key: cord-310606-msmh7d8m authors: Westerhuis, B. M.; Aguilar-Bretones, M.; Raadsen, M. P.; de Bruin, E.; Okba, N. M. A.; Haagmans, B. L.; Langerak, T.; Endeman, H.; van den Akker, J. P. C.; Gommers, D. A. M. P. J.; van Gorp, E. C. M.; Rockx, B. H. G.; Koopmans, M. P. G.; van Nierop, G. P. title: Severe COVID-19 patients display a back boost of seasonal coronavirus-specific antibodies date: 2020-10-12 journal: nan DOI: 10.1101/2020.10.10.20210070 sha: doc_id: 310606 cord_uid: msmh7d8m Severe acquired respiratory syndrome coronavirus-2 (SARS-CoV-2) is the cause of coronavirus disease (COVID-19). In severe COVID-19 cases, higher antibody titers against seasonal coronaviruses have been observed than in mild cases. To investigate antibody cross-reactivity as potential explanation for severe disease, we determined the kinetics, breadth, magnitude and level of cross-reactivity of IgG against SARS-CoV-2 and seasonal CoV nucleocapsid and spike from 17 severe COVID-19 cases at the clonal level. Although patients mounted a mostly type-specific SARS-CoV-2 response, B-cell clones directed against seasonal CoV dominated and strongly increased over time. Seasonal CoV IgG responses that did not neutralize SARS-CoV-2 were boosted well beyond detectable cross-reactivity, particularly for HCoV-OC43 spike. These findings support a back-boost of poorly protective coronavirus-specific antibodies in severe COVID-19 patients that may negatively impact de novo SARS-CoV-2 immunity, reminiscent of original antigenic sin. targeted treatment to prevent or overcome severe disease 5 . Therefore, there is an urgent 48 need for detailed insight in the SARS-CoV-2 immune response in the context of a CoV-49 experienced immune system. 50 For other viruses, several mechanisms have been described on how pre-existing 51 immunity to antigenically related viruses affect the outcome of infection. First, for influenza 52 and dengue viruses the immunological imprint of the initial infection dominates the host 53 response towards subsequent related infections for life 8, 9 . This mechanism, termed "original 54 antigenic sin" (OAS), relates to the propensity of adaptive immunity to preferentially fight 55 5 A strong SARS2-N and SARS2-S IgM and IgG response is mounted in the vast majority 80 of infected individuals targeting the receptor binding domain (SRBD), the outer-(S1) and inner 81 domains (S2) to various extent 19 . In severe COVID-19 cases this response is dominated by IgG 82 targeting the S2 domain 16, 19 . There was no evidence of substantial IgG cross-reactivity towards 83 SARS-CoV-2 in healthy controls 16, 19 . In contrast, COVID-19 patients displayed elevated serum 84 IgG titers against the majority of seasonal CoV N and S antigens which suggests that the IgG 85 responses against these viruses are correlated by undefined mechanisms. Strikingly, high 86 serum OC43-SECTO IgG titers are associated with COVID-19 disease severity, which might be 87 related to stretches of amino acid sequence homology in S2 16, 17 . Together, these studies 88 suggest that high antibody titers to seasonal CoV may associate with poor disease outcome 89 in COVID-19 patients and support the hypothesis that the immunological background is key. 90 A limitation of serological studies is that monoclonal IgG cross-reactivity patterns cannot be 91 defined and the functional contribution of seasonal CoV-specific IgG clones to immune 92 protection or immunopathogenesis remains largely unknown. 93 At the clonal level, it was shown that pre-existing SARS-CoV-2 cross-reactive IgG 94 memory B-cells predominantly target the SARS-CoV-2 S2 with limited virus neutralization (VN) 95 potential 20 . Besides a shared epitope that has been described in S2 of SARS-CoV-2 and OC43 21 , 96 little is known on seasonal CoV IgG cross-reactivity patterns with SARS-CoV-2 and their 97 functional contribution. Pre-existing CoV-specific B-cells may provide protection from SARS-98 CoV-2 infection by rapid production of cross-reactive antibodies that aid in viral clearance, 99 e.g. by a faster rise of cross-neutralizing antibodies from memory recall compared to initiating 100 a primary humoral response 22,23 . However, based on observations in other viral infections, 101 pre-existing immunity may also promote pathology 7 . 