key: cord-0728632-6203masw
authors: Vitte, Joana; Diallo, Aïssatou Bailo; Boumaza, Asma; Lopez, Alexandre; Michel, Moïse; Allardet-Servent, Jérôme; Mezouar, Soraya; Sereme, Youssouf; Busnel, Jean-Marc; Miloud, Tewfik; Malergue, Fabrice; Morange, Pierre-Emmanuel; Halfon, Philippe; Olive, Daniel; Leone, Marc; Mege, Jean-Louis
title: A granulocytic signature identifies COVID-19 and its severity
date: 2020-09-17
journal: J Infect Dis
DOI: 10.1093/infdis/jiaa591
sha: cbed6ef86bb0765b027bad443ed234b43f9145f2
doc_id: 728632
cord_uid: 6203masw

BACKGROUND: An unbiased approach of SARS-CoV-2-induced immune dysregulation has not been undertaken so far. We aimed to identify previously unreported immune markers able to discriminate COVID-19 patients from healthy controls and to predict mild and severe disease. METHODS: An observational, prospective, multicentric study was conducted in patients with confirmed COVID-19: mild/moderate (n=7) and severe (n=19). Immunophenotyping of whole blood leukocytes was performed in patients upon hospital ward or intensive care unit admission and in healthy controls (n=25). Clinically relevant associations were identified through unsupervised analysis. RESULTS: Granulocytic (neutrophil, eosinophil and basophil) markers were enriched during COVID-19 and discriminated between mild and severe patients. Increased counts of CD15 (+)CD16 (+) neutrophils, decreased granulocytic expression of integrin CD11b, and Th2-related CRTH2 downregulation in eosinophils and basophils established a COVID-19 signature. Severity was associated with the emergence of PDL1 checkpoint expression in basophils and eosinophils. This granulocytic signature was accompanied by monocyte and lymphocyte immunoparalysis. Correlation with validated clinical scores supported pathophysiological relevance. CONCLUSION: Phenotypic markers of circulating granulocytes are strong discriminators between infected and uninfected individuals as well as between severity stages. COVID-19 alters the frequency and functional phenotypes of granulocyte subsets with the emergence of CRTH2 as a disease biomarker.

The hallmark of COVID-19, the infection caused by the severe acute respiratory syndrome coronavirus 2 (SARS CoV-2), is the occurrence, in 10-20% of patients, of a sudden deterioration 7-10 days after the onset of symptoms, increasing the risk of acute respiratory distress syndrome, of intensive care unit need and ultimately of death [1] . Studies exploring the immune response suggested that SARS-CoV-2 may induce unique patterns of immune dysregulation [2] [3] . To our knowledge, a systematic approach of SARS-CoV-2-induced immune dysregulation at the phenotype level has not been undertaken so far. Single-cell RNA sequencing of peripheral blood mononuclear cells evidenced phenotypic remodeling affecting innate and adaptive populations [4] . Our aim was to establish a comprehensive, unsupervised map of circulating immune cells in COVID-19 patients using a first-in-class flow cytometry approach for rapid whole-blood assessment. The primary objective was the identification of immunophenotypic patterns most accurately associated with COVID-19 diagnosis and severity. Among the large number of phenotypic markers of circulating immune cells modulated by SARS-CoV-2, those related to granulocyte lineage (neutrophils, basophils and eosinophils) were strong discriminators between infected and uninfected individuals as well as between different degrees of disease severity. Beside SARS-CoV-2 associated lymphopenia, changes in frequency and activation of granulocyte subsets may be predictive of clinical worsening during A c c e p t e d M a n u s c r i p t 7

This open multicenter prospective observational study was conducted in the intensive care unit (ICU) of North Hospital of Marseille and the COVID-19 ward unit of European Hospital of Marseille.

Patients admitted to ward and ICU with confirmed SARS-CoV-2 infection were included in the study if they fulfilled the criteria: i) age 18 or older and ii) a positive SARS-CoV-2 reverse transcriptase-polymerase chain reaction (RT-PCR) in nasopharyngeal swabs or tracheal aspiration. Exclusion criteria were preexisting treatments interfering with immune functions, pregnancy and missing clinical or laboratory data.

Demographic, clinical and laboratory data (arterial blood gas analysis, complete blood count, biochemistry, virology) including SARS-CoV-2-related symptoms, date of disease onset, organ support, and medications were collected for each patient upon admission to ICU or conventional ward. The same data were collected on the day of blood sampling. At day 28 after COVID-19 diagnosis, the duration of mechanical ventilation, length of ICU and hospital stays, and ICU and hospital mortality rates were also recorded. The Simplified Acute Physiology Score II (SAPS II) (5] , the Sepsis-related Organ Failure Assessment (SOFA) [6] , the National Early Warning Score 2 (NEWS2) [7] , and the World Health Organization (WHO) progression scale [8] [9] were calculated at admission and on the day of blood sampling.

