key: cord-0787693-okf4tufv authors: Legros, V.; Denolly, S.; Vogrig, M.; Boson, B.; Rigaill, J.; Pillet, S.; Grattard, F.; Gonzalo, S.; Verhoeven, P.; Allatif, O.; Berthelot, P.; Pelissier, C.; Thierry, G.; Botelho-Nevers, E.; Paul, S.; Walzer, T.; Cosset, F.-L.; Bourlet, T.; Pozzetto, B. title: A longitudinal study of SARS-CoV-2 infected patients shows high correlation between neutralizing antibodies and COVID-19 severity date: 2020-09-01 journal: nan DOI: 10.1101/2020.08.27.20182493 sha: 5a42f0a8ff7143fc6fe2270572095b2e2fb0f118 doc_id: 787693 cord_uid: okf4tufv Understanding the immune responses elicited by SARS-CoV-2 infection is critical in terms of protection from re-infection and, thus, for public health policy and for vaccine development against the COVID-19. Here, using either live SARS-CoV-2 particles or retroviruses pseudotyped with the SARS-CoV-2 S viral surface protein (Spike), we studied the neutralizing antibody (nAb) response in serum specimens from a cohort of 140 SARS-CoV-2 qPCR-confirmed patients, including patient with mild symptoms but also more severe form including those that require intensive care. We show that nAb titers were strongly correlated with disease severity and with anti-Spike IgG levels. Indeed, patients from intensive care units exhibited high nAb titers, whereas patients with milder disease symptoms displayed heterogenous nAb titers and asymptomatic or exclusive outpatient care patients had no or poor nAb levels. We found that the nAb activity in SARS-CoV-2-infected patients displayed a relatively rapid decline after recovery, as compared to individuals infected with alternative coronaviruses. We show the absence of cross-neutralization between endemic coronaviruses and SARS-CoV-2, indicating that previous infection by human coronaviruses may not generate protective nAb against SARS-CoV-2 infection. Finally, we found that the D614G mutation in the Spike protein, which has recently been identified as the major variant now found in Europe, does not allow neutralization escape. Altogether, our results contribute to the understanding of the immune correlate of SARS-CoV-2 induced disease and claim for a rapid evaluation of the role of the humoral response in the pathogenesis of SARS-CoV-2. can be detected for up to 36 months [11, 12] . It is therefore urgent to evaluate the nAb response 78 elicited by SARS-CoV-2 infection, the factors associated with its robustness, and its persistence. 79 Here, the nAb activity of serum specimens from a cohort of 140 SARS-CoV-2 qPCR-confirmed 80 patients was quantified. We show that nAb titers are strongly correlated with disease severity. 81 Importantly, we were able to quantify the persistence of nAb activity in patients, which indicated 82 a relatively rapid decline of nAbs after recovery. We also showed the absence of cross-protection 83 conferred by previous infection by endemic coronaviruses. Finally, we found that the D614G 84 mutation in the Spike protein, that has recently been identified as the major variant now found in 85 A total of 140 patients followed at the University Hospital of Saint-Etienne were enrolled 94 between March and May 2020. All patients were sampled by nasal swab and tested positive for 95 In the ICU and HOS groups, 3-4 sera were sampled at three periods of follow-up after the onset 100 of symptoms: 0 to 15, 16 to 30 and > 30 days. In the EOC group, 2 sera were sampled from 13 to 101 62 days after the onset of symptoms. 102 Time post-onset was defined as the time after onset of the first symptoms. 103 no differences of clinical symptoms between these two groups, which differed essentially with 168 their severity. In the EOC group, moderate symptoms were mainly fever, cough and asthenia. 169 170 Pseudoparticles neutralization is correlated with wild-type SARS-CoV-2 neutralization. 171 Neutralization of live virus in in vitro assays is considered as the gold-standard method for the 172 assessment of nAbs. However, for SARS-CoV-2, it requires BSL-3 facility and it is time-173 consuming. Hence, we developed a SARS-CoV-2 pseudoparticle assay, named SARS-CoV-2pp, 174 to quickly, safely and reliably assess nAb activity. This assay was compared to a classic 175 neutralization assay performed with wild-type SARS-CoV-2 in BSL-3 facility. To identify 176 unspecific neutralizing activity, each serum tested with SARS-CoV-2pp was tested in parallel 177 with RD114pp ( Figure S1 ), i.e., pseudoparticles coated with the glycoprotein gp70 of the 178 endogenous feline retrovirus RD114 [15] . 