key: cord-0295170-oyt74tbk
authors: Saker, K.; Pozzetto, B.; ESCURET, V.; Pitiot, V.; Massardier-Pilonchery, A.; Mokdad, B.; Langlois-Jacques, C.; Rabilloud, M.; alfaiate, d.; Guibert, N.; Fassier, J.-B.; BAL, A.; Trouillet-Assant, S.; Trabaud, M.-A.
title: Evaluation of Commercial Anti-SARS-CoV-2 Neutralizing Antibody Assays in seropositive subjects
date: 2022-01-05
journal: nan
DOI: 10.1101/2022.01.04.22268652
sha: e29a049db78efa08fd9953070f7968edc7be5d23
doc_id: 295170
cord_uid: oyt74tbk

The virus neutralization test (VNT) is the reference for the assessment of the functional ability of neutralizing antibodies (NAb) to block SARS-CoV-2 entry into cells. New competitive immunoassays measuring antibodies preventing interaction between the spike protein and its cellular receptor are proposed as surrogate VNT (sVNT). We tested three commercial sVNT (a qualitative immunochromatographic test and two quantitative immunoassays named YHLO and TECO) together with a conventional anti-spike IgG assay (bioMerieux) in comparison with an in-house plaque reduction neutralization test (PRNT50) using the original 19A strain and different variants of concern (VOC), on a panel of 306 sera from naturally-infected or vaccinated patients. The qualitative test was rapidly discarded because of poor sensitivity and specificity. Areas under the curve of YHLO and TECO assays were, respectively, 85.83 and 84.07 (p-value >0.05) using a positivity threshold of 20 for PRNT50, and 95.63 and 90.35 (p-value =0.02) using a threshold of 80. However, the performances of YHLO and bioMerieux were very close for both thresholds, demonstrating the absence of added value of sVNT compared to a conventional assay for the evaluation of the presence of NAb in seropositive subjects. In addition, the PRNT50 assay showed a reduction of NAb titers towards different VOC in comparison to the 19A strain that could not be appreciated by the commercial tests. Despite the good correlation between the anti-spike antibody titer and the titer of NAb by PRNT50, our results highlight the difficulty to distinguish true NAb among the anti-RBD antibodies with commercial user-friendly immunoassays.

and different variants of concern (VOC), on a panel of 306 sera from naturally-infected or 27 vaccinated patients. The qualitative test was rapidly discarded because of poor sensitivity and 28 specificity. Areas under the curve of YHLO and TECO assays were, respectively, 85.83 and 29 84.07 (p-value >0.05) using a positivity threshold of 20 for PRNT 50 , and 95.63 and 90.35 (p-30 value =0.02) using a threshold of 80. However, the performances of YHLO and bioMérieux 31 were very close for both thresholds, demonstrating the absence of added value of sVNT 32 compared to a conventional assay for the evaluation of the presence of NAb in seropositive 33

subjects. In addition, the PRNT 50 assay showed a reduction of NAb titers towards different 34 VOC in comparison to the 19A strain that could not be appreciated by the commercial tests. 35

Coronavirus disease 2019 is an emerging disease caused by severe acute 47 respiratory syndrome coronavirus 2 (SARS-CoV-2), and since late 2020, vaccines against 48 SARS-CoV-2 have been available worldwide. In recent months, a large number of 49 commercial immunoassays have been developed for the detection of specific anti-SARS-50

CoV-2 antibodies (1,2). However, the presence of anti-SARS-CoV-2 antibodies does not 51 indicate whether the antibodies are able to neutralize the virus that has been reported to have a 52 role in the protection from COVID-19 both in animals and humans (3). The gold standard for 53 assessing the ability of antibodies to prevent the virus from entering into susceptible cells is 54 the virus neutralization test (VNT) (4), but it requires a biosafety level 3 laboratory and takes 55 approximately 10 days to complete. This has led to the development of SARS-CoV-2 56 surrogate virus neutralization tests (sVNT) that are more simple and rapid; these are based on 57 the competition between patient antibodies and the angiotensin converting enzyme 2 (ACE2) 58 receptor protein for binding to the spike receptor binding domain (RBD) that mediates the 59 entry of the virus into susceptible cells (5). 60

These competitive immunoassays, which can be conducted using qualitative 61 immunochromatographic cassettes or quantitative automated or manual enzyme-linked 62 immunosorbent assay (ELISA) platforms, allow rapid and easy processing of large numbers 63 of samples in conventional serological laboratories. However, the performance of these newly 64 developed commercial sVNT assays by comparison to classical serological assays detecting 65 anti-RBD IgG and/or to the reference plaque reduction neutralization test 50% (PRNT 50 ) 66 performed with live virus has been poorly evaluated up to now (6-9). Moreover, previous 67 studies have evaluated the specificity using seronegative and/or prepandemic serum which do 68 not inform if commercial sVNT can differentiate serum with or without neutralizing antibody 69 in seropositive samples. The aim of the present study was to evaluate the performance of three 70 . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. ; https://doi. org/10.1101 org/10. /2022 First of all, we compared performance of sVNT assays with VNT (clade 19A) on 81 and 246 121 convalescent samples for the three sVNT and the two quantitative sVNT respectively. As the 122 YHLO assay provided the best results among the investigated sVNT, the added-value of this 123 test compared to a commercial serological assay detecting anti-RBD IgG ( . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 5, 2022. CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. Figure 1 , Table 2 ). The combination of a good AUC but a very low 177 specificity observed with the positive threshold indicated by the manufacturers led us to 178 determine the best-fit cut-offs for the YHLO and TECO assays using the Youden index. 179 (Figure 2A) , and between the VNT (PRNT 50 ) and the bioMérieux assay 195 . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. The YHLO ( Figure 3A ) and bioMérieux ( Figure 3B ) assays underestimated NAb titers for all 207 variants with respect to the 19A strain, with notably a marked decrease for the Beta. 208

