key: cord-0848470-jaidc1oo authors: Embregts, C. W. E.; Verstrepen, B.; Langermans, J. A. M.; Boszormenyi, K. P.; Sikkema, R. S.; de Vries, R. D.; Hoffmann, D.; Wernike, K.; Smit, L. A. M.; Zhao, S.; Rockx, B.; Koopmans, M.; Haagmans, B. L.; Kuiken, T.; GeurtsvanKessel, C. title: Evaluation of a multi-species SARS-CoV-2 surrogate virus neutralization test date: 2021-05-07 journal: nan DOI: 10.1101/2021.05.07.21252267 sha: 97d3f18dfc024e739f00db257b3740163a06dfd6 doc_id: 848470 cord_uid: jaidc1oo Assays to measure SARS-CoV-2-specific neutralizing antibodies are important to monitor seroprevalence, to study asymptomatic infections and to reveal (intermediate) hosts. A recently developed assay, the surrogate virus-neutralization test (sVNT) is a quick and commercially available alternative to the 'gold standard' virus neutralization assay using authentic virus, and does not require processing at BSL-3 level. The assay relies on the inhibition of binding of the receptor binding domain (RBD) on the spike (S) protein to human angiotensin-converting enzyme 2 (hACE2) by antibodies present in sera. As the sVNT does not require species- or isotype-specific conjugates, it can be similarly used for antibody detection in human and animal sera. In this study, we used 298 sera from PCR-confirmed COVID-19 patients and 151 sera from patients confirmed with other coronavirus or other (respiratory) infections, to evaluate the performance of the sVNT. To analyze the use of the assay in a One Health setting, we studied the presence of RBD-binding antibodies in 154 sera from nine animal species (cynomolgus and rhesus macaques, ferrets, rabbits, hamsters, cats, cattle, mink and dromedary camels). The sVNT showed a moderate to high sensitivity and a high specificity using sera from confirmed COVID-19 patients (91.3% and 100%, respectively) and animal sera (93.9% and 100%), however it lacked sensitivity to detect low titers. Significant correlations were found between the sVNT outcomes and PRNT50 and the Wantai total Ig and IgM ELISAs. While species-specific validation will be essential, our results show that the sVNT holds promise in detecting RBD-binding antibodies in multiple species. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) likely originates from an infection experiments show that many more animal species, including non-human primates 55 [13, 14] , ferrets [15, 16] , rabbits [17] , hamsters [18, 19] , and human angiotensin-converting 56 enzyme 2 (hACE2) transgenic mice [20] are permissive to the virus, while other animals 57 including pigs and chickens are not [21, 22] . The large number of permissive species and the 58 potential risks of additional (reverse) zoonotic events clearly indicate that a One Health 59 approach is required to gain insights into the circulation of SARS-CoV-2 in humans and 60 epidemiologically connected animal host populations, which is essential for the prevention or 61 mitigation of further spread. 62 Assays to reliably detect SARS-CoV-2 specific antibodies across species are urgently needed, 63 for example to investigate seroprevalence and asymptomatic infections, for vaccination studies 64 in humans and animals, and for the identification of natural reservoirs and intermediate hosts. are indicative for exposure, however, quantifying neutralizing antibodies is more informative. 69 were included to test the specificity of the assay. A set of possible cross-reactive sera from 124 MERS-CoV infected macaques and dromedary camels was included as well. 125 126 The use of human specimens was approved by the Erasmus MC medical ethical committee 128 (MEC approval: 2014-414), which allows the use of clinical data and left-over material from 129 the specimen delivered to our laboratory for diagnostics, unless patients have declared they 130 opted out of this scheme. 131 Animal sera were obtained as left-over material from various infection experiments or field 132 studies (mink and cats). Specific approval was obtained for each set of sera and can be found 133 in the referred articles. Additional non-human primate sera were obtained from various 134 experiments that were approved by the Dutch Central Committee for Animal Experiments 135 (license: AVD5020020209404). Mink and cat sera were obtained by a certified veterinarian 136 during a SARS-CoV-2 outbreak at a mink farm in the Netherlands. 