key: cord-0770825-0kdzthm2 authors: Grzelak, L.; Velay, A.; Madec, Y.; Gallais, F.; Staropoli, I.; Schmidt-Mutter, C.; Wendling, M.-J.; Meyer, N.; Planchais, C.; Rey, D.; Mouquet, H.; Glady, L.; Hansmann, Y.; Bruel, T.; De Seze, J.; Fontanet, A.; Gonzalez, M.; Schwartz, O.; Fafi-Kremer, S. title: Sex differences in the decline of neutralizing antibodies to SARS-CoV-2 date: 2020-11-15 journal: nan DOI: 10.1101/2020.11.12.20230466 sha: aaaa596112e9333a98254f00248cd52f335c76b0 doc_id: 770825 cord_uid: 0kdzthm2 The evolution of SARS-CoV-2 humoral response in infected individuals remains poorly characterized. Here, we performed a longitudinal study of sera from 308 RT-qPCR+ individuals with mild disease, collected at two time-points, up to 6 months post-onset of symptoms (POS). We performed two anti-S and one anti-N serology assays and quantified neutralizing antibodies (NAbs). At month 1 (M1), males, individuals > 50 years of age or with a body mass index (BMI) > 25 exhibited higher levels of antibodies. Antibody levels decreased over time. At M3-6, anti-S antibodies persisted in 99% of individuals while anti-N IgG were measurable in only 59% of individuals. The decline in anti-S and NAbs was faster in males than in females, independently of age and BMI. Our results show that some serology tests are less reliable overtime and suggest that the duration of protection after SARS-CoV-2 infection or vaccination will be different in women and men. The duration of humoral immune responses to SARS-CoV-2 is poorly characterized and 44 debated. The male sex, greater age and a higher body mass index (BMI) are risk factors for a 45 more severe disease. It is well established that severe COVID-19 patients produce higher 46 antibody titers than asymptomatic or mildly symptomatic individuals 1-4 . However, some studies 47 reported stable antibody levels within the first three months of recovery, whereas others 48 showed a rapid decrease in convalescent patients regardless of disease severity 1-3,5-8,9,10 . 49 Cross-sectional analyses of 10,000-30,000 individuals showed a relative stability of the 50 humoral response 7,11 . Anti-Spike (S) antibody amounts correlate with virus neutralization 51 capacity, since S is the main, if not the unique, target for neutralizing antibodies. Neutralizing 52 antibody titers also vary depending on the time post-onset of symptoms (POS) and the severity 53 of disease 1-4 12 . Very little is known about the influence of characteristics such as sex, age, 54 body mass index (BMI) on the longevity and efficiency of anti-SARS-CoV-2 antibodies, 55 particularly in mildly symptomatic individuals, who represent the majority of COVID-19 cases. 56 We analyzed the longitudinal antibody response in a monocentric cohort of 308 RT-57 qPCR confirmed staff from Strasbourg University Hospitals. The inclusion criteria are indicated 58 in supplemental Fig. 1 . The cohort included 75% females, with median age of 39 years (Table 59 1). The participants were nurses, doctors, caregivers and administrative staff. Contact with a 60 COVID-19 patient, within or outside of the hospital, was reported by 37% of individuals, and 61 290 participants (94%) had mild symptoms consistent with COVID-19 (dry cough, fever, 62 dyspnea, anosmia or ageusia…). Sixteen (5%) participants were hospitalized for moderate 63 disease. None progressed to severe or critical illness. The median (interquartile range (IQR)) 64 time from onset of symptoms to PCR testing was 3 (1-5) days. All individuals were sampled 65 twice. The first blood sampling (M1) was performed at a median (IQR) of 31 (24-38) days POS 66 (range: 11-58) and the second one (M3-6) at a median (IQR) of 107 (92-130) days POS (range: 67 78-172). 68 Seropositivity rates in this cohort were estimated based on time-points of collection, with 69 four different assays (Fig. 1) . We used the flow cytometry based S-Flow assay to measure 70 anti-Spike IgG, because it displays high sensitivity (>99 %) and specificity (100%), as 71 previously determined with pre-pandemic and SARS-CoV-2 positive samples 13,14 . All 72 participants had anti-SARS-CoV-2-antibodies by S-Flow at M1 and only 3 participants became 73 negative at M3-6. The second assay was a lateral flow assay (LFA, Biosynex TM ), to detect anti-74 Spike IgG and IgM. This test was slightly less sensitive than the S-Flow, but stable over-time, 75 with 85% individuals seropositive for IgG at the two time-points. IgM detection was higher at 76 M1 (93% of seroprevalence) and decreased to 79% at M3-6, likely reflecting the contraction of 77 the IgM response. Measurement of anti-N IgG antibodies with an ELISA (EDI TM ) gave similar 78 3 results than the LFA at M1, but only 59% of individuals remained positive at M3-6. This sharp 79 decrease may reflect a lower abundance of anti-N-specific antibodies in mild disease, a 80 different kinetic of the anti-N response, or a lower sensitivity of the test 15 . It has recently been 81 reported that differences in the sensitivity of ELISA tests, including those detecting anti-N 82 antibodies, may be dependent on the days POS 12, 16 . 