key: cord-0751711-4s1m5vwo authors: Wang, J.; Kotagiri, P.; Lyons, P.; Mescia, F.; Bergamaschi, L.; Turner, L.; Al-Lamki, R.; Morgan, M. D.; Calero-Nieto, F. J.; Bach, K.; Mende, N.; Wilson, N. K.; Watts, E. R.; Cambridge Institute of Therapeutic Immunology and Infectious Disease - NIHR Covid BioResource,; Chinnery, P.; Kingston, N.; Papadia, S.; Stirrups, K.; Walker, N.; Gupta, R. K.; Toshner, M.; Weekes, M.; Nathan, J. A.; Walmsley, S.; Ouwehand, W. H.; Kasanicki, M.; Gottgens, B.; Marioni, J. C.; Smith, S. G.; Pober, J. S.; Bradley, J. R. title: Factor V is an immune inhibitor that is expressed at increased levels in leukocytes of patients with severe Covid-19 date: 2021-01-15 journal: nan DOI: 10.1101/2021.01.14.21249801 sha: 241ce49bc81abfaadfb9c512e6ce9d05e9124940 doc_id: 751711 cord_uid: 4s1m5vwo Severe Covid-19 is associated with elevated plasma Factor V (FV) and increased risk of thromboembolism. We report that neutrophils, T regulatory cells (Tregs), and monocytes from patients with severe Covid-19 express FV, and expression correlates with T cell lymphopenia. In vitro full length FV, but not FV activated by thrombin cleavage, suppresses T cell proliferation. Increased and prolonged FV expression by cells of the innate and adaptive immune systems may contribute to lymphopenia in severe Covid-19. Activation by thrombin destroys the immunosuppressive properties of FV. Anticoagulation in Covid-19 patients may have the unintended consequence of suppressing the adaptive immune system. Introduction 61 62 Dysregulation of both the immune 1 Immunological responses include T cell lymphopenia, which we have found can persist for months 64 after the initial illness. Coagulopathy is an important cause of morbidity and mortality in patients 65 with Covid-19, and a marked increase in circulating FV activity has been reported in patients with 66 severe Covid-19, associated with increased risk of thromboembolism 3 . 67 68 Production of FV by T cells 4 and monocytes 5 has been previously reported. We report that circulating 69 leukocytes are a source of FV in patients with Covid-19, and propose that neutrophil, monocyte and 70 Treg derived Factor V may be an important determinant of the dysregulated immune response to 71 SARS-CoV-2. 72 73 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. Analysis of the transcriptome of peripheral blood cells from healthy controls and patients with 77 Covid-19 express FV, and expression is increased in patients with more severe disease for at least 60 78 days after onset of symptoms ( Figure 1A) . Weighted gene co-expression network analysis was performed to create distinct modules comprising 90 of non-overlapping co-expressed genes. The module containing FV also contained genes strongly 91 expressed in neutrophils ( Figure 2A, Supplementary Figure 3 ). In addition, FV was a "hub gene" 92 meaning its expression closely mirrored that of the module eigengene, which is a single expression 93 profile summarising all genes within the module. This FV module eigengene expression correlated 94 with severity of Covid-19 ( Figure 2B ). FV module expression is similar to healthy controls in 95 asymptomatic or mildly symptomatic individuals in the community. In hospitalised patients with mild 96 disease FV module expression is elevated at presentation, but declines as patients recover. In 97 patients with severe disease FV module expression is elevated at presentation and increased levels 98 persists for several weeks. Analysis of peripheral blood cells for expression of other coagulation 99 factors showed expression of Factor XIIIa and low levels of Factor XII but not other coagulation 100 factors (data not shown). 101 To determine whether increased peripheral blood cell FV mRNA levels correlate with FV protein 102 expression we assayed plasma FV levels and performed liquid chromatography -mass spectrometry 103 on neutrophil lysates from healthy controls and patients with severe Covid-19. There was a modest 104 but statistically significant correlation between FV gene expression and FV protein expression ( Figure 105 3A; p = 0.023, R = 0.17). Proteomic analysis showed significantly higher levels of FV in neutrophil 106 lysates from patients with severe Covid-19 compared to healthy controls ( Figure 3B is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; derived enzymes 10,11 . Thus leukocyte derived FV may suppress local T cell proliferation at sites of 160 infection. In support of this plasma FV levels correlated with biomarkers of haemostasis, whereas T 161 cell counts correlated with FV module gene expression in circulating leukocytes. The liver may be the 162 predominant source of circulating FV, and it is noteworthy that the liver displays adaptive immune 163 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Flow cytometry data were analysed using FlowJo (Tree Star, USA). Graphs and statistics were 231 generated using GraphPad Prism software. Results were presented as mean ± s.e.m.as indicated. 232 Differences between two groups were compared using two-tailed student's t-test. 233 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; https://doi.org/10.1101/2021.01.14.21249801 doi: medRxiv preprint Library preparation and RNA-Seq processing 235 RNA was quantified using RNA HS assay on the Qubit, and libraries prepared using the SMARTer® 236 Stranded Total RNA-Seq it v2 -Pico Input Mammalian kit (Takara) with 10ng of RNA as starting input. 237 Library quality and quantity were validated by capillary electrophoresis on an Agilent 4200 238 TapeStation. Libraries were pooled at equimolar concentrations, and paired-end sequenced (75bp) 239 across 4 lanes of a Hiseq4000 instrument (Illumina) to achieve 10 million reads per samples. 