The performance of the qualitative Dynamiker assay was found to be poor, both in terms of 211 sensitivity and in terms of specificity. The two other quantitative sVNT assays evaluated in 212 the present study were found to be more sensitive but their specificity was extremely low 213 since, at the manufacturers' cutoff, most samples (TECO assay) or all of them (YHLO assay) 214 from convalescent individuals with no detectable NAb using the live virus neutralization 215 assay were positive for NAb with sVNT. It can be postulated that part of the antibodies 216 detected by these ELISA are able to interfere with the interaction between ACE receptor and 217 the viral RBD but not to prevent cell entry of the virus; this may be related to the 218 affinity/avidity of antibodies that is reported to be low after primary infection or first vaccine 219 dose (11, 12) , and is likely to be even more the case for the population included herein who 220 . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. ; https://doi.org/10.1101/2022.01.04.22268652 doi: medRxiv preprint were sampled 6 months after infection. Despite this low specificity, these assays correlated 221 with the live virus neutralization assay as also found in other studies (5,6,13-15). The low 222 specificity could also be attributed to a lack of sensitivity of live VNT but it is rather unlikely 223 that decreasing the threshold below 20 would be clinically relevant, such low titer having 224 little chance to be protective in vivo (16). It seems thus preferable to increase the sVNT cutoff 225 to improve specificity, and go closer to the protective threshold. Our data from ROC curves 226 would indicate that, for the detection of NAb, a threshold of 70 IU/ml and 30 AU/ml should 227 be applied for the TECO an YHLO assays, respectively. However, these data have been 228 obtained from infected subjects late after infection, at a time where antibodies are decreasing 229 (17). This could explain the low frequency of sera with detectable NAb using the VNT, and 230 also the discrepancy in terms of specificity between our results and previous ones using 231 samples earlier after infection (18) (19) (20) (21) . With time, waning of antibodies could have more 232 impact on the blocking of infection than interference with ACE binding. The study of Von 233

Rein et al (22) suggested that the correlation between sVNT and VNT was greater at higher 234 level of neutralisation titer and they concluded that sVNT are only useful when inhibition was 235 above 50%, which is more consistent with our data. Most of the previous studies used the 236 cPASS assay from GeneSript,showing the correlation of competitive immunoassay to live 237 VNT, with good sensitivity and specificity compared to VNT (5, 6, 9, 14, 15, 19, 20, 22, 23 Other studies have compared VNT and sVNT with assays detecting IgG binding to RBD or S 242 proteins and observed a correlation between them (6, 8, 9) . Fisher et al (6) found that the 243 correlation between sVNT and antibody binding assay is better for samples with high than 244 low PRNT, while in the study of Nandakumar et al (9) the correlation between VNT and 245 . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. antibody binding assays was lower than between sVNT and antibody binding assays and in 246 that of Gillot et al (8) the correlation between sVNT and VNT exceeded that between VNT 247 and antibody binding assay. Despite the high performance of the two competitive automated 248 immunoassays evaluated in our study, taking VNT as gold standard, we did not demonstrate 249 any added value of sVNT compared to serological assay detecting anti-RBD IgG for 250 evaluating the presence of Nab in seropositive subjects. These results highlight the difficulty 251 to distinguish the Nab among anti-RBD IgG using a standard immunoassay. This difficulty 252 could be further extrapolated considering the antibodies able to neutralize the SARS-CoV-2 253 variants. Using serum collected one month post full vaccination in patients with high Nab 254 titers, we confirmed the diminution of Nab titers against different SARS-CoV-2 variant 255 compared to initial strain. Nevertheless, the RBD coated in these competitive sVNT is not 256 adapted to virus evolution and are not able to detect the decrease of NAb titers To date, VNT 257 remains the only way to detect Nabs against VOC. Taking together, from our data and those 258 previously published, the predictive value of surrogate neutralization assays is still not 259 obvious in all population (infected and/or vaccinated, after priming or boost immunization, 260 early versus late after immunization). In addition, sVNT are not able to predict neutralization 261 of variant, and thus improvements are needed before they can be considered equivalent to 262 VNT to detect NAbs able to protect from infection. 263

The respective suppliers kindly provided all the serological kits used in the present study. 266

ACKNOWLEDGMENTS 268 269

This study was supported by Hospices Civils de Lyon and Fondation des Hospices Civils de 270

Lyon. The respective suppliers kindly provided all the serological kits used. We thank the 271 staff of the occupational health and medicine department of the Hospices Civils de Lyon who 272 . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. 

. CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. ; https://doi.org/10.1101/2022.01.04.22268652 doi: medRxiv preprint Positivity was established according to manufacturers' instructions. Sensitivity and specificity data were those described in the instruction for utilization sheet from each manufacturer.

Abbreviations: Ab: antibodies, Ig: immunoglobulin, ELISA: enzyme-linked immunosorbent assay, CLIA: chemiluminescence immunoassay, ELFA: enzyme-linked fluorescent assay, RBD: receptor binding domain, CI: confidence interval, AU: arbitrary unit, UI: unit international.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 5, 2022. ; https://doi.org/10.1101/2022.01.04.22268652 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 5, 2022. . CC-BY-ND 4.0 International license It is made available under a 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 January 5, 2022. ; https://doi.org/10.1101/2022.01.04.22268652 doi: medRxiv preprint

Neutralizing antibody titers (PRNT 50 ) [Anti-SARS-CoV-2 neutralizing antibodies