Patients were classified as mild/moderate (WHO grade 4 and 5, hereafter termed "mild") depending on the presence of oxygen supply, while those receiving high-flow oxygen A c c e p t e d M a n u s c r i p t 8 therapy (WHO grade 6) or invasive mechanical ventilation (WHO grade 7-9) were considered as severe.

Samples of healthy blood donors (HBD) group, serving as controls, were received from Etablissement Français du Sang (EFS), Marseille, France.

The study was conducted in accordance with the Declaration of Helsinki and the French law on research involving humans. It was registered with the French ANSM under the reference ID-RCB: 

All antibodies and reagents were from Beckman Coulter (Villepinte, France). Blood (4 mL) was collected by venipuncture on EDTA-anticoagulated tubes, stored and delivered at room temperature to the Immunology laboratory. Multiparametric flow cytometry was used for immune cell enumeration and immune phenotyping less than 24 hours after blood collection. Each immune phenotyping panel ( Table 1 ) was provided in a pre-mix dry antibody cocktail completed in some cases by the addition of liquid conjugates prior to sample addition. Staining of leukocytes for enumeration was performed by the addition of 100 µl whole blood to the IM Count tube followed by 15 min incubation at room A c c e p t e d M a n u s c r i p t 9 temperature. Lysis of red blood cells was achieved with 2 ml Versalyse (Beckman Coulter) followed by a 15 min incubation prior to acquisition. Immune phenotyping followed a similar protocol except for the incubation (20 and 10 minutes, respectively, in the dark).

Lysed cells were washed with 3 ml PBS and the cell pellet re-suspended in 0.5 ml PBS 1X, 0.1% formaldehyde. Acquisition was done with a Navios flow cytometer (Beckman Coulter).

Multiparameter flow cytometry data files were analyzed using the Kaluza software, version A c c e p t e d M a n u s c r i p t

During the study period, 55 confirmed COVID-19 cases were referred to the participating centers. Among them, 19 patients were admitted to the ICU (severe group) and seven to the conventional ward (mild group). Twenty-five HBD served as control group (Supplementary Figure 1) . Demographic data are presented in Table 2 . Differences were observed between the mild and the severe group. Elevated body mass index and hypertension were more frequent in the severe group than in the mild group (p = 0.005 and 0.03, respectively).

Severity scores, including the WHO progression score and the SOFA score, were significantly higher in the severe group than in the mild group. C-reactive protein was increased in the severe group as compared with the mild group whereas eosinophils and monocytes were significantly decreased between patient groups ( Table 2) . Lymphopenia, defined as a lymphocyte count of less than 1 giga/L, was found in 85% of COVID-19 patients comprising 71% of the mild group and 89% of the severe group, a non-significant difference ( Table 2) .

An unsupervised analysis of circulating leukocyte subsets and immune phenotypic markers yielded more than 100 significant discriminators between COVID-19 patients and controls, with FDR p-values less than 0.05. Further analysis was arbitrarily restricted to the 25 most discriminant markers (Fig. 1a) . PCA of these 25 markers effectively discriminated COVID-19 patients from controls (Fig. 1b) . Eleven of 25 were granulocyte-associated markers, followed by lymphocyte, NK and dendritic cell (DC) variables (Fig. 1a) . Enrichment in granulocyteassociated markers affected the three granulocytic lineages: neutrophils, eosinophils, and basophils. There was a significant increase in the frequency of CD15 + granulocytes (mainly A c c e p t e d M a n u s c r i p t 11 comprising neutrophils), an increase in the frequency of CD15 + CD16 + neutrophil subset, and a decrease in the frequency of basophils in COVID-19 patients, as compared with controls ( Fig. 2 a-b) . Two prominent function-associated membrane antigens were modulated in COVID-19 patients as compared to HBD: CD11b (M subunit of integrin CD11bCD18, also known as complement receptor 3, CR3), whose expression was decreased at the surface of neutrophils and basophils ( Fig. 2 c) , and CRTH2 (CD294), a receptor for prostaglandin D2 (PGD2), whose expression was decreased on basophils and eosinophils (Fig. 2 d) . Hence, SARS-CoV-2 infection was characterized by changes in frequency of granulocyte subsets and alteration of their functional phenotypes with the emergence of CRTH2 as a biomarker for COVID-19.

We wondered if the granulocyte signature displayed specific changes associated with disease severity (Fig. 3a) . Unsupervised analysis followed by PCA of the best markers discriminating between mild and severe COVID-19 patients (Fig. 3b ) evidenced the predominance of granulocytic markers (8 out of 19 with a FDR p value less than 0.05). Some of the markers discriminating COVID-19 patients from HBD also discriminated mild from severe patients. Neutrophil subset frequency was one of these shared markers. The frequency of CD15 + granulocytes and CD15 + CD16 + neutrophils was significantly increased in the severe group (p = 0.002), while the levels of expression of both CD15 and CD16 were decreased in the severe group as compared with the mild group (Fig. 4a) . Another shared marker was eosinophil CRTH2 expression, which was profoundly decreased in the severe group (Fig. 4b) . Hence, COVID-19 severity was associated with a more profound imbalance of granulocyte subsets and functional markers of the disease.

