23 . We 24 and others 25 have reported that SARS-CoV-2 infection is associated with emergence of polyarthritis or flares of synovitis in RA patient in sustained disease remission, suggesting potential shared pathogenic mechanisms of COVID-19 and RA.

RA immunopathogenesis and therapeutic targets 23 are well understood and thus might be informative for COVID-19 therapeutic strategies. For example, dexamethasone is a drug of choice for flares of arthritis and recent efficacy studies of dexamethasone 26 highlights opportunities to understand alveolar macrophage-based mechanism of resolution of COVID- 19 . We recently delineated the functional biology of pathogenic and inflammation resolving synovial tissue macrophage clusters and their roles in pathology or resolution of RA synovitis 5 . Thus, we raised the hypothesis that functionally equivalent macrophage clusters in the lung govern inflammation and resolution of COVID-19 pneumonitis. We propose that this perspective can offer new insight into the role of innate immunity in the immunopathogenesis and resolution of COVID-19.

Genomics single cell RNA sequencing data of myeloid cells from bronchoalveolar lavage fluid (BALF) of Liao et al (2020) 22 (8.4.2) . Kruskal-Wallis test with Dunn's correction for multiple comparisons was used. Differential Expression Analysis. In order to identify cluster markers, the Seurat function FindAllMarkers was used with each dataset individually. The "test.use" function was used to determine genes differentially expressed between clusters within each dataset using MAST. As recommended, for DE comparison the non-batch normalized counts were used. For identification of cluster markers, we specify that any markers identified must be expressed by at least 40% of cells in the cluster ('min.pct' parameter 0.4). We use the default values for all other parameters. Genes are considered significantly DE if the adjusted p-value (< 0.05) by Bonferroni Correction and multiple test correction (multiple by number of tests/clusters).

Dataset Comparison. Integration. Following individual clustering analysis, dataset integration was performed using the same method as described above, integrating 8993 common genes between ST and BALF data. These "integrated" batch-corrected values were then set as the default assay and the gene expression values are scaled before running principal component analysis. Clustering and Dimensional Reduction. To prevent bias in clustering, we reduced cell number of BALF data to 32139 random cells to match ST data. The data was re-scaled (ScaleData) before the first 30 principal components (PCs) were used for UMAP visualization.

Cells were coloured by their original identity to illustrate how clusters of each dataset overlay.

Comparative analyses. A count matrix with the average expression of all common genes by each cluster was generated (AverageExpression) before hierarchical clustering (dist, hclust) and plotting a dendrogram. Principal component analysis (PCA) was also performed on this pseudobulk expression matrix (prcomp). These initial analyses allowed for the identification of similar clusters between BALF and ST data. Shared marker genes between similar clusters were identified by comparing marker gene lists of each cluster using Venny (2.1) To visualise pseudobulk expression of the overlapping marker genes, the pheatmap (1.0.12) package was used to visualize shared genes as a heatmap using a custom script. Comparison of the pseudobulk of expression marker genes unique and common to all clusters similar between datasets was performed by Pearson correlation (ggscatter, add = "reg.line", conf.int = TRUE, cor.coef = TRUE, cor.method = "pearson"). These were performed based on the 916, 721 and 986 unique and common genes of BALF FCN1 pos /ST CD48 high S100A12 pos , BALF (Fig.1a-c) . Among the 9 phenotypically distinct clusters of ST macrophages that differed in distribution between health and RA, the CD48 high S100A12 pos and CD48 pos SPP1 pos clusters were expanded in RA patients with active disease 5 (Fig.1d-f ). To test the relationship between COVID-19 BALF and RA ST macrophage clusters, we integrated the macrophage scRNAseq datasets from both COVID-19 BALF 22 (Fig.1a-c) and RA ST 5 (Fig.1d-f ). The The most striking pathogenic similarities were between the COVID-19 pneumonitis FCN1 pos and the active RA synovitis CD48 high S100A12 pos clusters. They share 238 top marker-genes ( Fig.1k ) that include up-regulation of inflammatory and prothrombotic pathways. These consisted of IFN pathway (e.g. IFITM2, IFITM3 and ISG15), inflammation-triggering alarmins (S100A8/9/12A), B-cell activation factors (e.g. BAFF-TNFSF15B), promotors of IL-1β and TNF production (e.g. CARD16 and LITAF), prothrombotic factors (e.g. FGL2) and integrins mediating cell migration and adhesion (e.g. ITGB2 and ITGX). They also share receptor expression profiles e.g. TNFR2 (TNFSF1), G-CSF receptor (CSFR3) and the complement receptor (C5aR) that may increasingly render them susceptible to pro-inflammatory mediators to escalate inflammation ( Fig.1l and Supplementary Fig.2) . The hierarchical analysis and comparison of cluster markers also indicate a shared transcriptional profile of COVID-19 Supplementary Fig.4) . These include the complement pathways (e.g. C1q, that facilitates uptake of apoptotic bodies), high expression of genes of retinoic acid production (e.g. ALDH1A1 and RBP4) driving T-reg differentiation 32 , and B7-related co-inhibitory molecule VSIG4 which inhibits T-effector cells 33 , together suggesting a primary role of this cluster in governing lung immunity. While the functional contribution of FABP4 pos macrophages to the resolution of SARS-CoV-2 infection is yet to be established, we found that their counterpart TREM2 pos ST macrophage clusters produced inflammation-resolving lipid mediators and induce a repair phenotype in tissue stromal cells that maintain disease remission. These homeostatic responses are driven by MerTK, a member of the immunosuppressive tyrosine kinase receptor TAM family (namely Tyro, Axl and MerTK respectively) 5 . TAM receptors and their ligands GAS6 or PROS1 form a homeostatic brake on inflammation and autoimmunity [34] [35] [36] . In addition, PROS1 is an essential inhibitor of blood coagulation preventing thrombosis 37 . Lung-resident FABP4 macrophages uniquely express Axl rather than MerTK. Their Axl is constitutively ligated to GAS6 38 39 and is key in preventing exacerbated inflammation e.g. during influenza virus infection 38 

What is already known about this subject?