179 We first compared the neutralizing activity of the sera or plasma from our cohort of COVID-19 180 patients using both live virus and pseudoparticle assays. For live virus, the neutralization was 181 assessed by ID50 (serum dilution that inhibits 50% of the infectivity) whereas for pseudoparticle 182 assays, it was expressed as % of neutralization at the serum dilution of 1/100, both relative to a 183 no serum condition. The nAb activity of each serum sample from the 140 patients as well as 184 negative sera, i.e. sera that were collected before the emergence of the outbreak and sera from 185 patients previously infected with other coronavirus (see below), were blindly quantified using 186 these two detection methods. Similar results were obtained ( Figure 1A) , as indicated by the high 187 Spearman's rank correlation (rho=0.75). Hence, neutralization assays based on SARS-CoV-2 188 pseudoparticles can be reliably used to quantify nAb activity. 189 190 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint To explore potential humoral protective immune responses generated after SARS-CoV-2 192 infection, we compared the nAb activity of serum samples from patients in function of the 193 severity of their symptoms. The nAb activity of their serum samples was evaluated using both 194 SARS-CoV-2pp and wild-type virus. The cut-off for neutralization, i.e., 35%, was set using the 195 mean neutralization of a 1/100 dilution of the negative sera + 2 standard deviation [19] for 196 SARS-CoV-2pp. For wild-type virus, the cut-off for the ID50 was placed at the 1/10 dilution 197 (first tested dilution) as all negative sera were found below this threshold. 198 We found that most patients from the different groups developed nAbs, although the neutralizing 199 activity was highly variable ( Figure 1C ). Moreover, 44 patients (31.4%) exhibited a robust 200 neutralization allowing over 90% neutralization of SARS-CoV-2pp ( Figure 1B, total) . 201 Strikingly, we found that ICU patients were particularly prone to display high nAb activity as 202 compared to other groups with milder disease symptoms, such as HOS and EOC patients ( Figure 203 1C). Indeed, only one patient in ICU did not develop a nAb response at the time of sampling, 204 whereas for the HOS and EOC patients, the nAb activity was lower and more heterogenous. 205 Specifically, 21.9% and 25% of patients in HOS and EOC categories respectively, did not 206 develop nAbs at the time of sampling ( Figure 1C -right). In addition, 34.5% of HOS patients had 207 serum with no neutralizing activity at 1/100 and this number was increased to 70.7% in EOC 208 patients ( Figure 1C -left) . 209 Concerning the HOS patients, we sought to classify the patients according to the severity of their 210 symptoms. Accordingly, we defined as 'severe', the patients who had a respiratory rate >30/min 211 and/or blood oxygen saturation <92% and/or lung lesions observed by CT scan and/or intensive 212 oxygen therapy. The HOS patients were classified as 'moderate' if they did not fulfill the above 213 criteria. However, we did not find a significant correlation (p=0.0981) between the severity of 214 the symptoms of the HOS patients and their nAb titers, even though the % neutralization was 215 higher in serum of patients with more severe disease forms ( Figure 1D ). This indicated that the 216 severity of the symptoms may not be the only factor that explains the diversity of nAb activity in 217 the HOS patients. Additionally, we found that neither the age and gender criteria nor the Ct of 218 the first positive RT-qPCR titers (or the timing of sampling for qPCR) were associated with 219 stronger or weaker neutralization (Table S1 ). This excluded an impact of the viral load or a bias 220 due to the variability of the set-up of sampling in the heterogeneity of nAb activity observed in 221 these patients. 222 Overall, these results indicated that the nAb activity was highly correlated with the severity of 223 the symptoms, suggesting either that a robust humoral response was generated only in patients 224 with severe symptoms or that the humoral response may contribute to the aggravation of the 225 disease. