Limited Neutralization of Authentic Severe Acute Respiratory Syndrome Coronavirus 2 Variants Carrying E484K In Vitro

Abstract

As vaccination campaigns against coronavirus disease 2019 (COVID-19) are ongoing, the majority of the world´s population remains unimmunized, with many at risk for severe disease. Therapeutic and prophylactic agents with proven efficacy are both still urgently needed. Immunoglobulin (Ig) G1 monoclonal antibodies (mAbs) prevent viral attachment and entry into human cells by blocking attachment to the angiotensin-converting enzyme 2 (ACE2) receptor. However, several severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) “variants of concern” (VoCs), which emerged in late 2020, have been associated with increased transmissibility or immune evasion, and it is of particular importance to evaluate the effectiveness of mAbs against these variants. In particular, substitutions E484K and K417N in the spike protein have been associated with immune escape, or increased binding to the ACE2 receptor (eg, N501Y) [1].

To evaluate whether variants harboring these substitutions might be also effectively neutralized is critical for effective treatment. So far, studies have been conducted with artificial pseudoviruses, indicating that the B.1.351 (beta), P.1 (gamma), and P.2 (zeta) variants are expected to be resistant to therapeutic mAbs. Furthermore, these variants have been proposed to be partially resistant to neutralization by serum samples obtained from convalescent patients with COVID-19, as well as by vaccine-elicited serum samples obtained from individuals after immunization with BNT2b2 or messenger RNA (mRNA) 1273. However, because in real-world settings additional substitutions that might determine replication efficiency define each VoC, evidence is still lacking that evidence is still lacking that data generated using artificial pseudovirus models also apply to naturally occuring (authentic) viruses. Hence, in the current study we analyzed the ability of bamlanivimab, casirivimab, and imdevimab to neutralize authentic SARS-CoV-2 VoCs, including B.1.1.7 (alpha), B.1.351 (beta), and P.2 (zeta), in infectious cell cultures. We also analyzed vaccine-elicited serum samples after immunization with BNT162b2 and mRNA1273, and convalescent serum samples for their ability to neutralize authentic SARS-CoV-2 variants.

Bamlanivimab and REGN-CoV-2 are IgG1 mAb preparations used to treat COVID-19. Bamlanivimab (LY-CoV555) has been demonstrated to accelerate the decline in viral load [2] and has been authorized by the Food and Drug Administration for emergency use in early mild to moderate COVID-19. REGN-CoV-2 consists of 2 potent mAb REGN10933 (casirivimab) and REGN10987 (imdevimab), both binding noncompetitively in different regions of the spike protein [3]. It is also approved by the Food and Drug Administration for emergency use. Class 1 mAb REGN10933 has an ACE2 blocking property, while class 3 mAb REGN1087 binds outside the receptor-binding domain (RBD).

Emerging SARS-CoV-2 VoCs and “variants of interest,” referred to as B.1.1.7 (alpha), B.1.351 (beta), P.1 (gamma), and P.2 (zeta), initially observed in the United Kingdom, South Africa, and Brazil, respectively, are in the process of fixation in the population. Mutations in the spike protein’s receptor binding domain, in particular N501Y, were associated with increased infectivity due to enhanced receptor binding [4]. Further studies have shown that of E484K represents an immunodominant site on the RBD because it reduced the naturalization capacity of human convalescent serum samples by >100-fold [1–4]. These amino acid substitutions have also been observed to evade the antibody response elicited by infection with other SARS-CoV-2 variants, or vaccination. So far, it is still being debated whether mAb preparations are equally effective against these variants.

In the current study, we examined the ability of bamlanivimab, casirivimab, imdevimab, vaccine-elicited serum samples after immunization with BNT162b2 and mRNA1273, respectively, and convalescent serum samples to neutralize authentic SARS-CoV-2 variants B.1.1.7 (alpha; mutations include N501Y and Δ69/70), B.1.351 (beta; mutations include E484K and N501Y), and P.2 (zeta; mutations include E484K in the absence of a 501 mutation). These isolates were obtained from travelers from Great Britain, South Africa, and Brazil, respectively. Two SARS-CoV-2 isolates collected in early 2020 were also tested (B [FFM1] and B.1 [FFM7]) [5, 6].

METHODS

SARS-CoV-2 IgG levels were quantified using the SARS-CoV-2 IgG II Quant kit (Abbott Diagnostics) (Table 1). The mAb working solutions, convalescent serum, or serum samples from BNT162b2- and mRNA1273-vaccinated individuals were serially diluted 1:2, incubated with 4000 TCID₅₀ median tissue culture infective dose)/mL for each SARS-CoV-2 isolate, and subjected to cell-based SARS-CoV-2 neutralization assay. The corresponding sample dilution resulting in 50% virus neutralization titer was determined. After 3 days of incubation, cells were evaluated for the presence of a cytopathic effect.