137 The 50-percent plaque-reduction neutralization test (PRNT50) was used as the gold standard in 140 this study and was performed as described before [38] . The PRNT50 titer was defined as the 141 reciprocal value of the highest serum dilution resulting in 50% plaque reduction. Serum titers 142 of ≥ 20 were defined as SARS-CoV-2 seropositive. 143 The pseudotype VN assay was performed as described recently [37] with some minor 144 modification: serum samples were twofold diluted (starting at 1:8) and mixed 1:1 with SARS2-145 titers were defined by the reciprocal of the highest dilution that resulted in >50% reduction of 149 luciferase activity (IC50 titer). (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Spearman's correlation coefficients were calculated on the sVNT inhibition percentages and 175 the log2-transformed PNRT50 titers or the ODratio for the Wantai Ig or IgM in SPSS 27 (IBM). 176 Correlation was considered significant with p values < 0.05. 177 The performance of the commercial sVNT was evaluated by determining the correlation 180 between the PRNT50, a gold standard assay, and the sVNT. Although the initial commercial 181 sVNT guidelines included a cut-off of 20%, a recent validation paper now recommends a 182 positivity cut-off at 30% of inhibition [29] . We therefore evaluated the performance of the 183 sVNT by both a 20% and 30% cut-off (Table 1) . For the discussion of our results, we will focus 184 on the evaluation with the 30% cut-off. Using the serum panel of PCR-confirmed COVID-19 185 patients we found an overall sensitivity of 91.3 and a corresponding specificity of 100%. We 186 found a strong increase in sensitivity of the assay with increasing PRNT50 titers; sensitivity rose 187 from respectively 50% and 74.1% in the low-titer groups of 20 and 40, to 91.4% and above for 188 titers of 80 and higher (Table 1 , Fig. 1A ). 100% sensitivity was reached for sera with titers of 189 160 and above. In line with the expected rise in titer during the course of disease, we found that 190 assay sensitivity increased from 88.2 to 91.0 and 96.6 when comparing the periods between 1-191 10, 11-21 and >21 dpd (Table 1) . Overall, a significant (p < 0.001, with an Spearman's r of 192 0.68) correlation was observed between the two serological tests. However, the variation within 193 PRNT50 groups shows that the sVNT results should be interpreted with care, since high 194 inhibition in the sVNT is not directly translatable to high PRNT50 titers (Fig. 1A ). Performing 195 a 30%-inhibition titration of each sample would allow a more accurate comparison of both 196 assays. However, this would drastically decrease sample throughput and increase costs, which 197 The results of the sVNT also showed significant correlation (p < 0.001) with OD ratios of the 199 The specificity of the sVNT was further investigated using a serum panel containing sera of 210 individuals diagnosed with other coronaviruses or other (respiratory) viruses or diseases. Using 211 this panel, we confirmed that the sVNT is 100% specific, as we did not find cross-reactivity 212 with any of the tested sera (Fig. 1B) . Two samples were found to have an inhibition between 213 20-30%, one serum of a HCoV-229E patient and one of an adenovirus patient. Both samples 214 tested negative in the PRNT50. 215 In parallel to the human sera, we assessed the performance of the sVNT in an elaborate panel 216 of animal sera that included experimental model species, but also (suspected) reservoir species 217 ( Fig. 2A) . Specificity of the assay was assessed using a panel of control sera from naïve animals. 218 Sera from MERS-CoV infected cynomolgus macaques and dromedary camels were included 219 to assess possible cross-reactivity. Similar to the results obtained with the human validation 220 serum panel, the sVNT showed a good performance in general, with a sensitivity of 93.9% and 221 a specificity 100%. For the rhesus macaques, ferrets, rabbits, hamsters, cats, cattle and mink we 222 observed a 100% accurate detection of (the absence of) RBD-specific antibodies in the sera of 223 SARS-CoV-2 positive and negative animals. 