83 We then quantified the neutralizing activity of the sera, to assess the potential protection 84 of the humoral immune response. While different tests are currently available, we selected a 85 neutralization assay based on lentiviral particles pseudotyped with the Spike protein. We and 86 others previously demonstrated a strong correlation of neutralization titers obtained with single-87 cycle pseudovirus and infectious SARS-CoV-2 14,16,17 . We tested a single dilution of the sera 88 (1:100) to calculate the neutralization activity. We previously reported a correlation between 89 the percentage of neutralization of the lentiviral pseudotypes measured at this non saturating 90 1:100 dilution, and the titers obtained by performing serial dilutions of the sera 14 . We defined 91 a threshold of a positive neutralization at 20% (ID20 neutralization), based on background 92 signal obtained with pre-pandemic sera (specificity of the test: 97.9%, not shown). With this 93 threshold, 95% and 84% of the sera were positive at M1 and M3-6, respectively ( Fig. 1 ). Applying more stringent thresholds (50% or 80% neutralization) confirmed a decline of the 95 neutralization activity over time (Extended data Fig. 1 ). 96 We next assessed the dynamics of the immune response by comparing antibody levels 97 at different time points POS (Fig. 2) . To provide a quantitative measurement of anti-S IgG with 98 the S-Flow, we defined a Binding Unit (BU), which corresponds to the fluorescent signal 99 standardized by establishing a dose-response curve with a reference anti-Spike human 100 monoclonal antibody. A longitudinal analysis demonstrated a slight but significant decrease of 101 anti-S IgG amounts between M1 and M3-6 ( Fig. 2A) . This decline was also visible with 102 neutralization activity and anti-N IgG levels of sera ( Fig. 2A) . We did not use the LFA in this 103 analysis, since it does not provide quantitative results. As expected 14,16 , we observed a 104 correlation between neutralization activity and anti-S or anti-N IgG in the sera (not shown). Plotting the median values of anti-S, neutralizing and anti-N antibodies at different time 106 intervals confirmed this slow decline and further showed large inter-individual variations ( Fig. 107 2B). 108 We then sought to determine whether variations of the antibody response may be 109 attributed to the biological or clinical characteristics of the participants. We analyzed the 110 associations between antibody levels (anti-S IgG, neutralizing activity or anti-N IgG) and the 111 sex, age, BMI and type of symptoms. This analysis was performed at M1 and M3-6. We also 112 calculated the slope of the curves between the two time-points, to assess the impact of the 113 participants' characteristics on the evolution of the immune response. Levels of anti-S and 114 neutralizing antibodies were higher in males than in females at M1 (Fig. 2 ), in line with previous 115 reports of a more robust induction of the immune response (cytokines and antibodies) in male 116 patients 18, 19 . This difference was no longer visible at M3-6. Accordingly, the slope of antibody 117 decline was significantly steeper in males (Fig. 2) . A multivariate analysis showed that anti-S 118 and neutralizing antibodies were higher at the first time-point (Extended data Table 1 ) and 119 declined more rapidly in males ( Table 2) , independently of other parameters. There was no 120 significant difference between males and females regarding the levels and decline of anti-N 121 IgG. 122 The majority of individuals showed a decline whereas other displayed stable antibody 123 amounts. We categorized the subjects, based on the stability of the humoral response, into 124 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. ; 4 "sustainer" and "decayer" groups 20 (Extended data Fig. 2 Fig. 128 2B). We also noted that female subjects were in higher proportion sustainers than decayers 129 compared to males (Extended data Fig. 2C ), in line with our observation that antibodies persist 130 for longer periods of time in women (Fig. 2B) . 131 Categorization of the participants by age (≤30, 30-50 and >50 years old) and by BMI (17-132 25, ≥25) further showed that older participants and those with a high BMI had higher antibody 133 titers at M1, as seen with anti-S, neutralization and anti-N antibodies (extended data Fig. 3 ). 134 However, the decline of antibody levels occurred at the same rate, regardless of age or BMI 135 ( Table 2 and extended data Fig. 3) . 