102 . 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 October 12, 2020. Reactivity of secreted IgG in oligoclonal culture supernatants towards an array of CoV N and 139 S proteins was analyzed using protein microarray analysis (PMA) as described elsewhere 26 . 140 Paired sera of the respective PBMC samples were included in the analysis. Microarray slides 141 were scanned using a Powerscanner (Tecan, Switzerland). Background was determined for 142 each spot and mean fluorescence intensity signal (MFI, range 1 -65,000) of two spots was 143 calculated for each serum or culture supernatant. For B-cell culture supernatants a cut-off 144 was set at 1,000 MFI for all antigens based on the highest signal of non-reactive cultures. 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 October 12, 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 October 12, 2020. . https://doi.org/10.1101/2020.10.10.20210070 doi: medRxiv preprint All patients mounted a strong SARS-CoV-2 immune response shown by rising IgG titers against 174 SARS2-SECTO, SARS2-S1 and SARS2-SRBD which peaked around 4 weeks post onset of clinical 175 symptoms ( Figure 1A and Supplementary figure 1A for individual patients). These patients 176 were assumed never exposed to SARS-CoV and MERS-CoV. Nevertheless, there was a strong 177 IgG response towards all SARS-CoV antigens. Given the high structural similarities between N 178 and S of SARS-CoV and SARS-CoV-2 16,28 , likely a cross-reactive response was mounted. The 179 increase in MERS-SECTO IgG, but not MERS-N and MERS-S1 IgG and highest amino acid 180 homology within S2 with SARS-CoV-2 suggested this cross-reactive response targeted MERS-181 S2 17 . 182 Besides the SARS-CoV-2 IgG response, serum IgG reactivity towards various, but not 183 all seasonal CoV antigens, increased in longitudinally analyzed COVID-19 patients ( Figure 1A 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 October 12, 2020. . were stimulated oligoclonally in vitro at a limiting density that was empirically tested to give 196 a clonal B-cell response. Supernatants were individually screened for IgG reactivity towards 197 all available CoV antigens using PMA. The number of reactive B-cells was normalized for the 198 number of screened B-cells in order to compare frequencies between different samples 199 ( Figure 1B) . Overall, the frequency and total reactivity of in vitro stimulated peripheral blood-200 derived B-cells was representative for the total serum IgG reactivity ( Figure 1B 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 October 12, 2020. . patterns that significantly contribute to the induced immune response ( Figure 2B ). The level 220 of cross-reactivity as calculated by Pearson correlation test varied slightly over time due to 221 the limited number of clones per sample. Overall, the cross-reactivity patterns between 229E-222 N, NL63-N and HKU1-N became more apparent and statistically more significant when all 223 clones from three time points (n=125) are combined for analysis ( Figure 2B ). OC43-N and 224 MERS-N clones displayed no cross-reactivity and cross-reactivity between SARS-N and 225 seasonal CoV was considered non-substantial (Figure 2A/B) . 226 S-reactive IgG clones showed an increasing mostly SARS-CoV-2 type-specific response 227 with increasing numbers of detected SARS2-SECTO and SARS2-S1 B-cell clones. Strong SARS2-228 SECTO and SARS2-S1 cross-reactivity indicated that mostly shared epitopes within S1 were 229 targeted in this patient ( Figure 2C and D). The frequency and number of detected SARS2-SRBD 230 clones remained more or less stable, yet cross-reactivity towards SARS2-SECTO and SARS2-S1 231 increased over time suggesting maturation of the SARS-CoV-2 specific response. The initial 232 lack of cross-reactivity of SARS2-SRBD towards SARS2-SECTO and SARS2-S1 suggested that the 233 epitope(s) targeted by the SARS2-SRBD clones at day 9 were sterically hindered or displayed in 234 an alternative conformation in the larger SARS-CoV-2 S proteins on PMA ( Figure 2C ). The 235 increasing cross-reactivity suggests clones that target epitopes displayed on the outer surface 236 of S are naturally selected. 