224 For the cynomolgus macaques two sera of SARS-CoV-2 infected animals were found to have 225 an inhibition between 20% and 30%, and a PRNT50 titer of 40 and 80 (Fig. 2B ). However, low 226 antibody titers were expected since these animals only showed a short period of viral shedding 227 with low levels of viral RNA in nose and trachea. 228 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. In agreement with the human cross-reactive serum panel, no cross-reactivity was detected in 229 serum from MERS-CoV-infected rabbits and dromedary camels. The sVNT showed a 230 sensitivity and specificity of 100% in animal sera with a PRNT50 titer of 160 and above. In 231 contrast to the panel of human sera, no clear linearity was detected in the panel of animal sera 232 and large differences were observed between species. Interestingly, serum samples from ferrets, 233 rabbits and cattle with relatively low PRNT50 titers (80 and below) showed a high inhibition in 234 the sVNT. Especially SARS-CoV-2-infected ferrets, where only two animals reached a PRNT50 235 titer of 80, had an inhibition of above 92%. While these high levels of inhibition in sera with a 236 relatively low PRNT50 is beneficial for detecting RBD-binding antibodies in a qualitative 237 manner, species-specific determination of the optimal serum dilutions is essential when the data 238 is to be interpreted (semi-)quantitatively. The control sera of the ferrets, rabbits and cattle had 239 negative sVNT outcomes, indicating that the high inhibition levels were not due to background 240 or aspecific binding. Furthermore, the wide range of species-specific endemic coronaviruses 241 complicates the design of specific serological tests [41, 42] , and cross-reactivity needs to be 242 examined for every targeted species. While our data shows that the sVNT detects RBD-binding 243 antibodies in nine animal species, it clearly indicates that more elaborate validation is required. 244 Validations should include higher number of sera per species, a panel of potentially cross-245 reactive sera, and should aim at determining optimal serum dilutions and cut-off levels. 246 247 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 May 7, 2021. ; https://doi.org/10.1101/2021.05.07.21252267 doi: medRxiv preprint Our results show moderate to high sensitivity and high specificity of the sVNT for detecting 249 RBD-binding antibodies, with a 100% accuracy in sera with a PRNT50 titer 160, for both human 250 and animal sera. sVNT results should be interpreted rather qualitatively than quantitatively, 251 since the results only show partial linearity with the PRNT50 titers. 252 Despite the low sensitivity in detecting low titers, the sVNT still has potential use. The 253 possibility for high sample throughput makes the sVNT a suitable assay for large 254 seroprevalence studies that aim at detecting high titers, for example in vaccination trials or in 255 large scale initial testing of potential animal reservoirs. While the required titer for complete 256 protection is still under investigation, studies have shown that with a PRNT50 titer of 80 and 257 above, no infectious virus could be detected in the respiratory tract [43] . It is thus to be expected 258 that threshold titers for complete protection will be in this range or higher, and as a consequence 259 the sVNT can be a valuable assay to assess protection in a qualitative manner. However, the 260 sVNT does not serve as a full replacement of gold standard tests that use authentic virus, given 261 that it lacks the sensitivity to detect low titers and only targets RBD-binding antibodies. 