key: cord-0824591-pos0jxrw authors: Schiffner, J.; Backhaus, I.; Rimmele, J.; Schulz, S.; Moehlenkamp, T.; Klemens, J. M.; Zapf, D.; Solbach, W.; Mischnik, A. title: Long-term course of humoral and cellular immune responses in outpatients after SARS-CoV-2 infection date: 2021-06-25 journal: nan DOI: 10.1101/2021.06.24.21259218 sha: 7b4a955114cfc070556923d0335de4bd31480ef5 doc_id: 824591 cord_uid: pos0jxrw Characterisation of the naturally acquired B and T cell immune responses to SARS-CoV-2 is important for the development of public health and vaccination strategies to manage the burden of COVID-19 disease. We conducted a prospective, longitudinal analysis in COVID-19 recovered patients at various time points over a 10-month period in order to determine how circulating antibody levels and interferon-gamma (IFN-{gamma}) release by peripheral blood cells change over time following natural infection. From March 2020 till January 2021, we enrolled 412 adults mostly with mild or moderate disease course. At each study visit, subjects donated peripheral blood for testing of anti-SARS-CoV-2 IgG antibodies and IFN-{gamma} release after SARS-CoV-2 S-protein stimulation. Anti-SARS-CoV-2 IgG antibodies were identified in 316/412 (76.7%) of the patients and 215/412 (52.2%) had positive neutralizing antibody levels. Likewise, in 274/412 (66.5 %) positive IFN-{gamma} release and IgG antibodies were detected. With respect to time after infection, both IgG antibody levels and IFN-{gamma} concentrations decreased by about half within three hundred days. Statistically, IgG and IFN-{gamma} production were closely associated, but on an individual basis we observed patients with high antibody titres but low IFN-{gamma} levels and vice versa. Our data suggest that immunological reaction is acquired in most individuals after infection with SARS-CoV-2 and is sustained in the majority of patients for at least 10 months after infection. Since no robust marker for protection against COVID-19 exists so far, we recommend utilizing both, IgG and IFN-{gamma} release for an individual assessment of immunity status. Infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) leads to various symptoms, 62 including cough, fever, cold, and loss of smell and taste. The course of the disease varies in symptoms and severity, 63 from asymptomatic infections to severe pneumonia with lung failure and death. Manifestation indices are 64 estimated to be 55-85% 1 . About 48 % of patients are women, and 52 % are men. In Germany, 2.6 % of all persons 65 with confirmed SARS-CoV-2 infections died in connection with a COVID-19 illness. The main risk factors for 66 death are age and comorbidities like diabetes or obesity. The diagnosis is based on clinical grounds and proven by 67 virus detection through RT-PCR in respiratory samples. In the present study, we analysed the long-term course of the immune response with respect to serum IgG 74 antibodies and the capacity of peripheral blood cells to produce interferon-gamma (IFN-γ) upon viral S-Protein 75 specific stimulation. Based on previous experience of our group 9 with the low diagnostic significance of IgA and 76 IgM antibodies in the long-term course of infection, we deliberately determined only IgG antibodies. The study was performed from May 2020 until January 2021. None of the study participants had received a 78 79 80 81 Study area The study was performed on patients which were notified as index cases to the Health Protection Authority of the 84 City of Luebeck/Germany (approx. 220,000 inhabitants, population density approx. 1,000/sqm). With the 85 exception of two major outbreak-related periods in December 2020 and mid-January 2021 the city has mostly been 86 a low-incidence region, when compared to Germany as a whole (Fig. 1 ). Study population The data presented here were obtained from the sera of patients that were notified to the local Health Protection 90 Authority as being SARS-CoV-2 positive by PCR irrespective of the clinical manifestation. All of them recovered 91 from the disease without hospitalization. In total, 1,279 patients were invited by e-mail to participate. 111 of these 92 were first infected in the "first wave" of the pandemic between February 27, and July 31, 2020. The results of the 93 initial antibody profiling from these patients have been published elsewhere. 15 No-response n = 843 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. It can be clearly seen that the antibody levels decline over time. There was a moderate but significant negative 220 linear relationship between IgG ratio and the time passed since the positive SARS-CoV-2 diagnosis (r = -0.3, p < 221 0.001) (Fig. 3 ). When we sought for factors that were associated with the presence of antibodies, it became clear that disease 223 severity was positively associated, whereas reported comorbidities had no impact on antibody development (Model 224 I; OR: 1.69, 95% CI: 1.10 -2.84; Table 2 ). As could be seen in Fig. 3 237 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 June 25, 2021. ; https://doi.org/10. 