A c c e p t e d M a n u s c r i p t 12 However, severe disease was associated with the emergence of specific markers. Severe patients differed from mild ones with respect to functional markers of eosinophils and basophils. At the surface of both basophils and eosinophils, the expression of checkpoint inhibitors such as PDL1 was significantly higher in the severe group than in the mild group ( Fig. 4c-d) . Such prominent changes in surface expression of functional granulocytic markers prompted us to ask whether granulocyte alterations correlated with clinical scores. We found that both WHO and SOFA scores correlated positively with innate immune checkpoints such as PDL1 expression on basophils and eosinophils, and negatively with neutrophil CD11b and eosinophil CRTH2 expression (Fig. 5) . The level of correlation between immunophenotypic markers and clinical scores was similar to that of clinical scores between them (WHO versus SOFA: R 2 = 0.567).

The granulocytic signature of COVID-19 was not isolated since it was associated with a decreased representation of CD4 + T cells, CD8 + T cells, and plasmacytoid dendritic cells (Supplementary Fig. 2 ).

The upregulation of checkpoint inhibitors was not restricted to the granulocyte lineage: PDL1 expression on monocytes and NK cells, and PD1 expression on T cells were also increased in the severe group (Supplementary Fig. 3 

A c c e p t e d M a n u s c r i p t 13

This study was undertaken as a holistic description of immune cells and markers from COVID-19 whole blood samples. Alterations of lymphocyte subsets have been widely reported [4;10-11]. Here, a multiparametric flow cytometry approach using whole blood samples allowed us to assess the features of the cells involved in the innate response, beyond the lymphocyte response. As opposed to monocytes and lymphocytes, granulocyte investigation requires freshly isolated whole blood samples. A combination of dry antibody panels optimized for whole blood investigation and the detection of rare events [12] enabled the simultaneous study of more than 100 phenotypic markers. This unbiased approach showed that changes in the frequency of granulocyte subsets and alteration of their functional phenotypes characterize patients during the course of COVID-19.

In previous studies, the neutrophil-to-lymphocyte ratio was used to predict the degree of disease severity in patients with early-stage COVID-19 [13] [14] . Eosinopenia was reported in severe patients [15] [16] and was also present in our study population. We show here that the increase in neutrophil counts is characterized by the emergence of cells involved in the inhibition of immune responses. At the neutrophil level, the increase in absolute numbers was due to CD15 + CD16 + neutrophils, which may have pro-inflammatory properties [17] . Neutrophils express predominantly the glycophosphatidyl inositol-linked CD16b isoform, also known as low affinity IgG receptor FcRIIIb, which acts as a suppressive FcR receptor [18] . Low fucosylation of anti-SARS-CoV-2 antibodies [19] suggests that increased CD16 + neutrophils in severe patients might contribute to persistent inflammation through synergistic mechanisms. The strength of our study is the translational dimension. We explored the association between immune status and disease severity assessed with validated scores in two distinct, well-characterized patient groups. There were marked clinical differences between the mild and the severe group, notably the recourse to invasive mechanical ventilation required in 89% of the latter. We found a continuum between the decreased counts and surface marker Box plots summarizing (A) the differences observed in basophils and SS hi CD15 + granulocytes frequency, and the differential expression of (B) CRTH2 on basophils and eosinophils and (C) CD11b on basophils and SShiCD15+ granulocytes observed between controls and COVID-19 patients. M a n u s c r i p t 24 A c c e p t e d M a n u s c r i p t 25 A c c e p t e d M a n u s c r i p t

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Complex Immune Dysregulation in COVID-19 Patients with Severe Respiratory Failure

Dysregulation of immune response in patients with COVID-19 in Wuhan, China

A single-cell atlas of the peripheral immune response in patients with severe COVID-19

A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study

The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine

The National Early Warning Score 2 (NEWS2)

World Health Organization. Infection prevention and control during health care when novel coronavirus (nCoV) infection is suspected

WHO Working Group on the Clinical Characterisation and Management of COVID-19 infection. A minimal common outcome measure set for COVID-19 clinical research

A novel one-step extracellular staining for flow cytometry: Proof-of-concept on sepsis-related biomarkers

Neutrophil-to-lymphocyte ratio as a predictive biomarker for moderate-severe ARDS in severe COVID-19 patients

Neutrophil to lymphocyte ratio as prognostic and predictive factor in patients with coronavirus disease 2019: A retrospective cross-sectional study

Patients of COVID-19 may benefit from sustained Lopinavircombined regimen and the increase of Eosinophil may predict the outcome of COVID-19 progression

Eosinophil responses during COVID-19 infections and coronavirus vaccination

The proportion, origin and pro-inflammation roles of low density neutrophils in SFTS disease

Functional diversification of IgGs through Fc glycosylation

Symptomatic SARS-CoV-2 infections display specific IgG Fc structures. medRxiv

A c c e p t e d M a n u s c r i p t