262 Our evaluation shows that the sVNT also has potential use for detecting RBD-specific 263 antibodies in animal sera, but we observed large species-dependent differences in sensitivity of 264 the test. While in some species we observed high sVNT results in sera with low PRNT50 (ferrets, 265 rabbit, cattle), sera with low to moderate PRNT50 from other species resulted in negative or low 266 sVNT results (cynomolgus macaques). More elaborate species-specific validations are required 267 to determine the true potential of the sVNT. 268 269 270 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. [6] OIE, COVID-19 Events in animals, (2020). https://www.oie.int/en/scientific-302 expertise/specific-information-and-recommendations/questions-and-answers-on-303 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 May 7, 2021. ; https://doi.org/10.1101/2021.05.07.21252267 doi: medRxiv preprint 2019novel-coronavirus/events-in-animals/ (accessed December 10, 2020). 304 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 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 May 7, 2021. ; https://doi.org/10.1101/2021.05.07.21252267 doi: medRxiv preprint Table 2 . Sensitivity and specificity analysis of the sVNT using an animal serum panel of nine 538 different species, including SARS-CoV and MERS-CoV positive animals and control animals. 539 540 Seropositive Seronegative 20% 30% 20% 30% Cynomolgus macaques 11/13 9/13 10/10 10/10 Rhesus macaques 9/9 9/9 10/10 10/10 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 May 7, 2021. ; https://doi.org/10.1101/2021.05.07.21252267 doi: medRxiv preprint From people to panthera: Natural sars-331 cov-2 infection in tigers and lions at the bronx zoo Reverse Zoonosis of COVID-19 : Lessons 334 From the 2009 Influenza Pandemic Comparative pathogenesis of COVID-19, MERS, and 341 SARS in a nonhuman primate model Respiratory disease and virus 346 shedding in rhesus macaques inoculated with SARS-CoV-2 Susceptibility of ferrets, cats, dogs, and other domesticated 356 animals to SARS-coronavirus Pathogenesis and transmission of SARS-CoV-2 in golden hamsters Syrian hamsters as a small animal model for SARS-CoV-2 infection and 372 countermeasure development An evaluation of COVID-19 serological assays informs future diagnostics 406 and exposure assessment Establishment and 410 validation of a pseudovirus neutralization assay for SARS-CoV-2, Emerg Rapid quantification of SARS-CoV-2-414 neutralizing antibodies using propagation-defective vesicular stomatitis virus 415 pseudotypes A nanoluciferase SARS-CoV-2 for rapid neutralization testing 419 and screening of anti-infective drugs for COVID-19 A 423 SARS-CoV-2 surrogate virus neutralization test based on antibody-mediated blockage 424 of ACE2-spike protein-protein interaction CoV-2 surrogate virus neutralization test for detection of antibody in human, canine, 430 cat and hamster sera Comparison of potency assays to 433 assess SARS-CoV-2 neutralizing antibody capacity in COVID-19 convalescent 434 plasma Validation and clinical 437 evaluation of a SARS-CoV-2 surrogate virus neutralisation test (sVNT) Multi-species ELISA 442 for the detection of antibodies against SARS-CoV-2 in animals Serological survey of SARS-CoV-2 for experimental, domestic, companion and wild 446 animals excludes intermediate hosts of 35 different species of animals Intranasal fusion inhibitory lipopeptide prevents direct contact SARS-CoV-2 457 transmission in ferrets A human monoclonal 461 antibody blocking SARS-CoV-2 infection Severe Acute Respiratory Syndrome Coronavirus 467 2-Specific Antibody Responses in Coronavirus Disease Patients Potent neutralizing antibodies against multiple epitopes on SARS-CoV-2 spike Two-Way Antigenic Cross-Reactivity between Severe Acute Respiratory 481 Syndrome Coronavirus (SARS-CoV) and Group 1 Animal CoVs Is Mediated through 482 Antigenic Site in the N-Terminal Region of the SARS-CoV Nucleoprotein Antigenic Cross-Reactivity between the Nucleocapsid Protein of 485 Severe Acute Respiratory Syndrome (SARS) Coronavirus and Polyclonal Antisera of 486 Antigenic Group I Animal Coronaviruses: Implication for SARS Diagnosis [2] Shedding of infectious virus in hospitalized patients 493 with coronavirus disease-2019 (COVID-19): duration and key determinants SARS-CoV-2 neutralizing human antibodies protect against lower respiratory tract 499 disease in a hamster model 553 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.