136 There was no association of reported clinical signs, except anosmia/ageusia, to the 137 amount of antibodies at M1 nor with their evolution overtime (extended data Table 1 and Table 138 2). This likely reflects the homogeneity of symptom severity in this cohort, as all participant 139 suffered from a mild-to-moderate disease. A minor fraction of patients were temporarily 140 hospitalized (n=16). As previously reported 14,21 , the antibody levels at M1 were higher in 141 hospitalized individuals but they decreased at the same rate than non-hospitalized patients 142 (extended data Table 1 and extended data Fig. 4 ), Multivariate analyses indicated that high 143 antibody levels at M1 were associated with a more rapid decline, independently of any other 144 parameters (Table 2 ). This is in line with a recently reported link between antibody persistence 145 and disease severity 20 . 146 In conclusion, we performed here a longitudinal analysis of the humoral immune 147 response in a cohort of 308 RT-PCR confirmed SARS-CoV-2 infected patients with mild 148 disease. We report that antibody levels and neutralizing activity decline over the 172 days of 149 analysis. Neutralizing antibody titers decreased twice as fast as anti-S IgG, with half-lives of 150 19.9 and 41 weeks, respectively. Assessing the humoral response on the long-term is critical 151 to evaluate immune protection at the population level. Commercially available assays have 152 been validated with sera collected from acutely or recently infected individuals. We show here 153 that some of these assays are not sensitive enough and not suited for long-term analyses. This 154 may explain some discrepant results in the literature regarding the stability or waning of 155 antibody titers in convalescent patients. We further report sex differences in the longevity of 156 the immune response, with males displaying higher antibody levels shortly after infection, but 157 a steeper decrease. Multiple studies have demonstrated that women develop more robust 158 responses to infections and vaccination and are more sensitive to auto-immune diseases than 159 men 22,23 . This may be linked to sex hormones, X chromosomal and environmental factors. SARS-CoV-2-infected women mount significantly more robust T cell activation than male 161 patients 18 , which will also impact the duration of the response. Future work will help 162 determining whether the sex differences reported here are amplified over longer periods of 163 time, and may be linked to differences in antigen persistence 12 . It will also be of interest 164 extending our analysis on antibody longevity to other categories of persons, including 165 asymptomatic individuals who represent the majority of SARS-CoV-2 cases, patients who 166 recovered from more severe forms of COVID-19, and volunteers engaged in vaccine trials. 167 Whether future vaccines might provide a longer protection in women than in men remains an 168 outstanding question. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. ; https://doi.org/10.1101/2020.11.12.20230466 doi: medRxiv preprint S-Flow assay. A first version of the S-Flow assay was previously described 14 and was adapted 214 by using 293T cells stably expressing the codon-optimised SARS-CoV-2 spike protein (293T 215 Spike cells). Stainings were also performed on control (293T Empty) cells. The specificity and 216 sensitivity of this assay were assessed using 253 pre-pandemic samples and 377 RT-PCR 217 confirmed SARS-CoV-2 cases. The sensitivity is 99.2% with a 95% confidence interval of 218 97.69%-99.78% and the specificity is 100% (98.5-100). As previously described, 5x10 4 cells 219 were plated in a 96-well round bottom plate and 50 µL of a 1:300 dilution of the serum in MACS 220 Buffer was added. The mix was incubated for 30 minutes at 4°C. The cells were then washed 221 with PBS and stained with anti-IgG Alexa Fluor 647 secondary antibody (dilution 1:600; 222 Thermo Fisher Scientific) for 30 min at 4°C. Cells were fixed for 10 min with 4% 223 paraformaldehyde and data was acquired on an Attune NxT instrument (Life Technologies). 224 Results were analysed with FlowJo 10.7.1 (Becto Dickinson & Company). After testing 253 225 pre-pandemic samples, the positivity of a sample was defined as a specific binding above 40%. 226 The specific binding was calculated as follow: 100 x (% binding 293T Spike -% binding 293T 227 Empty)/ (100 -% binding 293T Empty). (1 -mean (luciferase signal in sample duplicate)/mean(luciferase signal in virus alone)).A 240 titration with a human anti-S monoclonal antibody (mAb 48) was added to each plate to verify 241 the reproducibility of the assay. 