240 The quality of raw reads was assessed using FastQC 242 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). SMARTer adaptors were trimmed, 243 along with sequencing calls with a Phred score below 24 using Trim_galore v.0.6.4 244 (http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/.) Residual rRNA reads were 245 depleted in silico using BBSplit (https://github.com/BioInfoTools/BBMap/blob/master/sh/bbsplit.sh). 246 Alignment was performed using HISAT2 v.2.1.0 19 against the GRCh38 genome build achieving a more 247 than 95% alignment rate. A count matrix was generated in R using featureCounts (Rsubreads -248 packages) and converted into a DGEList (EdgeR package), for downstream analysis. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; https://doi.org/10.1101/2021.01.14.21249801 doi: medRxiv preprint Modules are not comprised of a priori defined gene sets but rather are generated from unsupervised 260 clustering. The eigengene of the module is then correlated with the sample traits and significance 261 determined. A signed adjacency matrix was generated, and a soft thresholding power chosen to 262 impose approximate scale-free topology. Modules identified from the topological overlap matrix had 263 a specified minimum module size of 30. Significance of correlation between a clinical trait and a 264 modular eigengene was assessed using linear regression with Bonferroni adjustment to correct for 265 multiple testing. Modules were annotated using Enrichr and Genemania. Genes with high 266 connectivity termed "hub genes" were identified based on a module membership of 0.8 or above 267 and were selected to have a correlation with the trait of interest >0.8. 268 Correlation 269 The relationships between multiple features were quantified using Pearson's correlation (Hmisc is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; https://doi.org/10.1101/2021.01.14.21249801 doi: medRxiv preprint ion exchange. Purified proteins were analysed by reducing and non-reducing SDS-PAGE, A280 to 336 determine concentration, size exclusion and mass spectrometry to confirm identity. 337 338 339 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; https://doi.org/10.1101/2021.01.14.21249801 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; https://doi.org/10.1101/2021.01.14.21249801 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; https://doi.org/10.1101/2021.01.14.21249801 doi: medRxiv preprint 394 395 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; hirudin had no effect on their own and in combination with construct 1. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. Factor V module membership vs gene significance for severe Genes comprising the Factor V module are graphed. The x axis represents the correlation of a given 450 gene's expression with the module eigenvalue. The higher the correlation, the more representative 451 the module expression is of the gene. The Y axis represents the correlation between a gene's 452 expression and disease status (HC versus severe Covid-19 at 0-12 days post symptom onset). The 453 higher the correlation, the more able a given gene can distinguish severe Covid-19 from health. The 454 red lines demarcate the "hub genes" which represent genes that strongly model the module and 455 disease status. The genes illustrated in Figure is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 15, 2021. ; Immune asynchrony in COVID-19 pathogenesis 467 and potential immunotherapies COVID-19 and Thrombotic or Thromboembolic Disease: Implications for 469 Antithrombotic Therapy, and Follow-Up: JACC State-of-the-Art Review Marked factor V activity elevation in severe COVID-19 is associated with 472 venous thromboembolism The serine protease cofactor factor V is synthesized by lymphocytes Characterization of coagulation factor synthesis in nine human primary cell 476 types Hospitalized COVID-19 Patients and Venous 478 Thromboembolism: A Perfect Storm COVID-19 and Thrombotic or Thromboembolic Disease: Implications for 480 Antithrombotic Therapy, and Follow-Up: JACC State-of-the-Art Review Proteomic and Metabolomic Characterization of COVID-19 Patient Sera Compartmental immunophenotyping in COVID-19 ARDS: A case series Neutrophil extracellular traps infiltrate the lung airway, interstitial, 487 and vascular compartments in severe COVID-19 SARS-CoV-2-triggered neutrophil extracellular traps mediate COVID-19 489 pathology COVID-19 patients receiving prophylactic or therapeutic anticoagulation: a systematic review 494 and meta-analysis Molecular 496 mechanisms of antithrombin-heparin regulation of blood clotting proteinases. A paradigm 497 for understanding proteinase regulation by serpin family protein proteinase inhibitors Pharmacology of 500 anticoagulants used in the treatment of venous thromboembolism Warfarin and the vitamin K-dependent gamma-carboxylation 503 system Immunomodulating activity of 505 heparin Point of Care Nucleic Acid Testing for SARS-CoV-2 in Hospitalized Patients: 507 A Clinical Validation Trial and Implementation Study Graph-based genome alignment and 509 genotyping with HISAT2 and HISAT-genotype EmptyDrops: distinguishing cells from empty droplets in droplet-based 511 single-cell RNA sequencing data Bayesian demultiplexing of pooled single-cell 513 RNA-seq data without genotype reference A step-by-step workflow for low-level analysis of 515 single-cell RNA-seq data with Bioconductor TNFR2 ligation in human T regulatory cells enhances IL2-induced cell 517 proliferation through the non-canonical NF-kappaB pathway