Serum Neutralizing Activity against B.1.1.7, B.1.351, and P.1 SARS-CoV-2 Variants of Concern in Hospitalized COVID-19 Patients

The recent spreading of new SARS-CoV-2 variants, carrying several mutations in the spike protein, could impact immune protection elicited by natural infection or conferred by vaccination. In this study, we evaluated the neutralizing activity against the viral variants that emerged in the United Kingdom (B.1.1.7), Brazil (P.1), and South Africa (B.1.351) in human serum samples from hospitalized patients infected by SARS-CoV-2 during the first pandemic wave in Italy in 2020. Of the patients studied, 59.5% showed a decrease (≥2 fold) in neutralizing antibody titer against B.1.1.7, 83.3% against P.1, and 90.5% against B.1.351 with respect to the original strain. The reduction in antibody titers against all analyzed variants, and in particular P.1 and B.1.351, suggests that previous symptomatic infection might be not fully protective against exposure to SARS-CoV-2 variants carrying a set of relevant spike mutations.


Introduction
One year ago, the Director-General of the World Health Organization declared the first pandemic caused by a coronavirus named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1]. Neutralizing antibodies, targeting the viral spike (S) protein and its receptor-binding domain (RBD), are considered a surrogate of protection against COVID-19 [2], the disease caused by SARS-CoV-2, although no formal correlates of protection have been established so far. In fact, available COVID-19 vaccines that have shown high efficacy are developed with the concept that the S protein is the immunodominant antigen. These vaccines are designed based on the Wuhan strain, the original strain that since January 2020 has been at the origin of the worldwide pandemic.
In the last months, different variants of concern (VOCs) [3] of SARS-CoV-2 emerged around the world, initially as occasional isolates; however, in some settings, the most efficient variants are rapidly replacing the original Wuhan strain.
One new variant emerged in the United Kingdom (UK), affecting people under 60 years of age. Retrospective analyses have dated the first identification as occurring in September 2020 in South East England [4,5]. This variant, called B.1.1.7, spread in several countries around the world (i.e., Italy, Denmark, the United States, the Netherlands, Australia, Iceland), and it has been associated with 50% increased transmission [6] and increased risk of death [4, [7][8][9]. Seventeen mutations/deletions in the viral genome characterize the B.1.1.7, including eight in the S protein (∆69-70, ∆Y144, N501Y, A570D, P681H, T716I, S982A, D118H) [5,8]. Three mutations are of biological significance, namely ∆69-70, N501Y, and P681H. The first one is related to an increased viral infectivity and a potential impact on PCR assays targeting the S gene [4, 8,10]; however, this mutation is not restricted to this variant [10]. The N501Y is a mutation in the RBD resulting in an increased binding affinity to its cellular ACE-2 receptor [7,8,11]. The last one, P681H, is supposed to be related to an improved transmissibility of the virus [12].
Two new variants emerged in Brazil, both originating from the B.1.1.28 clade [13]. The first variant (named B.1.1.28.2), detected in October 2020, was characterized by the E484K mutation in the S protein, related to a possible escape from neutralizing antibodies [14]. The second variant (named B.1.1.28.1) was reported by the National Institute of Infectious Diseases in Japan after sampling of Brazilian travelers to Japan. Since this variant presents other substantial mutations such as H655Y, L18F, D138Y, and N501Y, in addition to the E484K [7,13,15], it was classified as a "variant of concern" [16]. Notably, the E484K mutation is in common with the B.1.351 variant first identified in the Republic of South Africa but not with B.1.1.7 [13]. The P.1 and B.1.351 variants share common mutations in the S protein, increasing the possibility of evasion of the humoral response and enhanced transmissibility [13,17].
As of 2 March 2021, B.1.1.7 accounts for 54.0% of Italian cases nationwide, ranging from 0 to 93.3% between regions, and is becoming the most widely present variant in the country. The P.1 variant is less prevalent for now, accounting for 4.3% of all new local COVID-19 cases, while B.1.351 is involved in just 0.4% of new cases [18]. Recently, a new VOC emerged, the SARS-CoV-2 B.1.617.2, also named Delta, first detected in India and now dominant in the UK [19,20]. The new variant has overtaken B.1.1.7 in the UK, and it is estimated that it will represent 90% of all viruses circulating in Europe by the end of August [19,21]. Additionally, it seems to be characterized by a higher risk of hospital admission and increased transmissibility [19,20,22]. Here we assess the neutralization activity against B.1.1.7, P.1, and B.1.351 variants in a panel of human serum samples from hospitalized infected COVID-19 patients, previously tested by 2019-nCov/Italy-INMI1, clade V strain (Wuhan strain).

