High sera levels of SARS‐CoV‐2 N antigen are associated with death in hospitalized COVID‐19 patients

The presence of free severe acute respiratory syndrome coronavirus 2 (SARS‑CoV‑2) nucleocapsid‐antigen in sera (N‐antigenemia) has been shown in COVID‐19 patients. However, the link between the quantitative levels of N‐antigenemia and COVID‐19 disease severity is not entirely understood. To assess the dynamics and clinical association of N‐antigen sera levels with disease severity in COVID‐19 patients, we analyzed data from patients included in the French COVID cohort, with at least one sera sample between January and September 2020. We assessed N‐antigenemia levels and anti‐N IgG titers, and patient outcomes was classified in two groups, survival or death. In samples collected within 8 days since symptom onset, we observed that deceased patients had a higher positivity rate (93% vs. 81%; p < 0.001) and higher median levels of predicted N‐antigenemia (2500 vs. 1200 pg/mL; p < 0.001) than surviving patients. Predicted time to N‐antigen clearance in sera was prolonged in deceased patients compared to survivors (23.3 vs 19.3 days; p < 0.0001). In a subset of patients with both sera and nasopharyngeal (NP) swabs, predicted time to N‐antigen clearance in sera was prolonged in deceased patients (p < 0.001), whereas NP viral load clearance did not differ between the groups (p = 0.07). Our results demonstrate a strong relationship between N‐antigenemia levels and COVID‐19 severity on a prospective cohort.

(p < 0.001), whereas NP viral load clearance did not differ between the groups (p = 0.07).Our results demonstrate a strong relationship between N-antigenemia levels and COVID-19 severity on a prospective cohort.antigens or viral RNA, with the gold standard being the reversetranscriptase polymerase chain reaction (RT-PCR) performed on nasopharyngeal (NP) swabs or other respiratory samples. 1 RT-PCR techniques are sensitive and allow an estimation of the viral load with the cycle threshold (CT) value.Several works have demonstrated a link between NP viral load and disease severity. 2However this association has been questioned by other studies. 3,4Indeed, many biases can interfere in the quantitative evaluation.
Direct detection of SARS-CoV-2 by the RT-PCR has also been performed on blood samples and it seems to be associated with the severity of the disease, although discrepancies exist. 5,6In addition, other works have evaluated the detection of N-antigen in blood.The detection of this circulating antigen in the serum appears to provide a good sensitivity for clinical diagnosis 7,8 even in patients with negative NP RT-PCR, suggesting it might reflect viral replication in the lung. 7[11][12] Thus, N-antigen detection and quantification could provide an easy-to-use, scalable, and inexpensive diagnostic tool that would also provide a prognostic assessment.However, the kinetic of this marker has not been evaluated and compared to other markers such as NP viral loads.In this work, we used samples from hospitalized patients presenting various degrees of disease severity to assess the SARS-CoV-2 N-antigen dynamics and association with disease severity.

| Study population
We selected 320 patients that were included in the French COVID cohort (clinicaltrials.govNCT04262921) between January 25 and September 2, 2020, when the historical strain of SARS-CoV-2 was the dominant variant in Europe.We had at least one serum per patient collected within 60 days since symptoms onset (SSO).
Patients were classified according to their clinical outcome that is, (I) Survival for recovery (with or without ICU transfer) or (II) Death (for patients who died within 60 days).

| N-antigen and viral load assays
N antigenemia levels were determined with a certified in-vitro diagnostic (CE-IVD) ELISA assay, COV-Quanto ® (AAZ, Boulogne-Billancourt, France), as previously described. 7The use of standards made of recombinant N antigen allowed the quantification of Nantigenemia levels.The range of linearity was between 2.89 and 150 pg/mL and the limit of quantification was 2.97 pg/mL.

| Anti-SARS-CoV-2 IgG titers
Anti-N IgG titers were assessed using the multiplex immunoassay V-PLEX panel on the Meso Scale Discovery (MSD) platform, according to manufacturer recommendations.Standards were ranging from 0.004 to 200 units/mL.

| Viral dynamic model
We reconstructed N-antigen kinetics by considering this biomarker as a plasmatic proxy for NP viral load.We used a target-cell limited model described previously 2,13 to which we added a transfer rate to the plasmatic/extracellular compartment, yielding to a N-antigenemia.This transfer rate also allows for consideration of the extra-pulmonary infection producing N-Antigen.The model is given by the following equations: ELISA was performed as previously described. 14In brief, ELISA plates were coated with recombinant human interferon-α (rhIFN-α) or interferon-ω (rhIFN-ω) and incubated with 1:50 dilutions of plasma samples from the patients or controls., respectively (Table 1).

| N-antigenemia kinetics
We used a target-cell limited model with a transfer rate to the plasmatic compartment to reconstruct the viral dynamics of N-antigen concentrations with a nonlinear and saturable effect of anti-N IgG on its clearance.
We estimated the basic reproduction number R 0 to 13.