All relevant ethical guidelines have been followed and approved by the Ethik-Kommission des Fachbereiches Medizin der Goethe Universitaet Frankfurt (no. 250719). All necessary patient/participant consent has been obtained, and the appropriate institutional forms have been archived.

RESULTS

The variant B.1.1.7 (alpha), as well as B and B.1 isolates from early 2020 (FFM1 and FFM7), could be efficiently neutralized by bamlanivimab, casirivimab, and imdevimab (titer, 1/1280, respectively) (Figure 1A). However, with bamlanivimab, no neutralization effect could be detected against either B.1.351 (beta) or P.2 (zeta), both harboring the E484K substitution. Spike protein alignments confirmed that only E484K substitution occurs exclusively in the 2 strains that could not be neutralized by bamlanivimab (GenBank accession nos. MW822592 [B.1.351; FFM-ZAF1/2021] and MW822593 [P.2; FFM-BRA1/2021]). Imdevimab. binding outside the ACE2 RBD, was able to neutralize the virus without decline in efficiency. However, for casirivimab a severe drop in neutralization capacity was observed for B.1.351 (beta) (titer, 1/20), and a considerably reduction in the neutralization capacity against P.2 (zeta) (titer, 1/320) (Figure 1A).

Figure 1.

Neutralization titers against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants. The monoclonal antibody (mAb) solutions (A) and serum samples from (B) convalescent or (C) vaccinated individuals were serially diluted (1:2) and incubated with 4000 TCID₅₀ (median tissue culture infective dose)/mL of the indicated SARS-CoV-2 variant. Susceptible cells were subsequently inoculated and analyzed for a cytopathic effect formation after 72-hour incubation. Values represent reciprocal dilutions of SARS-CoV-2 microneutralization titers resulting in 50% virus neutralization (NT₅₀). The indicated mAb solutions were used in physiological concentrations according to the manufacturer’s instructions (Regeneron Pharmaceuticals and Eli Lilly and Company, respectively). Mean values from 2 replicates are shown, each determined with Caco2 and A549-AT cells. Colored dots represent serum samples from individuals vaccinated with messenger RNA 1273 (red) or BNT162b2 (blue). Statistical significance, compared with SARS-CoV2 B (FFM1) or B.1 (FFM7), respectively, was determined using a 1-tailed, paired Student t test; sample 21 (from an individual vaccinated with BNT162b2 after convalescence) was excluded from statistical analysis. Control serum samples were tested negative for SARS-CoV-2 antibodies. *P < .05); **P ≤ .01; NS, not significant.

Open in new tabDownload slide

These data show that B.1.351 exerts a 1000-fold reduction of the SARS-CoV-2 neutralizing activity of bamlanivimab and casirivimab and confirms previously published observations with artificial pseudoviruses [7]. In addition, P.2 (zeta) virus was resistant to bamlanivimab and partly resistant to casirivimab (>4-fold). However, using the clinically used combination of REGN-CoV-2, full neutralization was observed, indicating unrestricted effectiveness of a therapeutic treatment.

To determine the neutralization efficiency of convalescent and vaccine-elicited serum samples, SARS-CoV-2–specific antibodies were quantified (Table 1). Antibody titers were detected for the vast majority of the tested samples, with an average of 524.0 (standard deviation, 1256.0) binding antibody units (BAU)/mL for convalescent serum samples, 2135.9 (1868.4) BAU/mL for BNT162b and 2927.6 (1711.4) BAU/mL for mRNA1273 vaccine-elicited serum samples. Comparable to SARS-CoV-2 variants from early 2020 (B and B.1), convalescent (Figure 1B and Supplementary Figure 1A) and BNT162b2 or mRNA1273 vaccine-elicited serum samples, respectively (Figure 1C and Supplementary Figure 1B), efficiently neutralized SARS-CoV-2 B.1.1.7 (alpha).

A moderate but significant decrease (average of all vaccine-elicited serum samples, 1.27-fold) in the neutralization efficiency of serum samples from vaccinated individuals was observed for SARS-CoV-2 B.1.1.7 (alpha) compared with SARS-CoV-2 B (P = .02). Compared with B.1 (FFM7), a slightly higher neutralization efficiency (0.54-fold) could be observed (P = .02). All tested serum samples were significantly less efficient against SARS-CoV-2 B.1.351 (beta) and P.2 (zeta) (Figure 1B and 1C and Supplementary Figure 1). Testing of convalescent serum samples revealed 6.27–6.74-fold lower neutralizing activity against B.1.351 (beta) and 5.09-fold lower activity against P.2 (zeta), compared with the variants from early 2020 (Figure 1A and Supplementary Figure 1A).

