key: cord-0253291-m9sjumxh
authors: Montomoli, E.; Apolone, G.; Manenti, A.; Boeri, M.; Suatoni, P.; Sabia, F.; Marchiano, A.; Bollati, V.; Pastorino, U.; Sozzi, G.
title: Timeline of SARS-CoV2 spread in Italy: results from an independent serological retesting
date: 2021-07-19
journal: nan
DOI: 10.1101/2021.07.14.21260491
sha: 1f2430efe7b2982aa40cfc0ba78774843d09115a
doc_id: 253291
cord_uid: m9sjumxh

The massive emergence of COVID19 cases in the first phase of pandemic within an extremely short period of time suggest that an undetected earlier circulation of SARS-CoV-2 might have occurred, as documented by several papers in different countries, including a few that reported positive cases even earlier the first cases identified in Wuhan. Given the importance of this evidence, an independent evaluation was recommended. Here we report the results of SARS-CoV-2 antibodies blind retesting of blood samples collected in the prepandemic period in Italy, and in control samples collected one year before, by two independent centers. Results suggest the presence of SARS-CoV-2 antibodies in some samples collected in the prepandemic period, though the detection of IgM and/or IgG binding and neutralizing antibodies is strongly dependent on the different serological assays and thresholds employed, while being absent in control samples collected one year before. These findings highlight the importance of harmonizing serological assays for testing SARS-CoV-2 virus spreading and may contribute to a better understanding the future virus dynamics.

analyzed clinical samples of asymptomatic and symptomatic subjects. Positive SARS-CoV-2 RT-PCR results were reported in France in a patient with pneumonia on December 27 th , 2019 (5) and in a child with suspected measles and a woman with extended skin dermatosis on November 2019, in Milan, Italy (6, 7) .

Large seroprevalence studies, in USA and in Europe support an earlier than expected circulation of the virus. Retrospective SARS-CoV-2 serological testing of 7,389 routine blood donations collected in nine U.S. states from December 13, 2019-January 17, 2020 suggested that the virus was present as early as December 13-16, 2019 (8) .

Using serum samples routinely collected in 9144 adults from a French general population-based cohort, 353 participants with a positive anti-SARS-CoV-2 IgG test were identified. Notably, 13 participants with positive ELISA-S had been sampled between November 5, 2019 and January 30, 2020 and were confirmed by neutralizing antibodies testing. In these positive subjects, the authors identified symptoms, history of possible exposure, or specific events compatible with early SARS-CoV-2 infection (9).

The Italian Ministry of Health accomplished a large SARS-CoV-2 seroprevalence study in a representative sample of 64,660 subjects collected between May 25 th and July 15 th , 2020. A global prevalence rate of 2.5% was reported, with a peak in the Lombardy region (7.5%) and in particular in Bergamo Province (24%) (www.salute.gov.it). According to these numbers, the true number of Italians who had been in contact with the virus would be approximately 1.5 million, many of which asymptomatic, an estimate almost 5 times higher than the official figures reported suggesting an under surface SARS-CoV-2 circulation.

We investigated the presence of SARS-CoV-2 Receptor-Binding Domain (RBD)-specific antibodies in blood samples of 959 asymptomatic individuals enrolled in the SMILE prospective lung cancer screening trial (clinicaltrials.gov ID: NCT03654105) between September 2019 and March 2020, across all the Italian regions. SARS-CoV-2 IgM/IgG RBD-specific antibodies were detected in 111 of 959 (11.6%) subjects, starting from September 2019 (14%), with a cluster of positive cases (>30%) on the 2nd week of February 2020 and the highest number (53.2%) in Lombardy (10) . Twelve subjects with samples collected on July 2019 were also included in the study and resulted all negatives (data not shown). The publication of our report generated a lively debate on the possibility that the virus circulated months earlier in Italy without the surveillance programs were able to identify any signs of its presence. To validate these findings, we were offered to blindly retest our samples in an external WHO affiliated laboratory (Erasmus Medical Center, Rotterdam) by using different serological assays. Here we report the results of this crossvalidation study.

The sample series included: 29 plasma samples collected in the SMILE trial (clinicaltrials.gov ID: NCT03654105) starting from July 23, 2019 until February 17, 2020 (10); 29 plasma samples from another lung cancer screening cohort (bioMILD, clinicaltrials.gov ID: NCT02247453) collected from July 14, 2018 until February 23, 2019, matched by date of collection, sex, age, smoking habits. The Institutional Review Board and Ethics Committee of Fondazione IRCCS Istituto Nazionale dei Tumori of Milan approved the study. All eligible subjects provided written informed consent. Nine samples from asymptomatic and symptomatic convalescent COVID19 patients with positive molecular (RT-PCR) swab collected from the . CC-BY-NC-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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint National Institute for Biological Standard and Control (NIBSC), the University of Siena and the University of Milan, as well as 2 commercial (BioIVT) negative control samples were also included (Table 2) .