K
E Y W O R D S antigenemia, COVID-19, hospitalized, serological marker, severity 1 | INTRODUCTION Diagnostic of COVID-19 infection relies on several assays detecting either severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) × exp ^{(−exp ^{( − × ))}} Where Ag is N-antigenemia, R the transfer rate, V i the infectious viral particles, V ni the noninfectious viral particles, E 1 the elimination rate of N-antigen in absence of antibodies, E max the maximal elimination rate of N-antigen mediated by IgG, and IgG 50 the antibody titer required to achieve 50% of E max .Kinetics of IgG were described using a Gompertz sigmoid function, with a being the maximal concentration of IgG, b and c dimensionless parameters.For a comprehensive understanding of this model, additional details can be found in the supplementary methods.Serum samples were screened for auto-antibodies against 18 targets in a multiplex particle-based flow cytometry assay, in which magnetic beads with differential fluorescence were covalently coupled to recombinant human proteins.Patients with a fluorescence intensity (FI) of >1500 for IFN-α2 or >1000 for IFN-ω were tested for blocking activity, as were patients positive for another cytokine.

A
total of 1053 samples from 320 patients collected during the first two waves in France were analyzed.Patients were mainly male (66%) and were 63 years old in median [IQR: 52−71].238 patients were discharged alive, including 94 patients that had been admitted to ICU, while 82 patients died within 2 months postinfection.Median delay between symptom onset and hospitalization was slightly different between surviving and deceased patients: 8 days [IQR: 6−10] and 7 days [IQR: 4−10] 6 and the loss rate of infected cells to 0.78 day −1 .Viral particles (estimated at 3.7 × 10 5 particles/day) 2 produced were transferred to the plasma compartment at a rate R equal to 10 −4 day −1 .Elimination rate of N-Antigenemia in the absence of IgG was estimated at E 1 = 0.13 day −1 , whereas the maximal elimination rate mediated by IgG was estimated at E max = 0.7 day −1 .This caused the half-life of N-antigen in the plasmatic compartment to decrease from 5 days to at most 0.8 days due to IgGmediated elimination.We estimated IgG 50 (the concentration of IgG required to achieve 50% of this effect) at 2.54 A.U/mL, and the maximal predicted concentration of IgG was estimated at 5.4 A.U/mL.The entire N-antigen course was reconstructed for patients until clearance.Our model predicted N-antigenemia to peak 2.4 days after symptom onset [IQR: 1.1−6.9days] and a time to N-antigenemia clearance of 19.8 days SSO [IQR: 17.3−23.9days] (Figure 1A,B).

3. 3 |
N-antigenemia levels across severity groups Overall, ninety percent of patients tested positive for N-antigenemia, with a median concentration of 2.4 log 10 pg/mL [IQR: 1.96−3.43]on their first positive sample.T A B L E 1 Epidemiological and clinical description of the 320 patients with SARS-CoV-2 included.

F I G U R E 1
Evolution and dynamic of sera N-antigen levels according to patient's outcome.(A, B) Evolution of sera N-antigen levels (A, red) and anti-N antibodies (IgG) titers (B, blue) according to the delay since symptom onset (SSO).Triangles represent data under the Limit of Quantification (LoQ) (2.97 pg/mL for N-Antigen and 500 pg/mL for anti-N IgG).(C) Predicted sera N-antigen levels (log 10 pg/mL) at day 8 SSO according to survival status.(D) Predicted time to N-antigen clearance in sera according to survival status.(E, F) Predicted time to viral clearance in nasopharynx (E) and in sera (F) according to survival status, in the subset of patients with both sera and nasopharyngeal samples.In panels C to F, bars indicate median, patients discharged alive are shown in green, and deceased patients are shown in the dark red.Wilcoxon−Mann−Whitney tests were performed, with: ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 and ns: p > 0.05.