237 Strikingly, there was a dominant OC43-S ECTO -specific response that increased over 238 time. This response significantly cross-reacted with HKU1-SECTO and SARS-SECTO, but not 239 SARS2-SECTO ( Figure 2D ). This suggests the outgrowth of OC43-SECTO clones is not driven by 240 SARS-CoV-2 cross-reactivity. The limited cross-reactivity between OC43-SECTO and OC43-S1 241 suggested these clones predominantly target S 2 . The correlated response between the paired 242 . 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 October 12, 2020. . 12 S1 and SECTO antigens of NL63-, HKU1-, MERS-, SARS-and SARS-CoV-2 validated availability of 243 shared epitopes in both antigens. 244 The total number of detected SECTO-reactive clones (n=363) outnumbered S1-reactive 245 clones (n=85) and showed greater cross-reactivity between different CoV compared to the S1. In order to determine the overall pattern of IgG cross-reactivity between N and S of different 260 CoV in severe COVID-19 patients we pooled all available N (n=2039, Figure 3A ) and S-reactive 261 IgG clones (n=2699, Figure 3B ) from all 17 included donors. For N, Pearson correlation analysis 262 validated that 229E-N, NL63-N and HKU1-N IgG clones most substantially cross-reacted (range 263 R=0.52 -0.72, all p<0.0001, Figure 3C ). Additionally, weak yet highly significant correlations 264 between OC43-N and other seasonal CoV strains were observed (range R=0.066 -0.14, all 265 p<0.003). Conversely, no significant cross-reactivity was detected between seasonal and 266 . 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 October 12, 2020. . https://doi.org/10.1101/2020.10.10.20210070 doi: medRxiv preprint 13 emerging CoV indicating different epitopes to be involved for these recent zoonotic strains 267 ( Figure 3C ). 268 The combined analysis of all S-IgG clones showed complex cross-reactivity patterns, 269 including endemic seasonal and epidemic zoonotic strains ( Figure 3D ). All SARS-CoV and SARS-270 CoV-2 S antigens positively correlated (range R=0.14 -0.80, all p<0.0001). Additionally, all 271 corresponding S1 and SECTO antigens positively correlated (range R=0.35 -0.80, all p<0.0001) 272 validating availability of shared epitopes in paired antigens for all CoV strains. 273 Our cohort showed there was broad cross-reactivity within S1 between 229E-, NL63-, 274 HKU1-and OC43-CoV but not MERS-CoV. For MERS-CoV significant cross-reactivity was 275 restricted to SECTO, suggesting cross-reactive epitopes are mostly located in S2 ( Figure 3D ). 276 277 Given the synergistic outgrowth of OC43-SECTO and SARS-CoV and SARS-CoV-2 S-reactive B-280 cells in severe COVID-19 patients ( Figure 1C ) one would presume that SARS-CoV-2 cross-281 reactivity drives in vivo expansion of OC43-SECTO-specific B-cells. However, cross-reactivity 282 between OC43-SECTO and SARS-CoV-2 S antigens showed significant weak negative 283 correlations (range R=-0.038 --0.088, range p=4,7x10 -6 -0.014) suggesting that OC43-S ECTO 284 specific clones are unlikely to cross-react with SARS-CoV-2 S in severe COVID-19 patients 285 ( Figure 3B ). Indeed, the majority of OC43-SECTO clones (752 out of 920 clones, 82%) did not 286 cross-react with SARS-CoV-2 indicating these were initiated prior to SARS-CoV-2 infection. 287 Due to lack of a pre-sample we cannot draw firm conclusions on pre-existence of SARS-CoV-288 2 cross-reactive clones in this cohort. 289 . 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 October 12, 2020. To determine what circulating S-specific B-cells contribute to VN, the normalized S-305 reactive B-cell counts were correlated with paired serum SARS-CoV-2 VN titers. Similar to 306 serum IgG titers, SARS2-S ECTO , SARS-S 1 and SARS2-S 1 correlated with serum VN titers. 307 Contrasting to serum titers, OC43-S 1 B-cell numbers showed a positive correlation with VN. 308 Notably, the relatively low frequency of OC43-S1-reactive B-cells (<0.01% of screened B-cells) 309 suggested the potential functional contribution of these clones to VN in vivo was limited. 