COVID-19 clinical trials testing anti-inflammatory drugs used for the treatment of RA e.g.

anti-IL-6R, anti-GM-CSF and dexamethasone support the concept of some common COVID- 19 and RA pathogenic and resolution mechanisms. Underlying shared innate cellular and molecular mechanisms are unknown and if dissected could provide novel therapeutic targets to resolve severe COVID-19 pneumonitis.

• Lung BALF macrophage clusters in severe COVID-19 (FCN pos and FCN pos SPP1 pos ) share pathogenic mechanisms with pro-inflammatory macrophage clusters that drive synovitis (CD48 high S100A12 pos and CD48 pos SPP1 pos ). • Dexamethasone inhibits inflammation via macrophages' MerTK function.

• COVID-19 pneumonitis is driven by macrophage clusters similar to those that drive chronic arthritis, and importantly, that resolution of COVID-19 might be accelerated by engaging similar macrophage-resolving pathways as found in remission arthritis, providing putative candidate inflammation-resolving therapeutic pathways (e.g MerTK) that might prevent/resolve severe COVID-19 pneumonitis. 

Blood vessel injury may spur disease's fatal second phase

The trinity of COVID-19: immunity, inflammation and intervention

COVID-19 severity correlates with airway epitheliumimmune cell interactions identified by single-cell analysis

Lung disease in rheumatoid arthritis

Distinct synovial tissue macrophage subsets regulate inflammation and remission in rheumatoid arthritis

Cytokine release syndrome in severe COVID-19

Breadth of concomitant immune responses prior to patient recovery: a case report of non-severe COVID-19

Heightened innate immune responses in the respiratory tract of COVID-19 patients

Exuberant elevation of IP-10, MCP-3 and IL-1ra during SARS-CoV-2 infection is associated with disease severity and fatal outcome

Effective treatment of severe COVID-19 patients with tocilizumab

Should we stimulate or suppress immune responses in COVID-19? Cytokine and anti-cytokine interventions

COVID-19 and rheumatology: first steps towards a different future

Monocyte-derived alveolar macrophages drive lung fibrosis and persist in the lung over the life span

Two distinct interstitial macrophage populations coexist across tissues in specific subtissular niches

Three Unique Interstitial Macrophages in the Murine Lung at Steady State

Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage

Identification of a nerve-associated, lungresident interstitial macrophage subset with distinct localization and immunoregulatory properties

Single-Cell Transcriptomic Analysis of Human Lung Provides Insights into the Pathobiology of Pulmonary Fibrosis

Lung Surfactant for Pulmonary Barrier Restoration in Patients With COVID-19 Pneumonia

Single-cell landscape of bronchoalveolar immune cells in patients with COVID-19

Immunopathogenesis of Rheumatoid Arthritis

Comparative analysis of synovial inflammation after SARS-CoV-2 infection

Case of acute arthritis following SARS-CoV-2 infection

Covid-19: Low dose steroid cuts death in ventilated patients by one third, trial finds

Reactome pathway analysis: a high-performance in-memory approach

UNC1062, a new and potent Mer inhibitor

Locally renewing resident synovial macrophages provide a protective barrier for the joint

Osteopontin: A novel regulator at the cross roads of inflammation, obesity and diabetes

Synovial tissue macrophages: friend or foe? RMD Open

Retinoic acid can enhance conversion of naive into regulatory T cells independently of secreted cytokines

VSIG4, a B7 family-related protein, is a negative regulator of T cell activation

TAM receptors are pleiotropic inhibitors of the innate immune response

Identification of the product of growth arrest-specific gene 6 as a common ligand for Axl, Sky, and Mer receptor tyrosine kinases

Differential TAM receptor-ligand-phospholipid interactions delimit differential TAM bioactivities

Blood clots and TAM receptor signalling in COVID-19 pathogenesis

AXL receptor tyrosine kinase is required for T cell priming and antiviral immunity

The Axl receptor tyrosine kinase is a discriminator of macrophage function in the inflamed lung

Biology of the TAM receptors

Diversification of TAM receptor tyrosine kinase function

Diagnosis and Management of Rheumatoid Arthritis: A Review

GM-CSF-based treatments in COVID-19: reconciling opposing therapeutic approaches

GM-CSF blockade with mavrilimumab in severe COVID-19 pneumonia and systemic hyperinflammation: a single-centre, prospective cohort study

Mechanistic immunological based classification of rheumatoid arthritis

Gene-expression profiles and transcriptional regulatory pathways that underlie the identity and diversity of mouse tissue macrophages

Protective Role of the MER Tyrosine Kinase via Efferocytosis in Rheumatoid Arthritis Models

Tampering with the immune system

Coagulopathy of COVID-19 and antiphospholipid antibodies