242 Statistical analysis. Baseline characteristics between men and women were compared using 243 a chi-square test for categorical variable and student's t-test for continuous variables. Correlations between antibody measures at M1 were estimated. Factors associated with BU 245 and neutralization levels at M1 were investigated using linear regression models, while factors 246 associated with IgM and IgG positivity at M1 were investigated using logistic regression 247 models. The difference in BU and neutralization levels between M1 and M3-6 was then 248 estimated and standardized by the time interval between the two timepoints. Factors 249 associated with these standardized differences were investigated using linear regression 250 models. For these analyses, factors that were associated with the outcome with a p-value 251 <0.15 in univariate analysis were introduced in the multivariate model. A p-value <0.05 was 252 considered statistically significant. Subjects were divided into "sustainer" or "decayer" 253 categories, for anti-S IgG, anti-N IgG and neutralizing antibody responses. To determine 254 appurtenance at one the two categories, we calculated the fraction: antibody value at M3-6 255 divided by value at M1. If the fraction was ≥ 1, the donor was considered as "sustainer" for this 256 antibody while it was considered as "decayer" if the fraction was <1. The half-life of decayers 257 was then extrapolated from the equation of the segment formed by the two timepoints. It is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. ; Fig. 1. Temporal evolution of anti-SARS-CoV-2 antibodies. A. The number of individuals displaying anti-S (S-Flow or Biosynex TM tests), anti-N (ELISA N EDI TM ) or neutralizing antibodies were plotted at Month 1 (M1) and M3-6 post onset of symptoms (POS). The percentages of positive cases are indicated in the bars. Neutralization positivity was defined as a neutralizing activity against lentiviral pseudotypes higher than 20%, at a serum dilution of 1:100 (defined as Inhibitory dose (ID20)). Differences between time-points were analyzed with a Chi-square test, **** p <0.0001. B. The level of antibodies defined as Binding Units for anti-S IgG (S-Flow assay), percentage of neutralization, and Optical Density (OD) for anti-N IgG (ELISA) were plotted against the number of days POS. Pink and purple points stand for M1 and M3-6 time-points, respectively. Each grey line connects the two timepoints from a same donor. The black line represents the median of all samples for each time-point. Paired Wilcoxon test was performed between M0 and M3-6, **** p<0.001. C. Donors were grouped by days POS in 14-or 29-day categories. BU, neutralization percentage and OD are shown. Results are medians with interquartile range. . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. ; https://doi.org/10.1101/2020.11.12.20230466 doi: medRxiv preprint Table 2 . Multivariate analysis of antibody evolution. A univariate analysis of the temporal evolution of antibody levels was first performed for each of the indicated category. When the p value was < 0.15, a multivariate analysis by linear regression was performed for the indicated categories. For each type of antibodies (anti-S, anti-N, neutralizing), the factor effect with its 95% confidence interval (95% CI) and the associated p-value are indicated. Non-significant (ns) corresponds to an analysis with a univariate p-value higher than 0.15. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 15, 2020. ; https://doi.org/10.1101/2020.11.12.20230466 doi: medRxiv preprint Clinical and immunological assessment of asymptomatic SARS-CoV-2 291 infections Persistence and decay of human antibody responses to the receptor binding 293 domain of SARS-CoV-2 spike protein in COVID-19 patients Rapid Decay of Anti-SARS-CoV-2 Antibodies in Persons with Mild Covid-295 Dissecting antibody-mediated protection against SARS-CoV-2 Loss of Anti-SARS-CoV-2 Antibodies in Mild Covid-19 Longitudinal observation and decline of neutralizing antibody responses in the 301 three months following SARS-CoV-2 infection in humans Robust neutralizing antibodies to SARS-CoV-2 infection persist for months Cross-Sectional Evaluation of Humoral Responses against SARS-CoV-2 Spike Persistence of serum and saliva antibody responses to SARS-CoV-2 spike antigens 307 in COVID-19 patients A comprehensive, longitudinal analysis of humoral responses specific to four 309 recombinant antigens of SARS-CoV-2 in severe and non-severe COVID-19 patients Repeated cross-sectional sero-monitoring of SARS-CoV-2 in New York 312 City Evolution of Antibody Immunity to SARS-CoV-2. Biorxiv Serologic responses to SARS-CoV-2 infection among hospital staff with 316 mild disease in eastern France A comparison of four serological assays for detecting anti-SARS-CoV-2 318 antibodies in human serum samples from different populations Distinct antibody responses to SARS-CoV-2 in children and adults across 320 the COVID-19 clinical spectrum Longitudinal analysis of serology and neutralizing antibody levels in 322 COVID19 convalescents Serological Assays Estimate Highly Variable SARS-CoV-2 Neutralizing 324 Antibody Activity in Recovered COVID19 Patients Sex differences in immune responses that underlie COVID-19 disease 327 outcomes Convergent antibody responses to SARS-CoV-2 in convalescent 329 individuals Quick COVID-19 Healers Sustain Anti-SARS-CoV-2 Antibody Production Virological assessment of hospitalized patients with COVID-2019 SeXX matters in immunity Implications of Sex Differences in Immunity for SARS Evaluation of the performance of SARS-CoV-2 serological tools and their 338 positioning in COVID-19 diagnostic strategies We thank the patients for participation to the study, Maaran Michael