Study Population
Sera of 42 COVID-19 patients (35 recovered and 7 with fatal outcome), hospitalized at Humanitas Gavazzeni (Bergamo, Italy) during the first epidemic wave that occurred in Italy between March and May 2020, were included in the present study. Subject characteristics and study procedures were described in detail elsewhere (approval number 17373-Ethics Committee of the University of Siena; approval number 17/20-Ethics Committee of Humanitas Gavazzeni) [23].
For the purpose of the present study, sera available for each patient at three time points were selected: the hospital admission sample (baseline), the sample showing the highest neutralizing antibody titer against 2019-nCov/Italy-INMI1 strain (hereafter referred to as wild-type (wt) strain) found in the previous study [23] and defined as "peak", and the last sample available during hospital stay. For 22 patients, the "peak" sample was the last sample available during the hospital stay. Samples of deceased/recovered patients were pulled together for the purpose of this study, as in the original study no difference was found between the 2 groups in terms of neutralizing antibody titers with respect to the wt strain.
Adherent subconfluent cell monolayers of VERO E6 were prepared in growth medium (DMEM High Glucose containing 2% FBS, 2 mM L-glutamine, 100 units/mL penicillinstreptomycin) in 175 cm 2 flasks or 96-well plates for propagation or titration and neutralization tests of SARS-CoV-2, respectively.
Cells were seeded in a 175 cm 2 flask at a density of 1 × 10 6 cells/mL. After 18-20 h, the subconfluent cell monolayer was washed twice with sterile Dulbecco's phosphate-buffered saline (DPBS). After the DPBS was removed, cells were infected with 3.5 mL of DMEM 2% FBS containing the SARS-CoV-2 virus at a multiplicity of infection (MOI) of 0.01. After 1 h of incubation at 37 • C in a humidified atmosphere with 5% CO 2 , 50 mL of DMEM containing 2% FBS was added. The flasks were observed daily, and the virus was harvested when 80-90% of the cells manifested cytopathic effect (CPE). The culture medium was centrifuged at +4 • C and 469× g for 8 min, aliquoted, and stored at −80 • C.

SARS-CoV-2 Viruses
The SARS-CoV-2 (nCoV strain 2019-nCov/Italy-INMI1 strain) wt virus was purchased from the European Virus Archive goes Global (EVAg, Spallanzani Institute, Rome). Notably, the used strain did not carry the S protein amino acid change D614G.
The B.

Virus Neutralization Assay
The virus neutralization (VN) assay was performed as previously reported [24]. Briefly, after heat-inactivation for 30 min at 56 • C, serum samples, starting from 1:10 dilution, were mixed with an equal volume of SARS-CoV-2 (B.1.1.7, P.1, and B.1.351) viral solution containing 100 tissue culture infective dose 50% (TCID50). After 1 h of incubation at room temperature, 100 µL of virus-serum mixture was added to a 96-well plate containing an 80% confluent Vero E6 cell monolayer. Plates were incubated for 4 days at 37 • C and 5% CO 2 in humidified atmosphere and then checked for presence/absence of CPE by an inverted optical microscope. A CPE higher than 50% indicated infection. The VN titer was expressed as the reciprocal of the highest serum dilution showing protection from viral infection and CPE.

Statistical Analysis
All statistical analyses were performed using GraphPad Prism version 6.01 for Windows (GraphPad Software, San Diego, CA, USA [25]. For each variant, the mean fold decrease in antibody levels with respect to the wt strain was calculated along with its standard deviation (SD). Antibody levels were expressed as log and statistically evaluated with respect to the wt strain using a paired t-tests. Statistical significance was set at p < 0.05, two-tailed.