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 October 12, 2020. Severe COVID-19 patients displayed an evolving SARS-CoV-2 specific IgG response that 325 correlated with the increasing SARS-CoV-2 neutralization potential of serum during disease 326 progression. Additionally, these patients displayed a strong rise of pre-existing seasonal CoV-327 specific clones that did not correlate with SARS-CoV-2 neutralizing titers, which is suggestive 328 of OAS. Based on this mechanism, one would expect that cross-reactivity to SARS-CoV-2 329 would drive outgrowth of pre-existing seasonal CoV-specific B-cells. However, here we 330 observed a strong outgrowth of clones that selectively targeting seasonal CoV. Using Pearson 331 regression analysis, we were able to distinguish rare from substantial cross-reactivity 332 patterns. Although clones were identified that showed SARS-CoV-2 cross-reactivity, these had 333 a minor contribution to the overall SARS-CoV-2 response. 334 For N, very limited cross-reactivity was observed between seasonal and emerging CoV 335 strains. Even though we could not confirm this for SARS-CoV-2, the high level of SARS2-N and 336 SARS-N sequence homology (88.6%) and reported strong correlation between serum SARS2-337 . 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 October 12, 2020. only a minor fraction cross-reacted with OC43-S1 (66 out of 920 clones, 7%). This suggests the 344 majority of OC43-SECTO clones target S2, which corresponds with previous studies 16, 21 . The 345 fraction of OC43-SECTO clones that cross-reacted with SARS-CoV-2 was limited (168 out of 920 346 clones, 18%) but for the clones that did, reactivity was increasingly directed towards SARS2-347 S1 and SARS2-SRBD and not S2. As IgG targeting S1 and SRBD confer the strongest VN potential, 348 these OC43-SECTO clones contribute to evolving SARS-CoV-2-specific IgG response 23,31 . 349 OC43-S2 clones potentially have low-affinity reactivity towards SARS-CoV-2, that is 350 below the detection limit of protein microarray, which drives expansion of OC43-SECTO clones. 351 This is of interest because low affinity antibodies, with consequentially a poor neutralizing 352 potential, can mediate engagement of Fc receptors (FcRs) on innate immune cells and 353 promote viral entry through a mechanism termed antibody-dependent enhancement (ADE). 354 This may either result in productive infection and increased viral spread -this was shown for 355 flaviviruses 32 -or promote inflammation and tissue damage, as was described for SARS-356 CoV 33-35 . However, the level of OC43-SECTO clones that cross-react with SARS2-S2 remained 357 stable over time, as determined by the OC43-SECTO clones that selectively cross-react with 358 SARS2-SECTO (Week 1, 7%; week 2, 4 %; week 3, 4% and week >4, 5%). If OC43-SECTO clones 359 were selectively expanding due to undetected low affinity cross-reactivity, one would expect 360 that clones that have detectable cross-reactivity to have a growth advantage. The reason for 361 . 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 October 12, 2020. . https://doi.org/10.1101/2020.10.10.20210070 doi: medRxiv preprint this is that the clonal selection of B-cells is driven by the affinity of the B-cell antigen receptor 362 and SARS-SECTO-specific TH-cell-mediated linked recognition of antigens, the presence of which 363 has been confirmed in an largely overlapping patient cohort 25 . 364 We find no evidence of a functional contribution of OC43-SECTO reactive IgG to SARS-365 CoV-2 VN. OC43-S1 specific clones did correlate with SARS-CoV-2 VN titers although these 366 clones represent only a minor fraction of OC43-CoV specific clones. While it is possible that 367 OC43-SECTO specific IgG restrict viral replication by indirect mechanisms, they may also play 368 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 October 12, 2020. . 