1101 The odds of having antibodies with neutralizing capacity were almost two-fold higher in participants who had a 238 moderate to severe COVID-19 disease course as compared to those with none or mild disease (Model I; OR. 1.96, 239 95% CI: 1.18 -2.27; Table 3 ). Similar to IgG SARS-CoV-2 antibodies, elapsed time since the confirmed COVID-240 19 diagnosis was inversely associated with the probability of expressing neutralizing antibodies with a clear 241 weaning after 60 to 119 days and even more after 120 days (Model I, OR. 0.15, 95% CI: 0.08 -0.31). Individuals 242 for which the time since SARS-CoV-2 testing was more than 120 days ago were less likely to have neutralizing 243 antibodies (OR: 0.14; 95% CI: 0.07 -0.28) compared to those with a diagnosis made between 0-59 days apart 244 from the time of the serological survey. For female participants, although not statistically significant, the odds ratio 245 for having neutralizing antibodies was slightly decreased (Model II; OR, 0:94; 95%CI: 0.58 -1.52). The odds 246 ratios in Model II did not significantly change when adjusting for age and sex, suggesting that both variables are 247 not confounding the association between disease course, time since rt-PCR testing and having SARS-CoV-2 248 neutralizing antibodies. The mean antibody ratio was highest in the first 1-3 months in patients with a severe disease course. In 250 asymptomatic or mild symptomatic patients, the highest mean antibody level, although at a lower level, was also 251 observed 1-3 months post COVID-19 infection, with a declining trend in the subsequent time windows (Table 4. ) 252 The data thus clearly show that antibody expression is related to the severity of disease and that antibody levels 253 fade continuously within approx. 300 days (Fig. 3, Table 4 ). Looking at the correlation between antibody levels and IFN-γ concentrations, a heterogeneous picture emerged 264 (Fig. 5) . While in most cases both values were concordant, there were a considerable number of cases with high 265 to very high IFN-γ levels and low antibody levels and vice versa. Looking at the cloud of dots, it is striking that 266 there appears to be a population of patients whose cells produce extremely high levels of IFN-γ (> 2,500 mIU/ml), 267 regardless of the antibody response (Fig. 4) or the time post SARS-CoV-2 diagnosis (Fig. 3) . In summary, in our SARS-CoV-2 PCR-positive cohort, the antibody profile was heterogeneous. In most cases, In the present study, we determined the development of specific humoral and cellular immune responses in 280 outpatients recovered from SARS-CoV-2 infection, mostly with mild to moderate disease (Fig. 2) . Therefore, we 281 chose to determine serum IgG antibodies and IFN-γ release in response to the viral spike (S) glycoprotein in view 282 of the time that had passed since the infection. In the present study about 22 % of the patients had no detectable antibodies. These data substantiate our previous 284 finding from the first wave of the pandemic in early 2020 9 . On a population basis, the clinical severity of the 285 disease was positively correlated with the level of SARS-CoV-2 neutralising antibodies (Tables 3, 4), as had been 286 shown previously by others 12 . On an individual level, there was a great variability between patients. Thus, the 287 individual level of antibodies is not of diagnostic value, for example for the assessment of patients with long-288 COVID syndrome, which is often associated with chronic fatigue 13 . As expected, in many cases the antibody levels 289 faded over time. When compared to the first three months after infection, the mean antibody levels decayed steadily 290 and approximately halved within 300 days (Fig. 3) . Although our data do not allow a meaningful calculation of 291 the half-life, the finding is in accordance with reports from others 14, 15 , who calculated the half-life of neutralizing 292 IgG antibodies with 140 to 220 days. Interestingly, after COVID vaccination neutralisation capacity was different from natural infection 16 . Very few published data sets compare antigen-specific B-cell and T-cell immunity. We therefore examined 295 interrelationships between IgG antibody levels and IFN-γ release. Like with the antibody kinetics, the IFN-γ values 296 decayed over time after infection with similar kinetics (Fig. 4) which is in line with other works 17 . Unexpectedly, 297 however, we saw a substantial number of patients with low antibody levels and extremely high IFN-γ levels and 298 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 