key: cord-0806723-0sd45kwq
authors: Becker, M.; Dulovic, A.; Junker, D.; Ruetalo, N.; Kaiser, P.; Pinilla, Y.; Heinzel, C.; Haering, J.; Traenkle, B.; Wagner, T.; Layer, M.; Mehrlaender, M.; Mirakaj, V.; Held, J.; Planatscher, H.; Schenke-Layland, K.; Krause, G.; Strengert, M.; Bakchoul, T.; Althaus, K.; Fendel, R.; Kreidenweiss, A.; Koeppen, M.; Rothbauer, U.; Schindler, M.; Schneiderhan-Marra, N.
title: Immune response to SARS-CoV-2 variants of concern in vaccinated individuals
date: 2021-03-10
journal: nan
DOI: 10.1101/2021.03.08.21252958
sha: 62d8133b4dae602c15cf00e6180b866ff80a69c5
doc_id: 806723
cord_uid: 0sd45kwq

The SARS-CoV-2 pandemic virus is consistently evolving with mutations within the receptor binding domain (RBD) being of particular concern. To date, there is little research into protection offered following vaccination or infection against RBD mutants in emerging variants of concern (UK, South African, Mink and Southern California). To investigate this, serum and saliva samples were obtained from groups of vaccinated (Pfizer BNT-162b2), infected and uninfected individuals. Antibody responses among groups, including salivary antibody response and antibody binding to RBD mutant strains were examined. The neutralization capacity of the antibody response against a patient-isolated South African variant was tested by viral neutralization tests and further verified by an ACE2 competition assay. We found that humoral responses in vaccinated individuals showed a robust response after the second dose. Interestingly, IgG antibodies were detected in large titers in the saliva of vaccinated subjects. Antibody responses showed considerable differences in binding to RBD mutants in emerging variants of concern. A substantial reduction in RBD binding and neutralization was detected for the South African variant. Taken together our data reinforces the importance of administering the second dose of Pfizer BNT-162b2 to acquire high levels of neutralizing antibodies. High antibody titers in saliva suggest that vaccinated individuals may have reduced transmission potential. Substantially reduced neutralization for the South African variant highlights importance of surveillance strategies to detect new variants and targeting these in future vaccines.

Since the initial outbreak in Wuhan, China in late 2019 9,10 , SARS-CoV-2 has evolved into a global pandemic, with more than 112 million infections and 2.5 million deaths (as of February 26 th , 2021) 11 , impacting severely on mental health 12,13 and global economics 14 . In response, the scientific community has made unprecedented progress, resulting in the generation of multiple vaccines, using a variety of different approaches 8,15,16 , such as the Pfizer BNT-162b2 vaccine, which encodes a full-length trimerized spike protein 17 . In parallel, SARS-CoV-2 is continually evolving which might impact its infectivity 2 , transmission 3,18,19 and viral immune evasion 20,21 . To date, advanced genomic approaches have identified thousands of variants of SARS-CoV-2 with multiple RBD mutations circulating due to natural selection 4,22 . The variability of RBD epitopes is of specific concern as such mutations might reduce vaccine efficacy, increase viral transmission or impair acquired immunity by neutralizing antibodies 2,23,24 . For the pandemic to be brought under control, herd immunity must be achieved through vaccination. However, there is discourse about how long antibodies generated during the first wave persist, with some studies suggesting seroreversion between two to three months 25 , while others find antibodies present for up to seven or eight months post-infection [26] [27] [28] . Alarmingly, antibodies generated during the first wave also appear to have reduced immunoreactivity and neutralization potency towards emerging variants 27 .

As the virus is known to continually mutate, particularly the emerging UK (B.1.1.7) 3 , South African (B.1.351) 5 , Brazil (P1) 29 , Mink (Cluster 5) 6 and Southern California (hereon referred to as "LA" (B1.429) 7 variants are of concern. The UK variant may have an increased risk of transmission 30 and potentially increased mortality 18, 30 . It further exhibits reduced neutralization susceptibility 21,31 , which is most substantially related to a subset of RBD-specific monoclonal antibodies 31 . The N501Y mutation appears to mediate increased ACE2-RBD interaction 1 and is known to be critical for SARS-CoV-2 infection in vivo in mice 32 . Similarly, the South African variant which is now spreading globally, has two escape mutations within the RBD (K417N and E484K) 5 in addition to the N501Y mutation. The combination of these three point mutations might result in both a higher infection rate and reduced capacity of neutralizing antibodies . 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint 5 produced against variants without RBD-mutations of concern (hereon referred to as "wildtype"). In light of these developments, and in spite of increasing data provided by vaccine companies, it remains unclear whether vaccines formulated against the original Wuhan strain of the virus will remain effective against new and emerging variants such as UK or South Africa.