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint 2B), the correlation was lower between anti-N IgG and nAbs (Spearman's ρ = 0.5765, Figure 237 2A). Such differences were particularly obvious when we compared the distribution of the IgG 238 among each category of patients. Indeed, while we found a low overlay in the profiles of anti-S 239 IgG between EOC and ICU patients, which appeared similar to those of nAbs (compare Figure 240 1C vs. Figure 2C left), we detected a strong overlay in the value of anti-N IgGs between EOC 241 and ICU patients (compare Figure 1C vs. Figure 2C right). This suggested that anti-S IgG values 242 are more relevant to mark the presence and levels of nAb activity. Indeed, our results indicate 243 that 95% of sera with above 124 AU/mL of anti-S IgGs were able to strongly neutralize SARS-244 CoV-2pp (>90% neutralization). 245 Overall, these results indicated that the anti-S IgG response could be used as a surrogate of 246 neutralizing activity in individuals. 247 Next, we evaluated the kinetics of the anti-S IgG and nAb activities in our cohort of COVID-19 250 patients. Indeed, beside the diversity of clinical forms, one key feature of our cohort is the serial 251 serum sampling for most patients, allowing to estimate the persistence of humoral factors. Both 252 anti-S IgGs and nAb activities were generally detectable at 5-7 days post-onset of symptoms in 253 patients who developed a humoral response, and rapidly increased to reach a peak before 254 declining ( Figure 3A , 3B). Hence, the kinetics of nAb titers as well as of anti-S IgG in each 255 group were modelized with a second order polynomial ( Figure 3A and 3B), which proved our 256 best model among different ones tested (see parameters indicated in Figure 3 ). This regression 257 analysis indicated that, instead of plateauing after the peak, the nAb activity could rapidly 258 decrease, with an estimated half-life of 26 days. Additionally, the general tendency for anti-S 259 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint IgGs seemed to follow the same pattern, with a peak and half-life similar to those calculated for 260 the nAb activity, further confirming the correlation between anti-S IgG and neutralizing activity. conferred by the D614G mutation, we used the SARS-CoV-2pp assay, which is particularly 279 suitable to compare the nAb activity of serum specimens against pseudoparticles harboring or 280 not this mutation. Yet, we found that that D614G mutation did not affect nAb activity of the 281 serum samples from our cohort, as shown by similar profiles of neutralization ( Figure 5) , 282 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. forms. Our cohort is composed of 140 patients, one third of whom are ICU patients, which 305 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint represent the largest clinical cohort published to date in which neutralizing activity has been 306 evaluated. Those characteristics allowed us to address important questions such as correlations of 307 nAb activity with disease severity, IgG response, or kinetics of antibody levels. Here, we 308 confirmed the heterogeneity of the humoral response in COVID-19 patients. While we found that 309 parameters, such as age or gender, were not associated with nAb activity, we show that among 310 patients with mildest disease forms, many did not exhibit robust neutralizing activities ( Figure 311 associated with viral clearance, it seems that a robust nAb activity does not confer protection 319 against disease progression in COVID-19 patients. One possibility is that disease severity is 320 correlated with higher viral loads and hence, with more antigens available for induction of 321 antibodies. Yet, no correlation between viral loads and nAb activity could be identified in our 322 cohort (Table S1 ), though the assessment of the viral loads by the first positive qPCR could be 323 unreliable, due to parameters that can influence the quality of sampling [27] . Alternatively, a 324 robust humoral response may be a feature of an overall exaggerated immune activation in severe 325 SARS-CoV-2 infection. Indeed, antibodies could mediate additional immune functions that may 326 have both protective and pathological consequences. In addition, humoral responses have been 327 shown to be corelated with cytokines and chemokines levels in COVID-19 patients [28] that are 328 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. activity be associated with protection against subsequent re-infections, the determination of anti-340 S IgG titers, by e.g., ELISA, may be useful to discriminate protected from unprotected 341 individuals, in particular for patients with the most moderate forms, since they are less prone to 342 develop robust neutralizing activity. In our cohort, 95% of sera with anti-S antibody values 343 above 124 AU/mL were associated with robust neutralization (90% neutralization or more), 344 suggesting that anti-S antibody determination could be used as an evaluation of nAb activity, 345 rather than for anti-N IgG assays. 