Discussion
Recently, different variants of SARS-CoV-2 with mutations in the S protein have emerged, raising concerns about the protection elicited by natural infection or conferred by vaccination.
In this study, we evaluated the neutralizing activity against B.1.1.7, P.1, and B.1.351 VOCs in a panel of human serum samples from hospitalized patients infected by SARS-CoV-2 between March and May 2020 during the first wave of the pandemic in Italy [23].
Although we assessed the neutralizing activity at three time points for each patient, namely hospital admission (baseline), neutralizing antibody titer peak, and the last time point available (at discharge or decease), here we mainly discuss the last one since it may represent the "antibody baggage".
In our study, 59.5% of the patients had a decrease (≥2-fold) in neutralizing antibody titer against B.1.1.7, and 21.4% showed a 4-fold decrease, as shown in a similar study [26]. On the contrary, other studies have reported that samples from convalescent or hospitalized patients are able to neutralize the B.1.1.7 variant, either maintaining similar levels of neutralizing activity or exhibiting a modest decrease compared to the original antibody titer [27][28][29]. We can hypothesize that the decrease in neutralizing antibody titer found in our study may be ascribed to the wt virus used, lacking D614G mutation. In general, the modest change in the potency of neutralization against this variant could be related to the N501Y mutation that, even though related to an increased binding to the ACE-2 receptor, does not appear to have a significant implication for neutralizing activity [28].
Regarding the P.1 variant, 83.3% of the patients involved in the study showed significantly reduced neutralizing activity (a ≥2-fold decrease), and 16.7% had a decrease higher than 4-fold, almost twice that reported for B.1.1.7. The reduction in neutralizing activity was even higher against the B.1.351 variant, with 90.5% of patients showing a fold decrease of ≥2 with a mean decrease almost 3 times higher than that against the P.1 variant. These findings are consistent with other studies [27,28,30,31] in which a reduced neutralizing potency of antibodies against these variants has been reported, although convalescent and vaccinated sera seem to better neutralize P.1 than B.1.351. The reduced neutralization could be ascribed to the E484K mutation shared between P.1 and B.1.351 and related to a possible escape from neutralizing antibodies. Considering that some mutations of P.1 are in common with B.1.351, such as the E484K mutation, it is very likely that a similar or higher reduction in neutralizing antibodies may be caused by variants carrying the same mutation. The immunoevasion could lead to reinfection of subjects who recovered from a previous infection [16] and could reduce the protection induced by vaccination and/or natural infection [27]. However, recent studies have proven that boosting pre-existing immunity against SARS-CoV-2 with one dose of mRNA vaccine induces a strong neutralizing response even against divergent variants (i.e., B.1.351) [8,32]. In Italy, subjects naturally infected by SARS-CoV-2 receive just one dose of vaccine from 3 to 6 months after infection [33]. In addition, a recent study [34] reported that neutralizing antibodies after natural infection by SARS-CoV-2 can last for at least 9 months.
Notably, one serum sample was not able to neutralize the B.1.1.7 variant, three were not able to neutralize the P.1 variant, and nine were not able to neutralize the B.1.351 variant, suggesting that symptomatic subjects previously infected by SARS-CoV-2 wt virus might not be fully protected against the emerging variants.
A key strength of the study is that the neutralizing activity has been assessed using authentic live SARS-CoV-2 viruses and not surrogate VN assay.
This study has some limitations. First, the cohort only included hospitalized patients, which may not be representative of the general population. The lack of follow-up samples after discharge does not allow the assessment of the full breadth of cross-neutralization potential that can be reached following more extended affinity maturation.
Overall, our findings provide evidence of a remarkable lower neutralization capacity of SARS-CoV-2 antibodies acquired during the first wave of the pandemic against the B.1.351 variant, and to a lesser extent against P.1 and B.1.1.7. This might suggest the possibility of reinfection or that previously infected individuals may be partially protected against current and new emerging variants with relevant mutations. However, the immune response induced by natural infection may protect from severe disease [35]. The emergence of the Delta VOC is also raising concern; however, so far, it appears that vaccines remain effective, especially after two doses [21,22,36]. Our study highlights the importance of evaluating SARS-CoV-2 pre-existing immunity against emerging variants as a tool to foresee the immune escape and the extent of vaccine efficacy.