To understand this, we characterized the antibody response post vaccination with the Pfizer BNT-162b2 vaccine in both serum and saliva and then investigated the presence and efficacy of neutralizing antibodies against emerging variants of concern (UK, South Africa, Mink and LA). 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint

To analyze the humoral response generated by vaccination, SARS-CoV-2 reactive antibody titers in serum samples from vaccinated, convalescent (hereon referred to as "infected") and un-infected (hereon referred to as "negative") individuals were measured using MULTICOV-AB 33 (Fig. 1) . Descriptions of all groups of donors can be found in Extended Data Table 1 .

Vaccinated individuals had not been previously infected with SARS-CoV-2 as demonstrated by the absence of anti-Nucleocapsid IgG and IgA (Fig. 1a) . As expected, there was typical variation in antibody titers reflecting individual immune responses ( Fig. 1a and b) . When comparing between vaccine doses ( Fig. 1c and d) , all vaccinated subjects showed an enhanced antibody response with increasing time after the first dose and a further significant boost after the second dose. This boosting effect was so pronounced that it reached the upper limit of detection for MULTICOV-AB, as confirmed by a dilution series (Extended Data Fig. 1) .

To expand our understanding of the immune response of vaccinated individuals, we analyzed their saliva for IgA and IgG antibodies. The saliva of infected and negative individuals served as controls. Infected individuals had significantly higher levels of IgA than negative (p-value 0.0008) or vaccinated individuals (p-value 0.03), with no significant difference seen between vaccinated and negative individuals (p-value 0.23) (Fig. 2a) . Conversely, the IgG response in saliva of vaccinated individuals was significantly higher than either infected (p-value <0.0001) or negative individuals (p-value <0.0001) (Fig. 2b) . These results were verified using a second antibody test measuring IgG in saliva (Extended Data Fig. 2) . We also identified that vaccination with Pfizer BNT-162b2 does not appear to offer any cross-protection against other endemic coronaviruses (Extended Data Fig. 3) , as seen by absence of change in antibody titers following vaccination.

Having determined the humoral response induced by Pfizer BNT-162b2, we then examined how emerging RBD mutations present in different variants of concern impact antibody binding.

For this, we included RBD mutants for the UK (501Y), South African (417N, 484K and 510Y), is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint (Fig. 3a) . In contrast, a varied and reduced immune response was visible for the South African variant in both groups (Kendall's tau 0.844) (Fig. 3b) . Both the Mink and LA variants had a similar response as the wild-type variant (Extended Data Fig.   4 ). Having seen that antibody binding responses were reduced in the context of RBD mutants in the South African variant, we examined its neutralizing potential on samples from vaccinated individuals using a virus neutralization test (VNT) 34 , employing a patient-derived South African variant of the SARS-CoV-2 virus. Despite detectable variation, the VNT revealed substantially reduced neutralization for the South African variant for sera obtained from vaccinated and infected individuals (Fig 4a) . We further confirmed these findings using an ACE2 inhibition assay (Fig. 4b) . Here, we additionally observed increased neutralization capacities for both wild-type (Fig. 4c ) and the South African variant (Fig. 4d) , in all samples derived from vaccinated individuals following the second dose, which was further confirmed by the NeutrobodyPlex 35 (Extended Data Fig. 5) . Overall, our results showed individual differences in neutralization capacity for RBD mutations found in variants (Extended Data Fig 6) , with minimal to no change in neutralization for the UK, Mink or LA variants. Conversely, neutralization capacity versus the South African variant was severely compromised.

. 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint

RBD mutants are particularly important to track and study due to the role of the RBD:ACE2 interaction site in virus transmission and neutralization 1,18 and their potentially increased infectivity 3 or lethality 30,36 . This tracking has led to the identification of several variants of concern, notably the UK 3 , South African 5 , Mink 6 and LA 7 variants. We initiated this study to reveal the vaccine-induced immune response and most importantly to shed light on the still controversial question of how efficiently antibodies can bind and neutralize SARS-CoV-2 variants of concern. The presence of large titers of IgG antibodies within the saliva of vaccinated individuals far exceeded those seen in convalescent individuals. This was both surprising and welcome as it could indicate that vaccination might confer a sterilizing immune response in the oral cavity and thereby lower virus transmission. Focusing on antibody response, we examined in detail the effects of RBD mutations observed in emerging variants of concern. While only minor differences were detectable for the UK, Mink and LA variants, a substantial reduction in RBD binding antibodies was observed for the South African variant.