346 We also addressed the question of the stability of the nAb levels, owing to the availability of 348 several serial samples for most of the patients used in the study. We confirmed a tendency for a 349 decrease of the nAb activity as well as of anti-S IgG after reaching a peak, as previously shown 350 by others [32], indicating that for some patients, the nAb activity may be very transient ( Figure 351 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint 3). Importantly, for ICU patients, we were able to give a more precise estimation of the 352 persistence of the neutralizing activity, with a preliminary estimation of 26 days for its half-life. 353 Thus, our model predicts that the SARS-CoV-2 nAb activity is likely to rapidly vanish. These 354 results, which need to be confirmed by studying alternative COVID-19 cohorts and more Finally, we also analyzed 9 serum samples from patients infected with alternative coronaviruses 371 that cause mild symptom in adults, including respiratory illnesses, and enteric and neurological 372 diseases [34] . The cross reactivity between other coronaviruses and SARS-CoV-2 immunity is 373 key to understand, as it might influence the severity of the disease or the response to a vaccine 374 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint [35] . In line with this, T cell reactivity against SARS-CoV-2 was observed in unexposed people 375 [36] . Interestingly, by investigating the cross reactivity of circulating Abs, we found that none of 376 the 9 samples displayed cross-reacting nAbs against SARS-CoV-2 infection. Furthermore, cross-377 neutralization against SARS-CoV-2 can be induced by sera from convalescent SARS-CoV 378 patients, which is likely due to high homology between these two viruses, although it seems to be (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . a n t i N (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted September 1, 2020. . https://doi.org/10.1101/2020.08.27.20182493 doi: medRxiv preprint COVID-19 404 diagnosis and management: a comprehensive review SARS-CoV-2 Infection Leads Neurological Dysfunction CoV-2: Apprehensions, opportunities, and the path forward Rapid generation of neutralizing antibody responses in COVID-19 patients. medRxiv Diagnose Novel Coronavirus Disease (COVID-19) Early detection of SARS-CoV antibodies in COVID-19 patients as a serologic marker of infection Neutralizing antibody responses to 424 -2 in a COVID-19 recovered patient cohort and their implications Antibody responses to SARS-CoV COVID-19 patients: the perspective application of serological tests in clinical practice The time course of the immune response to 430 experimental coronavirus infection of man Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of the COVID-19 Important role for the transmembrane 442 domain of severe acute respiratory syndrome coronavirus spike protein during entry Human serum 445 facilitates hepatitis C virus infection, and neutralizing responses inversely correlate with viral 446 replication kinetics at the acute phase of hepatitis C virus infection Epidemiology and clinical features of COVID-19: A review of current 449 literature COVID-19 and Factors Associated with Severe Clinical Course: A Systematic Review and 452 Comorbidity and its impact 454 on 1590 patients with COVID-19 in China: a nationwide analysis Spike on Viral Infectivity and Antigenicity. Cell. 2020 Serologic responses 464 to SARS-CoV-2 infection among hospital staff with mild disease in eastern France Antibody responses to SARS-CoV-2 in 467 patients of novel coronavirus disease 2019 SARS-CoV-2 in COVID-19 Inpatients and Convalescent Patients Asymptomatic and Mild COVID-19 Patients SARS-CoV-2 T cell immunity: Specificity, function, durability, 475 and role in protection Potential preanalytical and analytical vulnerabilities in 477 the laboratory diagnosis of coronavirus disease 2019 (COVID-19) Profiling serum cytokines in COVID-19 483 patients reveals IL-6 and IL-10 are disease severity predictors hyperimmune immunoglobulin for people with COVID-19: a living systematic review Structural insights into coronavirus entry Longitudinal 491 evaluation and decline of antibody responses in SARS-CoV-2 infection SARS-coronavirus in SARS patients and their clinical significance Human Coronavirus-229E, -OC43, -NL63, and -HKU1 Pre-existing immunity to SARS-CoV-2: the knowns and unknowns Supplementary Material