VisMederi, an independent laboratory part of the CEPI consortium (https://cepi.net), used

proprietary IgG-RBD/IgM-RBD Elisa assays and a qualitative Micro-Neutralization CPE-Based assay (MN) with the aim to increase the sensitivity of the test (11, 12) . The assays were qualified and validated following the guideline issued by the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) (www.ich.org). A cut-off value of each plate was obtained by multiplying 3 times the "BLANK" optical density (OD) signal derived by six micro-wells containing sample diluents and secondary HRP-antibody, but no analyte (12) . By Vismederi standards, criteria of RBD-specific IgM and/or IgG above the cutoff value, with or without the presence of neutralizing antibodies, indicate presence of SARS-CoV-2 specific antibodies (12) .

Erasmus tested the samples with a multiplex IgG protein array including three antigens: S1, ecto and nucleocapsid protein (NP). By Erasmus standards, criteria of IgG triple antigen positives, with confirmation of neutralization assay using Plaque Reduction Neutralization test (PRNT), were needed before scoring a sample as positive. For IgM detection, Erasmus adopted the Elisa IgM commercial kit Wantai using cut-off as suggested by the manufacturer (13, 14) .

Specificity of both VisMederi and Erasmus assays was validated across the most common HCoV sera (12, 13) .

For both the VisMederi and Erasmus assays, the IgM results were calculated by relating each specimen OD value to the respective cut-off value of the plate. Results were thus expressed as OD ratio and considered positive when > 1. For correlation analysis of nonparametric data, the . CC-BY-NC-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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint Spearman rho coeficent (r) with respective 95% confidence interval and two-tailed p-value was calculated. To compare data distributions, p-values were calculated using the Mann-Whitney test. For the analysis and the graphic design the GraphPad Prism software (version 5.2) was adopted.

According to Vismederi of the 29 SMILE samples, 7 cases were positive for IgG, 16 cases were classified as positive (n=13) or borderline (n=3) for IgM , of which 6 samples were also positive in qualitative microneutralization assay and 6 cases were negative (Table 1) .

According to the Erasmus Medical Center none of the SMILE series showed IgG triple antigen positivity (3 samples had S1 or S1+NP positivity only), 2 samples were positive and one was in grey zone for IgM detection using Wantai cut-offs ( Table 1) . Out of the 12 tested samples for the PRNT assay, 1 IgM positive sample has also neutralizing antibodies (Table 1 ). In 9 additional samples, classified as IgM positives by VisMederi, a signal was also detected by Erasmus, but below the threshold level of the Wantai test ( Figure 1A ). These differences, especially in IgM outcomes, would seem to derive from a different setting of the cut-off values used to attribute positivity or negativity to anti SARS-CoV-2 antibodies. In fact, the ELISA cut-off of VisMederi was established internally through a blind study of symptomatic and asymptomatic subjects positive to molecular swab (15). On the other hand, the Wantai ELISA commercial Kit uses a different cut-off set up for diagnostic purposes in symptomatic patients. Despite this, a significant correlation among IgM values (VisMederi vs. Erasmus) was observed (Spearman r=0.6130; p=0.0004), when OD ratios were considered as continuous variables ( Figure 1A) .

. CC-BY-NC-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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint

To better assess assays specificity, 29 plasma samples collected from July 2018 until February 2019 from subjects matched to those of the SMILE study and supposed to be negative were also included. Indeed, all these samples resulted negatives by both VisMederi and Erasmus To have better insights about the assays performances and the concordance between the two laboratories, 11 among positives convalescent patients (3 symptomatic and 6 asymptomatic) and negatives controls (n=2) were included in the study ( Table 2) 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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint 28 th , 2020 (Lombardy) were IgG positive for VisMederi and IgG S1 and IgG S1+NP, respectively, by Erasmus. Additional IgM positive cases could have been detected also by Erasmus by lowering the cut-off of the commercial IgM assay. The older among these putative additional IgM positive samples was collected on September 3 th , 2019 in the Veneto region, one of the first and mostly affected COVID19 regions.

Noteworthy, particularly for samples from asymptomatic SARS-CoV-2 infected subjects, where the antibody concentration is known to be very low and are rarely able to develop neutralizing antibodies (16) (17) (18) , the agreement among the different tests used in the two laboratories is poor, highlighting the issue of sensitivity of commercial serological tests.

Conversely, samples from symptomatic COVID19 convalescents tested positive for IgG and showed neutralizing activity by both the laboratories. If it is true that high specificity is an important parameter in SARS-CoV-2 seroprevalence studies, even sensitivity is an equally important parameter that should be considered when testing asymptomatic infected subjects.