These findings were confirmed at a functional level by a VNT using patient-derived viral isolates, which showed a significant decrease in the neutralizing capacity of sera from This study is limited by the sample size, the restriction to only two time points after vaccination for data analysis, and the lack of paired saliva and serum samples for our infected and negative groups. However, we want to note that our sample size examined is similar to or larger than most other sample sets used to examine immune response to mutants in detail 37 , and that we examined paired serum and saliva samples for all of our vaccinated subjects. Furthermore, we performed VNT assays comparing wild-type to the South African variant with authentic virus isolates on human cells, in contrast to utilizing a pseudotype neutralization assay 37 or genetically engineered wild-type variant 38 . Notably, we focused our study on a detailed 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint characterization of the humoral immune response, as a proxy of an individual's immune response, as T-cell immunity has already been extensively studied 26, 39 . Future work should investigate the antibody response and their persistence from different vaccines (i.e.

AstraZeneca-Oxford, Moderna) against similar or newly emerging RBD mutants over a longer timeframe with increased sample size.

Viewed in a larger context, the impaired RBD-binding capacity to mutations in emerging variants of concern highlights the importance of updating current vaccines accordingly. 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint

No statistical analysis was used to determine sample size. The experiments were not randomized and the investigators were not blinded during either experimentation or data is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

The pCAGGS plasmid encoding the receptor binding domain (RBD) of SARS-CoV-2 was kindly provided by F. Krammer 41 . RBDs of SARS-CoV-2 variants of concern were generated by PCR amplification of fragments from wildtype DNA template followed by fusion PCRs to introduce described mutation N501Y for the UK variant and additional mutations K417N and 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 March 10, 2021. Amplified fragments were assembled by subsequent fusion PCR using forward primer RBDfor and RBDrev. RBD mutation L452R as recently reported for the SARS-CoV-2 variant of concern identified in Southern California (referred to in this manuscript as "LA"), was introduced using primer RBD-for and reverse primer L452Rrev 5´-CGG TAC CGG TAA TTG TAG TTG CCG CCG -3´ for amplification of fragment 1 and forward primer L452Rfor 5´-GGC AAC TAC AAT TAC CGG TAC CGG CTG TTC CGG AAG -3´ for fragment 2. Both fragments were subsequently fused using primers RBDfor and RBDrev. DNA coding for mutant RBDs (amino acids 319-541 of respective spike proteins) were cloned into Esp3I and EcoRI site of pCDNA3.4 expression vector with N-terminal signal peptide (MGWTLVFLFLLSVTAGVHS) for secretory pathway that comprises Esp3I site. All expression constructs were verified by sequence analysis.

Coupling of RBD mutant antigens was done by Anteo coupling (#A-LMPAKMM-10, Anteo Tech Reagents) following the manufacturer's instructions. Briefly, 100 μL of spectrally distinct populations of MagPlex beads (1.25x10 6 ) (Luminex) were activated in 100 μL of AnteoBind Activation Reagent for 1 hour at room temperature. The activated beads were washed twice with 100 μL of Coupling Buffer using a magnetic separator. Following this, 50 μg/mL of antigen (diluted in Coupling buffer) was added to the beads and incubated for 1 hour at room temperature. The beads were then washed twice with 100 μL Coupling buffer and blocked for 1 hour at room temperature in 0.1% BSA in Coupling Buffer. After washing twice in storage buffer, the beads were stored at 4°C until further use. 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint

MULTICOV-AB, a multiplex immunoassay which simultaneously analyses 20 antigens was performed as previously described 33 on all samples. In addition to the antigens presently included in MULTICOV-AB (Extended Data Table 3 ), RBD mutants from variants of concern were also included for measurements of all serum samples. All RBD mutants except the Mink variant (#40592-V08H80, Sino Biological) were produced in-house. Both IgG and IgA were measured for all serum samples. To adapt MULTICOV-AB to analyze antibodies in saliva, the dilution factor was changed from 1:400 for serum to 1:6 for saliva. Saliva samples were diluted into assay buffer inside a sterile workbench. 25 μL of diluted sample was then added to 25 μL of 1X Bead Mix 33 using a 96-well plate (#3600, Corning). Samples were incubated for 2 hours at 20°C, 750 rpm on a Thermomixer (Eppendorf), after which the unbound antibodies were removed by washing three times with Wash Buffer (1x PBS, 0.05% Tween-20) using a microplate washer (Biotek 405TS, Biotek Instruments GmBH). Bound antibodies were detected using either 3 μg/mL RPE-huIgG (#109-116-098, BIOZOL) or 5 μg/mL RPE-huIgA (#109-115-011, BIOZOL) by incubation for 45 mins at 20°C, 750 rpm on a Thermomixer.