Serum samples from vaccinated individuals were on average 4.35-fold and 2.67-fold (B and B.1, respectively) less effective against variant B.1.351 (beta) and 2.84-fold and 2.86-fold (B and B.1, respectively) less effective against variant P.2 (zeta). Considering the antibody titers, the efficacies of both mRNA1273 and BNT162b2 vaccine-elicited serum samples were comparable in our study. In 1 serum sample obtained from an individual who was vaccinated after convalescence (sample 21) a loss of neutralization efficiency was observed despite high antibody concentrations (Table 1).

Discussion

The COVID-19 pandemic continues to set an extraordinary burden on world health. While the proportion of protectively vaccinated individuals is steadily increasing, there are still only limited treatment options available. Furthermore, it is still unclear how long vaccinations will last and whether they are effective against all variants. Because the SARS-CoV-2 variants B.1.1.7 (alpha) and B.1.351 (beta) tested in this study are currently displaced by others (such as B.1.617.2; delta), there will be a constant need to test the available protective and therapeutic options. Previous studies, conducted predominantly with pseudotyped viruses, have shown that the sensitivity of B.1.1.7 (alpha) to neutralization by convalescent serum samples is slightly lower compared with preceding viral isolates; however, the neutralization sensitivity of variant B.1.351 was shown to be significantly reduced [8]. Our in vitro findings using authentic SARS-CoV-2 confirm that, in contrast to vaccine-elicited serum samples, bamlanivimab and casirivimab may not provide efficacy against SARS-CoV-2 variants B.1.351 (beta) and P.2 (zeta), both harboring the E484K substitution. Because imdevimab was able to efficiently neutralize both variants, therapeutic treatment with the REGN-COV-2 combination remains effective. In agreement, previous studies using artificial pseudoviruses reported that LY-CoV555 and REGN10933 are ineffective against B.1.351 (beta) but is still effective against B.1.1.7 (alpha) [7, 9].

This and previous work revealed a lower neutralizing activity against E484K harboring variants B1.351 (beta) and P.2 (zeta), which may facilitate reinfection with emerging variants [8, 10, 11]. Amino acid substitutions hinders spike proteins to be bound by antibodies, resulting in reduced protection against SARS-CoV-2 infection [12]. Hence, a higher antibody titer is needed, which might be provided by a second vaccination dose inducing the formation of a critical amount of neutralizing antibodies [13, 14]. Even a single immunization already increased the neutralizing titers of convalescents, although with reduced efficiency for B.1.351 (beta) [8]. Of note, immune response after vaccination also includes a broad T-cell repertoire, which might be effective despite resistance to antibody-mediated immunity [15]. Hence, viral escape of T-cell immunity is unlikely. However, whether this also applicable for tested and future SARS-CoV-2 variants has to be further investigated.

Our study has several limitations. While we used authentic SARS-CoV-2 variants rather than pseudotyped surrogate models, only a single representative isolate was studied for each variant. Hence, a broader analysis with more isolates should be expanded in future studies. We tested convalescent serum samples and vaccine-elicited serum samples (mRNA1273 and BNT162b2) for their neutralization efficiency; however, the number of serum samples was relatively small and most of the convalescent serum samples originated from individuals infected with non-VoCs. In future, it would be interesting to study serum samples from individuals infected with VoCs and thus with corresponding antibodies. Serum samples from individuals vaccinated with vector vaccines and heterologously vaccinated should also be included.

We conclude that confirmation of the SARS-CoV-2 variant, including screening for E484K, may be needed before initiating mAb treatment with bamlanivimab and to ensure efficacious and efficient use of the antibody product. Variant-specific mAb agents may be required to treat emerging SARS-CoV-2 VoCs. To efficiently neutralize VoCs carrying E484K, a high antibody titer is needed to induce formation of a critical amount of neutralizing antibodies.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Acknowledgments. We are thankful for the numerous donations to the Goethe-Corona-Fond and for the support of our SARS-CoV-2 research.

Potential conflicts of interests This work was supported by the Deutsche Forschungsgemeinschaft (grant WI 5086/1–1 to M. W. and the Federal Ministry of Education and Research (COVIDready; grant 02WRS1621C to M. W.). S. H. has received research support from Roche Diagnostics and a speaker’s fee from Sanofi Genzyme. T. W. has received speaker and consultancy fees from Gilead Sciences, Merck Sharp & Dohme, and Janssen Pharmaceuticals. V. M. C. reports patent PCT/EP2021/064352 pending to the German Center for Neurodegenerative Diseases and Charité-Universitätsmedizin Berlin. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References