June 25, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 vice versa (Fig. 5) . Although unlikely, it cannot be completely ruled out, that the observed T-cell reactivity was 299 due to pre-existing memory T-cells recognizing the common cold coronaviruses, as has been described before 18, 300 19 . Existing T cells might be an explanation for cross reactions in an IGRA assay which could also be seen in our 301 study 20 . To explain this dichotomy, a longitudinal in-depth analysis of the precise numbers and types of IFN-γ 302 producing cells will be necessary, especially the characterization of T memory cells. Since our data and those from others show that determination of antibodies alone is not predictive for protection 304 against SARS-CoV-2 disease, simultaneous determination of IFN-γ may be a valuable adjunct and may also 305 predict the time-point for possibly necessary booster vaccinations on an individual basis. Strengths and limitations 308 The current study has few limitations. Longitudinal data for each subject, with at least three time points per subject, 309 would be required for more precise understanding of the kinetics of durability of SARS-CoV-2-specific 310 antibodies. Nevertheless, the current cross-sectional data describe well the dynamics of spike-specific antibodies 311 over 10 months and IFN-γ release by blood T-lymphocytes at one study point. This study was not sufficiently 312 powered to control for many variables simultaneously. This study shows that on average 9.8 months after detection of SARS-CoV-2 in nasopharyngeal specimen, the Table 1 . Sample characteristics (n=412) 397 398 Table 2 . Factors associated with the presence of SARS-CoV-2 IgG antibodies (n=412) 399 400 Table 3 . Factors associated with the presence of SARS-CoV-2 neutralizing antibodies (n=412) 401 402 Table 4 . Mean antibody ratio since confirmed COVID-19 diagnosis and disease severity 403 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 June 25, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 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 June 25, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 (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 June 25, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 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 June 25, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 (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 June 25, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 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. (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 June 25, 2021. ; (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Legend: 1.0 = reference category; OR: Odd Ratio; CI: Confidential Interval; Table 2 presents the unadjusted (Model I) and the age and sex-adjusted odds ratio (OR) (Model II) of seropositivity (Yes/No) by participant characteristics. The odds ratio was calculated to describe the risk of different groups in positive ELISA serums compared with non-positive ELISA sera. n = 412. 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 June 25, 2021. ; https://doi.org/10.1101/2021.06.24.21259218 doi: medRxiv preprint Legend: 1.0 = reference category; OR: Odd Ratio; CI: Confidential Interval; Table 3 presents the unadjusted (Model I) and the age and sex-adjusted odds ratio (OR) (Model II) of having neutralizing antibodies (Yes/No) by participant characteristics. The odds ratio was calculated to describe the factors of different groups in positive neutralizing antibody assays compared with negative neutralizing antibody assays. 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 June 25, 2021. ; https://doi.org/10.1101/2021.06.24.21259218 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 June 25, 2021. ; https://doi.org/10. 1101 Prevalence of Asymptomatic SARS-CoV-2 Infection : A Narrative Review Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor Antibody responses CoV-2 in patients with COVID-19 Mapping 338 Neutralizing and Immunodominant Sites on the SARS-CoV-2 Spike Receptor-Binding Domain by Structure-339 BNT162b1 elicits human antibody and TH1 T cell responses Longitudinal Isolation 344 of Potent Near-Germline SARS-CoV-2-Neutralizing Antibodies from COVID-19 Patients Immune Responses in Patients With Severe Acute Respiratory Syndrome Coronavirus 2 Infection Return-to-work, disabilities and 361 occupational health in the age of COVID-19 Immunological memory CoV-2 assessed for up to 8 months after infection Naturally acquired 365 SARS-CoV-2 immunity persists for up to 11 months following infection Antibody Response in Persons with Past Natural Infection Interferon-369 gamma release assay for accurate detection of SARS-CoV-2 T cell response -2 T cell epitopes in unexposed humans Characterization of