Investigation of intra-host SARS-CoV-2 diversity in a cohort of patients admitted in CCU for severe COVID-19. (A) Number of iSNVs and mean allelic frequency in each sample according to post-symptom delay. Red dot corresponds to the sample obtained from patient#2 sequenced at day 9 after CCU admission (8 days after remdesivir treatment). (B) Distribution of all iSNVs across the SARS-CoV-2 genome. Circles represent non-synonymous mutations, squares synonymous mutations, and triangles mutations in untranslated region (UTR). Non-synonymous mutations are showed with the Wuhan-Hu-1 genome as reference. (C) Medical history and evolution of SARS-CoV-2 viral load and PaO₂/FiO₂ ratio (red line) obtained from the NPS of the patient#2. A lower PaO₂/FiO₂ ratio indicates a more severe hypoxemic state. The patient had a medical history of diabetes, hypertension, dyslipidemia, as well as peripheral arterial occlusive disease and received four SARS-CoV-2 vaccination doses. Computed tomography (CT) scan was obtained 1 day prior to CCU admission and showed a “crazy paving” pattern with severe involvement (>50% of the lung parenchyma). A chest CT scan performed on day 10 after CCU admission revealed the appearance of a fibrotic-like pattern. Interestingly, patient#2 tested negative for SARS-CoV-2 on day 32 after CCU admission and simultaneously tested positive for influenza A virus (FluA) that can be associated with the alteration of immune antiviral defense caused by the presence of anti–IFN-β antibodies. He was discharged and returned home on day 36. The patient remained alive 90 days after inclusion. (D) Evolution of the nasal IFN-I score according to symptom onset during the CCU stay (n = 16 patients). A mixed-linear regression model was used with post-symptoms delay (days) as a fixed effect and individual slope and intercept as random effects. The grey line represents the predicted value of the fixed effect. The grey area represents the normal range defined in uninfected healthy volunteers. The P value indicates the statistical significance of the fixed effect. R-squared represents the conditional R-squared. (E) Association of viral load kinetics with nasal IFN-I score kinetics during CCU stay (n = 16). The evolution of viral load and nasal IFN-I score during CCU stay was assessed by calculating the differential between samples obtained at 7 days and 1 day after the inclusion (n = 10) or 3 and 1 days after the inclusion (n = 4). For 2 patients, only the sample from day 1 was available, and thus they are not represented in this figure. Grey lines connect the trajectories of both parameters for each patient. Red line/dots depicted values for patient with anti–IFN-β antibodies. (F) Representation of the luciferase activity under different IFN-I cytokines conditions, in the presence or absence of a patient’s plasma (each point corresponds to one patient). One plasma sample was missing for IFN-β and IFN-ω assay, and two plasma samples were missing for the IFN-α2 assay. The grey zone corresponds to neutralizing activity defined as the median plus 3 standard deviations of the control conditions luciferase activity (cells only), corresponding to a value of 3.67 log₁₀ luciferase activity. Red dot corresponds to the sample obtained from patient#2. Patient suffering from autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APS-1) (positive control for anti–IFN-α and -ω autoantibodies, n = 1) and polyclonal antibody neutralizing all of IFN-α, -β, and -ω IFN (n = 1) (Tebubio, ref product 39000-1, Le Perray en Yvelines, France) were used as positive controls for IFN-I neutralization. The experiments were performed in duplicate, and the neutralizing activity of IFN-β was confirmed in an independent experiment. CCU, Critical Care Unit, ICU, Intensive Care Unit.