By Erasmus standards, criteria of IgG triple antigen positives, with confirmation of neutralization assay, before scoring a sample as positive are used because individual antigens do occasionally have some low level reactivity. This is based on in house validation data for surveys of pre-outbreak sera to set a high specificity value for studies in which it is important to be able to have conclusive results. Indeed the percentage of triple positive sera increases starting at 7 days post infection indicating lower sensitivity in recently (<7 days) infected subjects.

Accordingly, by using this approach only few symptomatic COVID19 patients were identified while asymptomatic patients and SMILE subjects did not reached the triple antigens cut-off. On the other hand, IgM antibodies and neutralizing activity were found in a SMILE sample collected on February 5 th , two weeks before the first officially declared COVID19 Italian patient.

. CC-BY-NC-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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint Lombardy was the first region to have been affected by the COVID-19 outbreak, with a death rate almost six times higher than in the rest of Italy. A recent large phylogenetic study on 346 SARS-CoV-2 patients in Lombardy allowed the identification of seven SARS-CoV-2 lineages and the presence of local transmission clusters within three of them, suggesting that the virus was circulating undetected for some time before first detection and confirming a central role of Lombardy in the SARS-CoV-2 epidemic (19) . Similar conclusions were reached in a previous report in a smaller number of sequenced samples (20) . In contrast, no evidence of 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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint isolation in January 2020 (23). Recently, Kumar et al. reconstructed the mutational history of SARS-CoV-2 using a so called 'mutation order approach' (MOA) (24). From their analysis of more than 174,000 genomes, major mutational fingerprints revealed useful to identify and track the novel coronavirus spatiotemporal evolution. The progenitor genome identified differed from that of the first coronaviruses sampled in China by three variants, implying that none of the earliest patients represent the index case or gave rise to all the human infections. However, multiple coronavirus infections in China ,USA and Europe harbored the progenitor genetic fingerprint in January 2020 and later, suggesting that the progenitor was spreading worldwide months before.

We acknowledge some limitations in this re-test study related to the limited sample size, the highly specific selection of screening participants (heavy smokers ≥ 30 pack years and ≥55 years old), possibly not representative of the general population, and the intrinsic experimental variability of the immunoenzymatic assays employed in the different laboratories. Moreover, we cannot exclude that other confounding conditions, such as preexisting immunity against other agents, might have contributed to the SARS-CoV-2 positivity in our assays. Nonetheless, cross reactivity towards the most common HCoVs was ruled out.

As pointed out in the recent WHO report (www.who.int) and in other commentaries (1, 2) , studies from different countries suggest that SARS-CoV-2 was growing under the surface for some time before the first diagnosed case in Wuhan. Suggestive evidence from our previous published study (10) and conflicting results from the current retesting exercise, despite adding important signals in this direction, do not allow us to accept or discard this hypothesis. Indeed findings of these studies are only partially confirmed due to the heterogeneity of methods utilized and the risk of non-specific signals in serological assays. Despite that, the report underlines the . CC-BY-NC-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 July 19, 2021. ;  importance of investigating these potential early events, to solve the still unanswered questions about the origin and timing of the current pandemic and may contribute to a better understanding of the future virus circulation dynamics. . CC-BY-NC-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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint 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 July 19, 2021. ; https://doi.org/10.1101/2021.07.14.21260491 doi: medRxiv preprint

Where did COVID come from? Five mysteries that remain

Nowcasting epidemics of novel pathogens: lessons from COVID-19

SARS-CoV-2 has been circulating in northern Italy since December 2019: Evidence from environmental monitoring

SARS-CoV-2 was already spreading in France in late

Evidence of SARS-CoV-2 RNA in an oropharyngeal swab specimen

COVID-19-related dermatosis in November 2019: could this case be Italy's patient zero?

Unexpected detection of SARS-CoV-2 antibodies in the prepandemic period in Italy

Extremely potent human monoclonal antibodies from COVID-19 convalescent patients

Comparative analyses of SARS-CoV-2 binding (IgG, IgM, IgA) and neutralizing antibodies from human serum samples

An evaluation of COVID-19 serological assays informs future diagnostics and exposure assessment

Accurate serology for SARS-CoV-2 and common human coronaviruses using a multiplex approach

Mapping Neutralizing and Immunodominant Sites on the SARS-CoV-2 Spike Receptor-Binding Domain by Structure-Guided High-Resolution Serology

Antibody Binding to SARS-CoV-2 S Glycoprotein Correlates with but Does Not Predict Neutralization. Viruses [Internet

Weak correlation between antibody titers and neutralizing activity in sera from SARS-CoV-2 infected subjects

Genomic epidemiology of SARS-CoV-2 reveals multiple lineages and early spread of SARS-CoV-2 infections in

Early phylogenetic estimate of the

The authors thank the Department of Viroscience, Erasmus MC for performing SARS-CoV-2 serology on the shared samples, for providing results and for critical discussion of data.