Following another washing step, beads were re-suspended in 100 μL of Wash Buffer and reshaken for 3 mins at 20°C, 1000 rpm. Plates were then measured using a FLEXMAP3D instrument (Luminex) using the same settings as 33 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint then coated into 96well Costar microtiter high binding plates (#3590, Corning) and blocked at 4°C overnight with The Blocking Solution (Candor Bioscience GmBH), at room temperature on a microplate shaker set to 700 rpm. Before each of the following steps, wells were washed with PBS/ 0.1% Tween20. Saliva samples were diluted using The Blocking Solution (1:3 -1:729) and 100 μL added to each well. Plates were then incubated at room temperature for 1 hour. For detection, 1:20,000 biotinylated anti-human IgG (#109-065-008, Jackson Immuno

Research Laboratories) and 1:20,000 Streptavidin-HRP (#109-035-098) were added and incubated for 1 hour and 30 mins respectively. For visualization, TMB was added and the reaction was stopped using 1M HCl. The plate was read at 450 nm and 620 nm using a microplate reader (CLARIOstar, BMG LABTECH). Data is presented as concentration in ng/mL as estimated by a respective dilution series of highly pure human IgG (#31154, ThermoFisher). 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint 210211_SAv was detected by Western blotting, using sera from a convalescent patient. 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint samples. 25 μL of diluted plasma samples were then added to 25 μL of 1X BeadMix 33 per well of a 96-well plate (#3642, Corning). In addition to the standard bead mix used in MULTICOV-AB, all bead coupled RBD mutants were included. As a control, 500 ng/mL ACE2 was also used. Samples were incubated on at 21°C, 750 rpm for 2 hours on a Thermomixer. Samples were then washed using a microplate washer with Wash Buffer (1X PBS + 0.05% Tween20). 

The NeutrobodyPlex was performed as previously described 35 on all serum samples using a Nanobody concentration of 2 nM.

Data analysis and figure generation was performed in RStudio (Version 1.2.5001), running R (version 3.6.1) with the additional packages "beeswarm" and "RcolorBrewer" for data depiction purposes only. The type of statistical analysis performed (when appropriate) is listed in the figure legends. Figures were generated in Rstudio and then edited for clarity in Inkscape (Inkscape 0.92.4). Mann-Whitney U test was used to determine difference between signal distributions from different sample groups using the "wilcox.test" function from R's "stats" library. Kendall's τ coefficient was calculated in order to determine ordinal association between the observed antibody responses towards RBD mutant and wild-type proteins using the "cor" 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint function from R's "stats" library. Linear regression was performed to assess reduction in ACE2 neutralization observed for RBD mutants compared to wild-type proteins using the "lm" function from R's "stats" library. Pre-processing of data such as matching sample metadata and collecting results from multiple assay runs was performed in Excel 2016. GraphPad Prism version 8.4.0 was used to process VNT data.

Data relating to the findings of this study are available from the corresponding authors upon request.

Custom analysis code in R and required input files are available from the corresponding authors upon request. 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 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 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint 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 March 10, 2021. 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint 29 Extended Data Table 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint 30 Extended Data Table 3 

Full list of antigens included as standard in MULTICOV-AB (minus controls), their manufacturer, and if available, their category number. Full details on all NMI produced antigens can be found in 33 .

. 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint

The second vaccination appears to plateau the serum antibody response at the Upper Limit of Detection for MULTICOV-AB. To confirm this, 11 samples consisting of 10 paired samples from five vaccinated individuals and one negative individual were examined in a dilution series.

The three sera with the highest response maintained a similarly high response for the initial dilution, indicating a plateau in the range of >40,000 MFI, therefore suggesting that even for donors with high responses in the first sample, the second vaccination strongly increased the antibody response. A uniform curve shape for all samples (including the negative control) confirmed reliability of the generated data. 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 March 10, 2021. 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 March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint

SARS-CoV-2 variant B.1.1.7 is susceptible to neutralizing antibodies elicited by ancestral Spike vaccines. bioRxiv

Adaptation of SARS-CoV-2 in BALB/c mice for testing vaccine efficacy

Exploring beyond clinical routine SARS-CoV-2 serology using MultiCoV-Ab to evaluate endemic coronavirus cross-reactivity

We thank Johanna Griesbaum for technical assistance, Florian Krammer for providing us with 

Vaccinated individuals did not have an increased antibody response towards S1 proteins of is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted March 10, 2021. When compared to wild-type (wt), RBD mutants for both the Mink (a) and LA (b) variants of concern resulted in similar response. RBD mutant antigens were generated (LA) or purchased (Mink) and added to MULTICOV-AB to measure the immune response towards them from vaccinated (N=45) and infected (N=35) sera, compared to the wild-type RBD. A linear curve (y=x) is shown as a dashed grey-line to indicate identical response between wild-type and mutant. Kendall's tau was calculated to measure ordinal association between the mutant and wild-type. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted March 10, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint 36 Extended Data Figure 6 

To d etermine the effect variants of concern had upon neutralization potential, an ACE2 competition assay was developed. RBD mutants for all variants of concern included within this is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted March 10, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted March 10, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted March 10, 2021. ; https://doi.org/10.1101/2021.03.08.21252958 doi: medRxiv preprint