Viral genomic, metagenomic and human transcriptomic characterization and prediction of the clinical forms of COVID-19.

COVID-19 is characterized by respiratory symptoms of various severities, ranging from mild upper respiratory signs to acute respiratory failure/acute respiratory distress syndrome associated with a high mortality rate. However, the pathophysiology of the disease is largely unknown. Shotgun metagenomics from nasopharyngeal swabs were used to characterize the genomic, metagenomic and transcriptomic features of patients from the first pandemic wave with various forms of COVID-19, including outpatients, patients hospitalized not requiring intensive care, and patients in the intensive care unit, to identify viral and/or host factors associated with the most severe forms of the disease. Neither the genetic characteristics of SARS-CoV-2, nor the detection of bacteria, viruses, fungi or parasites were associated with the severity of pulmonary disease. Severe pneumonia was associated with overexpression of cytokine transcripts activating the CXCR2 pathway, whereas patients with benign disease presented with a T helper "Th1-Th17" profile. The latter profile was associated with female gender and a lower mortality rate. Our findings indicate that the most severe cases of COVID-19 are characterized by the presence of overactive immune cells resulting in neutrophil pulmonary infiltration which, in turn, could enhance the inflammatory response and prolong tissue damage. These findings make CXCR2 antagonists, in particular IL-8 antagonists, promising candidates for the treatment of patients with severe COVID-19.