18 which individuals were infected or reinfected cannot be deferred. Potentially, the strong and 386 dominant OC43-SECTO response observed in severe COVID-19 patients relates to antigenic 387 seniority, but this cannot be confirmed. Regardless of the unknown timeline of preceding 388 infections, the outgrowth of seasonal CoV-specific clones is similar to the back-boost that was 389 initially described for influenza 11 . Here, using antibody landscapes of hemagglutination 390 inhibition titers it was shown that infection or vaccination induced a broad increase in titers 391 towards preceding viruses 11,13 . Based on the antigenic distance it was calculated that the back 392 boost extended beyond cross-reactivity 11,13 . However, this mechanism was never validated 393 by antigen binding at a clonal level or shown for any other infectious agents. To our 394 knowledge, we present the first data on a back boost of IgG clones towards a different virus 395 family and substantiate this mechanism with clonal reactivity. 396 For influenza, back-boost is considered to have a positive contribution to vaccine 397 response by helping maintain immunity towards a broad range of influenza viruses and 398 allowing improved coverage by preemptive vaccination updates 11 . Contrastingly, in OAS, 399 clones initiated by prior infections are shown to negatively impact de novo type-specific 400 response and worsen clinical outcome 8-10,13,38 . Whether the SARS-CoV-2-specific response is 401 hampered by back-boost of seasonal clones, similar to what has been described for OAS, is of 402 interest. Immune dominance of IgG reactivity towards seasonal CoV and the late peak SARS-403 CoV-2 response in these severe COVID-19 patients, 4 weeks after onset of clinical symptoms, 404 suggests the initiation of the SARS-CoV-2 response may be hampered by proliferation of 405 seasonal CoV-specific clones. 406 In summary, the detailed insights in kinetics and cross-reactivity patterns of N and S-407 reactive IgG presented in this study will aid in the interpretation of serological assays and 408 expands our understanding of how the humoral immune system responds towards a novel 409 . 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 October 12, 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 October 12, 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 October 12, 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 October 12, 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 October 12, 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 October 12, 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 October 12, 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 October 12, 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 October 12, 2020. 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 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 October 12, 2020. 2 4 6 8 10 12 14 16 229E NL63 HKU1 OC43 MERS SARS N 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 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 October 12, 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 October 12, 2020. Week 1 n=3 Week 2 n=10 Week 3 n=11 Week 4 n=3 Week 5 n=2 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 October 12, 2020. . https://doi.org/10.1101/2020.10.10.20210070 doi: medRxiv preprint A novel coronavirus from patients with pneumonia in China Clinical features of patients infected with 2019 novel coronavirus in 434 Clinical characteristics of 138 hospitalized patients with 2019 novel 436 coronavirus-infected pneumonia in Wuhan, China Epidemiological and clinical characteristics of 99 cases of 2019 novel 439 coronavirus pneumonia in Wuhan, China: a descriptive study Risk factors associated with acute respiratory distress syndrome and 442 death in patients with coronavirus disease 2019 pneumonia in Wuhan, China The potential danger of suboptimal antibody responses in Antibody-dependent 447 enhancement and SARS-CoV-2 vaccines and therapies Original antigenic sin in dengue 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 102 104 106 108 110 112 114 116 118 120 122 124 126 128 130 132 134 136 138 140 142 144 146 148 150 152 154 156 158 160 162 164 166 168 170 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 102 104 106 108 110 112 114 116 118 120 122 124 126 128 130 132 134 136 138 140 142 144 146 148 150 152 154 156 158